Hermes Homebrew Fault Finding

Post your awesome projects here for everyone to marvel at!
User avatar
Prem
Posts: 25
Joined: Tue Apr 11, 2017 11:57 am
Location: UAE
Contact:

Hermes Homebrew Fault Finding

Postby Prem » Mon Oct 09, 2017 4:47 am

I homebrewed Hermes and tested Rx and Tx part. Rx part found ok. But my Tx gives some issue as highlighted below:

• Audio monitored thru head phone plugged to Hermes jack, with Tune button pressed and with SSB thru mic plugged also clean and sound good.
• Tone monitored in nearby Rx is clean with Tune button pressed
• With MOX pressed, even without Mic plugged in S meter show RF FWD power which vary with drive slider (even upto 200 plus mW) . This comes when MIC is enabled. If MIC disabled this signal not there !
• The same case when I use VAC and my external mic signal thru audio interface.
• With Mic plugged and SSB signal monitored in another receiver is distorted.
• With advice from some other Hams in our yahoo group, I have ordered a replacement for Audio Codec IC.
• Please advice me will this IC replacement will be solution or else you can advice some further simple fault findings.

73s

Prem, A65BK
User avatar
w-u-2-o
Posts: 518
Joined: Fri Mar 10, 2017 1:47 pm

Re: Hermes Homebrew Fault Finding

Postby w-u-2-o » Mon Oct 09, 2017 11:37 am

Prem,

I do not believe replacing the CODEC will fix your problem. The reason I do not believe replacing the CODEC will work is because you wrote:

With MOX pressed, even without Mic plugged in S meter show RF FWD power which vary with drive slider (even upto 200 plus mW) ...The same case when I use VAC and my external mic signal thru audio interface.

The VAC audio path does not use the CODEC. Therefore, if the CODEC was the problem, it would not happen using the VAC audio path.

The audio path with the front panel microphone is as follows:

MIC > CODEC > FPGA > Ethernet > PC > PowerSDR Software > PC > Ethernet > FPGA > DAC

The audio path with VAC is as follows:

MIC > PC > PowerSDR Software > PC > Ethernet > FPGA > DAC

The audio path for TUN and Two-Tone is as follows:

PowerSDR Software > PC > Ethernet > FPGA > DAC

If your TUN signal is clean, and if your Two-Tone signal is clean, but your MIC audio is dirty using the VAC path where there is no CODEC IC involved, this is a classic sign of RFI getting into your microphone and/or PC audio interface. Or it could be a bad microphone element or bad microphone connector/cable.

73,

Scott
User avatar
Prem
Posts: 25
Joined: Tue Apr 11, 2017 11:57 am
Location: UAE
Contact:

Re: Hermes Homebrew Fault Finding

Postby Prem » Mon Oct 09, 2017 4:42 pm

Hi Scott,

Thank you very much for detailed explanation of various audio paths.
Please note that my monitored audio thru headphone with "MON" enables is always clean. So we can expect MIC and associated cabling is ok.
Also I had tried the Hermes RF out direct without any PA connected. So this mW RF power can contribute to RFI ?

As you advised, even I was not convinced that the Codec IC can resolve my issue.
Is the audio filter in Mic line so crucial ? I had used the smaller available equivalent part. If the filter is not functioning , possible the reason for our fault ?
But in VAC, this filter is not there in picture !!
With the current test result, can we presume my FPGA part is ok ?

Please note that when either MIC button or VAC button is not enabled, this faulty FWD power not seen.
Also note that at this time if I look into MIC S meter reading not there !!

Can you advise some trouble shoot guidance ?

73s

Prem, A65BK
Last edited by Prem on Mon Oct 09, 2017 5:19 pm, edited 1 time in total.
User avatar
w-u-2-o
Posts: 518
Joined: Fri Mar 10, 2017 1:47 pm

Re: Hermes Homebrew Fault Finding

Postby w-u-2-o » Mon Oct 09, 2017 5:18 pm

The path for MON using all radio connections for MIC and speaker is

MIC > CODEC > FPGA > Ethernet > PC > PowerSDR Software > PC > Ethernet > FPGA > CODEC > Speaker

If this path sounds 100% OK, then the problem could be anywhere on the RF path from the FPGA on out, i.e.:

...FPGA > DAC > AMP

It could, theoretically, be a bad FPGA (unlikely), a corrupt firmware load (possible), a bad connection between the FPGA and DAC, or something to do with the DAC or downstream of the DAC.

However, if the MON output as noted above seems 100% OK with voice, TUN and Two-Tone signals, do TUN and Two-Tone signals truly look OK at the RF output? Have you put an oscilloscope on the RF output?
User avatar
w-u-2-o
Posts: 518
Joined: Fri Mar 10, 2017 1:47 pm

Re: Hermes Homebrew Fault Finding

Postby w-u-2-o » Mon Oct 09, 2017 5:27 pm

P.S. moving this thread to the homebrew forum, since it is not a digital mode interfacing topic.
User avatar
Prem
Posts: 25
Joined: Tue Apr 11, 2017 11:57 am
Location: UAE
Contact:

Re: Hermes Homebrew Fault Finding

Postby Prem » Mon Oct 09, 2017 5:30 pm

Hi Scott,

I have never connected an oscilloscope at RF out. I am not that much a skilled guy to do all these. Moreover I do have only a PC based Hantek 20 MHz rated toy unit !
I had done hand soldering and so the possibility of any bad connection cant be ruled out !
I had redone the firmware upload after seeing this issue. But no luck.
Keeping in mind that my MON signal is clean, can you suggest the signal path ( IC pin outs) to be rechecked for any bad connection ?

73s

Prem, A65BK.
User avatar
w-u-2-o
Posts: 518
Joined: Fri Mar 10, 2017 1:47 pm

Re: Hermes Homebrew Fault Finding

Postby w-u-2-o » Mon Oct 09, 2017 6:32 pm

A 20MHz Hantek is plenty fine for looking at your RF output signal. Set PowerSDR to the "Extended" region and you can transmit on any frequency you wish. Set your VFO to 1MHz, use TUN to look at the output. You might also look at Two-Tone as well.

73,

Scott
User avatar
Prem
Posts: 25
Joined: Tue Apr 11, 2017 11:57 am
Location: UAE
Contact:

Re: Hermes Homebrew Fault Finding

Postby Prem » Tue Oct 10, 2017 2:49 am

Hi Scott,

Please note that the faulty FWD S meter reading disappear when I disable MIC button .in PSDR.

I have sent three short video screen capture for my scope screen. Please review and advise me further.


73s

Prem, A65BK
User avatar
w-u-2-o
Posts: 518
Joined: Fri Mar 10, 2017 1:47 pm

Re: Hermes Homebrew Fault Finding

Postby w-u-2-o » Tue Oct 10, 2017 1:32 pm

Prem,

It is most unfortunate that you did not post the videos here for all to appreciate. You could have used Youtube to publish them.

First, and most important, a mic level of -70dB is completely normal for the output of the CODEC with no microphone plugged in. This degrades a few dB when the 20dB boost setting is active. Since you already reported that MON sounds fine coming out of the radio speaker/headphone jack, this is just more evidence that rules out the CODEC as the source of any problems you may be experiencing. The noise level is fine, and since MON sounds fine, that means both sides of the CODEC are working properly.

It is also unfortunate that you did not embrace my guidance to put the radio region setting to "Extended" and use a 1MHz output. Given the low bandwidth of your scope, it is impossible judge the quality of the sine wave (we hope it's a sine wave) coming out of the Hermes using TUN at 14MHz.

I'd recommend you repeat the experiment at 1MHz. Assuming you have at least a half watt rated dummy load (resistor), adjust your output drive level to 250mW, that should represent approx. 5V peak on the oscilloscope (with the scope input in a high impedance setting, not 50 Ohms). Then adjust your timebase to 100nS per division. This should then show 1 entire RF cycle on the oscilloscope display. With those settings you can judge whether or not the RF output looks reasonably proper in the time domain (is a well-formed sine wave).

If you have an appropriate attenuator or coupler, you can also feed the TX output back into the RX input and, with DUP mode turned on, use the radio itself as a spectrum analyzer to look at the spectral purity of your own TUN transmission. At 250mW, or +24dBm, 60dB of attenuation would be sufficient to knock that signal down to around -36dBm, which is S9+37. A large signal, but well within the capability of the radio.

If the TUN signal looks good in the time and frequency domains then obviously we have to look elsewhere than the DAC or amp for the problem.

An audio recording of the "bad audio" might also be useful as a diagnostic aid.

73,

Scott
User avatar
w-u-2-o
Posts: 518
Joined: Fri Mar 10, 2017 1:47 pm

Re: Hermes Homebrew Fault Finding

Postby w-u-2-o » Tue Oct 10, 2017 1:36 pm

One other thought that might be even easier: are your PA gain drive calibration settings correct? With a -70dB audio noise level (normal) coming out of the CODEC, power out of a Hermes, which is rated for 500mW, and assuming no other processing turned on (no COMP, etc.), the output RF power level should be close to zero. If your PA gain is set way too high, this could possibly cause the effect you are experiencing.

73,

Scott
User avatar
Prem
Posts: 25
Joined: Tue Apr 11, 2017 11:57 am
Location: UAE
Contact:

Re: Hermes Homebrew Fault Finding

Postby Prem » Tue Oct 10, 2017 2:54 pm

Hi Scott,

I have repeated the test on 1 Mhz and posting three pics here. This time I did not do any video. I am really disappointed to see my waveform. I don't know If I ado anything wrong way !!

Regarding PA level setting , I will add one pic.

73s

Prem, A65BK
Attachments
Two tone .JPG
Two tone .JPG (838.24 KiB) Viewed 253 times
Tune .JPG
Tune .JPG (96.15 KiB) Viewed 253 times
Mox without Mic plugged .JPG
Mox without Mic plugged .JPG (862.58 KiB) Viewed 253 times
User avatar
Prem
Posts: 25
Joined: Tue Apr 11, 2017 11:57 am
Location: UAE
Contact:

Re: Hermes Homebrew Fault Finding

Postby Prem » Tue Oct 10, 2017 2:58 pm

I am posting Pic to show my PA gain settings .

Please note that even though the Mic S meter reading is ok without Mic, I can see FWD Power S meter reading which vary from 0 to say 250mW , when I move MIC gain or Drive slider !!

I have uploaded a video in youtube to show the faulty FWD power seen even without Mic Plugged in !!!

https://youtu.be/HIMPnosnV58

73s

Prem, A65BK
Attachments
20 m band.JPG
20 m band.JPG (88.45 KiB) Viewed 244 times
Wattmeter].JPG
Wattmeter].JPG (114.16 KiB) Viewed 253 times
PA setting .JPG
PA setting .JPG (119.29 KiB) Viewed 253 times
Last edited by Prem on Tue Oct 10, 2017 4:13 pm, edited 1 time in total.
User avatar
w-u-2-o
Posts: 518
Joined: Fri Mar 10, 2017 1:47 pm

Re: Hermes Homebrew Fault Finding

Postby w-u-2-o » Tue Oct 10, 2017 4:12 pm

Well, Prem, as they said in Apollo 13, "Houston, we have a problem!" I pretty much thought this would be the case, and look how easy and handy that scope was to find it to be true.

We don't care about two-tone or about MOX or CODECs or any of that stuff anymore until you can find the problem that is causing that non-sine wave output when you invoke TUN.

The next debug step is to take a look with the scope right at the analog output of the DAC, or as close as you can get to it. The easiest place is probably at the input to the RLP-40 low pass filter, Pin 2 of FL1 if I'm using the correct schematic. This is on the output side of the DAC output transformer and will be easier to probe then the differential outputs of the DAC. Again, use TUN at 250mW. If the waveform still looks the same at the input to the first LPF, then you've got it narrowed down to the DAC or transformer. If the waveform doesn't look the same at that point in the circuit (same shape, not same amplitude) post what it does look like.

However, assuming it still looks the same, the next debug step after that is to reduce your drive level down to just a few milliwatts. If the shape you see there persists with TUN, and I think it will, then I'm guessing you somehow soldered it together in a way that's got a couple of DAC data bits swapped or shorted, either at the FPGA pins or at the DAC pins, or both. If the waveform doesn't persist then post what it does look like when you drop the drive level.

If you continue to get the results I think you are going to, you will likely need to do a close inspection of your soldering of the DAC data lines at both the FPGA and DAC pins under high magnification, and perhaps the DAC output transformer as well.

73,

Scott
User avatar
Prem
Posts: 25
Joined: Tue Apr 11, 2017 11:57 am
Location: UAE
Contact:

Re: Hermes Homebrew Fault Finding

Postby Prem » Tue Oct 10, 2017 4:19 pm

Hi Scott,

Please have a look at the waveform when I go to 20 m band and higher. It is almost sine . At hands below 40m this ugly waveform comes.
Please consider this pic also.

Also please tell me the capacitor value , grounded from both "+" input points of Am in IC - U7.
In my sketch it is C129 & C228 with value 47pF. Is this value critical. I think I have used 33pF instead of 47pF !!! Any serious issue ? I have marked as some confusion on this value in my sketch while assembling. Dont remember what exactly happened that time !
Possibly the MTO/parts list which I was following had different value for these Cap !! ??

Now during a detailed check in RF stage, I noticed the C33 (100nF) was wrongly fixed with 22ohm resistor. I replaced it. But no change in results !!


73s,

Prem
Attachments
40 m Band .JPG
40 m Band .JPG (836.85 KiB) Viewed 243 times
20m band 2.JPG
20m band 2.JPG (89.49 KiB) Viewed 243 times
20 m band.JPG
20 m band.JPG (88.45 KiB) Viewed 243 times
User avatar
w-u-2-o
Posts: 518
Joined: Fri Mar 10, 2017 1:47 pm

Re: Hermes Homebrew Fault Finding

Postby w-u-2-o » Tue Oct 10, 2017 10:07 pm

Prem,

I don't know which revision of the Hermes board you have, nor which version of the Hermes schematic and BOM go with it. My only reference is here. I recommend you follow the documentation you have precisely.

The scope photos you have posted are not at all relevant. Your 20MHz scope will not give a good representation of such high freq's. At any rate, all of them show essentially the same problem.

I suggested some debugging steps in my previous post. If you could work with those it would be most helpful. If we see the same incorrect waveform at the output of the DAC transformer then we know that whatever mistakes you may have made in the RF section are not the root cause of the main problem.

73,

Scott
User avatar
Prem
Posts: 25
Joined: Tue Apr 11, 2017 11:57 am
Location: UAE
Contact:

Re: Hermes Homebrew Fault Finding

Postby Prem » Wed Oct 11, 2017 6:53 am

Hi Scott,

I was using V18 schematic got from another Ham friend.
Can you advise me if the cap C37 and C77 seen in schematic V15 (as you shared) is critical ? I think I had used 33pF instead of seen value in dgm as 47 pF !! There is no 47pF listed in BOM. There are some errors in BOM.
Let me do the suggested fault finding steps this evening and update you.

73s

Prem, A65BK
User avatar
w-u-2-o
Posts: 518
Joined: Fri Mar 10, 2017 1:47 pm

Re: Hermes Homebrew Fault Finding

Postby w-u-2-o » Wed Oct 11, 2017 10:17 am

I honestly don't know if those capacitor values are critical, but I would think not, and they are not the source of your main problem regardless.

Please do the debugging steps at 1MHz. The scope settings you used previously at 1MHz worked very well. You will have to change the volts/division setting as required.
User avatar
Prem
Posts: 25
Joined: Tue Apr 11, 2017 11:57 am
Location: UAE
Contact:

Re: Hermes Homebrew Fault Finding

Postby Prem » Wed Oct 11, 2017 5:21 pm

Hi Scott,

Please have a look at test result @ 1 Mhz Tune, measured at R7 terminal , which is signal out from DAC.
So it is not at all Sine, almost same signal which we saw at final Ant out point !

I have physically inspected for any bridge between DAC out pins. Could not see anything strange. Can I get the DAC IC Pin voltages to verify ?
The signal from FPGA (14 nos) can measure some square wave on scope ?
Please advise further please.

73s
Prem, A65BK.
Attachments
Signal at DAC Out @R7 @1Mhz.JPG
Signal at DAC Out @R7 @1Mhz.JPG (809.71 KiB) Viewed 213 times
User avatar
w-u-2-o
Posts: 518
Joined: Fri Mar 10, 2017 1:47 pm

Re: Hermes Homebrew Fault Finding

Postby w-u-2-o » Wed Oct 11, 2017 9:24 pm

Prem,

All of the voltages on the DAC pins are straightforward and available from the schematic. They are either 3.3 or GND, "SLEEP" is floating. REFIO is an analog voltage that varies with the RF drive setting, I don't know what the range is. On the digital inputs they are whatever is coming out of the FPGA, i.e. 3.3 or GND.

Did you try lowering drive and seeing if the waveform stayed the same, but simply varied in amplitude? I think that is what will happen.

I'm grasping at straws now. I think I've reached the limit of what I can debug. We know it has something to do with the DAC and/or the FPGA connections to the DAC.

Are all of the connections on T1 correct?

Or, possibly...you have a 16-bit Hermes board, right? Did you by chance load the 14-bit firmware by accident (really grasping at straws)?

73,

Scott
User avatar
Prem
Posts: 25
Joined: Tue Apr 11, 2017 11:57 am
Location: UAE
Contact:

Re: Hermes Homebrew Fault Finding

Postby Prem » Thu Oct 12, 2017 1:27 am

Hi Scott,

Low drive rest yet to do and will update you.
Firmware I took from web only. How to identify if it is 14 or 16 bit ?

All connections found ok at dac side.
Let me see voltages also.

Rds

Prem, A65BK
User avatar
w-u-2-o
Posts: 518
Joined: Fri Mar 10, 2017 1:47 pm

Re: Hermes Homebrew Fault Finding

Postby w-u-2-o » Thu Oct 12, 2017 1:41 am

Prem,

All you need to do is look at the firmware section of this forum. However, assuming you have a regular, 16-bit Hermes card (the kind that is in an ANAN-10, not an ANAN-10E), you will find what you need here.

73,

Scott
User avatar
Prem
Posts: 25
Joined: Tue Apr 11, 2017 11:57 am
Location: UAE
Contact:

Re: Hermes Homebrew Fault Finding

Postby Prem » Thu Oct 12, 2017 4:17 am

Hi Scott,

I have made 16 Bit regular one only. Also the regular rbf file seen in web page. Hope 14 bit version will be listed separately.
Do you expect a damage in my FPGA or DAC chips ?

73s,
Prem, A65BK.
User avatar
w-u-2-o
Posts: 518
Joined: Fri Mar 10, 2017 1:47 pm

Re: Hermes Homebrew Fault Finding

Postby w-u-2-o » Thu Oct 12, 2017 7:03 pm

You can't hurt anything if you load the wrong firmware, if that's what you are asking.

I don't think anything is broken, just put together wrong. The output of the DAC, while wrong, is very clean, and is wrong in a very "obvious" way. you can clearly see the sign wave. It just seems like the most significant bit goes wrong.

73,

Scott
User avatar
Prem
Posts: 25
Joined: Tue Apr 11, 2017 11:57 am
Location: UAE
Contact:

Re: Hermes Homebrew Fault Finding

Postby Prem » Fri Oct 13, 2017 2:20 am

Hi Scott,

How we know if firmware loaded got corrupted or wrong ?
I did not get any error while loading firmware.
I use programmer to load firmware.
Will the error in boot loader will create this issue ?
I had not even seen error during boot loader uploading also. That I did first time using my Altera programmer unit.
Can you please guide me further ?

73s

Prem, A65Bk
User avatar
w-u-2-o
Posts: 518
Joined: Fri Mar 10, 2017 1:47 pm

Re: Hermes Homebrew Fault Finding

Postby w-u-2-o » Fri Oct 13, 2017 2:58 am

Prem,

There is no way to know if the firmware got loaded wrong other than to try it out. It would seem that the firmware is working well enough. All of the gross functionality is there.

It still looks like there is a problem with the most significant bit of the data word going across to the DAC. An open or short on the board.

Have you ohm'd out those paths (with the power off)?

73,

Scott

Return to “PC's & Homebrew Hardware”