Ext. amplifier protection methods

[Paul Loughnane]

Hi all

After some initial networking issues, I'm happy to say I have a stable system in place for some time now and have been enjoying mostly Ft8 operation. I've had a few SSB contacts, but with so many people running high power, it's been a challenge with 5W.

I've just completed a nice 150W amp with a couple of old bipolar transistors I'd lying around (SD1487) combined with a high power switchable filter board which seems to work well (so far only into a dummy load), but would prefer some kind of SWR protection before putting it on the air.

As there's no hardware ALC on Hermes, are there any other good options - I see there's a TX_ENABLE line back to the FPGA. Could I put the reverse power level through a comparator and assert that line above a certain reflected power level for a crude shut-down? How quick would that be versus how quick such a system would need to be to provide any protection?

If there's no good easy answer, I'll likely just be as careful as I can and use it anyway

Regards

Paul

AD7QO

[nekv...@gmail.com]

Thanks for asking this question, Paul. I would like to see the concept extended to account for a variety of "telemetry" that could be monitored and/or acted upon. I'm making useful progress on a PA which will require substantial metering: temp, supply voltage and current, output power, reflected power and perhaps more. AFAIK, amps like the Hardrock-50 don't send data back to HL; they monitor and act on it locally. It would be helpful to have an explainer on how to get the telemetry back to HL - probably via I2C - and how to augment either the FPGA code or the control software or both to process, display and/or act on the telemetry.

Brian K1LI

[Duncan Clark]

Hi Paul,

Maybe use the W6PQL amplifier control board linked to the high power
switch, couplers and the LED Bar Graphs?

https://www.w6pql.com/amplifier_control_board.htm

https://www.w6pql.com/high_current_solid-state_dc_switch.htm

https://www.w6pql.com/led_bar_graph_meter.htm

https://www.w6pql.com/1_8_to_54_mhz_dual_coupler.htm

https://www.w6pql.com/Sampling_RF_Power.htm

All very homebrewable from his schematic and pcb layouts or as you are
in the US way cheaper to buy direct than us here in Europe.

Homebrewing his VHF OCXO for transverter use is neat as well.

https://www.w6pql.com/vhf_ocxo.htm

Good luck.

Duncan





In message <01e7958e-2daf-40bd...@googlegroups.com>,
Paul Loughnane writes

>--
>You received this message because you are subscribed to the Google
>Groups "Hermes-Lite" group.
>To unsubscribe from this group and stop receiving emails from it, send
>an email to hermes-lite...@googlegroups.com.
>To view this discussion on the web visit
>https://groups.google.com/d/msgid/hermes-lite/01e7958e-2daf-40bd-aebe-dd246ad1a981n%40googlegroups.com
>.

--
Duncan Clark
G4ELJ

[Matthew]

I wanted to be able to switch off a homebrew power amplifier if temperature or drain current got too high, without PC software needing to control/support this. I also added control for 2 fans that could be turned on/off from the PC.

I ended up designing and building this:

https://github.com/m5evt/hl2_extpamon

The board connects to the i2c2 lines on the HL2 and PC software (that has been programmed to support it) can then read/display the telemetry (temperature and drain current). I imagine it wouldn't be too hard to setup a Node-RED program to interrogate this PCB module (via the HL2) through port 1025. You would need to study the HPSDR protocol and the HL2 Python module.

I've been using this for around 4 months now. During a storm (first night of CQWW CW) my antenna broke and SWR went high, the module tripped the power, I was alerted to this via my PC software that supports this module. I went out to fix the antenna in the middle of the night (in the middle of the storm) and carried in the contest. I had previously blown an MRF101 in a similar circumstance due to what I believe to have been excess Id, replacing a FET in the middle of a contest is a lot more time consuming and expensive.

73 Matthew M5EVT.

[Paul Loughnane]

Yes, would be nice especially for remote operation which is one of the big benefits of Hermes - in my case my equipment is in a detached garage while I operate from the comfort of home! However, the problem I see, and has been mentioned to me by another group member is that reliable protection really needs to happen locally with a tight closed loop in hardware to shut down the amp before it goes outside it's safe operating parameters. A closed loop over ethernet is likely not fast or reliable enough, so monitor remotely but act locally I think, at least for those conditions that can kill an amp.

[Paul Loughnane]

Matthew

I like the solution but would need to redesign somewhat for my amp as I'm using 12V bipolar transistors so much higher current than the FET's @ 50V. Also, is simply shutting of the voltage safe if drive continues to be applied to the device? Or better to shut off everything - Vcc, Vbias and drive? What about simply cutting the Vbias, would that save the transistors I wonder, or the amp would continue to operate just not in a linear fashion any more (showing my RF design knowledge ignorance here)? I could also easily cut off the PTT line, but is that fast enough given that the switchover relays need to then de-energize before drive is removed?

The W6PQL option mentioned by Duncan above is also something I'll look at for more ideas

Paul

AD7QO

[Steve Haynal]

Hi Paul,

You can use the "TX Inhibit" at CN8 to disable TX externally. This is pulled high to 3.3V. Just ground this line to disable TX. You should use a FET or other protection to ensure the voltages on CN8 do not exceed 3.3V or go below 0V. The TX inhibit is standard in all gateware. It effectively act immediately (a few internal clock ticks of 76.8MHz) and turns of the PA bias, the preamp, the DAC, etc.

73,

Steve

kf7o

[Matthew]

Hi Paul,

Yes, it sounds like you may need to tweak the design, but it is probably just component changes to fit the same PCB footprints. The FET I selected to do the high side switching was spec'ed for 25 A if I remember correctly. The PCB track width is quite large for the main supply voltage too. I selected a common footprint of current sense resistor so you should be able to select a suitable alternative. I wrote the code such that parameters such as current sense R value etc. were in a header file. If you were interested in pursuing this, I could look at modifying the code so that the current sense resistor value could be configured via the hermeslite.py script and saved in the microcontroller EEPROM.

Answering your question about shutoff etc. I was a little misleading in my answer. I kill the main 50 V supply to the amplifier. Within the amplifier board, the Vbias is derived from this. You are correct, it would be best to shut off the drive power from the HL2 as well. I have a spare IO line on the board that I designed to connect to CN8 on the HL2 (I think I also wrote the code to enact this too).

It sounds like your setup is similar to mine. My radio is not nearby when I operate, so monitoring and control like this is key for me.

73 Matthew M5EVT.