Topics

Only noise when connected to the antenna #v6


Jonathan Kayne, KM4CFT
 

Hi all,
I recently soldered up what was basically a copy of the uBitx V6. I used the schematics and pictures of the board to essentially create a copy of the uBitx v6 in KiCad. I found a few problems during this process but I could not figure out why receive isn't working. The odd thing is that when I connect the system to an RF function generator and insert an AM signal, it comes right through. The moment I hook it up to an antenna, I get nothing but noise.
Now I know people were saying I should have done it Manhattan Style. Yes I could have but I wanted to learn RF PCB design. 
I do get the beeps when I turn the encoder, but no tone when I try to adjust frequency or BFO.

I also wanted to share a link to the project, so that maybe someone here could look over my PCB designs. perhaps I missed something. Also, I put jumpers between each section so it is a modular construction (just on a PCB)

Here is a link to my project: https://sites.google.com/a/vt.edu/amp_lab/projects/ubitx-workshop
I have the KiCad files there so please take a look at it.

Wondering if there is any tests I can do to ensure signals are doing what they should? There is little to no information about how to do any real tests on the modules, so quite lost. I have access to a full RF lab with Spectrum Analyzers, frequency counters, network analyzers, etc.

Thanks,
-Jonathan Kayne, KM4CFT


Arv Evans
 

Jonathan  KM4CFT

Mixed thoughts about this.  Your asking for support of something that is not a real uBITX so
we have no idea how close it is to the original design.  But...it is fashioned after the uBITX design
and thus might be supported here. 

The BITX and later versions of iBITX are almost modular in design.  That is not apparent unless
you understand how each section works.  I would start with the block diagram that has been
published several times on this forum.  That provides signal flow for tracing.  The voltage charts
that have also been published here could be used to troubleshoot a particular section once
you have found where the signal stops. 

Some may complain that your one-off is taking away from profitability of the HF Signals product.
Others may note that your are just proving that the overall design is repeatable, and that Farhan
has put the uBITX V.x designs into open-source or public domain so that anybody can copy it.

Arv  K7HKL
_._


On Thu, Feb 27, 2020 at 2:53 PM Jonathan Kayne, KM4CFT <jzkmath@...> wrote:
Hi all,
I recently soldered up what was basically a copy of the uBitx V6. I used the schematics and pictures of the board to essentially create a copy of the uBitx v6 in KiCad. I found a few problems during this process but I could not figure out why receive isn't working. The odd thing is that when I connect the system to an RF function generator and insert an AM signal, it comes right through. The moment I hook it up to an antenna, I get nothing but noise.
Now I know people were saying I should have done it Manhattan Style. Yes I could have but I wanted to learn RF PCB design. 
I do get the beeps when I turn the encoder, but no tone when I try to adjust frequency or BFO.

I also wanted to share a link to the project, so that maybe someone here could look over my PCB designs. perhaps I missed something. Also, I put jumpers between each section so it is a modular construction (just on a PCB)

Here is a link to my project: https://sites.google.com/a/vt.edu/amp_lab/projects/ubitx-workshop
I have the KiCad files there so please take a look at it.

Wondering if there is any tests I can do to ensure signals are doing what they should? There is little to no information about how to do any real tests on the modules, so quite lost. I have access to a full RF lab with Spectrum Analyzers, frequency counters, network analyzers, etc.

Thanks,
-Jonathan Kayne, KM4CFT


Tom, wb6b
 

Hi,

Your write up is pretty cool. The fact that you get noise when you connect an antenna is Hugh. Means most things work. 

Other than encouraging you to start at the antenna and work your way back, I have no idea. I would suggest checking that all the frequencies used in the various mixer stages are what you are expecting. If you use a unmodulated test signal you should be able to tune across that frequency and hear the beat note go from high to low then high again. 

Try a nearby AM station if one is around and test against that.

Farhan is likely not bothered at all about someone making a learning project out of his design.

Tom, wb6b


Curt
 

Jonathan

Nice work on your build. Try injecting a small signal into various stages, such as before and after the lower frequency crystal filter. Then work your way forward. Possibly a open connection somewhere in the path, which might explain why generator is detected by capacitive coupling somewhere.

If the bfo is badly misaligned not much may get 5hrough the chain.

Important to determine if LO signals are strong enough to drive the mixers, so I suggest checking these with a scope. A bad connection in LO path is a possibility, along with proper relay action.

Curt wb8yyy


Jonathan Kayne, KM4CFT
 

Curt,

Thanks for the info. I tried going through the signal path and everything is running as it should be! I found out that the reason my sound was not sounding as it should was because of the same relay that was causing problems before. The relays I installed were latching and I plan on replacing it with non-latching. I didn't realize that the baseband goes through that same relay before going into the final audio amplifier. I soldered a jumper in place and the audio came through clear as day.

Now for the next issue: I am trying to figure out why my display has problems. To get it to work, I have to disconnect SDI from the touch section of the ILI9341 during bootup and then reconnect it afterwards. For some reason the touch controller is causing problems for the display SPI line and I don't know why. This of course is more of a code issue I assume and I would expect this issue to arise if someone tried replacing the LCD in a genuine Raduino.
Any Ideas?

73,
-Jonathan Kayne, KM4CFT


Jonathan Kayne, KM4CFT
 

Not sure why, but even after doing this fix I get no receive when I hook up to antenna. I should easily be able to hear WWV or FT-8 but nothing. All I can hear is a high pitched whine that only goes away when I disconnect the demodulator from the audio amplifier section. I know that the signal is propagating through to at least the demodulator stage (after the 11 MHz filter). It was pretty hard to see if the 11 MHz signal was being converted to audio at that point with the spectrum analyzer or oscilloscope.


Jonathan Kayne, KM4CFT
 

Something I am wondering about is how I can tell if the 11 MHz filter is working properly. It seems to attenuate the signal quite a bit and when I inject a 14 MHz sine wave into the antenna line, this is the only stage where I don't see any change at the output. 
Don't know if this would be a cause but I used BFR93W transistors for both amplifier stages. I looked at the reason behind this and it seemed like it was because 2N3904 wasn't good enough at the first stage, so I don't see how that could be the issue though. 
Also, how important is it that the crystals are matched in the filter? I was pretty sure that the Cohn design was supposed to make the filter more tolerant to frequency difference.

Just trying to through out ideas as to what it could be but if anyone who has experience with this knows of things that can cause screw ups in the design please let me know.


Curt
 

Jonathan

Funny thing we learn more when we have issues.

The crystals do not need to be precisely matched for this filter, except if 2 or more styles with very different characteristics were used. First, there are a lot of connections, you could sniff each junction to see if its open or shorted. Also its fair to bypass a filter in signal tracing, just with a wire.

Now if the oscillator after this stage is not working or way off in frequency,  the crystal filter would appear to be open circuit in its network behavior.

Do use a signal generator or even touching a large enough antenna at the front and back of the filter to check the difference. The loss of a working filter should be low relative to audio level with or without it.

I don't expect your transistor choice to matter much. Farhan uses those bfr93 transistors to increase 45 MHz gain relative to various sources of leakage, that plagued the different frequency plan of the earlier versions.

Let me know as you need anything further. BTW I work with many Hokies up here.

73 curt wb8yyy



Curt
 

The audio squel may be unintended leakage, or related to 11 Mhz oscillator way out of alignment.

Maybe someone else could comment on touchscreen issues you cite. If you had a raduino that might be easier to integrate, if someone in the US could ship you a discarded one.


Jonathan Kayne, KM4CFT
 

Okay, so I went through and recorded the gains and losses of each section. I did this because my RF mentor said that I should be able to easily hear a -90 dBm signal. Unfortunately this was a bit of an issue as the noise floor of the spectrum analyzer was above that.
Regardless, here is what I found going from antenna to audio:
LPF + 1st conversion: -9 dBm
Preamp 1: +18 dBm
2nd conversion: -8 dBm
BPF: -8 dBm
Preamp 2: +20 dBm
SSB filter: -8 dBm
final conversion: unknown

The (de)modulator stage does not seem to have a specified conversion loss so I am not sure what to do about that. Looking at the signal (with an oscilloscope) out of this looks like a noisy mess even without any incoming RF signal with tons of harmonics. I have attached a scope capture:

Turning on and off the function generator does not seem to affect this wave.

Something I have been considering is adding on another amplification stage between the LPF and the 1st conversion to bring the incoming signal from -90 dBm to -60 dBm because I suspect the signal would be falling below the noise floor.


Jonathan Kayne, KM4CFT
 

Here is a capture of the (de)modulator from the spectrum analyzer. I Injected a 14 MHz sine at -60 dBm. On the uBitx v6 schematic this would be at TP20.
Again, turning on or off the function generator seems to not affect anything shown here and I suspect this is crossover from the other mixers and LO.


Curt
 

Jonathan

Nice work getting used to a nice spectrum analyzer. And doing some calculations. Keep in mind the bandwidth at the receive output is about 3 kHz.

I suggest starting near the back of receiver and going forward. Turn the volume all the way down, place a signal within the IF of the 11ish MHz xtal filter, both before and after. BTW if you have a leftover xtal, easy to make a generator that outputs approx s9 and s3 signal levels, I homebrewed one from norcal qrp website. Or use whatever signal generator thats there, but highly attenuated. if that seems good, work your way toward front. Not a bad idea to check voltage bias of each Rx path amplifier.

Solder connections can look good and be bad.

Thinking ahead, for transmit you should find or make a dummy load. 4 2 watt 200 ohm resistors in parallel etc.

All those mixers have transformers, always possible the insulation isn't off a wire. When you check continuity, never touch wire but a connected pad is better than edge of solder.

I presume you checked frequency of each oscillator and its presence.

curt


Evan Hand
 

Jonathan,

Are you sure the windings of the transformer are the correct polarity?  If you had one of the balanced mixer side of the transformer maybe not mixing properly.  You might also want to check the mixer diodes.

The above is assuming (as I had read in prior emails) that you have traced the signal all the way up to the BFO product detector stage.

Just ideas
73
Evan
AC9TU.


Jonathan Kayne, KM4CFT
 

I have published a YouTube video that goes through my findings. I was also trying to explain how the circuitry works. I show spectrum analyzer readings as I explain it so it might help.
https://youtu.be/pQb-sM8slCg

Curt, I did indeed check each oscillator and they were all working as expected. In fact that was the first thing I checked when I started troubleshooting. I also observed that the LO would appear at the output of the mixer due to nonlinearities of the mixer. 

Evan, I have tested up to the product detector stage. It is of course possible that one of the transformers isn't working properly, but I will say that I checked the product detector by inserting a 1 kHz -50 dBm (that's the lowest the function generator could do as the Agilent I was using for testing is 100 kHz minimum and had to switch to the lower frequency SG) into the audio side and it mixed the baseband up to 11 MHz. The weird thing about that was that the signal appeared to be amplified and not attenuated (by about +28 dBm if I remember correctly) maybe that signal generator wasn't as accurate as the Agilent...

-Jonathan, KM4CFT


Jonathan Kayne, KM4CFT
 

I think I might have the reason that the display is causing this issue. In the code the touch control CS pin is pulled high after initializing the TFT portion. That pin is probably low on startup so the touch controller probably thinks the TFT commands are for it resulting in both touch and TFT to respond simultaneously and effectively jumbling up the SPI line. Might be something to fix in the code as it is a good idea to make sure all SPI CS lines are high before sending anything!

-Jonathan, KM4CFT


Allen Hill
 

Hi Jonathan, I watched your YouTube vid and you did an excellent job stepping through the receiver. If you're still having issues, I have an offer for you. I have a mountain cabin within driving distance of VT. I could spend the nite at my place and drive to meet you at VT to help you work through the uBitx. 

I'm a retired engineer with tons of RF lab experience. I have a uBitx and while not professing to be an expert, I can easily make my way around the circuits. I could spend a couple of hours or a full day with you. Being retired, I can make just about any day that works for you (even a day during the weekend).

I'll send you a separate email with more contact info. I'd be glad to help a fellow ham out.

Allen Hill
KI4QCK


Jonathan Kayne, KM4CFT
 

It seems that somehow there is a bunch of noise making it into the radio. Just running it through the 1st conversion seems to bring the noise floor up enough for any reasonable signal to have adverse affects on the performance as seen by these two pictures! This was at the antenna connection so somehow noise from the 1st conversion is bleeding into the circuit making it too noisy to use. 
Any ideas on how to fix?


Dr. Flywheel
 

More than likely, your IF oscillator frequency is off and the Xtal filter attenuates the incoming signals as a result. Try calibrating the BFO oscillator.

73's
Ron
N7FTZ

On Mon, Mar 2, 2020 at 3:03 PM Jonathan Kayne, KM4CFT <jzkmath@...> wrote:
It seems that somehow there is a bunch of noise making it into the radio. Just running it through the 1st conversion seems to bring the noise floor up enough for any reasonable signal to have adverse affects on the performance as seen by these two pictures! This was at the antenna connection so somehow noise from the 1st conversion is bleeding into the circuit making it too noisy to use. 
Any ideas on how to fix?


Jonathan Kayne, KM4CFT
 

If I trace the signal through and there is noise appearing in the first stage then I don't think it's a matter of alignment. Plus I have tried that numerous times and it never helped. 
At this point I am literally connecting the spectrum analyzer to the antenna connector. Disconnect the antenna switching section from the rest of the circuit and it's clean. Connect the LPF plus 1st conversion to the circuit and the noise floor goes above the -90 dBm signal. I need to figure out what is raising the noise floor at the first stage which should fix the issue. 


Dr. Flywheel
 

Are you using your SA Tracking Generator to check the behavior through the signal path? Did you measure the actual frequencies of the three OSC outputs?

---Ron
N7FTZ

On Mon, Mar 2, 2020 at 4:02 PM Jonathan Kayne, KM4CFT <jzkmath@...> wrote:
If I trace the signal through and there is noise appearing in the first stage then I don't think it's a matter of alignment. Plus I have tried that numerous times and it never helped. 
At this point I am literally connecting the spectrum analyzer to the antenna connector. Disconnect the antenna switching section from the rest of the circuit and it's clean. Connect the LPF plus 1st conversion to the circuit and the noise floor goes above the -90 dBm signal. I need to figure out what is raising the noise floor at the first stage which should fix the issue.