When SiLabs built these chips, SSB and CW were not front burner issues
for them and were not included in the standard DSP firmware on the device.
Fortunately, SiLabs did included a way to download custom DSP code
into RAM on the chip, and eventually some manufacturers did sell
reasonably cheap SW radios using these chips that support SSB and CW.

I haven't looked at writing DSP code for it, I'm sure it's not trivial.
If you want to play with the parts, here's the webpage Paulio mentioned:
? ??

I have an XHDATA D-808 on my nightstand which is doing well.
Replaces a string of a half dozen others that just didn't measure up
to my beloved 40 year old Sony ICF-2002, now aging and crippled.

The XHDATA has hit the floor dozens of times with no apparent damage,
faring better than the Sony in that regard, in part because it is half the weight.
SSB and CW reception is acceptable.
Has an internal 18650 li-ion battery that can be charged through a USB-micro jack.
User interface is better than most of these radios (most of which are atrocious).
My primary issue was identifying the buttons in the dark,
a dab of epoxy on the "FM" and the "5" buttons fixes that.

One sorry mis-feature to look out for on these radios: Soft Muting.
They assume if the signal is weak that you won't want to be bothered with that nasty static,
so they mute the audio for you.? Most of the SiLabs based radios I tried have this "feature",
and no way to defeat it.? Infuriating!? Fortunately, the XHDATA doesn't do this.

I've had my XHDATA for a couple years now, the market may have changed some.
Not exactly a top of the line communications receiver, but for under $100 I'm not complaining.

Jerry, KE7ER

Functionally for about 99 percent the 4735 and 4732 are the same. The main difference is the package. The 4732 is in a 16 pin SOIC which makes it easier to work with. It truly is a radio on a chip. You do have to add your favorite flavor of audio amp. In my radio I used a class d amp and get Stereo output. I control it with an Arduino Uno. The VFO and BFO are code based. PU2CLR wrote a library for these chips sets. He has tons of info on his Github page.
73 de W9WC
Paul

On Wed, Mar 31, 2021, at 09:45 AM, Jerry Gaffke wrote:

So, not as a solution but as a way of understanding the problem I might try the following steps:
1)? Remove L31, terminate that mixer port with 50 ohms to ground.? (There's no signal coming from the IF amp in CW mode.)
2)? Add a 6dB pad (150,33,150 ohms like R101,102,103) between T2 and L4 (perhaps? remove L4,C204 and tack it in there, avoids cutting traces)
3)? Replace the mixer with something else, perhaps an ADE-1.? Could be a resonance inside that mixer?

These steps make sense to me, especially the first one.? I did try killing the power to the BiDi amp and that had no effect.

The one point that needs to be made is that I am not the only one experiencing this issue.? It is also weird that the other v6 I worked on did not have this problem.? I am searching for the reason as much to learn as for others to be able to implement a fix if they want CW on 40meters.? As stated in another email, 40meters SSB seems to be the highest output with a two-tone audio inpout.

Thank you for the input.? I will post when I get any results.
73
Evan
AC9TU

Evan,

That tp2 scope trace for 40m looks pretty bad.
It's ok to have harmonics present, but they shouldn't nearly overpower the fundamental like that.?

What does tp13 look like, is there possibly something coming in from the 45mhz IF amp?
Probably not the problem, your tp2 trace looks somewhat regular from cycle to cycle,
so is mostly 7mhz plus harmonics.? But worth looking at.

>? Could this be caused by a 3rd order mixing product?? 3rd harmonic - fundamental would give the second harmonic.? Why only on 40meters?

Mixers are weird, I haven't figured them out well enough to predict what awful thing they might do next.
Could well be something like what you suggest going on, and/or some inadvertent resonance.?

I don't think it's a solution here as we want to keep the rig simple,?
but will point out that the only port properly terminated at 50 ohms is the clk2 port.
The filter at L31,L32 does allow low frequency energy to be terminated by what I think is the 50 ohm output
impedance of that bidi amp, but blocks high frequency stuff above 45mhz which will reflect back into the mixer.
The filter at L1,2,3,4 does the same for stuff above 30mhz.? (Actual cutoff frequencies are considerably higher,
up by a factor of perhaps 1.5x).

So, not as a solution but as a way of understanding the problem I might try the following steps:
1)? Remove L31, terminate that mixer port with 50 ohms to ground.? (There's no signal coming from the IF amp in CW mode.)
2)? Add a 6dB pad (150,33,150 ohms like R101,102,103) between T2 and L4 (perhaps? remove L4,C204 and tack it in there, avoids cutting traces)
3)? Replace the mixer with something else, perhaps an ADE-1.? Could be a resonance inside that mixer?

Jerry, KE7ER

Thanks OM Raj!! Will revert after checking!
73

Jerry,

Sorry, I rethought the prior email and it is the 2nd harmonic that is causing the issue.? Somehow it is canceling some of the square wave that should be the 40meter signal.? That is also why the scope is reading 14.448MHz instead of 7.065MHz.? Going back to theory, there should only be odd harmonics for a square wave.

Could this be caused by a 3rd order mixing product?? 3rd harmonic - fundamental would give the second harmonic.? Why only on 40meters?

73
Evan
AC9TU

Jerry,

On thinking about the waveforms, could it be that the 3rd order harmonic on 40meters, being below the 30MHz filter cutoff is causing the issue?

More data, the TP12 readings for 80meters and 40meters:


73
Evan
AC9TU

On Tue, Mar 30, 2021 at 07:53 PM, Jerry Gaffke wrote:

I'm not 100% sure I understand what a "very high low harmonic" is.

Jerry,

Here are three screenshots from my scope:


They are from TP2 of the ?BITX.? The one that is not the correct measured frequency is supposed to be 40meters.? The frequency for 40meters was set at 7.065MHz on the ?BITX.

73
Evan
AC9TU

Hi Evan,

Yes. I see.

I'll check output on SSB and will look out for your findings.

Best,
N

Thanks everyone, sounds good to me. I will replace the regulator and add the dropping resistor both.
?
? ? ? ?Thanks, Mike

Got it, thank you again Evan.?

As it happens, I haven't tinkered much with CW yet, but I anticipate it in the future. Actually, despite having an Extra license, I have never so much as keyed a mic, not even on the local repeaters that I monitor. I am more interested in QRP,, antenna design and signal propagation under different atmospheric conditions. FT8 allows me to study and learn about these things without taking the time to be social. This WILL eventually change, because there are so many genuinely fascinating people to talk to, but for now, SSB, FM and CW are almost too personal. I want to be lost in the various sciences that are associated with amateur radio, not with the sociology, you know? ;)

Evan,

I'm not 100% sure I understand what a "very high low harmonic" is.
Operating a 7mhz, I assume these are 2'd and 3'rd harmonics, so 14 and 21 mhz.
I'd expect that on pretty much all bands at TP1, especially the odd harmonics as we are feeding a square wave.
However, the 30mhz LPF at L1,2,3,4 should take out anything much above 30mhz.

Keep an eye out for anything undesired that is lower than the 2'd harmonic,?
as that stuff is probably getting all the way out to the antenna.
Otherwise, so long as those harmonics are not as strong as the fundamental,
seems it should be ok.

Jerry, KE7ER

Jerry,

I tried putting a 2.2kOhm resistor in parallel with R104 and got the same results.? Evidently, it is not the diode bias that is the issue.

I did note that the waveform at TP1 and TP2 has very high low harmonics on the waveform as opposed to the other bands where there is just some higher-order ripple on the fundimental frequencies.

Will keep looking.
Thanks again.
73
Evan
AC9TU

Skip,

You correct in that there is still heat that needs to be dissipated.? The reason to go to the higher current rated device is that it can handle the higher current without generating as much heat, or can tolerate the higher temperatures.? I did specify that you do need to have a good heatsink on the regulator and that in most cases adding the dropping resistor will solve the supply issue.

73
Evan
AC9TU

Gamaliel,

Your statement is true IF you do not want to get serious with CW.? There are a couple of options that make the ?BITX operate more as most commercial rigs do on CW.? The most significant is that the dial displays the transmit frequency.? The stock will transmit either the sidetone above or below the dial frequency as delivered.

So If you are planning on mostly SSB and only causal CW use, then the stock software is fine.? That is my opinion, others may have a different view.
73
Evan
AC9TU

开云体育

Mike replacing the regulator on the Radunio will not solve the heat problem. The regulator will source at least 1 amp to run everything that needs 5 Volts. The main problem is that if your supply voltage is say 13 volts the regulator is dissipating the heat of the 8 volts times the amps used. I installed a dropping resistor between the 12 volt supply and the regulator to drop the voltage the regulator is seeing and this reduced the heat the regulator is expected to dissipate. Mine runs much cooler now.?

It *seems* - and someone please correct me, in the likely event that I am wrong, bilut it *seems* that the newest stock firmware is quite solid, and has incorporated many of the advantages of the other firmwares. Would it be true to say that, unless one wants to install a Nextion or Teensy or some other significant mod, it is likely best to stick with the stock? Naturally those wanting to roll up their sleeves and tinker should do so, that is what uBitx is about.?

But staying stock, at least until us new folks acclimate, is looking best, I think.

Mike,

I run a 2.8" display on a stock v5 with just an added heatsink to the regulator.? You could go the extra step and add a series resistor to drop the 12+ volts (I am running 13.8) to a lower value.? I would assume a 1 amp draw and you want at least 7 volts to the regulator.? That means you want to drop no more than about 5 volts.? At one amp that is 5ohms and 5watts.? I used 2 3watt 2ohm resistors.? Close enough.

The total power draw on receive is under 400ma, so the above is really worst-case estimates.? The total 5vols is probably under 200ma.? That means you could just add a series resistor to drop the input voltage and dissipate the power in the resistor rather than the regulator using the stock regulator.? So if the max current is 500ma, and the lowest supply to the rig is 10volts, then you can drop 3 volts across the resistor at 500ma = 6ohms at 1.5 watts.? This would also mean that if the supply voltage goes under 10volts the Raduino would start acting erratic.

The true belt and suspenders person would do both: replace the regulator AND install the dropping resistor.? UNLESS you were designing for portable operation THEN a buck/boost regulator might be a better option (with a power line filter on the output).

Lots of options.? Makes my head spin.
73
Evan
AC9TU

Nick,

If you scroll back through some of the earlier posts on this thread you will see that I have a very similar issue with my v6 that I just assembled in the last two weeks.? It is not as drastic as your reduction in power on 40meters, 5watts vs your 3.24watts.? I ran a test using a tone generator as the audio input and got different results.? 40meters was the highest output and then tapered down from there to 10meters being the lowest, as expected.? This points to an issue with unbalancing the mixer to get the clock2 to the power amp as suggested by Jerry KE7ER.? I will try reducing the value of R104 to maybe 1kohm (add a 2.2k in parallel).? I will post what I find out.

You can simulate an audio tone generator by playing a constant tone on a PC or mobile and holding the mic close to the speaker.

The good news is that SSB is not affected.? The bad news is if you wanted full output on CW.? Read the messages from Jerry for more information.? Thank you, Jerry!

73
Evan
AC9TU