Wednesday, August 28, 2024

Neon bar-graph VSWR/Power meter using the ИН-13 (a.k.a "IN-13") "Nixie" - Part 1

Figure 1:
Power/VSWR meter using
ИН-13 (a.k.a. "IN-13") neon bar-graph indicators.
Click on the image for a larger version.
Several years ago I bought some Soviet-era neon bar-graph displays - mainly because I thought that they looked cool, but I didn't have any ideas for a specific project.  
 
After mulling over possible uses for these things for a year or so - trying to think of something other than the usual audio VU meter or thermometer - I decided to construct a visual watt/VSWR indicator for amateur radio HF use.
 
* * *
 
I actually bought two different types of these bar-graph tubes:
  • The ИН-9 (a.k.a. "IN-9").  This tube is 5.5" (140mm) long and 0.39" (10mm) diameter.  It has two leads and the segments light up sequentially - starting from the end with the wires - as the current increases.
  • The ИН-13 (a.k.a "IN-13").  This neon bar-graph tube is about 6.3" (160mm) long and 0.39" (10mm) diameter.  Like the ИН-9 its segments light up sequentially with increasing current but it has a third lead - the "auxiliary cathode" - that is tied to the negative supply lead via a 220k resistor that provides a "sustain" current to make it work more reliably at lower currents.
Note:  It would be improper to refer to these as "Nixies" as that term refers to a specific type of numeric display - which these are not.  Despite this, the term is often applied - likely for "marketing" purposes to get more hits on search engines.

Figure 2:
A pair of ИН-13 neon indicator tubes.  These tubes are
slightly longer than than the
ИН-9 tubes and have three leads
Click on the image for a larger version.
For a device that is intended to indicate specific measurements, it's important that it is consistent, and for these neon indicators, that means that we want the bar graph to "deflect" the same amount anytime the same amount of current is applied to it.  In perusing the specifications of both the 
ИН-9 and  ИН-13 it appeared that the  ИН-13 would be more suitable for our purposes.

This project would require two tubes:
  • Forward power indicator.  This would always indicate the forward RF power as that was that's something that is useful to know at any time during transmitting.
  • Reverse power/VSWR.  This second tube would switchable between reverse power, using the same scale as the forward power display, and VSWR - a measurement of the ratio between forward and reverse power and a useful indicator of the state of the match to the antenna/feedline.
Driving the tubes
  
"Because physics", gas discharge tubes require quite a bit of voltage to "strike" (e.g. light up) and these particular tubes need for their operation about 140 volts - a "modestly high" voltage at low current - only a few milliamps (less than 5) per tube, peak.

Figure 3:
Test circuit to determine the suitability of various inductors and transistors
and to determine reasonable drive frequencies.  Diode "D" is a high-speed,
high-voltage diode, "R" can be two 10k 1 watt resistors in parallel and
"Q" is a power FET with suitably high voltage ratings (>=200 Volts)
and a gate turn-on threshold in the 2-3 volt range so that it is suitable
to be driven by 5 volt logic.  V+ is from a DC power supply that is
variable from at least 5 volts to 10 volts.  The square wave drive, from a
function generator, was set to output a 0-5 volt waveform to
make certain that the chosen FET could be properly driven by a 5 volt
logic-level signal from the PIC as evidenced by it not getting perceptibly
warm during operation.
Generating high voltage from a low is one of the aspects that I tackled in a previous project on this blog when I built a high voltage power supply for the Zenith Transoceanic:  You can read about that here - A microcontroller-based A/B Battery replacement for the Zenith TransOceanic H-500 radio, with filament regulation - link.
 
The method used for this project and the aforementioned Zenith radio is  boost-type converter as depicted in Figure 3.  The switching frequency must be pretty high -  typically in the 5-30 kHz range if one wishes to keep the inductance and physical size of that inductor reasonably small.

As in the case of the Zenith Transoceanic project, I used the PWM output of the microcontroller - a PIC - to drive the voltage converter with a frequency in the range of 20-50 kHz.  For our needs - generating about 140 volts at, say, 15 milliamps maximum, I knew (from experience) that a 220uH choke would be appropriate.  Figure 4, below, shows the as-built boost circuit.
Figure 4:
The voltage boost converter section showing the transistor/inductor, rectification/filtering and
voltage divider circuitry.

Description:
 
Q301 is a high-voltage (>=200 volt) N-channel MOSFET - this one being pulled from a junked PC power supply (the particular device isn't critical) which is driven by a square wave on the "HV_PWM" line from the microcontroller:  R301, the 10k resistor, keeps the transistor in the "off" state when the controller isn't actively driving it (e.g. start-up).  L301, a 220uH inductor, provides the conversion:  When Q301 is on, the bottom end is shorted to ground causing a magnetic field to build up and when Q301 is turned off, this field collapses, dumping the resulting voltage through D301, which is a "fast" high voltage diode designed for switching supplies - a 1N4000 series diode would not be a good choice in this application as it's quite "slow".
 
R304, a 33k resistor, is used to provide a minimum load of the power supply, pulling about 4.25 mA at 140 volts:  This "ballast" improves the ability of the supply to be regulated as the difference between "no load" (the neon bar-graphs energized, but with no "deflection") and full load (all segments of the tubes illuminated) is less than 4:1.  The resistive divider of R302 and R303 is used to provide a sample of the output voltage to the microcontroller, yielding about 2.93 volts when the output is at 140 volts.  The reader will, by now, likely have realized that I could have used R304 as part of the voltage divider - but since the value of this resistor was determined during testing, I didn't bother removing R302/R303 when I was done:  Anyway, resistors are cheap!
 
Setting the current:
 
Having the 140 volt supply is only the first part of the challenge:  As these tubes use current to set the "deflection" (e.g. number of segments) we need to be able to precisely set this parameter - independent of the voltage - to indicate a value with any reasonable accuracy.  For this we'll use a "current sink".
 
Figure 5:
The precision current sinks that drive the neon tubes precisely based on PWM-derived voltage.
Click on the image for a larger version.
 
Figure 5, above, shows the driving circuits for the two tubes using the "precision current sink".  Taking the top diagram as our example, we see that the inverting input of the op-amp (U401c) is connected to the junction of the emitter of Q401 and resistor R406.  As is the wont of an op amp, the output will be driven high or low as needed to try to make the voltage (from the microcontroller) at pin 10 match that of pin 9 - in this case, based on feedback from the sense resistor, R406.

What this means is that as the transistor (Q401) is turned on, current will flow from the tube, through it and into R406 meaning that the voltage across R406 is proportional to the voltage on pin 10.  It should be noted that current through R406 will include the current into the base - but this can be ignored as it will be only a tiny fraction (a few percent at most) of the total current.  It's worth noting that this circuit is insensitive to the voltage - at least as long as such current can be sunk - making it ideal for driving a device like the ИН-13 (or ИН-9) in which its intended operation is dependent on the current rather than the operating voltage.

At this point it's worth noting that the driving voltages from the microcontroller ("FWD_PWM" and "REV_PWM") are not plain DC voltages, but rather from the 10 bit PWM outputs of the microcontroller.  The use of a 10k resistor and 100nF (0.1uF) capacitors (R405 and C406, respectively) "smooth" the square-ish wave PWM into DC.
 
Q401 and Q402 were, again, random transistors that I found in scrapped power supplies, but since there's at least 70 volts drop across the tube, about any NPN transistor rated to withstand at least 80 volts should suffice.  It's also worth noting the presence of R407, which provides the "sustain" current on the "auxiliary" cathode.
Figure 6:
An exterior view of the tandem coupler module.
Visible is the top shield and the three feedthrough
capacitors used to pass voltage and block RF.
Click on the image for a larger version.

RF sensing

For sensing forward and reflected power I decided to use an external "sensing head" that was connected inline with the radio, on the "tuner" side of the feedline.  

For sensing power in both directions I chose the so-called "Tandem" coupler which consists of a through-line sampler in which a short length of coaxial cable carrying the transmit power (T1 in the diagram of Figure 7) passes through a toroidal core - using some of the original cable's braid grounded at just one end as a Faraday shield.  An identical transformer (T2) is connected across the first (T1) for symmetry.

When carefully constructed this arrangement has quite good intrinsic directivity and a wide frequency range.  Figure 6 shows the diagram of this section.

Figure 7:
Schematic diagram of the "Tadem" coupler.  A bidirectional coupler sends power to
separate AD8307 logarithmic amplifiers - one for forward and the other for reverse.
The outputs, expressed in "volts/dB" are sent to the microcontroller.
Click on the image for a larger version.

The RF sensing outputs of the second tandem coupler (T2) then goes through resistive voltage dividers (R606/R607 for the reverse sample and R603/604 for the forward sample) to a pair of Analog Devices AD8307 logarithmic amplifiers - one for forward power and the other for reverse - to provide a DC voltage that is logarithmically proportional to the detected RF power.  This voltage is then coupled through series resistors (for both RF and DC protection) R605/R608 and to the outside world using feedthrough capacitors.

The use of a logarithmic amplifier precludes the need to have range switching on power meter as RF energy from well below a watt to well over 2000 watts can be represented with only a few volts swing.  Looking carefully at Figure 6 one can see a label that notes that the response of the AD8307 is about 25 millivolts per dB - and this applies across the entire power range of a few hundred milliwatts to 2000 watts.

All of this circuitry is mounted in a box constructed of circuit board material and connected to the display unit with an umbilical cable that conveys power and ground along with the voltages that indicates forward and reflected power.

Figure 8:
An inside view of the Tandem Match (sense unit) showing
the coupling lines, internal shielding and AD8307 boards.
Click on the image for a larger version.
Figure 8 shows the as-built "sense unit" and the two coaxial sense lines are clearly visible.  As can be seen, the "main line" coupler is physically separated and shielded from the secondary sense line, using PTFE ("Teflon") feedthrough lines to pass the signals.

The AD8307 detectors themselves can be seen at the left and right edges of the lower half of the unit, built on small pieces of perfboard.  All signals - including the 12 volt power and the DC voltages of the output pass through 4000pF feedthrough capacitors to prevent both ingress and egress of RF energy which could find its way into the '8307 detectors and skew readings.

* * * * *

In a future posting (Part 2) we'll talk about the final design and integration of this project.


This page stolen from ka7oei.blogspot.com

[END]