My A600 PA project.
This project demands knowledge and skills. It is not a beginners project.
|The subjects on this page :
- My problems wit this project..
- My modifications for my A600 PCB v2.1
- The CMCs on cable connections.
In June 2020, i spotted a by Razvan Fatu designed
"600W 1.8 MHz to 70 MHz linear amplifier". All info at www.qrpblog.com
I purchased the PA kit v2.1 , and later the accompanying Low Pass Filter kit v1 (see my advise below for inserting components).
I expected that this amplifier should withstand high VSWR user faults, because these faults should not cause damage to the used "Extra Rugged" NXP FETS.
Q2 died several times, always showing a short circuit between gate and source.
Q1 or Q2 can be damaged, if the driving TRX is set to a higher
frequency band than the LPF in the PA.
For instance : Changing the TRX from 3.65 MHz to 7.1 MHz, while not changing the LPF from 80 to 40m..
BEFORE applying drive power, FIRST CHECK the F600 LPF correct frequency band setting !
Update 20210125, maximal drive power :
This balanced broadband amplifier can show 600W output on the meter, by driving it with more power than 1Wrms. But this is not advised.
A single ended NXP test circuit enables 320W max output and 80% efficiency. These results are only possible due to the use of a resonating output matching circuit.
But a balanced broadband amplifier is NOT able to deliver 640W without overdriving it severly, resulting in strong IMD and a higher chance of damaging the power FETs.
For this balanced broadband amplifier, it is NOT safe to drive to 600W output, nor possible to keep low IMD levels. It is therefore wise to limit the input power to Tr1 to 1Wrms.
Using 20dB input attenuation,
and a 100W transmitter, the
amplifier can deliver up to 400W at all SW bands, with less chance to be
A QRP transmitter delivering 1W @ 50 Ohms should drive this amplifier to about 400W output.
Power gain is 25 to 28 dB.
For best power transfer to my 60 Ohm
MFJ-264 dummyload, i used during these measurements :
the 14MHz LPF at 10MHz,
the 28MHz LPF on 18MHz, and
the 50MHz LPF at 25MHz.
My amplifier showed possibilities for improvements.
Problems in order of expected urgency.
1. A to low parallel impedance was measured over balun T5 (possibly damaging Q2).
2. When switching-"ON" the external 50Vdc supply, resistors R31/32 inject an up to +17V DC-pulse onto the FET gates circuits.
3. Q1/2 outputs were not forced to be balanced, and chokes L1 and L2 could show DC saturation.
4. The output of TR1 was floating, not forced to be balanced.
5. The amplifier bandwidth was not limited to the intended range.
After a lot of E-mail traffic between ON9BOG and me, again testing and again
damaging another Q2, the here below described modifications were developed. They resulted in (20210125)
- A more robust amplifier.
- Clean legal maximal 400W SSB output.
- Less needed drive (1Wrms).
- Efficiency : 400W SSB 55%.
TIP : By adjusting both FET gate bias voltages for each 0.9 Adc Idq, the best amplifier linearity was achieved (see NXP data).
This PA now fulfils my expectations. Performance is reliable, as long as the LPF is set to the correct band.
REM: see also modification F and G.
Modifications to prevent damages to Q1/2
in the order of importance
A. The attenuator shunt is not needed anymore.
The Maximal needed drive power after these mods is about 1Wrms.
When using a 100W TRX and a 20dB 100W input attenuator, the drive power is
limited to 1Wrms in order to protect Q1/2 and to prevent IMD (this 1W will mainly be
dissipated into the 11 Ohm gate resistors).
F. Improve output balance. Replace balun T5 by a better
higher-Z Common Mode Choke.
Improve the output balance further by replacing chokes L1/2
Then tune this
transformer to about 7.5 MHz resonance as follows :
input transformer TR1 output to be in balance :
Blocking common mode currents on cable connections.
=> Only #31 ferrite should be used. Other ferrite materials are not recommended. <=
To prevent harmonics radiations, install on all three power
cables FairRite ring cores Nr. 2631801202
with the maximal number of turns cable through its 19mm hole.
Best use the thinnest cable practical to enable more turns and better chokes. Install at :
- The 240V~ input from the mains (3.5Aac).
- The 240V~ output to the 50Vdc power supply, (3.5Aac).
- The 50Vdc coming form the power supply (16Adc)
In case of RF radiation problems, you could install on the coax cables at the back panel #31 ferrite cores :
TRX and ANT
Wind 8 turns RG300 in the 34mm hole of a big 61mm ring core 2631803802.
Use "TyWraps" to fix the coax in place. PTFE cable is slippery.
OR easier to wind and to insert between transmission line and PA :
Wind seven turns RG300 in the 19mm hole of FairRite clamps Nr. 0431173551.
"Transmit" command cable (with BNC
Use thin RG316 or RG174 coax, which enables more turns throughnto a 19mm core hole.
Wind at least 14 turns through the 19mm hole of a 29mm FairRite ring core Nr. 2631801202 core.
Cheapest an best :
TIP : Make an order at MOUSER.COM
for total more than 50 Euro.
You get rabat and free postage. This can result in huge savings.