Using a PIR to reduce wear and tear on a Nixie clock

"Does a lit-up Nixie tube in a forest wear out even if there's no-one to see it?"

Figure 1:
The "Black'n'wood" Nixie Clock (blue back-light turned on)
with the PIR (Passive InfraRed) sensor to the right.
With no detected movement in the room, the high voltage
supply turns off, reducing wear on the tubes.
Click on the image for a larger version.

This millenia-old riddle has a simple answer:  Yes.  Yes, it does.

This article has not so much to do with this specific model of Nixie clock, but rather adding a PIR (Passive InfraRed) sensor to turn off the display when there is no-one in the room to look at it.

They wear out!

Nixie clocks and other neon-glow displays (e.g. Panaplex), along with VFD (Vacuum Florescent), Numitron and CRTs have a "wear out" mechanism when they are operating:  In other words, when they are on, they are slowly degrading.

By limiting the "on" time of such displays only to when someone is likely able to see it one can prolong its overall useful life in many cases.  As many of these devices are no longer made, the supply of "new, old stock" tubes is very limited and what there is still available is becoming more expensive year upon year.

* * *

A few years ago - at a swapmeet - I picked up a "Nixie" ¹ clock - the "Black'n'wood" by Nocrotec.  It was in a plastic bag with loose parts, but for only $20 I couldn't resist!

Getting it home I found the problem:  One of the elements of the "10s of hours" tube was shorted to the anode and very visual close inspection revealed that two internal wires were touching each other.  A bit of "percussive repair" (banging it on the table) moved the two wires away from each other and the tube was once again usable.  I suspect that the problem was originally caused by the tube experiencing mechanical shock.

The current limiting resistor associated with this same element was burned, so replacing it returned the clock to full operation.

Over the next year or two the clock has continued to work fine - although the display got "glitchy" and began to dim - but the biggest clue was that the flashing colon neon lights were flickering but this was quickly traced to the failure of the main high voltage filter capacitor on the 180 volt supply:  These problems went away with the replacement of that capacitor - but I digress.

* * *

All of this brings me to the main topic of this article:  Reducing the wear and tear of the neon tubes.  

Sitting unused, in a box, many "vacuum" devices (I'll include neon indicators and other cold-cathode tubes in this category even though they are not strictly "vacuum" devices) have the property that laying on the shelf, they (usually) have little/no degradation over time.  There are many (now) century-old devices that have been sitting around that work just as well as they did when they were made - the caveat that they haven't been compromised in some manner (e.g. broken, corrosion, failure of a seal, etc.)

Like most "vacuum bulb" devices - which include thermionic tubes/valves (with a filament) and those without a filament - like neon indicators - there is a definite lifetime related to acceptable performance when they are operating.  For normal tubes/valves, the emission from the filament/cathode will inevitably drop over time - often due to gradual degradation of emissivity and the "work function" of the cathode.  "Cold cathode" devices (e.g. those without a filament) like neon indicators also suffer degradation - and the causes are broadly similar:  Degradation of the materials and subsequent contamination.

In the case of the neon indicators, one major cause of degradation is the inevitable "blasting" of atoms from the electrodes' surfaces (called "sputtering") where the metal gets liberated - only to redeposit elsewhere.  The most obvious result of this is that the inside of the of the glass envelope darkens, reducing the brightness of the brightness of the display - but even if the glass were to remain clear, this and other effects conspire to reduce the brightness overall.

Running neon indicators such as these "Nixies" at lower than maximum current will reduce these effects - but what about not running them at all?

Nixies are meant to be seen!

The entire point of a "Nixie" clock is that it is cool to look at - but what if no-one is there to see it?  If we operate a Nixie tube in a forest and no-one is there to see it, it still wears out!

The goal, therefore, is to turn off the display when no-one is in the room.

Turning off the display

First, we need to figure out how to turn off the display, presuming that the clock or other device has no obvious means of doing so (e.g. there is no "turn of the display" switch or pin).  Two options came to mind:

Approach #1:  Removing the power

Figure 2:
Interface with HV converter.  PNP transistor "Qa", when
its base is pulled low, injects current into Pin 5 of the
HV converter chip, effectively turning it off.
Click on the image for a larger version.

As the "Black'n'Wood" clock has a battery back-up, I first tried the most obvious thing:  Interrupt the power to the clock if no motion was detected.  This worked - in theory - but I soon noted that the clock was losing almost a minute every day.

After I'd first repaired the clock, I applied correction factors via its menu and got it to stay within a fraction of a second per day - but I had assumed that this was done in the clock module itself (as some Dallas/Maxim devices are equipped) but this was not so:  While the timekeeping module continued to run on the battery, the firmware on the clock itself - being powered down - was obviously not and the calibration that I'd applied was missing, explaining why it was keeping time so badly.

To be sure, I could have likely done something to "fix" this (e.g. trim the oscillator with a tuning capacitor, added a GPS module to auto-set the clock, etc.) but since the clock - while it was running normally - was very stable, I decided to try another approach.

Approach #2:  Turning off the high voltage

Figure 3:
A top-down view of the components added to the high
voltage switching converter IC to produce the circuit depicted
in Figure 2, above.
Click on the image for a larger version.

The clock itself runs on 12 volts DC and to get the 150-180 volts needed to drive the neon displays, there is an onboard voltage converter.  While no schematic seems to be publicly available for this particular clock, the various sections of its circuitry are easily identified by visual inspection:  A large-ish inductor adjacent to a high voltage capacitor flagged the location of the voltage converter and the chip next to those components.  

This clock uses two switching supplies - the first one converts the nominal 12 volts down to 5 volts for the logic, but the second one - near the high-voltage capacitors - is the one that produces the (approximately) 180 volts for the Nixies.  Both of these use a common type of switching supply controller chip - the MC34063 - and since the implementation of these chips is spelled out in the data sheets, we have some insight as to how they work.

Fooling the voltage converter into shutting down

Like most any voltage regulator or converter, it monitors its own output voltage - typically through a pair of resistors ("Rdiv" is one of them, depicted in Figure 2 - the other, not shown, would go between the chip and ground) that are chosen to divide the desired output voltage down something close to the chip's on-board reference voltage - in the case of the MC34063, 1.25 volts, which is applied to its pin #5:  If the voltage on this pin is lower than 1.25 volts, the switching converter adjusts the voltage higher but if the voltage is higher, it reduces the voltage.

Figure 4:
A side view of the high voltage switching converter IC
showing the components added to it to allow the
180 volt supply feeding the Nixie tubes to be turned off.
Click on the image for a larger version.

As can be seen from the diagram in Figure 2, I tacked a PNP transistor ("Qa") - and three resistors - across several of the pins of the MC34063 high voltage converter.  Using a hot soldering iron and "pre-forming" the shape of the components to match the locations of the needed IC pins, it's possible to attach this simple circuit directly to pins 5 and 6 of the high voltage switching converter IC without risk of damage to the chip or other, nearby components.

If the base of Qa, the PNP transistor (I used a 2N3906) is pulled to ground (via the 10k resistor, Rc), it will turn on - and with its emitter connected to Pin 6 of the MC34063 - its power supply pin - it will apply current, via the 3.3k resistor (Ra), to Pin 5 of the MC34063, dragging the voltage on this pin up.  When this happens, the MC34063 will "think" that the voltage is too high and effectively turn off.  If the base of Qa is allowed to float (nothing connected to it), this transistor is biased off by the 100k resistor (Rb) between the emitter and base and the high voltage converter will run normally (e.g. the display will be on).

Any converter will do

While this article shows the example using the MC34063, this sort of technique could be applied to about any switching-type of voltage converter.  Determining a bit about the circuit itself could be done simply by referring to the data sheet of the chip that was used - as was done here - but it could also be done with a bit of reverse-engineering.

It would have also been possible to find the power supply lead feeding the voltage converter - in this case, about 12 volts from the external power supply - and interrupt it, perhaps with a relay, a PNP transistor or a P-channel FET.

If you are using an "old-school" power supply that does NOT some sort of switching converter, perhaps consisting of a high-voltage winding, rectifier and capacitor to develop the high voltage for the tubes, your best option may be to use a relay to open the supply - preferably interrupting the pre-rectified AC side, directly.   At such voltages switching DC is best avoided due to the possibility of contact-damaging arcs:  Switching on the AC side (or between the rectifier and the first filter capacitor) is better in that the voltage falls to zero twice per cycle of the AC waveform and any arcing that does occur will extinguish at that time.

Getting the connection outside the clock

In perusing the manual for this clock I noticed that the 6 pin mini-DIN connector - intended for connection to an external GPS or DCF77 radio receiver - not only had ground (Pin 1) and power (Pin 2 for 5 volts), but also an unused pin (#4) that I verified to be floating - and to this I connected the end of the 10k resistor (Rc) to this pin with a flying lead inside the clock.  With the three needed signals (power, ground and the "disable" line) on the mini-DIN connector, I was ready to connect it to a sensor.

A PIR sensor to turn it off and on

A PIR (Passive InfraRed) sensor fits the bill for this task quite well - and they are inexpensive.  These devices use pyroelectric detectors to detect heat from warm, moving objects - which includes us humans - by focusing deep infrared energy onto a pair of sensing surfaces from an array of Fresnel lenses.  A moving object in the field of view will cause a difference in the pair that can reliably indicate that an object in view was in motion.

Figure 5:
This circuit was added to the output of the PIR to present
an open-collector to allow transistor Qa in Figure 2
to properly turn off when the HV was to be turned on.
Click on the image for a larger version.

The PIR sensor that I chose was found on Amazon - three of them for under US$10 - and it has exactly three connections:

• Power.  This particular PIR sensor was happy to operate from between 5 and 12 volts, having an onboard 3.3 volt regulator.
• Ground.  This is the negative supply and the reference to the output signal.
• Output.  This output pin goes "high" (to 3.3 volts) when motion is detected.

This sensor also has two potentiometer adjustments:

• Delay - Which is the amount of time the output will go "high" when motion is detected.
• Sensitivity - As the name implies, this sets the degree to which the device reacts to movement.

There's also a jumper:  The piece of paper that came with the PIR sensor implies that this determines if the output is "re-triggerable" (the default setting) or not.  Being re-triggerable means that motion will reset the delay timer whenever it's detected:  If it were not re-triggerable, the delay time would be reset only after the delay had expired and the output had turned off.  Clearly, we want to use the "re-triggerable" setting so that any movement simply extends the timer.

It turns out that moving the jumper on this board from its factory position stopped the unit from working at all and a quick bit of reverse engineering revealed that whoever designed this board simply connected it to the wrong place - probably due to poor reverse-engineering on the part of the "designer" of this (likely cloned) circuit board.  Fortunately, the wiring of the circuit is such that it is already wired as being re-triggerable, so we can leave it alone.

Figure 6:
Perhaps a bit messy, but this is the two-transistor circuit
depicted in Figure 5, tacked to the pins of the PIR module's
circuit board.  The three-conductor cable that connects to the
clock via the mini-DIN connector can just be seen.
Click on the image for a larger version.

One problem with the 3.3 volt output is that it is a logic output that is limited to 3.3 volts because it has both pull-down and pull-up transistors, internally.  As we discussed in the previous section, we need to ground the base of transistor "Qa" through the resistor to disable the high voltage and let it float to an unknown voltage to allow it to turn on.  Because the output is not an open collector or open drain, it cannot be pulled higher than approximately one diode drop above the 3.3 volt supply on the PIR chip:  This voltage is lower than the emitter voltage of the transistor that we added ("Qa") which means that Qa will always be turned on, always disabling the high voltage converter!

To fix this we need to provide an open-collector output - but preserving the polarity of the output - which is to say that we want it to be an open collector to allow the base of "Qa" to float high when movement is detected, but go to ground and turn on "Qa" when it is not.  To accomplish this, transistor "Qb" takes the "high-active" pulse from the PIR and inverts it - and then transistor "Qc" will invert it yet again, but this time with the needed open collector.  "Qb" and "Qc" can be practically any NPN transistor - I used 2N3904 types in this cicruit.

In experimenting with this PIR sensor module, I noted that when set to "maximum" the "on" time from the output was about 150 seconds - about 2.5 minutes.  I was able to iteratively adjust the "sensitivity" control incrementally upwards until I found a setting that reliably detected even slight motion in the room - but seemed not to randomly "false" trigger, the result being that when I was in the room - say watching TV - it would stay on most of the time, but be easily (re)triggered by even slight movements.

Figure 1 shows the clock with the PIR sensor next to it, sitting on the shelf below my TV.  I purposely set the PIR sensor back from the edge of the shelf - not just to line up with the front of the clock, but to obscure part of the view of the floor to reduce the probability that a cat would trigger it:  Since cats sleep most of the time, anyway, their occasionally triggering the PIR sensor isn't a big deal.

Reducing "wear-out"

For a "cold cathode" tube like a Nixie, turning the high voltage on and off is not a stress on the tube:  After all, simply changing the segments to show the time is also turning on/off parts of the tube.  With no voltage present, there is no electron bombardment on the elements within the tube and thus, it will not experience wear.

Powering down other devices in the absence of "viewers"

There are other types of "antique" displays that may benefit from having some sort of "human presence detector".  For example, a VFD (Vacuum Fluorescent Display) has a wear-out mechanism similar to a Nixie in that electron bombardment will gradually degrade the phosphors - and the cathode (filament) may also lose emission.

Similarly, if one has a "Scope Clock" - a vector-graphics clock that uses an oscilloscope tube to show the time - it, too, will wear out over time, the emission of the from the cathode will drop over time - not to mention possible burning of the phosphor.

Figure 7:
An example of a "Scope Clock" - a vector-graphic clock
display shown on an oscilloscope using a cathode-ray tube
(CRT).  In this photo, the CRT in a Cushman CE-50A
communications monitor is being used to demonstrate, but
an old, analog oscilloscope would work as a "permanent"
fixture and blanking it when no-one is looking would
extend the life of increasingly-rare CRTs.
Click on the image for a larger version

For these two examples, a bit of care should be taken in that while removing the high voltage source may partially remove the wear-out mechanism (e.g. degradation of phosphors) other steps would be required to mitigate the diminution of filament emission over time.  This could include turning off the filament - or at the very least, reducing its voltage, perhaps in steps, in the absence of the anode voltage.  If grid voltages happen to be present, those, too, should be carefully considered to see if they should be removed when the high voltage is turned off - but since these often share the same power supply, this problem may take care of itself.

In so-doing - and depending on the nature of the display tube - other precautions may also be required (e.g. removing all other voltage prior to powering down the filament) to avoid damage - and frequent power-cycling of the filament itself may be an issue:  These are potential issues that should be considered - but are beyond the purview of this article.

Footnote:

1. The name "Nixie" is a trademark of Burroughs Corp. to describe certain types of neon-glow indicators.  Like many trademarks, it's become "genericized" - as done in this article - and nowadays it's commonly used to denote all types of similar cold-cathode glow devices in which each digit is indicated by a separate element within the tube in the shape of the desired numeral or symbol - whether they were made by the original trademark holder or not.

 * * * * * * *

This article stolen from ka7oei.blogspot.com

[END]

 

Repairing a NanoVNA "V2" with nonsensical readings (Blown up input)

Figure 1:
The SAA-2N, a variant of the Version 2.  As noted in the text
I chose this version as it had N connectors and was suited
for my specific needs and has served well for several years.
Click on the image for a larger version.

The NanoVNA has become a ubiquitous device for the RF toolbox:  For less than US$200 you can get a device that is reasonably capable, accurate and covers a usefully-wide frequency range - from a few 10s of kHz into the GHz range, depending on the model.

The original NanoVNA - sometimes found for less than US$50 - works reasonably well up to about 250-ish MHz - the limit of the frequency range of its RF synthesizer (typically an Si5131).  Using harmonics for the higher ranges allows the unit to "work" higher than this, but with the dropping gain of the mixers and lower power of the higher-order harmonics diminish its usefulness and it pretty much "runs out of steam" by the time one gets to around 1 GHz.

The "Version 2" of the NanoVNA (which includes many variants, such as the "H2") improved on this using a different detector (often the Analog Devices AD8342 which is rated to work above 3 GHz) and two RF synthesizers - the original Si5351 and another device (usually the Analog Devices ADF4350 or similar) that takes over from the '5351 at the higher frequencies (usually above around 140 MHz) up through 3+ GHz.  The different architecture of the Version 2 necessitates greater complexity - which also implies increased vulnerability as we'll see.

Note:  This article will generically refer to the units thus-constructed as "V2" NanoVNAs - regardless of who actually makes/sells them.

Background

When I looked for an "upgrade" to the original NanoVNA I was looking for a device that would have durable connectors that would withstand hundreds of connect/disconnect cycles and be able to hang the unit from the interconnect cables themselves to measure gear "in situ".  This "need" was due to its expected use:  Being dragged around in the equipment box to repeater and radio sites, the houses of other amateurs, and being used on the workbench - and almost never above about 1.5 GHz - a frequency range for which I have other gear, anyway.  Having friends with other NanoVNA V2 variants, I've heard how physically fragile these devices are - mostly related to the supplied cables and the SMA connectors themselves, and how they may be physically attached to the circuit board.  All of this ruled out anything with SMA connectors.

For this reason I got the SAA-2N (metal case with "N" connectors) and have used it enough that the nickle plating is starting to wear off the threads of the N connectors, indicative of a number of connect/disconnect cycles (now in the hundreds) that would have trashed even the best-quality SMA connectors:  Had anyone else made a device as physically rugged as this, I'd have considered it - but as of the time of this writing, no-one else has done so!

The problem

One issue that can befall users of all NanoVNA "V2" variants is that of static sensitivity - particularly on the S11 port (e.g. "Port 1" or "CH0").  While some users on the forums suggest that this is due to "inferior" components of some of the clones, this is simply false:  It's a result of the way that the unit is designed - and the components necessary to execute it which are intrinsically more vulnerable to ESD (ElectroStatic Discharge) than simpler versions.  

Additionally - and speaking for myself -  I'm much more likely to use this unit - with its "N" connectors - "out in the field" where it may be exposed to hazards (ESD) than someone who gently rests their NanoVNA on their workbench for its entire life.  At least for me, this increases the probability that a unit that is more likely to be used out in the field, away from the "safe" workbench, will be damaged in some way - electrically or mechanically.

Specifically, the problem is the Maxscend MXD8641 RF switch found in several places on the "Version 2" NanoVNA variants (e.g. those that work past 3 GHz).  This is a SP4T (4-way) RF switch used to route the signals that, among other things, is used to switch between the Si5351 oscillator and the ADF4350 (or similar) - but one of these devices may also be found right on the S11 port - and that's where the problem lies.  (This device is not on the original NanoVNA, which seems to be more rugged in this respect.)  As the list price of the MXD8641 is well under US$0.07 in any sort of production quantity, it's unlikely that so-called "cheap counterfeits" are being used.

While the data sheet for this device indicates that it has good ESD protection, it's clear from the NanoVNA forums - and my personal experience and that of friends who also have a V2 NanoVNA - that this protection is NOT necessarily adequate for in-the-field, casual use.  If your NanoVNA never leaves the workbench, you may never have this sort of problem - but if you drag it out in the field like I do, you may well have run across situations where the NanoVNA becomes damaged.

Figure 2:
If the RF switch chip is has a "dead" short to ground,
you may end up with chaos of lines like this after you do a
"short-open-load" recalibration.
Click on the image for a larger version.

As it happens, a good-quality coaxial cable is capable of storing quite a bit of energy, stored as voltage.  With a typical 50 ohm coax having around 30pF per foot (100pF/meter) this, coupled with triboelectric effects - not to mention static on your person or on the antenna to which you might connect it - pose a hazard to sensitive electronics, and this is device is no exception.

Considering that the specifications for this part note that it should withstand 1kV from the "human body model" (100pF discharged through a 1.5k resistor) it's easy to see that discharging even a hundred or so volts from 3000pF of capacitance - represented by about 100 feet (28 meters) of coaxial cable - and not having any series resistance - could easily damage this part.  One can trivially develop hundreds of volts across a piece of coaxial cable just from handling - not to mention that which might be developed across an antenna from wind static!

On the NanoVNA forums I've seen people commenting on how they are suddenly getting nonsensical readings on their V2 (see Figure 2) or the apparent inability of the unit to maintain calibration - and this is indicative of possible damage to the MXD8641 found on the S11 port, likely caused by voltage discharge:  With a 0.047uF series coupling capacitor, it would not take much accidentally inject enough energy to damage that part!

My own NanoVNA V2 was, itself, damaged through discharge after having used it for several years without incident, apparently when connected to a fairly long length of coaxial cable - despite reasonable effort and care.  I have also encountered others with V2 NanoVNA variants that have also experienced similar issues and in all of the cases where I have been able to get the unit into my hands, it's been due to damage of the MXD8641.  One thing in common with these other failures is that these folks often - but not always - use their NanoVNA V2 places other than just on the workbench - to test antennas, cables, etc. out in the field:  I don't know about you, but I can detect a common thread here...

Comment:  

As can be seen in Figure 2, there is a 0.047uF monolithic ceramic coupling capacitor.  Ideally, a capacitor tester would be used to verify its integrity, as would an ohmmeter to assure that it's not shorted - even intermittently.

It is possible that if the board is flexed repeatedly (tightening of the connectors or other mechanical stress on the board) that this capacitor and/or the solder joints in the signal path have broken or become intermittent.  It would be a good idea to use an ohmmeter to "buzz" out continuity from the end of the cable to the "connector" side of the coupling capacitor, and also between the other (circuit-side) lead of the capacitor and the trace.

Sometimes, reflowing both ends (allowing time in between for the two solder connections to cool) will restore connectivity if the solder joint has cracked - but note that the capacitor itself may be mechanically intermittent and thus require replacement, anyway.

Diagnosis:

Having corresponded with several others who have had the misfortune of damaging their V2 NanoVNAs, there are several ways that the damage will manifest itself that may appear to be totally different, but are actually the caused by the same problem.

• Sudden high return loss everywhere.  Sometimes the discharge will cause the '8641 to simply fail dead-shorted and the user will see a very high VSWR on everything and if one configures the unit for an S21 (e.g. "through") reading, the apparent insertion loss is extremely high - indicated by a very high noise floor (after attempted recalibration) as compared to what it was when working.  If one attempts to recalibrate (Open/Short/Load) a completely nonsensical display like that in Figure 2 is likely to result after doing so.
  

• Unit will not seem to "hold" calibration well.  This seems to be the more common failure:  The user will suddenly find that the unit is out of calibration for no obvious reason - and this is often worse at higher (VHF/UHF) frequencies than at HF (30 MHz or lower).  Using the Open/Short/Load, the unit will seem to calibrate, but the noise floor in "S21" measurements will likely be higher - and the calibration will seem to drift all over the place over time and with temperature.

This second failure mode is a bit more sinister in that the user may, at first, presume that the loss of calibration was a result of user error.  What has actually happened was that rather than fail shorted, the '8641 will be damaged, often putting between 30 and 80 ohms of resistance between the input port and ground  - a value that will vary significantly with temperature and over time.  This is insidious in that the unit may "seem" to be working - but it will likely give incorrect readings, even if it seems to calibrate properly.  This causes two problems:

• Calibration drift.  The NanoVNA's source impedance is nominally 50 ohms - but placing a varying resistance across this - possibly in the 30-80 ohm range - will lower this, causing calibration to seem to vary despite frequent recalibrations.  As noted above, if an "S21" measurement is done (e.g. a "through insertion loss"), the shunting of the source signal (on CH0/Port 1) will reduce its amplitude - possibly significantly - and you may notice that the noise floor is higher than expected as the signal levels have been reduced.
  

• Misleading results.  What's worse is that even though the unit will seem to calibrate correctly using the Open/Short/Load, the fact that it is no longer sourcing or loading 50 ohms can cause misleading readings if you are trying to sweep a filter, antenna, cable or other device that is expecting a source impedance in the area of 50 ohms.  This may be proven by using a known-good 50 ohm cable that is significantly longer than the one with which you may have calibrated the unit and terminating it with a known good 50 ohm load:  Ideally, it should remain "flat"(low VSWR/return loss) but if the unit itself no longer sources/loads 50 ohms, this test will reveal something other than a flat response.
  

  

  Figure 3:
  The screwdriver tip is touching the "circuit" side of the
  0.047 uF coupling cap.  It is between this point and ground
  that one would test with a digital Ohmmeter to determine if
  the RF switch chip has been damaged.
  Click on the image for a larger version.
  

A quick check with an Ohmmeter will reveal the problem and figure 3 shows where a measurement may be made, on the "circuit" side of the blocking capacitor on the CH0/Port1 terminal. This photo shows the SAA-2N, but your model may have a slightly different layout, but it should be pretty easy to locate this capacitor and do a similar test on other variants.

If all is well, a typical digital Ohmmeter will likely read around 10k or higher with the unit powered down - but if you get anything lower - specially if it is in the hundreds of Ohms or lower - the MXD8641 is likely blown.  (Do NOT use the "diode test" function.  Be sure to check your Ohmmeter in both directions to rule out a reading a protection diode on the chip.)

As mentioned earlier, the MXD8641 is a very inexpensive part - but it's difficult to find from U.S. suppliers - and for this reason I sourced a strip of the MXD8641 on EvilBay with each part costing well under US$1.00, including shipping.  Replacing this part, however, is another matter as we'll see shortly.

A promising equivalent part would seem to be the Skyworks 13414-485LF (available from DigiKey) which is also a SP4T switch - in the same package and with the same pin-out, but despite what look like identical specifications, parameters and truth tables in the data sheet, the units that I got from DigiKey did not work in the 'V2 for reasons still unknown.

Figure 4:
Highlighted section shows the installed TVS - the 30 volt
0603-package part mentioned in the text.  It is laying on its
side, which was the easiest orientation for mounting.
Click on the image for a larger version.

Input protection

If one makes it standard practice to always place a resistive attenuator (say, 6-10 dB) on CH0/Port1 of their VNA - and does calibrations with it in place - this may reduce the probability of damage, but doing such may complicate certain types of measurements due to the added loss and reduction of signal levels.

Unfortunately, none of the V2 variants that I've seen have usually included overt protection in addition to that intrinsic to the MXD8641 - and this would likely be in the form of a low-capacitance TVS diode.  As it happens, there are a number of TVS diodes that have very low capacitance (0.5pF or less) that are specifically intended for protection of GHz-range devices (HDMI, USB3 and RF/antenna) which make the suitable candidates for use with a NanoVNA.

The device that I chose was the Inpaq EGA10603V12B0DG - a 12/30 volt, 0.2pF part in a 0603 SMD package (DigiKey P/N:  3526-EGA10603V12B0DGCT-ND).  There are other devices with low capacitance with even lower voltages (which would have been preferred) but these were available only in 0402 (or smaller!) packages, making their handling very difficult:  As it is, a 0603 package is about 1/4th of the size of a grain of rice!

With a fine tip and small-diameter solder, it's possible to add the TVS diode.  Figure 4 shows how this might be done on the SAA-2N:  Adjacent to the blocking capacitor, the green coating is scraped off the board to bare the ground plane and the TVS diode is installed - first, soldered to the "circuit" (and not the "connector") side of the capacitor and then the other end flowed on the bared ground.  I find it a bit easier to set the TVS on its side, but your mileage may vary.

You will note that the TVS is placed on the "circuit" side of the blocking capacitor rather than on the antenna terminals.  The reason for this is that if there is a voltage across the input (and thus across the capacitor) that is then shorted, the energy of the 0.047uF capacitor (plus any "ringing" from inductance/resonance) will be dumped into the circuitry.  By placing the TVS on the "circuit" side, it will be able to dissipate at least some of the energy dumped by the capacitor, placing an absolute limit on the voltage peak.

How much effect does this TVS have on the readings?  I've added one to a perfectly-functioning Version 2 NanoVNA and then re-checked the calibration:  Below several hundred MHz, there was no detectable effect of the device's small amount of capacitance - and even well into the 2+ GHz range, the effect was very minor.

As I rarely use my NanoVNA above about 1.5 GHz, this effect was acceptable - and the added protection against damage (that is, finding your NanoVNA non-functional at a radio site - possibly disrupting the planned activities - perhaps resulting in a wasted trip unless you have a spare VNA or equivalent!) is worth it. The 0.2pF capacitance of this device is probably less than stray capacitances in the circuit/layout, anyway.

Figure 5:
Location of the failed chip (U551) - the first active component
in the signal path on CH0/Port 1 of the VNA - and the one
most likely to be damaged.  The SAA-2N is shown above.
Click on the image for a larger version.

Note:  I haven't put similar protection on CH1/Port 2 as it's less-commonly used - and it's typically used in conjunction with a device under test that is less likely to produce ESD.  If you are measuring the loss of coaxial cable, however, it would be a good idea to dissipate any stored charge on it (shorting the center to the shield - or even touching across the two with your fingers) prior to connecting it to either terminal of the NanoVNA.

Replacing the MXD8641

If your 'VNA has already been damaged, it may be too late unless you have the equipment to do small surface-mount work.  For such work, it's not so much skill that is needed - but the right gear (hot-air rework, ceramic tweezers, etc.) and good magnification is required:  None of this is particularly expensive as a suitable hot-air rework station can be had new for well under US$100.

Using the hot air rework wand with a small-diameter blower tip (Figure 6) the chip is carefully heated and using the ceramic tweezers (which are fairly heat-insulative and non-conductive) the old, blown-up chip is lifted off - noting the orientation of the small dot on the original chip - which, on this board, corresponds with the "arrow" symbol seen at about the 7 o'clock position on U551 in Figure 5.  This image shows the location of this chip on the SAA-2N, but it will be located near the connector itself on other models - you may need to trace the signal path for other models.

Figure 6:
Heating the failed chip with a hot-air rework tool.
Be sure to note the orientation of the tiny dot on the chip
before you remove it!
Click on the image for a larger version.

Flooding the area with a drop of "no clean" solder flux from a pen, the replacement chip is dropped onto the spot where the original was - taking care to orient its dot properly - and the area is re-heated with the rework tool.  Gently nudging it back and forth, when the solder melts it will "stick" in place as surface tension does its work - but it may be necessary to make sure that it's centered on the pad - or even push down on it slightly.  If it seems to not be centering itself, let it cool a bit and apply a bit more flux and re-heat it:  The trick here is to let the flux and surface tension do the work!  It is unlikely that more solder will need to be added as the new chip was pre-tinned at the factory and there will likely be plenty still on the pads.

Letting the unit cool - and verifying that its dot is oriented in the same manner as the original part - the unit may be reassembled and tested:  I've had extremely good luck with this method and have never had a failure - and having replaced '8641's in several Version 2 NanoVNAs, I've been 100% successful - but then, I've been doing surface-mount work for quite some time.

Figure 7:
The repaired NanoVNA, recalibrated.  The flatness and level
of the noise floor (<=-70dBm) indicate that it's well matched
and that the losses are low between 100kHz and 1.5 GHz.
Click on the image for a larger version.

Conclusion

Without careful examination of your Version 2 NanoVNA it may not be possible to determine if it has the added protection of the TVS diode - although this may be noted in its documentation.  If it does not - and you regularly take it out in the field rather than having it sit comfortably on your workbench - it's worth considering adding it.

Note:  As much as I would like to help, I'm not going to get into the business of repairing/modifying NanoVNAs in the manner described, so please don't ask - sorry.  If you have a "blowed up" NanoVNA and don't have the gear/skill to do SMD rework, find someone who can.  At the very least, you now know why this issue might happen and this may prevent you from damaging other devices in the future.

 * * * * * * *

This page stolen from ka7oei.blogspot.com 

[END]

Ali Express "SDR TX/RX Switch" - A design not well thought out...

Figure 1: 
Front panel of the SDR TX/RX antenna switch showing the
the 3.5mm audio connectors and the red/green RX/TX LED -
which have been swapped to be in their proper location.
Click on the image for a larger version.

        IMPORTANT:        

If you have one of these devices, DO NOT connect it to your transceiver and second receiver (SDR) UNTIL you have read and understood the issues described here. 

Failure to understand how this device works may result in you blowing up your SDR when you transmit! 

I've you've been following this blog you'll note that I've used SDRs (Software Defined Radios) quite a bit - particularly for reception.  Transmitting in the vicinity of any receiver - or trying to use an outboard receiver in conjunction with a transmitter on the same antenna - is a bit problematic for several reasons:

• If the transmitter and receiver are in close proximity and on very nearby frequencies (e.g. on the same band) then it is (nearly) inevitable that the receiver WILL be overloaded when the transmitter is active.
  
• Unless the frequencies (transmit and receive) are very well separated AND both the receiver and transmitter have adequate filtering, the receiver will be overloaded by the transmitter.
• It is possible that even if you have a separate receive antenna, it may intercept enough energy from the transmitter to damage/destroy the receiver.  If the two are on the same band, this is more likely - but even if the receiver is being operated on a very different frequency range than the nearby transmitter and there is insufficient filtering at the receiver the receiver could sustain damage.
• It is often the case that one might have a single antenna on which there are two receivers (e.g. the receiver built into the transceiver and an outboard SDR receiver).  In this case one clearly must protect (e.g. disconnect) the outboard receiver when transmitting.

It's worth noting that most SDR receivers do NOT have particularly strong filtering in them:  Unlike an amateur transceiver - which may have separate filtering for each amateur band (or groups of bands) - this is rarely the case for wide frequency-range software-defined radios:  RTL-SDRs, SDRPlay, Funcube and others have either no band-specific filtering or rather broad (e.g. covering about an octave or even wider) filtering in them.

What this unit does

Some modern radios actually have external receive ports on them to allow you to "share" the RF while protecting the external receiver.  If your radio doesn't have that, there are/were several devices to allow this that may be found on the market (e.g. the MFJ-1708) but my attention was brought to an inexpensive unit (pictured above) that has appeared on the various seller web sites (Amazon, EvilBay, Ali Express, etc.) so I obtained one via a U.S. seller.

The description of this device is typical of those found on at the stores of Chinese sellers, curiously being both under and over-descriptive at the same time:  "160MHz 100W Portable SDR Transceivers Aluminum Alloy Box Device Radio Switch Antenna Sharer Practical Signal Equipment Accessory"

By the description, with this device it should be possible to connect your transceiver and SDR (receiver) to the same antenna, perhaps receiving using both (e.g. the addition of a waterfall to an older radio) without fear of damaging the SDR or the transceiver.

This device also has another feature:  To re-route audio when transmitting - which is probably the most usable feature of this device as it comes out of the box as we'll see.

As we'll see, this device doesn't quite work as you might think that it should.

"Documentation?  What documentation?!" 

This unit arrived in a package with (surprise!) no documentation at all - which was somewhat disappointing:  Sometimes one gets a (badly!) translated half-sheet of paper that hurts one's brain to parse - or even a URL to a page with... something... but not the case here.

From a practical standpoint, it's somewhat "self documenting" in the sense that if you ordered this device in the first place, you already had an idea as to what it was supposed to do, so it's possible to figure things out.  Referring to Figures 1 and 2 (the front and back panels) we have:

Front panel (Figure 1):

• LED on the left-hand side.  This LED is illuminated when the unit is in "Receive" mode - that is, the "SDR" rear-panel RF connector is connected to the "ANT" rear-panel connector.  (The PC board shows this as a green LED, but on mine the red and green were interchanged during assembly:  I swapped them back.)
• 3.5mm jack labeled "SDR".  This is a stereo (2-channel) audio jack and, during receive, both channels are connected to the "Audio Out" connector.  It is disconnected during transmit.
  
• 3.5mm jack labeled "AUDIO OUT".  This is a stereo (2-channel) audio jack that is intended to be connected to speakers.
• 3.5mm jack labeled "TRX".  This is a stereo (2-channel) audio jack that is intended to be connected to the transceiver during transmit.  Is is disconnected during receive.
• LED on the right-hand side.  This LED is illuminated when the unit is in "Transmit" mode - that is, the "TRX" RF rear-panel connector is connected to the "ANT" rear-panel connector.  (As noted, this should have been a red LED according to the marking on the PC board but mine was populated with a green LED, which I swapped.)
  

Figure 2:
Back panel of the SDR antenna switch.  SO-239 connectors
are for the radio (transceiver) and antenna with the SMA for
the SDR.  The 3.5mm PTT and power connectors are visible.
Click on the image for a larger version.

Rear panel (Figure 2):

• RADIO connector.  This is an SO-239 (female) UHF connector to which the transmitter/transceiver is to be connected.
• ANTENNA connector.  This is an SO-239 (female) UHF connector to which the antenna is to be connected.
•  PTT ("Push To Talk") connector.  This is a 3.5mm connector in which the center pin (tip), when grounded, will switch the unit from"Receive" to "Transmit" mode.  It's likely that a 3 or 4 wire 3.5mm cable was provided in which case only the tip (PTT) and sleeve (outer-most ring - ground) are needed.
  
• SDR connector.  This is an SMA connector to which the SDR (or other auxiliary receiver) is to be connected.
• 13.8 VDC connector.  This is a 2.1x5.5mm coaxial power connector (center positive) through which DC power is supplied.  This voltage is not critical and could be anywhere from 11.5 through 15 volts.

Also in the box my unit came with an SMA-SMA jumper, SMA-BNC adapter to adapt the SDR connector to BNC, three "audio" cables with 3-conductor 3.5mm connectors on each end, and a 12 volt switching supply (with a European "pin" plug) and a universal plug adapter:  The 12 volt switching supply seems to be the cheapest, meanest possible unit with no brand name and should NOT be trusted or used - but at least its DC cord is useful!  (In other words, do not use this power supply - particularly as it is unfiltered from an RF standpoint and it would be a really bad idea to use it on an RF receive device of any type.) 

How it actually works

As you would expect, the antenna is to be connected to the "ANTENNA" port.  When in receive mode, the "SDR" connector is also connected to the "ANTENNA" port - but the "RADIO" port is not!

What this means is that as shipped from the factory, if you connect your transceiver, antenna and SDR to the unit, when it's in receive mode, you will get no receive signals on your transceiver.  This is by design, apparently.

It is expected that the PTT connection on the back should be grounded when the transceiver is in transmit mode - and when this happens, the RADIO and ANTENNA ports will be connected to each other.  There is also an RF sensing circuit that is supposed to detect when the transmitter is producing RF, but this has its own issues as will be discussed later.

There is a jumper...

If you take the unit apart (via the four screws on the back panel) you'll see a jumper (J5 - see the schematic of Figure 3 and the photo of the board in Figure 4) and some awkwardly-worded text indicating that if you remove the jumper that you'll have "dual receive" - which means that the Radio's receiver and the SDR will be connected to the antenna at the same time.

This is technically true - but there are a number of "gotchas" here

 First, let's take a look at a reverse-engineered schematic of the unit, below:

Figure 3:
Reverse-engineered schematic diagram of the unit.
The parts designators are those shown on the silkscreen of the circuit board.
Click on the image for a larger version.

Circuit description:

DC power

The DC input power via J4 is protected with F1, a 500mA self-resetting fuse and D10, a diode for reverse polarity protection while L2, a 220uH inductor, isolates the connection at RF.  Capacitors C6 bypasses RF while C7, a 220uF electrolytic, provides smoothing/filtering - likely enough to even allow an AC power source (from a 12-ish volt transformer) to be used. 

It's worth noting that there are two "grounds" on this device:  The "antenna ground" of the rear panel and RF connectors and the "shack ground" of the DC power and audio which are isolated by L1, a 220uH inductor.  On paper, this isn't a bad idea - but on this unit there's a flaw:

The back panel being used to mount the connectors and it - and the entire case - is at "RF" ground - which would be fine as that would be the same "ground" as your radio.  The problem is that the circuit board's ground planes are not set back from the edges of the board meaning that it's possible that the green insulating coating could scrape off the board and contact the case it the mounting slot, connecting the two "grounds" together - perhaps intermittently.  (Practically speaking, most people would not be likely to ever have a problem.)

Oops.

Keying

J6 is the PTT input, activated by grounding the tip of the 3.5mm connection with the outermost sleeve being the "shack" (not antenna) ground.  Diode D6 blocks positive voltage and when the PTT is keyed, the gate of Q3, an N-channel MOSFET, goes low, de-energizing all of the relays.  When the PTT line is "un-grounded" capacitor C3 and R9 charge, preventing Q3 from re-activating the relays instantly, providing about 100msec or so of delay through the charging by R3.

Comments about the keying via the PTT port

This relay keying scheme assumes that there is either NO voltage, or that there is a POSITIVE voltage on the keying line from the radio when in receive that is greater than a few volts.  There are several caveats to this:

When powered up, the relays are energized whenever it's in "Receive" mode (PTT ungrounded or no transmit RF).  What this means is that if power is removed, it's as if it's in "transmit" mode.

• This connects the ANT to the RADIO port and the SDR port is grounded when in TX mode or powered down.
  
• Additionally, the front-panel AUDIO jack gets connected to the TRX jack. when in TX mode or powered down.
  

The keying line from the radio must go to GROUND when keyed.  If the keying line is shared with an amplifier, that amplifier CANNOT put a negative voltage on the keying line as that will hold the SDR switch in "Transmit" mode at best, damage the SDR switch in worst case - and in either case it would hold the amplifier in a "keyed" state.

If there IS a positive voltage on the keying line when "unkeyed" it must be at least 5 volts just to assure that Q3 will turn on reliably when the radio is un-keyed - and this lower voltage may affect the duration of the "un-key" delay.  If the voltage is less than 10 volts, the full delay caused by C3 and R3 may not occur.  The 100msec or so of "un-key"delay afforded by R3 and C3 is insufficient to prevent the relays from "chattering" during SSB and CW transmissions if RF sensing is used!

• If you are a CW operator, the unit will not switch back to receive mode as quickly as your radio might.  If you are a CW operator that prefers QSK (full break-in) you probably don't want to have this unit inline.
  

There is also an RF-sensing keying circuit:  Transmit RF is tapped from the RADIO RF line (from the transmitter) by C1, a 47pF capacitor and rectified by D1 and D2 and used to turn on Q1 - grounding it in the same way that grounding the PTT line does - which in turn keys the transmitter.  There is a fatal flaw in the design of this device exacerbated by RF sensing which I will discuss shortly.

Audio switching

Figure 4:
The circuit board, showing J5 in the center.
The power supply filtering is in the upper-right with the audio
relay (K3) on the left.
Click on the image for a larger version.

Relay K3 may be used to switch audio based on keying.  Let's assume that you are using a separate SDR with a computer to receive audio:  By connecting the computer speakers to the "Audio Out" connector and the computer audio output to the front-panel 3.5mm "SDR" jack the SDR audio will be muted when transmitting at which any audio from the radio connected to the front-panel 3.5mm "TRX" jack will be passed through to the speakers.

Practically speaking, you would probably never use the "TRX" jack to mute your radio's audio, but a more likely scenario is that if you are using an online WebSDR (see the WebSDR.org web site for a list) to listen on the air, you could use this to mute your speakers when you transmit to prevent your own transmitted audio from coming back with a delay and causing an echo.

RF Switching

This is where it gets a bit scary.  First, consider the configuration - from the factory - with jumper "J5" in place.

Remembering that when powered up, the relays are energized, you can see that when in "Receive" mode, the ANT port is connected directly to the SDR port - but you'll also note that the RADIO port is not connected to anything (e.g. floating).  When you transmit, the relays de-energize, connecting the RADIO port to the antenna and grounding the SDR port.  This means two things:

• There is no receive RF at the transceiver.  Most people that I know don't use their transceiver's receiver instead of the SDR, but use them at the same time - perhaps turning down the volume on the one not being used or, more likely, using their transceiver for audio and the SDR to display a waterfall.  Not having antenna RF to be able to receive anything on the transceiver is likely not what you really want to do.
  

• When using the RF sensing, the transmitter is connected to an open circuit before the relay switches.  This has several implications:
• If you don't have the PTT line wired to your radio, there will be a split second when RF first appears that the radio will see an infinite VSWR before the relays de-energize and the contacts close, connecting the radio to the antenna.  This repeated burst of infinite SWR can progressively damage a transmitter's finals, despite the SWR protection circuitry within the radio.
• When relay K1 does de-energize and connect to the antenna, it will be "hot switching" the relay contacts, which is to say that they will be carrying RF power at the instant that the antenna is connected.  This tends to burn contacts and shorten the life of the relay.
• The RF sensing circuit doesn't have adequate "hang time" to ride through word pauses and CW elements meaning that it will likely "chatter", repeatedly causing the hazards noted above.

As can be seen from the picture of the circuit board in Figure 4 there's a jumper, J5, on the board with the following somewhat confusing text:

J5 Usage
Open = Dual receive wheh [sic] RX
Short = Normal Operation

By removing J5, relay K1 is never energized meaning that it is always connected to the antenna:  This helps to mitigate the problem that - when RF sensing is used - that the transmitter is connected to "nothing" as it would be the case with J5 installed - but this also means that in receive mode, the transceiver and the SDR are connected in parallel. 

Simply paralleling two (nominally) 50 ohm devices (the transceiver operating in receive mode and the SDR) isn't a great idea - but it will generally work "OK", particularly if the SDR and the transceiver are tuned to the same frequency range.  When the transceiver is OFF or on a band  other than that to which the SDR is tuned may cause its filters to "suck out" RF and a loss of signal/sensitivity on the SDR.   

(Note:  A "properly-designed" device that shared the antenna for receive would likely include a built-in 2-way splitter which can reduce such problems, and it will also contain circuitry to protect the second receiver should excessive RF manage to find its way into it.)

The bad part here is that if you transmit - and, for some reason relay K2 doesn't de-energize instantly, as would be the case with RF sensing only - you will transmit directly into your SDR, likely destroying its front end.

Oops, again.

What this means is:

• If you remove J5, DO NOT operate the unit UNLESS you are using the PTT cable - which is to say DO NOT rely on RF sensing alone as transmit power will briefly enter the SDR's front end before the relay can switch.  The SDR is likely to be damaged due to the lack of RF power protection on that port.
  

• If your PTT cable accidentally becomes disconnected - or external keying is turned off in your radio's menu - you will transmit into the SDR and destroy its front end due to the inability of the RF sensing to act instantly and due to the lack of protection to the SDR.

The reason for this as as mentioned above:  Not only are the transceiver and SDR connected together without any protection circuitry, but also the RF sense needs to detect transmit power before it will activate - and by the time that it does, a brief burst of full transmit power may have found its way into your SDR.

A hardware bug

There is also a more subtle bug that I uncovered.  While testing the unit on the bench, I disconnected J5 - but was confused when the ANT and RADIO ports were not connected.  What was happening was that when I removed J5 - while the unit was powered up and in receive mode - enough current was flowing through LED D8 and resistor R4 to hold relay K1 closed.

Simply removing the power temporarily caused K1 to release - and there wasn't enough current to close it again, but if you are messing with the configuration, this "bug" could bite you, too!  I suppose that it's also possible that jarring the unit could cause the armature of K1 to hold in place - but I didn't try this.

The "fix" for this - if you want to bother with it - is to change resistor R4 to a 10k resistor:  This also tones down the TX LED's brightness a bit, too.

Overall comments

I get the sense that whoever designed this thing may have been copying the general idea from other, similar devices - but not really understanding what was being done, and why.  For example, the separate "grounds" implies an understanding that having them separated would be a good idea - but whoever laid out the circuit board made the "rookie mistake" of making it possible for the two "grounds" to be connected, anyway if the green coating on the board and the black anodization of the case wear through.

The description of this unit implies that it's useful up to 160 MHz.  I suspect that this is probably "true-ish", but the VSWR starts to climb when one gets to and above 6 meters (50 MHz) meaning that it's likely most useful at low power at these higher frequencies.

The cardinal - and unforgivable - sin has to do with the fact that if you want to use both your transceiver's receiver and your SDR simultaneously, you WILL want to remove J5 - but if you do - and you don't absolutely have the PTT connection working properly, you WILL blow up your SDR!

Fixing this problem is possible with the addition of some simple protection circuitry to allow the SDR to survive brief bursts of transmit power - perhaps the topic of a later post. 

AS IT IS, I WOULD NOT USE IT FOR ITS INTENDED PURPOSE, AS AN SDR ANTENNA SWITCH - at least not without significant modification.

What it IS useful for, out of the box

What it IS useful for is an audio switch to mute your computer audio when you transmit - as you might do when using a WebSDR or other remote receiver.  For this, I would:

• Connect the PTT to your radio's PTT
• Connect your computer's speakers to the "AUDIO OUT" jack
• Connect your computer audio output to the "SDR" jack

If you are hell-bent on not using the PTT cable, the RF sense may be useful, but you would also need to:

• Remove internal jumper J5
• Connect the transceiver to the "RADIO" port
• Connect the antenna to the "ANTENNA" port 
• DO NOT connect anything to the SDR port

As noted before, the time constant of C3 and R3 may not be enough and the relay may "chatter" - in which case R3 could be replaced with a higher-value resistor of, say, 330k - or with a 1 Meg potentiometer in series with a 47k resistor to allow "hang time" adjustment.  (Adjustment of R3 is preferable to increasing the value of C3 as the latter could also slow the activation time when RF is detected.)

Final comments

Other than to switch audio as described above, I wouldn't use this device as it is shipped for any other purpose without appropriate modification.  This makes this device a possible  "starting point" for another project (e.g. there are already some relays and a metal box!) - one to provide proper RF protection for the SDR and make the RF sensing more useful when using modes that have variable power levels (e.g. CW, SSB) which can cause the current design to "chatter".

* * * * * * *

This page stolen from ka7oei.blogspot.com

[END]

A short-term capacitor-based "UPS" for mini (NUC-type) PCs

The problem

Figure 1:
A "Beelink" small form-factor PC.  This unit sports a Ryzen
processor and runs from an external 19 volt supply.
Click on the image for a larger version.

Very small computers (so-called "NUCs" - a term that we'll use generically throughout) of recent manufacture are energy efficient and are increasingly used in lieu of full-size desk top PCs.  Many of these use external power supplies - often referred to as "bricks" - of the sort also used to power and charge laptops.

Even if one has a UPS (Uninterruptable Power Supply) attached to their computer - or especially in the case of a "whole house UPS" (e.g. Tesla Power Wall or equivalent) there are instances during which the transition between power grid going down and the UPS picking up the load may not be fast enough to prevent the computer from rebooting or just crashing and hanging.

For this article, we are looking at the case when the power supply for a "NUC" (small form-factor power supply) is incapable of "riding" through the aforementioned UPS transition.  In this instance at least part of the problem has to do with that unlike the power supply in a desktop computer - which are physically bigger and have a comparatively large reservoir of energy storage in the form of big filter capacitors - the small power supplies used for these small computers have a comparatively small energy reserve - and unlike a laptop, there is no onboard battery to serve as a backup.

Returning to the "whole house UPS" and, to a lesser extent its much smaller counterpart used to back up only critical gear, it may take more than 100 milliseconds for the power to resume after the grid is lost - depending on the nature of the outage:  Owning a Tesla Powerwall - and talking to others with this and similar (non-Tesla) systems - they all seem to share a common trait:  Sometimes they switch quickly enough that nothing reboots, but other times they take much longer to switch (sometimes more than 500 milliseconds) and many computers - even desktop PCs with more capacitive energy storage - fail to carry through the transition.

For this reason it might be reasonable to have a smaller (and, presumably a "fast") UPS to carry the computer through this transition although it seems a bit silly to have a UPS when one already has one for the entire house - but all it needs to do is to run for a few seconds, so even UPS batteries in poor condition will likely suffice.

In the case of a very small form-factor computer such as a NUC, we could contrive a means of providing power for just long enough for the UPS - whether desk-top or whole-house - to do its job.  It, too, needs only last long enough - perhaps a second or so.  If it's powered via a "brick" power supply this task is a bit easier and it is those devices with external power supplies that this article addresses.

Carrying through the interruption

Details about the replacement power supply The MeanWell power supply that works with this Beelink NUC  is P/N:  GST90A19-P1M  which may be found at Jameco Electronics and it is Jameco Part Number 2223486 (link) and from Digi-Key as Part Number 1866-2156-ND (link).  This unit is rated for 19 volts at 4.74 amps - much greater than the supply that is likely to have been supplied with the PC and it has the needed 5.5mm O.D./2.5mm I.D. coaxial power connector with center positive.  Other NUCs will have different power requirements and connector types and polarities so it is up to YOU to determine what might work for your computer. As noted, it has good power factor correction (PF of 0.9 or better) and produces very little to no radio frequency interference - unlike some power supplies of "unknown" brands.  As a bonus, it so-happens that this supply works perfectly with my older Asus ROG laptop as well!

In the specific case of the "NUC", these are often (but not always) powered via an external DC power supply.  In my case, I have a Beelink NUC using a Ryzen 5700 that is powered from a 19 volt supply.  In communicating with others who own this same mini PC it's clear that it's shipped with a wide variety of different power supplies from different manufacturers - some of them with ratings that seem a bit low for the expected power consumption of the computer - so I replaced it with a good-quality MeanWell unit (see sidebar) which not only has more robust ratings, but its input is power factor corrected - a very important consideration when powering it from a UPS! ^{See comment #1 at the end of this article}

In testing, neither the original or MeanWell power supply had enough reserve capacity to consistently carry it through a UPS transition - particularly if the computer was "busy" and consuming maximum power

The major reason why there is this concern is that this Beelink computer is located at the remote site of the Northern Utah WebSDR where power bumps and outages causing the load to switch to the UPS are very frequent - and occasionally, this causes the computer to "hang" (and not reboot!) requiring that the power outlet be remotely switched off - and back on.  As this is not a "public-facing" computer (it does WSPR monitoring) its outage may not be immediately noticed.  It's worth noting that the desktop-type computers have no issues with these transitions.

What to do?

Figure 2:
The Tecate SCAP PBLS-3.5/21.6 capacitor module.
This unit contains the necessary voltage equalization circuitry.
Click on the image for a larger version.

I did not want to put a separate mains-powered UPS on this computer and while I could have figured out a battery-based solution, this seemed overkill as I literally needed it to power the computer for less than one second - plus I didn't want to have batteries that would eventually "age out" and need to be replaced.  The obvious solution seemed to be the "supercapacitor" - devices with Farads of capacitance, capable of storing enough energy to power the computer for a very short period of time.

In perusing the DigiKey catalog I found at least two useful candidates:  One capacitor of 1.25 Farads with 540mΩ of internal resistance (Tecate P/N: SCAP,PBLS-1.25/21.6) and another of 3.5 Farads with 260mΩ of internal resistance (Tecate P/N: SCAP,PBLS-3.5/21.6), each rated for 21.6 volts - both suitable for use with a 19 volt supply.  These are actually capacitor modules, consisting of eight 2.7 volt capacitors of 10 and 20 Farads each, respectively, and containing simple circuitry to assure that the voltage across each of the internal capacitors was balanced.  It's worth noting that the voltage equalization circuitry itself will consume a small amount of current (perhaps as high as a few 10s of milliamps) - particularly as one approaches the maximum voltage rating and this must be considered in the design of the support circuitry.

It's important to note that these won't actually function as a UPS in the traditional sense:  These capacitors can store enough energy to power the computer for a short time - only for a few seconds at most - but this is more than enough to carry it through for the few hundred milliseconds of drop-out that might occur during a UPS transition. 

Using the supercapacitors

The problem with using a supercapacitor is that when they are discharged, they look like a dead short, meaning that you probably cannot simply tack them in parallel with a power supply:  To do so would stress the power supply - putting it into current limiting at best, possibly causing it to "trip out" and go offline, or in the worst case, damaging it - so provisions must be made to regulate the charging of the capacitor.  The diagram in Figure 3 shows the circuit surrounding the capacitor.

Figure 3:
Schematic of the supercapacitor NUC UPS.
A standard outboard power supply is used - typically the one supplied with the computer, but it could
be another unit - probably of better quality - as noted in the article.
Click on the image for a larger version.

How it works

For charging, we are using old and "newer old" tech here - R1 is a simple series resistor of 100 ohms with a power rating of 3-5 watts which will limit the current to around 200mA, tapering off gradually as the capacitor charges up.

In parallel with R1 is F2, a 100 milliamp self-resetting thermal fuse (e.g. "Polyfuse").  This device is really a thermistor and when "excess" current flows through it, it heats up and the resistance skyrockets, greatly reducing the current flow.  The way that it is used here means that when the power supply is first connected (and the capacitor is fully-discharged) there's a brief inrush of current until F2 "blows" (gets hot) at which point it takes only 15-20 milliamps to keep it in this state at which point R1 is handling most of the current.  As the capacitor charges and the voltage differential across R1 decreases, the current through the 100 ohm resistor will also drop - but F2 will also gradually cool down as the voltage across it decreases - but the current will also increase - but never more than approximately the 100 mA rating.

Figure 4:
Internals of the UPS.  The support circuit was constructed
on a small piece of prototype board (left) while the LEDs to
indicate the status are on the right.  The rear panel (far left)
has the power cable and coaxial power connector.
Click on the image for a larger version.
asfasdf

The use of F2, the 100mA fuse results in much faster charging of the capacitor than with the 100 ohm resistor alone.  In testing with a 3.5 Farad capacitor, it took about an hour for the capacitor's terminal voltage to be within a hundred or so millivolts of the power supply voltage with just R1, the 100 ohm resistor - but it took about 9 minutes with the addition of F2.  As an added bonus, when the capacitor is nearly fully charged (within a volt) the current through the 100 ohm resistor would be only about 10mA or so and the charging rate would slow to a crawl - plus the equalizing circuit within the capacitor module draws a few milliamps meaning that it will never get closer than 200-500 millivolts of the power supply voltage.

With the addition of F2 - and the fact that at this low voltage drop it will have cooled off and have a resistance of between 3 and 10 ohms - the capacitor's resting voltage will be within a few 10s of millivolts of the power supply rather quickly.  This is important as even a few hundred millivolts of extra charge on the capacitor will measurably extend the "run" time.  Attaining this sort of "full charge" could be done with a solid state circuit using FETs and op amps,  but it would be fairly complex:  This approach - with a single, inexpensive component - is nice and simple.  The use of F2 also overcomes the small current consumption of the capacitor module's equalization circuitry:  A few milliamps of current from this circuitry would drop the full-charge voltage by as much as a few hundred millivolts without F2.

A maximum charge current of 200-300mA seems reasonable as that would not put a significant amount of burden on the power supply - which must be able to power the computer and charge the capacitor at the same time.  I also considered the use of a simple transistor-type current limiter which would maintain a constant current until the capacitor got to within a volt or so of the supply voltage, but decided that it probably wasn't worth the added complexity - and I would still have required something like F2 to bring the capacitor right up to the supply voltage.

The "Charge" LED works by detecting the voltage crop across R1:  If it exceeds approximately 0.6 volts, Q1, a PNP transistor, is turned on, pulling its collector high, turning on LED2.  When this LED goes out, the capacitor will be within 0.5-0.6 volts of full charge.  The "Ready" LED (LED2) is in series with D2, a 15 volt Zener diode and it will start illuminating when the voltage across the capacitor exceeds about 17 volts for an old-tech AlGaInP LED (with a 2.1 volt threshold) or about 18 volts for a more modern GaN LED.  In a "standby" state, the "Charge" LED will have extinguished and the "Ready" LED will be on indicating the unit's readiness.  Neither of these circuits are perfect, but they give a "good enough" indication of the state of the device and let the user know that things are working.

Figure 5:
The completed UPS with the two LED indicators on
on the front panel.
Click on the image for a larger version.

An "ideal" diode - in real life

Parallel with R1 is a diode (D1) that is reverse-biased when the capacitor's voltage is lower than the supply voltage, preventing current flow other than through the resistor.  While I originally considered using an "ordinary" diode - which would have a voltage drop of about 0.6 volts for a standard silicon or around 0.4 volts for a high-current Shottky type - I decided to do something different:  Use an "ideal diode".

A voltage drop of 0.3-0.6 volts from a typical diode would represent an immediate voltage drop from the capacitor - and since the voltage on the capacitor will drop as it's discharged, the "diode drop" would represent less time that the computer could be powered by it, alone.  A hypothetical "ideal" diode would have zero voltage drop in the forward direction and block current in the reverse - and fortunately, something pretty close to that actually exists these days!

As it turns out, such a thing actually exists - and it is pretty inexpensive.  This implementation of an "ideal" diode is really a module with several components:  The specific modules that I used (which I got from Amazon - five for US$10) use the Diodes Incorporated DZDH0401DW chip along with an AGM30P05A P-channel FET along with a 100k and 1 Megohm resistor.  These "diodes" are rated for a maximum stand-off voltage of 26 volts and a steady-state current of 10 amps, but could probably handle 15 or even 20 amps for brief periods.

The way that these work is that the DCDH0401DW has a comparator that is used to detect the minute voltage drop between the "input" of the diode (the "+" side) and the "output" (the "-" side):  If the voltage on the input is higher than the output, the P-channel FET is turned on, allowing it to conduct from the input to the output.  If the voltage on the input is NOT higher than the output, the FET is turned off, preventing current from flowing from the output to the input.  The use of a P-channel FET allows the switch to be placed in the positive lead which permits the negative side of the power sources - the power supply and the super capacitor - to be connected together.  Incidentally, the FET is wired such that even if it weren't "on" at the moment that it might need to conduct, it's intrinsic diode would conduct, anyway, albeit with a 0.6 volt drop, but since the DZDH0401DW chip responds within a few microseconds at most, the FET would be very quickly turned on.

Figure 6:
The back panel of the supercap UPS.
The original power supply plugs into the jack while the
short cable needs to be just long enough to get to the
back panel of the PC.
Click on the image for a larger version.

When the FET is on, it's resistance is on the order of 5.5 milliOhms which means that if there's three amps flowing through it, less than 20 millivolts will be lost - about 1/30th of that of a standard silicon diode - and since there is so little voltage lost, there will be a similar fraction of heat being produced as well.

As you may have noticed in the schematic diagram of Figure 3, there are actually three connections to this "diode":  The anode, the cathode and ground - the ground being required because not only does the comparator/control chip need power, but the gate of the P-channel FET needs to be pulled negative with respect to its source.  The "overhead" current of the FET and comparator/control chip is only on the order of 175 microamps according to the data sheets so it's power consumption is practically negligible in our application.

The other components in the circuit include D2 - a 15 volt Zener diode along with LED1 and R2 for current limiting:  This LED will illuminate if the applied voltage exceeds about 17 volts and functions as a "Power" indicator.  Transistor Q1, a PNP, is connected across R1 via current-limiting resistor R4 and when the voltage drop across R1 exceeds about 0.6 volts, its collector will be pulled toward V+, causing LED2 to illuminate, indicating that the capacitor is charging.  When this LED goes out, this indicates that the capacitor is - at the very least - "mostly" charged.

The final component is F1 - a self-resetting thermal fuse (e.g. "polyfuse") which could have a rating of anything between 5 and 9 amps.  As the capacitor can deliver a large amount of current when shorted, this is provided as protection.  A "normal" fuse of 6-10 amps would suffice here, but I happened to have the polyfuse on hand.

Variations on a theme:  Backing up a 12 volt PC.

As noted, this unit was built using the 3.5 Farad capacitor - but it should be capable of doing its job with the lower-cost (and physically smaller) 1.25 Farad unit.

The described unit is also designed to be used with a NUC/PC that operates at 19 volts - a common voltage used by laptop computers.  Many of these small computers use 12 volts - and while one could possibly tack a small battery across the power supply, the use of a capacitor-based backup would mean that there would be no battery that would have to be checked/replaced on a routine basis.

The circuit depicted in Figure 3 - designed for 19 volts - would have to be modified slightly, as follows:

• D2, a 15 volt Zener, would be changed to a 9 volt device for a 12 volt bus.  This would better-represent the charge state of the capacitor for a 12 volt supply, causing it to illuminate once it had charged to better than about 11 volts.  The "Ready" LED would illuminate at voltages above that of the 9 volt Zener plus the LED's forward voltage.
  
• R1, a 100 ohm resistor for the 19 volt device would be changed to somewhere between 47 and 62 ohms but still a 5 watt device.
  
• The capacitor described is rated for 21.5 volts - which is probably overkill for a 12 volt power supply.  A 16 volt capacitor would be a better choice.  Additionally, a lower-voltage capacitor module will have commensurately lower internal resistance which improves efficiency - and for a 12 volt power supply where voltage droop due to Ohmic losses is arguably more important, it would be best to keep it below 400mΩ.   Possible capacitors for 12 volt use include:
• Tecate 2.0F, 13.5 volt, 360mΩ:  SCAP,PBLS-2.0/13.5
• Cornell-Dublier 2.5F, 18V, 334mΩ:  DSM255Q018W075PB 
• Tecate 5.6F, 13.5 volt, 185mΩ: SCAP,PBLS-5.6/13.5 
• It's worth mentioning that while a "12 volt" computer may operate from a supply voltage that is nominally 12 volts, it's worth checking to make sure that it's within the safe operating range of the capacitors that you choose.  For example, the Tecate capacitors listed above can operate safely only up to 13.5 volts, ruling out the use of a power supply that operates in that range - but the Cornell-Dublier capacitor with its 18 volt rating would work nicely over a slightly wider range.

Conclusion

Figure 7:
The supercap UPS, on the shelf next to the PC -
now in service at the Northern Utah WebSDR!
Click on the image for a larger version.

As can be seen from the photos, the capacitor and support circuitry was placed into a plastic enclosure:  The two LEDs were placed on the front panel and labeled while the back panel has a female coaxial power connector that matches that of the computer and power supply along with a short cord terminated with the same type of male power connector used by the PC - which happens to be the common "5.5mm x 2.5mm" type with the outside shell being negative.

To install the UPS, the PC was powered down and the device inserted into the power lead - the power supply plugging into the UPS and the short cable plugging into the PC.  After a bit less than 10 minutes, the "Ready" LED illuminated - followed soon after by the "Charge" light extinguishing - but since the charger is current-limited, the PC could be powered up immediately after installation - not needing to wait for it to fully charge.  Of course, any testing of the device to determine its ability to "ride through" an interruption should wait until the capacitor has fully-charged.

As can seen in Figure 7, the UPS was placed on the shelf next to the PC that it supports.  With the PC under a "moderate" load (about half of the maximum power consumption) the power supply was unplugged briefly to see if it would hold.  Interruptions of up to 1.5 seconds were tried with no disruptions of the PC with the capacitor being fully "recharged" to just a few 10's of millivolts of the maximum voltage in under two minutes due to the "shallow" discharge.  We chose not to try to see how long it really would hold the PC up, but with the UPS installed, we cycled the UPS several times and the PC happily rode through it - something that it would not do without.

In other words, success!

 * * * * * * *

Comment #1:

In this article, I mention that having a power factor corrected power supply is particularly important when running from a UPS.  If your UPS is running power supplies without power factor correction, it may well be that it will trip out due to overload at around half of its wattage rating:  The real clue is to closely look at your UPS's specifications and note that it has a "volt-amp" rating (which is more of a true indication of its capability) that is much lower than its "wattage".  What's worse is the "spiky" nature of the input current of a non power-factor corrected power supply which may put even more stress on a UPS - or even the circuit breaker/wiring - than the numbers would indicate. 

For example, using a "Kill-A-Watt" - a relatively inexpensive power analysis device - we measured the power factor of the originally-supplied power supply and found it to be 0.45.  What this means is that if computer were pulling 65 watts and the power supply were 85% efficient - which implies that it's needing about ( 65 / 0.85 = ) 76 "watts" - it would actually need to pull (76 / 0.45 = ) 170 volt-amps from the mains.  As the mains supply - which may be a UPS - must actually supply volt-amps, it must be capable of supplying that "170" value - which is more than 2.6 times the power that the computer is actually consuming.  The Mean Well power supply that we chose has a measured power factor of 0.94, so at an efficiency of 85%, it would be pulling only 80 volt-amps from the UPS - less than half of the load!

For more information about this, see the Wikipedia article about Power Factor (link) - and pay special attention to the section about "Non Linear Loads" which are what a typical, non power-factor corrected switching supply presents to the mains.  In these cases, the peak amperage can be several times higher than the average - and all power circuits must be able to supply these high peaks regardless of the average power, which is why a UPS, generator or even mains supply circuit must to be de-rated to accommodate devices with a poor power factor.

In other words:  If you don't use power-factor corrected power supplies on your UPS or generator, you won't be able to safely and reliably supply anywhere near its "wattage" rating - but if you do use only devices with good power factor, you will be able get much closer to its ratings without overloading it.

* * * * * * *

This page stolen from ka7oei.blogspot.com

[END]