HL2 Relays, RF3 loading and PS IMD3 tests.

[Pez]

Hello all,

I have had the HL2 for a few days and I am very impressed with it. I have 3 questions that I would like to ask the group: (I am using THETIS v2.9.0.8 HL2-Beta1).

1/ RX loading on RF3 port. I am using a -60dB sampler that is built into my amplifier to drive the HL2 RF3 port. In order to reduce the RX loading of the HL2, I have added a 470 ohm resistor across the sampler output to increase the impedance, and minimise this RX loading problem. However, I still see a couple of dB RX drop when I connect this to my RF3 port. 

Can I improve this by adding even more resistance? Maybe 1K, or 2.2K or something? I don't know if this will have other undesirable effects on the HL2 RX performance (or PureSignal operation), so I wanted to ask the group for some guidance on this.  

2/ Relay problem (PTT OFF) - Sometimes it looks like a relay that is used during PureSignal ON, does not disengage properly. Sometime when I stop TX, and turn PS-A off, the next time I PTT the radio will TX a very broad a dirty signal (tens of kHz wide), until I toggle PS-A ON and OFF again. Does anyone else have this issue? Is there any work around or setting to change?

I also hear two (2) relay clicks on PTT OFF with PS-A enabled - is that normal? 

3/ IMD3 results. Here are my IMD3 results. I would like to know what the group thinks, and ask if there is any room for improvement here? Is the process fully automatic or can I manually adjust for the best results somehow?

TX IMD3 Tests:  40m 

BAREFOOT 5W   : PureSignal OFF  - 36.2 dBc

BAREFOOT 5W   : PureSignal ON   - 47.3 dBc

PureSignal Barefoot = 11.1dBc improvement

OM2000A+ 400W : PureSignal OFF  - 28.3 dBc

OM2000A+ 400W : PureSignal ON   - 45.9 dBc

PureSignal Amplifier = 17.6 dBc improvement

Thanks in advance for any comments or advice. 

73

["Christoph v. Wüllen"]

There is also some info on that on my page dl1ycf.darc.de/hl2.htm

> Am 28.04.2023 um 10:47 schrieb Pez <perrin...@gmail.com>:
>
> Hello all,
>
> I have had the HL2 for a few days and I am very impressed with it. I have 3 questions that I would like to ask the group: (I am using THETIS v2.9.0.8 HL2-Beta1).
>
> 1/ RX loading on RF3 port. I am using a -60dB sampler that is built into my amplifier to drive the HL2 RF3 port. In order to reduce the RX loading of the HL2, I have added a 470 ohm resistor across the sampler output to increase the impedance, and minimise this RX loading problem. However, I still see a couple of dB RX drop when I connect this to my RF3 port.
>

I do not know how your -60dB sampler works, but in principle you need two resistors, one just across the RF3 port and the other "serial".
Typical values will still lead to a decrease of about -1dB in RX.

> Can I improve this by adding even more resistance? Maybe 1K, or 2.2K or something? I don't know if this will have other undesirable effects on the HL2 RX performance (or PureSignal operation), so I wanted to ask the group for some guidance on this.
>
> 2/ Relay problem (PTT OFF) - Sometimes it looks like a relay that is used during PureSignal ON, does not disengage properly. Sometime when I stop TX, and turn PS-A off, the next time I PTT the radio will TX a very broad a dirty signal (tens of kHz wide), until I toggle PS-A ON and OFF again. Does anyone else have this issue? Is there any work around or setting to change?
>

This is a known problem with the PS algorithm going crazy. Has nothing to do with the hardware.

> I also hear two (2) relay clicks on PTT OFF with PS-A enabled - is that normal?
>

I guess you hear the T/R relay and in addition one of the relays on the N2ADR filter board.

> 3/ IMD3 results. Here are my IMD3 results. I would like to know what the group thinks, and ask if there is any room for improvement here? Is the process fully automatic or can I manually adjust for the best results somehow?
>
> TX IMD3 Tests: 40m
>
> BAREFOOT 5W : PureSignal OFF - 36.2 dBc
> BAREFOOT 5W : PureSignal ON - 47.3 dBc
>
> PureSignal Barefoot = 11.1dBc improvement
>
> OM2000A+ 400W : PureSignal OFF - 28.3 dBc
> OM2000A+ 400W : PureSignal ON - 45.9 dBc
>
> PureSignal Amplifier = 17.6 dBc improvement
>
>

This is nearly the same as I have measured as well. I did not get better than -48 dBc,
and the conjecture is that this (in the case of an external PA) may be crosstalk from
the T/R relay.

For a "clean" PURESIGNAL feedback path, one would need two additional relays, namely

- one disconnecting the RX front end from RF3 upon RX
- one disconnecting the RX front end from the TR relay upon TX (and
grounding the RX side of the TR relay)

Yours,

> Thanks in advance for any comments or advice.
>
> 73
>
>

> --
> 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/4df2f42d-ae05-40b4-8ab6-dd2086dcf9cdn%40googlegroups.com.

[Pez]

Thank you Christoph for the excellent information! It is greatly appreciated and has helped a lot...

1 / - Regarding my RX loading issue, I had foolishly omitted the additional Series resistor in the circuit (placed after my existing 60dB sampler) because I didn't want any more attenuation... However, I now understand that this was causing the increased load on the RF3 port. I added a 120 ohm series resistor (before the 470 ohm output impedance resistor) and this has resulted in only 1.5dB loading, and has added 7.5dB attenuation to the RF sample output (now at a total of 67.5dB using my existing sampler in-line). This seems to all be working well. PS is happy enough. 

In your design, do you see reduced RX loading with higher series resistance? For example, if I was to replace my 120ohm series resistor (-7.5dB RF), and use a 150ohm series resistor (-13.6dB RF), would the 1.5db RX loading reduce some more? I have not tested this but I am wondering if you know the answer already?

2 / Thank you for confirming. I was worried about letting the smoke out! hihi ... Hopefully improvements will come for this issue in the future. 

3 / I would agree - the crosstalk is always present, so it will limit the effectiveness of PS to some degree, however the results are still excellent overall, so I can't complain! When you say:

For a "clean" PURESIGNAL feedback path, one would need two additional relays, namely

- one disconnecting the RX front end from RF3 upon RX
- one disconnecting the RX front end from the TR relay upon TX (and
grounding the RX side of the TR relay)

Is this something that requires Software support? Or is it a hardware only modification that can work independently of the software in use? I did a quick search and I didn't see much about this... But it sounds ideal, and it sounds like an add-on module just waiting for someone (someone much smarter than me) to design and sell these! :) 

Best 73

[Ed Grafton]

Might this add the hardware you want?

jimahlstrom/HL2IOBoard (github.com)

Ed

WA4SIX

[V85NPV]

Hello Pez ,

I think you are the one who post on the Hermes lite Facebook group . Since we have our discussion i came back with your RF3 Load issue and did some test on it .

For pure signal i am using mostly the metholody provided by DL1YCF http://dl1ycf.darc.de/hl2.htm

i add a step attenuator to my RF Sampler in order to cope with different output power : Details are here : 

https://f5npv.wordpress.com/pure-signal-rf-sampler/

i am using this configuration for more than 2 years now so far and my Hermes lite is using permantly PS with THETIS 2.8.12 or PIHPSDR

To come back with you RF3 issue , today i did some test with different load (non inductive resistors to RF3 or to the RF3 jumper from inside the Hermes lite i am using to connect the RF Sampler) and did not notice any disturbance or spurious attenuation regarding the RX. i even shortcut the both RF3 and ditto , no disturbance or attenuation due to RF3 spurious load or even shortcuting RF3.

According the HERMES LITE 2 Design , the RF3 input is ''protected'' with some insulation by T2 transformer , so i am not really surprise the RF3 load is providsing minimum disturbances to the RX path.

Maybe an investigation on T2 transformer may provide some answer because at this stage i am a bit puzzled. 

According the IMD and the RF level provided by the RF SAMPLER , i did initially some test to calibrate my step atttenuator with PIHPSDR which is very convenient and according the RF SAMPLER suitable level reported by PIHPSDR i found out level a magnitude from 10 to 22 are providing me the best results. The IMD enhancement are from 15 up to 20dB which is not bad at all. 

Best 73 Didier V85NPV

[Kristofer Danner]

Ugh, who decided a facebook group was a good idea? Dividing the discussions across platforms just puts less eyes on them.

[V85NPV]

Hello, 

not me for sure ? and i have no idea who has created this group .

i just say i ''think'' Pez was the one who post on this facebook group but not sure.

73s Didier

[r webby]

The trick with the loading is to make the output impedance of the coupler much _higher_ than 50 Ohms. However, this can lead to crosstalk problems. Since most users I know cannot hear the true noise floor due to man-made QRM, you may just as well bite the db reduction, and calibrate the S-Meter accordingly. In this case, it won't materially affect what you hear.

[Pez]

Hello all,

Thanks for the replies and the information - it has all been really helpful.  I have been having a great time with PS and the HL2 with Thetis, with fantastic on air reports too (other than the 1 in 100 'PS picket fence craziness' issue). I'm averaging around -48dBc post my amplifier, with between 1 and 1.5dB RX loading, so it's very worthwhile IMO. 

Thanks for the info @radiowe...  That is interesting that the higher impedance can cause more crosstalk - I had not really considered that point, however I am happy with the "compromise" I made so far. I have a VERY low noise floor at my QTH, so the 3dB RX loading matters a lot to me. The 1 to 1.5dB RX loading I have now, and the -48dBc IMD3 is a good workable solution for me. It's great! But I'm always looking to improve from here. 

FYI @kda... , the Facebook group has 742 members so far - there is no stopping it. hihi

Another question - the suggestion to use PS Single Cal doesn't really work well for me. I see a different dBc result on each Two-Tone PTT (at a guess, +/-6dBc) - why is that? It is like the correction is still happening each PTT after a single cal? My understanding was that it Single Cal is supposed to lock the correction in place.

73

[V85NPV]

Hello Pez,

You are right  3dB RX loading matters it is a lot and far too much . You will notice there is a post regarding PS, THETIS and many more .

https://groups.google.com/g/hermes-lite/c/S39CWXGFmyE

My understanding is THETIS has different forks and as far i know developpers are still finalizing the PS features. Myself i am using OPENSRD THETIS release 2.8.12 from 2021 and so far no issue . The single CAL feature is working as it should. PIHPSDR is working really well either (for me it is less complicated to setup).

Best 73s Didier

[Pez]

Thanks Didier, I am keeping an eye on that new PS issue thread. 

I am keen to try PIHPSDR also. Is there a specific HL2 fork that I need? Or, is the version I see at the following link the one to use? https://github.com/g0orx/pihpsdr#readme

I am interested in trying PIHPSDR, and also because I was told that remote control might be in development... That would be nice. 

73

[V85NPV]

Hello Pez,

Yes, this is the correct link for PIHPSDR and it will work perfectly with the Hermes lite. You will be able to find a fork from DL1YCF Christoph with interesting features here : https://github.com/dl1ycf/pihpsdr. Both PIHPSDR are working well (I am using both with a RASPBERRY PI3B).

As i said , PIHPSDR is very usefull to calibrate and setup the PS Feature and your RF Sampler. If you are looking for remote operation , James N2ADR has developped has developped a special release for QUISK regarding remote operation (https://groups.io/g/n2adr-sdr.)

Best 73s Didier

[Pez]

Fantastic - thanks again Didier!

73