Saturday, January 19, 2019

A transmit converter (and amplifier) for 630 and 2200 meters

There is a dearth of commercial equipment "out there" designed to allow operation on the new (to U.S. Amateurs) 630 and 2200 meter bands.
Figure 1:
Complete (except for antenna and matching network) 630
and 2200 meter transmit station.  The IF radio (a Yaesu FT-817)
is tuned to the frequency for 2200 meter WSPR operation -
136.0 kHz + WSPR audio offset.
Click on the image for a larger version.

There have been some attempts to use commercial amateur transceivers to produce transmit RF at these frequencies but due to the 630 meter band being less than 1/3rd the frequency of 160 meters, the filtering and circuitry within simply isn't designed for this - and that's if you can even get around the radio's inhibition to transmit outside its designed frequency range!

Here are a few different radios and their attempts to be used at these frequencies:
  • Flex 6000 series:  Several amateurs successfully use radios in the Flex 6000 series for receive and transmit on the 630 and 2200 meter bands - but with a caveat:  A low level transmit signal on these bands is available only from the transverter port and an external power amplifier and filtering is required.  I don't know to what degree earlier Flex radios may have supported 630 and/or 2200 meter operation.
  • Icom IC-7300:  Several have reported that the IC-7300 will seem to "go" down to 630 meters, but while this radio may coaxed to "tune" down here - and the wattmeter may even show output power - analysis has shown that not only is output at this frequency loaded with harmonics, but that attempted operation at this frequency may well stress other components (e.g. things get warm!)  Receive performance is reportedly rather poor, requiring strong band-pass filtering for 630 meters and (possibly) some receive signal amplification.  I am unaware of anyone who has successfully used this radio for transmitting on the 630 or 2200 meters bands.
  • Drake TR-7/A:  The TR-7 - a solid-state all band HF transceiver from the late 1970-early 1980s - has an "LF Input" pin on a rear panel connector which allows, with some external circuitry (amplifier, filtering) reception down to almost DC.  A slight modification of the radio can permit a transmit signal to be produced on this pin (in the sub-milliwatt range) down to a few 10s of kHz with appropriate amplification and filtering being required to make this useful.  Because this radio natively uses an analog VFO, a stable, outboard digital VFO is required to obtain the stability necessary for the narrow-band digital modes often used on these band. (I own a TR-7A and have done this in the past.)
  • Icom IC-735:  Some have reported the ability to "transmit" at 630 meters, but like the IC-7300 there is very little output at the desired frequency and there is the possibility of stressing components in the attempt.  Reception requires strong filtering and some amplification.
  • Elecraft K3/K3S:  The K3S can reportedly produce low power (approx. 1 mW) at 630 meters on its transverter port.  It would appear that doing similar for 2200 meters is not possible and that most K3S owners that operate 630/2200 meters seem to use transverters, anyway.  I do not know about the receive performance on these bands.  For more information about using the K3 at 630 meters, read this app note from Elecraft.
In short:
  • Even if the radio can be made to go into transmit mode at a frequency below 500 kHz, it is likely that it is producing very little power at these frequencies and is stressing transmit components:  The radios' power amplifiers simply cannot be used as-is.  In many radios, if they allow transmitting at all, the desired 630/2200 meter signal may be among harmonics and spurious signals, requiring good filtering if it is to be at all usable.
  • Many receivers are somewhat "deaf" at these frequencies - particularly at 2200 meters.  Even if they are not, strong band-pass filtering for the band of interest is usually warranted along with appropriate amplification, particularly if there are local AM (mediumwave) broadcast stations that can overload the front end.
  • If you have a radio that can tune below 500 kHz you may find that it is badly overloaded by local AM/Mediumwave broadcast signals.  A practical 500 kHz low-pass filter is described here:  Low-Pass filter for LF/MF (2200 meter and 630 meter) reception.  This filter works well for preventing AM broadcast station overload to a receiver and it may also be used for low-pass filtering in low-power (<1 watt) transmit circuits.

Off-the-shelf 630 and 2200 meter converters:

What all of the above means is that some sort of transmit converter may be warranted.  There are a number of transmit-capable converters out there designed for operation on one or both of these bands.

Here are a number of kits or pre-built units that are available for the 630 and/or 2200 meter bands.  I have no experience with any of these devices and cannot offer advice as to how well they might work - I will leave it up to you to do that! 
It is likely that there are more than the above transverters available and I will update this list if supplied information.  Again, I have not used any of the above and can make no specific recommendation.

A practical transmit converter:

As the name implies, a transmit converter takes another frequency - such as that produced by a conventional HF transceiver - and converts it to another frequency.  In my case I use an FT-817 - a low-power (5 watt) all-mode, all-band transceiver that is a favorite for VHF, UHF and microwave enthusiasts that use transverters.  Because of its small size, feature set and already-low output power, it is a natural to be used in this application.

If you don't have an FT-817 (or FT-818) on hand that can be modified to transmit "everywhere" you may have an HF transceiver that has a transverter output port that can produce a few milliwatts.  If your transceiver doesn't have a low-power transverter output you will either need to modify the transceiver to have one or use a 100 watt dummy load in conjunction with a 20 dB tap (or a 20 dB pad capable of handling 100 watts) to drop the power to a "safe" level.

I constructed my transverter from parts that were on-hand, but these parts are readily available:  A schematic diagram of the circuit may be seen below.

Figure 2:
Diagram of the transmit converter.  This circuit uses a 10 MHz local oscillator that is divided-by-two to yield a 5 MHz IF which can yield better overall frequency stability.
The circuit in the upper-right corner is used to convert a lower-output (3.3 volt) TCXO or OCXO to TTL level - see text.
Click on the image for a larger version


Circuit description:

Local oscillator:

The local oscillator frequency chosen for this converter is 5 MHz - a frequency band available on many HF transceivers that have been "opened up" to allow operation on the 60 meter amateur frequencies.  The choice of this frequency was also influenced by the convenience being able to use a readily-available 10 MHz oscillator, which could be a 10 MHz TCXO, an "ovenized" oscillator or 10 MHz from an available in-shack reference such as a GPSDO.

The use of a "low-ish" IF frequency like 5 MHz can also enhance the stability:  With many modern transceivers, a single, internal reference sets its frequency stability and accuracy and the lower the frequency, the greater the stability.  I took advantage of the availability of an inexpensive (<$20) EvilBay TCXO for my FT-817 to give it an overall stability that is better than one part per million over a wide temperature range.

Figure 3:
The transmit converter board.  The large can is the 10 MHz OCXO, the RF input and attenuator are in the lower-right corner and the mixer/transformers are in the bottom-center.  The driver amplifier is visible in the upper-left corner.
Click on the image for a larger version.

The output of the 10 MHz oscillator is amplified/buffered if necessary and then divided-by-two by U102 - a 74HC(T)7474 - a chip that is still readily available as a DIP part.  This divide-by-two step is necessary as the mixer requires a 50% duty cycle for best balance and efficiency.

Inexpensive, stable TCXOs are readily available with 1ppm ratings or better:  One such a part is the Taiten TXETALSANF-10.000000 (Digi-Key 1664-1262-1-ND) which, at the time of writing, costs $2.92 in single quantities and has a rated stability of 0.5 ppm.  This is a tiny 3.3 volt surface-mount device, but it can be easily adapted for this circuit:  The use of a device like this - with an output that is too low to drive TTL directly - would utilize the single-transistor converter seen in the upper-right corner of Figure 2.  Even though this is a 3.3 volt device, the 1.6-1.8 volt drop through a standard (not "ultra-bright) red LED from the 5 volt supply will yield the correct operating voltage.

For an example of using a small SMD TCXO like the Taiten device mentioned above, see the 20 February, 2018 entry of this blog - Better frequency stability for the QRL Labs ProgRock synthesizer - link.  Note that this article describes the use of a 27 MHz TCXO in the same, tiny SMD package as the 10 MHz TCXO noted above.


Switching mixer:

The heart of the converter is U201, a 74HC4066 quad bilateral switch, used as a commutating switching mixer.  While the popular FST3251 (or similar) could have been used, that chip is available only in a surface-mount package while the 74HC4066 is available in DIP and works at least as well in this application - much better than an integrated solution like the NE602.


On the input and output ports of this mixer are simple transformers used to assure a balanced signal in and out and these are trifilar-wound on small ferrite toroids.  For my version I used some FT37-43B toroids because they were on-hand,  but the more-common FT37-43 or FT50-43 could have been used instead with equal results.  The exact number of turns is not particularly important, but a general rule of thumb is for such a transformer's winding to have at least three times the inductive reactance as the operating circuit at its lowest operating frequency:  More inductance is better - within reason.

Because our lowest intended frequency will be 136 kHz, we would calculate the inductance thusly, designing for an inductance that yields at least 3 times the operating impedance at the lowest frequency (e.g. 3x 50 = 150 ohms):

Because:

Z = 2*Pi*F*L

Where:
   Z = Inductive reactance in ohms
   F = Frequency in Hz
   L = Inductance in Henries
   2*Pi = approximately 6.28

To get inductance we rearrange the equation as:

L = Z/(2*Pi*F)

So, for 137kHz and an assumed "Z" of 150 ohms (3x 50 ohms input/output), L =

150 / (6.28 * 136000) = 175uH

Let is now refer to a handy online toroid caculator - toroids.info.  If we have some FT-50-43 cores on-hand we can find this particular toroid, enter the desired inductance and we'll need 20 turns to get 175uH.  After this circuit was completed it was tested and found to provide useful output down to at least 60 kHz, indicating plenty of design margin.

Although a bit difficult to tell from the schematic, the "inside" windings of T201 and T202 are really two of the trifilar windings connected in series and this is used to quadruple the impedance seen by the switch U201 and minimize losses.  Practically speaking, this is probably unnecessary in this application, but it's easy to do.

Figure 4:
A close-up view of the mixer and other support components.  Right to left:  Input attenuator and high-pass filter, input transformer, 74HC4066 mixer (with 74HC74 mixer above it), output transformer and output low-pass filter.
The close-eyed observer will note that the 100 ohm, 2 watt resistors (bottom right, blue devices) are slightly browned from having accidentally set the FT-817 to 5 watts:  No real damage was done!
Click on the image for a larger version.

In some cases builders have been known to apply a mid-voltage DC bias (2.5 volts in this case) to the center of the input/output windings on such a mixer, but that was not done here as testing showed that it didn't seem to make a measurable difference in performance as either a transmit or receive conversion mixer.  If you don't use bias, make sure that these windings can "float" with respect to DC and the local ground.

On the input side may be seen a high-pass filter that nominally blocks signals below 5 MHz.  Perhaps this is overkill, but this was included to eliminate any signals below 5 MHz that might enter the mixer - specifically any local AM broadcast stations that might have strong enough signals to ingress the cable between transceiver and the converter - not to mention the (possibly) very strong MF/LF signal from the output amplifier driven by this converter that might re-enter the signal path and produce spurious signals!

Preceding the mixer is a simple 20dB attenuator pad that is used to reduce the nominal 1 watt from an FT-817 to about 10 milliwatts.  This attenuator was designed to be able to withstand the full 5 watts from the '817 in the event full power was accidentally used.  As noted in the text, a 5-watt 62 ohms non-inductive resistor is ideal, but I didn't have one so I used the resistor combination shown in the diagram, which is more than "good enough".

Following the mixer is a low-pass filter that removes signals above approximately 500 kHz - which includes leakage from the 5 MHz local oscillator and the mixer images in the 10+ MHz area.  Included in this circuit are R206 and C205 which form a crude diplexer to terminate those image frequencies while minimally affecting the desired LF/MF signals.

Bilateral use:

By this time the reader may have noticed that J201 and J202 are labeled as both inputs and outputs.  When this circuit was first built I envisioned making it usable as both a transmit and receive mixer - and this is possible because the signal path is completely passive.  In other words, if one connected a receiver tuned to the 5 MHz area to J201 and LF/MF signals to J202, it would function as a high-performance receive converter as well, albeit with the expect 6-8dB insertion loss of a passive mixer.

The only caveat with its use as transmit-receive mixer is the presence of the 20dB attenuator - but this isn't as much of a problem as one might think:   Receive antennas at LF/MF are typically amplified and the sub-microvolt sensitivity of modern HF receivers means that, in many cases, this additional 20dB of attenuation will not put the LF/MF noise floor below the receiver's noise floor.

Practically speaking, a relay could be inserted at this point, keyed by the transmitter to put the attenuator inline, which would eliminate this loss, but I chose to omit this circuit as it would have been inconvenient to wire this to the transmitter as well - plus it is likely that I would have accidentally transmitted into the mixer when it was in "receive" mode (e.g. no attenuator) and destroyed U201!

Ultimately, I decided to use other receive gear for 630 and 2200 meter reception rather than use this mixer:  An RFSpace SDR-14 is used on 630 meters and a SoftRock Ensemble II (the LF/MF version) along with a 192 kHz sound card is used for 2200 and 1750 meter reception, each sharing a connection from an low-pass filtered, active E-field whip.

Driver amplifier:

This converter will produce a few 10s of milliwatts of linear RF at most so some "help" is needed for driving an external amplifier.  A suitable driver amplifier is depicted in the schematic below:
Figure 5:
Transmit driver amplifier.  This amplifier is linear up to about 200 milliwatts.
Click on the image for a larger version.
This amplifier is based on the venerable 2N5109, a very linear UHF RF amplifier transistor designed for CATV amplifier use and it is still available in a through-hole case for a reasonable price.  This amplifier has moderate-gain (15-20dB) and presents a reasonable 50 ohm load to the mixer and has linear output to at least 200 milliwatts when powered from a 12 volt supply, producing nearly 500 milliwatts when saturated.

As noted in the diagram, the transistor should be heat-sinked as is it is running in the linear range and is pulling a fair amount of current when idle.

The output of this amplifier is intended to be passed along to a high-power amplifier, although it can be used directly if operating QRP (e.g. low power).  On 630 and 2200 meters many operators use amplifiers that are not linear because most of the communications uses modes that transmit only single tones (e.g. CW, JT-9, WSPR) where a nonlinear amplifier will suffice:  Linearity is usually traded for the higher power efficiency of a class D or E power amplifier.

Comment:
The use of a 20dB attenuator with 1 watt of RF output from the FT-817 yields approximately 10-15 milliwatts of drive power which, in conjunction with the amplifier depicted in Figure 5, can drive the amplifier described below to (more or less) saturation.

If you are using a radio with a "transverter output" that is markedly lower than 10 milliwatts, an additional amplification stage may be required to "max out" the power amplifier.

An example power amplifier:

In the figure below, a typical single-ended FET-type power amplifier that can be operated linearly is depicted schematically:

Figure 6:
Typical single-ended power amplifier with an example low-pass filter for 630 meter operation and an autotransformer-type matching network.  Not shown in the diagram is a series 10 ohm resistor between C401 and the gate of Q401 and a series 1k/1watt resistor and 0.1uF capacitor between the drain and gate of Q401 - these having been added to improve stability.
Click on the image for a larger version.
This amplifier has been designed to operate equally well on both 630 and 2200 meters - mostly by making sure that the coupling transformer (T401) and the coupling/decoupling capacitors are chosen appropriately for operation at 137 kHz.  A low-pass filter specific for 2200 meter operation is not shown, but links to proven designs may be found below.

The drive signal is applied via J401, which is capacitively coupled to the gate of Q401, a high-power N-channel switching FET with R401 offering a ground reference a bit of RF "swamping".    T401 transforms the lower impedance of the drain (10-15 ohms) to 50 ohms and this is coupled to the output via capacitors C410 and C411.  Capacitors C406-C408 together form a low-impedance RF bypassing network to remove RF from the power supply lead.

Comment:
Suitable FET devices that may be used for Q401 include, but are not limited to:
  • Infineon IPP17N25S3-100 - This device has a rating of 250 volts and 17 amps (Mouser P/N:  726-IPP17N25S3-100).  This device is useful with a power supply voltage of up to 20 volts.
  • D3 semiconductor D3S080N65B - This device has a rating of 650 volts and a current rating of 38.3 amps  (Mouser P/N: 488-D3S080N65B-U)  This device is more appropriate when operating the amplifier on a >20 volt supply.
Either of the above devices can tolerate a 33 volt supply (and an operating current of 3.4 amps) while producing 75+ watts into a wide variety of loads - but neither of them are completely impervious to abuse.  As noted above, use of the the 650 volt FET is recommended at higher supply voltages.

An optional bias supply is shown below the main circuit, using a 78L05 5 volt regulator as a stable voltage reference which is then made adjustable via potentiometer R402.  The bias is applied to the gate of Q401 via resistor R401, a 100 ohms, 2 watt resistor, the "cold" (non-RF) end of which is RF-grounded by C402 and C403.  If the bias supply is omitted,  the "bottom" of R401 would be connected directly to ground.
Figure 7:
The power amplifier portion, built on perforated prototype board, using a Hammond 1590D enclosure as the heat sink.  The layout allows the (relatively) easy replacement of the RF output transistor (upper-right corner of the board.)  Because one is likely to blow up the occasional output transistor, one should make it easily replaceable and keep several on hand.
Click on the image for a larger version.

Transistor Q401 is a high-power N-channel FET of the sort found in mains-powered switching power supplies and as such, it should have a voltage rating of at least 200 volts (a higher voltage rating like 400 volts is better!) and a current rating of at least 15 amps.  This transistor should be well heat-sinked:  The body of the Hammond 1590D enclosure has proven adequate for continuous duty operation with the amplifier operating into a reasonably-well matched load at power levels of up to 80-100 watts DC input:  Higher input power levels than this should be used with a "proper" heat sink and/or forced-air cooling.

A power supply voltage of 12-15 volts will produce RF power output in the 15-20 watt range while a 30 volt supply limited to 3.5 amps will yield 60-90 watts of RF power.

Particularly at this higher supply/output end of this range, a higher-voltage (>=400 volt) power FET is recommended to be able to withstand mismatch conditions that could occur if the antenna system is detuned.  It is strongly recommended that a current-limited power supply be used with its threshold current set just above the maximum current pulled by the amplifier when driven to full output into the intended load:  If a poor antenna match occurs, the transistor is somewhat protected and if there is a transistor failure, the damage to other components will be minimized.

In typical "non-linear" use the bias is either set to zero volts (R402's wiper to ground or R401 grounded) or increased such that there is only a few milliamps of FET no-signal idle current:  This latter condition slightly reduces the RF drive requirement and may yield slightly higher RF output.

Amplifier stability:

It is an unfortunate fact that while inexpensive power FETs can be used as inexpensive, high-power amplifiers that they are also easy to blow up when operated at radio frequencies.  As noted in the caption of Figure 7, several components were added to improve overall stability - namely the 10 ohm resistor in series with the gate and the RF drive and a series-connected 1k resistor and 0.1uF capacitor between the drain and gate.

This amplifier is powered from a current-limited adjustable bench-top supply (a Tenma 72-6628) that can produce 34 volts at a bit more than 3.25 amps.  Having strict current limiting goes a long way toward protecting the amplifier under fault conditions (mistuned, open or shorted antenna) and has likely prevented the need to replace the transistor several times - but care is still warranted.

On the air, the amplifier has been quite reliable - never having failed while in service unless something went amiss with the antenna or match system - but if "proper" reverse power protection had been included, it's likely that I'd still be running the original transistor.

Linear operation at reduced power:

If linear operation is desired it is strongly recommended that the power supply voltage be limited to around 18 volts as the amplifier circuit can become unstable at higher voltages (30 volts) when biased into the linear range, instantly destroying the gate-source junction of the FET.  Using an 18 volt supply, approximately 25 watts PEP of RF was produced with the transistor biased at about 200mA:  On the air, the audio report during a 630 meter SSB QSO was good and the observed spectra using a waterfall display appeared to be clean.


Similar RF amplifier circuits may be found at the "472kHz.org" Useful Links web page - see the "transmitting" section, near the bottom of that page.

The low-pass filter:


Figure 6, above, also depicts a 630 meter low-pass filter that adequately removes the 2nd and higher harmonics and this filter is shown in the figure below.

Figure 8:
Low-pass filter for 630 meters using 17 AWG wire wound on PVC forms oriented to minimize cross-coupling.  Silver-mica capacitors were used but high-quality polypropylene units will work as well.  If one attempts the use of ceramic capacitors, use only C0G/NP0 types with a 500 volt rating or greater.  Green PET insulating tape can be seen under the coils to provide insulation to the ground plane.  The filter was built into the lid of the amplifier's aluminum enclosure.
Click on the image for a larger version.

Additional harmonic suppression will occur in any practical antenna matching network (e.g. series loading coil) of reasonable "Q".

Other low-pass filter designs suitable for high-power 630 and 2200 meter transmitting are described by W1VD at his web site:
  
Other entries on related topics found at this site:
Other web sites that have information on 630 and 2200 meters: 

This list is by no means comprehensive.  Peruse the "links" sections on the sites below for even more information.
  • NJD Technologies - link  - This web page has a wealth of information related to 630 meter operation, propagation and reports of activity, plus lists of known-active operators on both 630 and 2200 meters.  This web site also has many links to others that have credible information on LF and MF band topics.
  • W1TAG's web site - link  - John, W1TAG, has long been an experimenter and operator on the MF and LF bands.  This site has details on equipment both for operating and measuring performance at these frequencies.
  • W1VD's web site - link - Jay, W1VD, has long been an experimenter on the LF/MF bands and this page offers a lot of information on equipment for transmitting and receiving on these bands.
  • Antennas by N6LF - link - The callsign gives you the clue that this guy likes LF/MF operation.  This page includes detailed information on LF/MF antennas and how to characterize/improve them.
This page stolen from ka7oei.blogspot.com

[END]

No comments:

Post a Comment





PLEASE NOTE:


While I DO appreciate comments, those comments that are just vehicles to other web sites without substantial content in their own right WILL NOT be posted!

If you include a link in your comment that simply points to advertisements or a commercial web page, it WILL be rejected as SPAM!