The 6M-160M HPSDR Transceiver Project Begins

[WA1GFZ]

Well if you are flashing overload  good chance the noise floor is well above -135 dBM.  HPSDR does need a nice preselector. My overload flashes once in a while but not when the band is quiet. Hope you don't have any BCI.

Hello This is Sgt. Andersen from the Marloboro Police Dept.....would you like to donate some SDR boards to a good cause?

[WA1GFZ]

Tom This all started when you wanted a good spectrum analyzer. Now you have one.

[K1JJ]

Got Penny transmitting on AM and ssb. Ran some two tone tests on ssb and appears to be TOO clean to be true at 100mW.  It looks like each band has to be calibrated to put out more than 10mW. I have about 400mW pep out on ssb and about 140mW pep on AM. That's plenty.  The % mod on AM is upwards of 200% the way it's set at 35% carrier on the setting.

The RX is working well now.  

The computer is still skipping with the pan-scopes on but smooth without them.

Maybe the mic input will work when the gain is all the way down, but I notice the ALC and mic meters are real close to red line, NG.

I have Penny keying PTT all by a single ground input.  If I have enough time, I'll try to get on with 100mW on AM and check into the QRP net and we can talk later on ssb too.

*** Update:  Phil/ VK6APH sent me a .bat file for the line input tonight. I sent you a copy to try too. As busy as they are designing this stuff, these HPSDR guys are really helpful.

T

[WA1GFZ]

Tom,
Remember 1/4 carrier power on AM so even if your transformer on Penny is right you will only get 100 mW. I think I'm driving erb at under 20 mW on the display. Yes you can get some real positive peaks if you want. So now after all this trouble over the past two days you get to see the TX. So tell me can Flex generate a signal that clean?
TNX for the file. I still have a real lot of extra gain with the condenser mic on penny. I could use some help setting up the EQ. I might get the JN or ZYL to listen to  me. The delay will make your head spin. I never looked at the ALC. I need to check it out. I'm running 25% on AM.
I find 3870 a good quiet test frequency in the afternoon.

[WA1GFZ]

Tom,
Batch files seem pretty easy to modify. Just have to know which one and where. When I first started playing with this stuff I had a Motorola DSP Eval module.  Pete WA1SOV gave me some code that I would modify to get different bandwidths. It really worked well but you had to reload code every time you changed bandwidths. It looked a lot like the DSP batch file Phil sent you.

[K1JJ]

Frank,

Use the "monitor"  function in the software and plug headphones into Mercury to hear yourself in real time without the delay. However, it does not work on AM. I will use a separate mod monitor for that, but will have to figure something out for the delay problem.

What do you SDR guys do on AM?  I bet Rob knows. He knows everything there is to know about using a Flex on AM.. :-)


Yes, the .bat files are EZ to work with and edit.

You can set the AM carrier level all the way to 50% and still see 170% modulation. It does not seem to mean "50%" of pep power like you'd think.  But I notice when the DRIVE slider is all the way up it produces some nasty spurs on the audio peaks. I wonder if the drive control is what changes the number of bits used, thus the desire to have it all the way up?  Where do you run your drive control?   Did you get into the transmitter section and play with the db gain for each band? It seems I needed to set it to 38db and do the calibrate function to get the power up there for both ssb and AM. Originally it was set at 10mW.   I'll fuss with it again today to get things optimized once the line level works with the .bat. 

Yes, keep an eye on the whole list of audio levels at the top right meter next to the S meter. Some are dangerously close to red line with the mic input all the way down, thus the new .bat mod attempt.

T

[WA1GFZ]

gain 38 sounds right. I need to check the manual on drive and gain
 My drive is not all the way up. Maybe I should put the HP141 on the kaz and monitor myself. SSB it is easy to watch the display and see the unwanted sideband come out of the noise floor. I need to determine how to watch the AM display to see when you get in trouble. I'm going to play with settings today to see the best settings. The power supply on the Erb final is not regulated and only does 24 volts. It makes nice positive peaks but the voltage isn't stable when you lean into it. I never noticed the ALC thing.

[Steve - K4HX]

Why can't you monitor your own spectrum with the HPSDR?

[K1JJ]

Steve,

The Merc RX samples RF directly from the Penny board at 500mW.   According to the Flex head guys, anyway,  there needs to be an attenuator put in and a sample manually tapped off the linear stage ahead. I have already placed connectors and relays for that on my rig.

Frank or Rob: (or Flex users)   How to you preset the AM TX bandwidth and ssb TX bandwidth so that it automatically adjusts for each mode? Under the setup/transmit area, I see the TX bandwidth and can adjust it for 2.8kc or +-4.5 Khz. However, it does not remember these settings and uses the previous setting whether on AM or ssb.   I hit apply and OK after selecting DX or default selections, but NG.

Now that I think about it, I really need to reread the Flex software manual...  There's other stuff I'm wondering about too... like how do you easily change freq when the panadapter is off without using the mouse thumbwheel? A simulated tuning knob on the screen might be nice too.

T

[K1JJ]

The .bat file to change the audio in from mic input to line level did not work.  So, I reduced the internal mic gain, preamp gain to minimum - and also brought the internal EQ down to -12db  and this simulates line input nicely. I now have the same input gain as the FT-1000D - perfect.

I found that the drive control for power is OK all the way up. The error was the internal two tone test puts out too much audio and causes spurs when the drive is too high. But running the rig at for drive is normally OK.

The signals on both AM and ssb look very clean with the line level audio, so think we're all set wid that w/o a .bat file.   I'm going to put the rig online with the station and play around with final adjustments. Maybe see ya on later for some tests.

I will be running 25mW AM carrier for the QRP Net, if I have the time to finish up here...   Into the Henry amp would make that about 400mW... wow!

T

[WA1GFZ]

Tom,
I checked my Penny on a power meter and getting just shy of 200 mW out on CW. with gain set at 38. I'm running drive pot at max. I think my transformer is backwards but the ALC seems to function fine. I have plenty of drive for the erb amplifier. I'm running about 125% positive peaks at 20 watts out. Just spent some time with Rob W1AEX on 3870 getting things dialed in on SSB and AM. We heard a carrier a couple times but not sure if it was you. Hope you can check in to the QRPP net. I found the TX monitors so looked at ALC. Rob thought it good to look at mic gain showing compressor action. Condenser mic seems to be working fine.

[W1AEX]

Frank or Rob: (or Flex users)   How to you preset the AM TX bandwidth and ssb TX bandwidth so that it automatically adjusts for each mode? Under the setup/transmit area, I see the TX bandwidth and can adjust it for 2.8kc or +-4.5 Khz. However, it does not remember these settings and uses the previous setting whether on AM or ssb.   I hit apply and OK after selecting DX or default selections, but NG.

Hi Tom,

Earlier today, Frank and I spotted some subtle differences in the software audio chain between the version of PowerSDR distributed by Flex and the modified version that the HPSDR guys are distributing, but I suspect the TX Profile procedure is the same.

The mode function and TX bandwidth function are separate, so it is necessary to select your desired mode, and then select the TX profile you wish to use for that session. When you have all your TX audio levels, TX EQ settings and TX bandwidth settings the way you want for AM, go to the "Setup" menu in the top left corner of PSDR and then select the Transmit tab. In the upper left you will see a "Save As" button that will allow you to keep that profile. I created four different AM TX profiles, one for 12kc, one for 10kc, one for 9kc and a last one I named AM Scratch that is just for messing around with.

Another setting worth exploring for AM is the AM Carrier Level which is also located on the Transmit tab. By reducing that below 100% (75% - 60% is a nice range to play with) it's possible to take advantage of the asymmetrical audio capabilities of your hardware and produce some impressive positive peaks.

For sideband, I created profiles from 2.7kc - 6.0kc in .1 steps so I can quickly look at what the station I am talking to is using and match right up to them.

Each TX profile you save will take a snapshot of your TX power level, TX audio level at each stage of the audio chain, TX EQ setting, and TX bandwidth setting. Each time you load one of these profiles, those settings will be faithfully retrieved. You can create as many profiles as you like, and you can delete profiles you wish to discard.

When you have things set up really well, if your version of PSDR allows it, you can then save your entire database to your computer's hard drive, allowing you to preserve every setting, including custom receive filters you design in PowerSDR. This allows you to import all your user settings into new versions of PowerSDR as you upgrade to newer versions.

Your 10mw signal was quite readable today on the AM QRP net. I believe you win the prize for lowest power! When you get your HPSDR squared away, it might be fun to play around with bandwidth and audio up on 6 meters. Having encountered your signal on that band a few times, I suspect your 10mw will be S-9 + 60dB at my location!

73,

Rob

[K1JJ]

Excellent, Rob. Thanks for the detailed explanation to getting the presets right. Yes, I saw those options, but didn't look closely yet. That's a nice feature to have - to be able to save complete session setting to call back later. Will do.

Glad you're getting zeroed in, Frank. I'll give you a listen next week.

I ran into some minor issues after the net with the rig, so wanted to focus in and not get on 3870.  It seems that when I PTT unkey, whether if the Mercury has an ant plug on the board connector or not, I get a spike that slams the S-meter to the end for an instant. What has happened is sometimes the software will lose its mind and push the AM carrier up to full output and look fuzzy on the scope, needing reboot.

OR, more common, when I unkey, the RX goes into a weird audio ramp up like a siren and then starts sweeping up and down in freq. The display then get frozen. If I turn off the computer, I can still hear a slight sound in the headphone, like the hardware is the cause.  It seems more sensitive to this when I run the dummyload and on AM. On the ant on ssb is doesn't seem to occur.

I'm thinking that maybe the unkey spike is clobbering the RX ADC input or something like that. When I power on and off and reboot the software, everything is fine again until a random PTT keying spike causes it again. Funny, but it started at the station position and never did it at the bench.

Anytime I turn off the RX I hear a short motorboating for a second and it dies.  Maybe I need some toroids or some kind of filtering.  It seems more stable when I key the MOX on the screen and manually key the ant realy. I dunno.

Other than that, the rig seems pretty close to operational. I ran some receiver tests and listened to the ssb bandwidth and it's CLEAN CLEAN!  When the TX filter is set at 3kc, it brick wall down. I ran a TX sweep with the audio gen and have never seen such a sharp filter in my life.

T

Notice I added a little fan. It definately keeps the PS and boards cooler:

[K1JJ]

After giving it some thought, I'll bet the spike I see in the RX on unkey is the inductive kick from that one relay connected to Penny's output keying. Even with the diode across the coil it's probably causing an arc.  I'll disconnect all three relays and see what happens.   The relay probably spikes the supply voltage and cause Penny to lock up in TX. That's probably the sweeping audio sound I hear in the RX and the trash I see on the scope when it happens.


Frank, do you see an S-meter kick up on unkey during AM operation?

I may need to go with those optical keying buffers.


T

[WA1GFZ]

Tom,
Too bad you didn't listen it this afternoon. Rob and I were on 3870 between 1:30 and almost 3:00. We set up my audio and shared set up ideas between Flex and HPSDR. I saved a couple set ups for AM and SSB. I don't know why SSB bandwidths track RX it is just a software decision. I don't think Merury monitors the TX signal since there is no connection between the TX and RX. I think is is some sort of digital interface to the display. The TX output signal is detected and fed back into software. I do know what to look for when in TX SSB. Your idea of looking at the ALC is a good idea in TX.
I'm not sure what your transient issue is. I'm still two switch operation but can tell you I have been out of sequence a number of times without any issues. BTW try the multiple RX mode. You can listen to two QSOs at once.
I had the QRP net on 3870 and the guys on 3.838. Both need to be on the same mode though. The range is limited by the sample rate.
We even dialed in my TX frequency offset which is quite easy.

[WA1GFZ]

Tom,
Low voltages like 3.3, 2.5 and 1.2 volts are very sensitive to power transients. This is my reason for wanting to use opto couplers as an interface.  You can cause interesting logic lockups if one supply gets out of range. I have even seen part failures when power come up wrong. Newer FPGAs are less sensitive.  Even transients fed back into the 12 volt supply can couple to the wrong place. This is my reason for having only linear power supply modules in my pandora. Everything is shielded or twisted with a ground lead. The chassis does not carry any power currents. I treat the chassis like it is an insulator so all power leads are twisted with a return back to the supply. May help to put a good size cap on 13 volts at the input of LPU.
LPU not my idea of a good power supply. You saw my pictures.

[K1JJ]

I'll check the monitor function now and see...


Update:

With the relays disconnected, I STILL get a spike on unkey that pins the S meter. What does your S-meter do when you unkey on AM???

I found that when I unkey and hear that sweeping racket in the RX, it's Penny locked up and sweeping the band. I pulled off its coax and the noise dropped way down, so it's Penny that stays keyed up and goes into a weird fuzz on the scope.  It gets ativated like that from the spike I think.

What ya think?

[WA1GFZ]

Rob,
I'm running 25% in the TX tab. I set the drive slider to maximum then go into the gain tab (which I don't think you have) and adjust the gain for maximum power in CW. There was an assembly error in a number of exciter boards when the output transformer was mounted in step down rather than step up. So I have 1.5 to 1 rather than 1 to 1.5. So now I'm at 1/2 output voltage or 1/4 power. I should get almost a watt out on 75 but actually only get 25 dBM less than 1/2 watt. I have to put 16 db of attenuation ahead of my MRI final so there is no motivation to increase the power. I was glad to see the ALC working fine today. I would parallel R14 on penny if I needed more ALC action.
I was getting under 150% modulation but had it a lot higher at one point.
I need to find out what is on the display during TX. The only connection to Mercury is relay leakage.

[WA1GFZ]

I hope you have a load on penny when you do this. The output stage might oscillate if it sees an open. NEVER STORT THE OUTPUT or good bye fianls.
Try doing it with the output set very low. Think about what is going on. Turn off TX the 13 volt cuttent drops a lot maybe setting up a transient in LPU. How about a big cap on 13 volts input to lpu? I would get rid of that thing.
The output stage is class A so if it still happens at low RF it could be a dc power issue.
I have no issues going from TX to RX

[K1JJ]

We posted at the same time . Yes, I willl add a big 13V cap there.

I checked the monitor function and it works FB on ssb. But has the delay, of course.  That IS a problem for monitoring. I guess the Flex crowd lives with it.

I just disconnected Pin 13, the Penny output that was driving the relay to gnd. I connected the relay to the input keying instead. That seems to have solved the feedback problem I had. I gots to get a linear supply and get rid of that LHU switcher crap and use twisted leads as you say.

Right now the only keying interface I have to Penny is a set of isolated contacts keying pin 1 to ground. Do I still need the optical isolation or are these contacts OK?

T

[K1JJ]

Again, what does your S meter do from transmit AM carrier to RX? Does it stay at the bottom of the meter or jump up to some value??

BTW, a 20,000 uf cap at the 13.8V input to the LHU helped somewhat. Seems more stable and there are no lock ups anymore. Tnx for the hot tip, OM.

[WA1GFZ]

Tom,
I don't have any issues going from TX to RX and never noticed the S meter take off. I'll check the next time I have it on. Your PTT input on a set of contacts should be fine as long as nothing else shares the contacts. The input to penny has an RC filter and diode protection. Sounds like you had a sequence problem.

[K1JJ]

I can key the boards using the mouse MOX on the software - with the boards disconnected from everything and the s meter still jumps up to full scale on unkey.

T

[W1AEX]

Rob,
I'm running 25% in the TX tab. I set the drive slider to maximum then go into the gain tab (which I don't think you have) and adjust the gain for maximum power in CW.

Frank:  Correct, there is no gain tab in the Flex version of PSDR. The AM carrier setting is interesting and I have run into guys running a very low percentage such as yourself. The potential is there for massive positive peaks. I found that the 60% - 75% range allowed me to easily and cleanly reach 150% positive peaks on the scope. If I go much higher, some receivers don't take kindly to it. Whatever you are doing, it sounded very clean today on AM with smooth and very well balanced audio. I believe you are using an electret, which is my choice as well.

Tom:  There is definitely latency when you monitor your audio in sideband through PSDR. The length of the delay can be minimized by lowering the TX buffer size at the sacrifice of filter sharpness. Regardless of processor speed, motherboard bus speed, interface speed, and PSDR buffer settings, there will always be latency due to all the audio processing that is taking place in real-time. I only use the PSDR monitor while setting up a new profile as I transmit into a dummy load. The audio lag is typically just enough to seem like an echo as I speak but it's still a good indicator of what the transmitted audio sounds like. There's no way I could carry on a conversation on the air with someone while listening to that. You can often tell when someone is monitoring themselves with a Flex in a QSO as they lose their train of thought and stumble while they speak due to the disconcerting latency. It takes the smugness out of their voices...

[K1JJ]

OK on the delay, Rob.   Well, I consider monitoring important to me for knowing instantly if we have a problem and also old DJ habits are tough to break.  The delay really is hard to deal with. Yes, you can hear it on the air when guys use it...

With this new lashup I still have the computer dropping out when at 50% CPU, when viewing the panadapter. I'm not in the mood to buy a new computer at the moment - big waste of $$. The two I have now are FB as-is.

Bottom line is it appears this rig will not be my main rig. It has its advantages and several disadvantages. I still don't know how it will hold up under QRO conditions. I may end up using it as a self-contained fun QRP rig and as a spectrum analyzer for testing.  I may go back to the FT-1000D, E-rig and 4X1 rigs and put this rig and computer next to my desk on the other side of the room as studio B, QRP.

I learned a lot and now have a spec analyzer, the main goal, for $550. Can't complain. Maybe in a few years when the SDR rigs are self contained, a generation or two past this technology,  with no external computer, I will take the plunge and get one as a main rig.

[fade to black]

More to come...

T