Georgy,

?

I did not mean to hijack your message thread.? I just wanted to measure the new firmware you provided but ran into a problem and then spun out of control.

I will stop posting here and will start a new thread if I discover anything about the AM modulation that I am seeing with my radio.

?

But I did want to share the measurement I made today because it may help someone use other microphones to make the radio sound better.? Since that is why you released the firmware, I figure it is OK.

?

I measured the microphone connector input impedance today.? Red is measured and blue is a simple model that matches well with the measurement

?

This schematic has a very similar impedance to the microphone input.? But be warned, the microphone signal pin has a bias of 6.2 V, so a bit of caution and thought is required before connecting anything.? I did not model any potential bias circuits, and I have no idea what the input circuit actually looks like.

?

I would think a microphone output impedance less than 600 ohms would work well.? My testing has been with a 50 ohm source impedance and should not be impacted by the radio input impedance.

?

Rainer,

I can potentially use white noise; my function generator is also an ARB.? I will have to rewrite my test. As of now it is using swept sin.

?

But you said "ALC effects"!? That is something I completely overlooked.? I am embarrassed to say don't even know how my radio is set up in terms of ALC.? I will check it out tomorrow.

?

I am using a function generator connected to the mic input, and the x6100 is connected to a Vector Signal Analyzer through a 40 dB 50W attenuator.

?

Thanks for the link to VA, I will check it out.? I like the picture on the homepage.

?

Cheers,

?

Dan W6DSW

Dan,

?

can you test with white noise? This should avoid any suspected resonance or ALC effects,

?

I plan to test the following when I get home from vacation:

?

PC with VA software -> USB -> X6100 internal soundcard -> 40dB power attenuator -> spectrum analyzer

?

VA: freeware, with interesting extras like RLC measurement option, based on a Ph.D. thesis at an Italian university.

?

Greetings,

?

Rainer DG1SMD

?

?

After some investigation I have found that there is a time varying gain of some sort from the microphone to the modulation output.? It is worse at lower frequencies but can be seen at all frequencies.

This plot shows a 165Hz tone in white and a 170Hz tone behind in red.

Zoom in on 170Hz


Here the envelope was extracted from the above data and plotted as dB from the mean signal amplitude.? Additionally, there are measurements of 250 Hz and 2.5 kHz.? You can see that the period of the modulation is not the same for each tone frequency I tested.? I do not know if it is related to the audio frequency or if the modulation frequency just drifts around.? In any case, my frequency response test just happens to be grabbing points off of this envelope and it was displayed as ripple vs frequency when in fact it is ripple vs time.?


I will play around with various mic gains and input levels to see if I can learn anything.? Also, the hand mic has a 220k resistor between the mic input and the base of a NPN transistor that acts as a switch for the power inside the hand mic.? I will try loading the mic line with a resistor to see if it somehow impacts the mic input.

Hmm... Just thinking, I should be able to see ~5Hz AM sidebands on these ~170Hz audio tones.? I hope it is not some sort of power supply stability issue in my radio.

I've always heard the tale as starting with a stone rather than an axe, but so applicable in many contexts! I have been known to relay that story when discussing feature creep with new in career software developers.

On Sat, Dec 28, 2024 at 12:24 PM, Dan Wertz wrote:

Now, that would be very cool if we could have custom TX audio filters!

?

?

On Fri, Dec 27, 2024 at 10:47 PM, Georgy Dyuldin wrote:

?

As far as I remember, biquad filter is used there. I'm not 100% sure, but looks like coefficients are precomputed (stored as a values within firmware). I wonder to extract them and to apply to some test signal on computer. Perhaps, it will be possible to design a filter with better parameters).

--

Georgy // R2RFE

?

Now, that would be very cool if we could have custom TX audio filters!

I initially thought so as well. ?I did measure the function generator, and it was extremely flat.

But after looking at several runs of the test using different drive levels, you can see that the ripples change run to run. ?Also, for the green trace below I cranked up the number of points in the FFT I am applying to the demodulated audio and there is a distinctly different pattern.? Tomorrow I will look closer at the demodulated waveforms to try and ferret out what is happening.? Hopefully it is something stupid in my test and not a "feature" of the radio

On Sat, Dec 28, 2024 at 02:36 AM, Dan Wertz wrote:

I found a 10uF tantalum cap I used as a DC block.? The 0.18uF I used previously was introducing some additional roll off.? I also changed to acquisition code to actually look at the FFT bin that has the signal.? Previously I was just taking the max of the entire FFT, but there is significant in band noise that was limiting the measurement floor.

I am not sure what is up with the fast passband ripple.? I may have some measurement issue.? Right now, I am trusting the function generator to be flat.? I will get a DMM on it to make sure.

It might be due to digital filter design on the firmware. At least, when I recorded signals - they all have some audible resonance close to cut-off frequency for any frequency from 160,200, 250, 300 Hz.

?

As far as I remember, biquad filter is used there. I'm not 100% sure, but looks like coefficients are precomputed (stored as a values within firmware). I wonder to extract them and to apply to some test signal on computer. Perhaps, it will be possible to design a filter with better parameters).

--

Georgy // R2RFE

I found a 10uF tantalum cap I used as a DC block.? The 0.18uF I used previously was introducing some additional roll off.? I also changed to acquisition code to actually look at the FFT bin that has the signal.? Previously I was just taking the max of the entire FFT, but there is significant in band noise that was limiting the measurement floor.

I am not sure what is up with the fast passband ripple.? I may have some measurement issue.? Right now, I am trusting the function generator to be flat.? I will get a DMM on it to make sure.

Thanks for this!? I am planning on connecting a nice microphone and footswitch to the X6100 and started doing some testing.? I noticed the 300Hz lower limit on the audio and searched to see if that was typical of SSB radios.? I stumbled on this thread, and I agree it sounds much nice to my ear with the 160Hz lower limit.? Though I am reading for DX contacts the below 300Hz does not really add anything to intelligibility.? I measured the frequency response of my radio tonight using your patched firmware and this is what I got.? I did have a 0.18uF DC block between the function generator and the mic input, so it is possible some of the low frequency roll off is from that. ?I have not measured the mic port input impedance, so I can't say.? I will track down a 1uF and remeasure to see if anything changes.? I am thinking I will have a couple EQ settings programmed, so I can have a pleasant sounding preset and a punchy sounding preset for poorer conditions.

https://github.com/gdyuldin/x6100_gui/discussions/23

--

Georgy // R2RFE

I wish I knew how to program.? Bluetooth keyboard messaging via JS8Call would be cool but the "heartbeat" alone has utility.

Georgy, I know you get a lot of suggestions but have you thought about JS8Call??

Sean,

Any SDR radio like the Hermes Lite 2, Square SDR or Radioberry has the ESSB capability and extra things like Pure Signal and CESSB depending on the PC program.

Gary

W9TD

I think eSSB is ok if used responsibly when bands are quiet and a little more audio clarity can help with long DX contacts. I would certainly never use it on 60 Meters as FCC rules are specific on that one. I would neither go above 4Khz as we are indeed abusing the limits of FCC rule 97.307(a).

Anything above 4KHz is wasted energy and more prone to interfere with adjacent channels.

I was considering the Yaesu FTDX-10 as this has the full 4KHz bandwidth, But $1,599.95 is a little pricy.

Sean,

Here in the US, ESSB is not actually allowed as FCC rule 97.307(a) states: "No amateur station transmission shall occupy more bandwidth than necessary for the information rate and emission type being transmitted, in accordance with good amateur practice." And later on in 97.307 in table 1 for 60meters, the USB emission code is 2K80J3E where the 2K8 stands for 2.8KHz bandwidth. Nevertheless, no one will probably get a fine for using ESSB.

Gary

W9TD

Georgy,

I plan to put together a simple proof of concept here on my X6100, and then put the code on GitHub.

I just ordered one of those USB->Ethernet adapters, because so far I have only connected via the serial interface. I think that for those tests it would be useful to have multiple ssh sessions open at the same time.

?

Franco

Would be nice to still keep the 160Hz for voice and 3500 would improve both voice and FT8. Maybe a bandwidth of 160 to 3500 would be a good compromise?