Gordon,

I think the first path to eliminate is the voltage doubler/HV link used to reverse bias the unused filters.? There is only the RFC to keep the signal path isolated.? If, when connecting that link, the second harmonic increases, then I would look to increase the RF suppression on that line.

If you still have the HV connected, then the solution is to use relays instead of the PIN diodes.

73
Evan
AC9TU

OK, so I had some time tonight and very carefully I dissected out? soldering until I was able to take the output from the final transformer and send it DIRECTLY to the highest low pass filter, and DIRECTLY from there to the SWR bridge/antenna output.? ?There are NO switching diodes involved this way.? ? ?I haven't thougth through exactly what happened to the T R switch (hope I didn't do anyting stupid!)




Using commercial pads and attenuators here is the spectral output:



The main fundamental is almost -18 dBm.? ?Subtract 43dB from that and you get -61dBm

There are a few IMD products that are just a tiny bit over the line, but for me the key measurements are:

2nd Harmonic:? ?-64 dB which is -46dBc -- clearly legal
3rd Harmonic:? -77 dBm which is -59dBc -- very legal.

This is the WORST CASE for that filter -- this is the same filter used for the upper 3 bands, and I am testing it at 21.010 MHz -- a WORST CASE test.? ??

Compare? this to measurements I made on 10/8/2023 (using the same attenuator system, but with the filter within the switching system), and the 2nd harmonic from my written notes was -26dBc and the 3rd Harmonic was -30 dBc!!!!? ?Tonight, by going DIRECTLY through the filter, the 2nd harmonic is 20dB more suppressed, and the 3rd harmonic is 30dB more suppressed.

So even with low-cost MOSFET parts, this is basically a legal amplifier WHEN DIRECTLY INTO THE FILTER.? ?And the associated circuitry is in some way making it a lot worse.? ? ?

I'm still not seeing any spectra from anyone other than me and one other user.? ? Anyone else able to confirm or refute these findings?? ?In medicine we call this "an N of 1" and ONE is not a very large sample.? ??

But it shows that we can have a clean signal if we can solve where the problems are.? ?That should be the next effort.? ? Steps in that direction are to first connect the "input" switching diode and see what happens, and then maybe the "output" switching diode, possibly witth a couple or three in parallel??

Gordon KX4Z

This project was never stated as a fully polished device but a work in progress that had unbeleivavble potential to be something very uniques and powerful. Also something that still needed polish and was highly customizable andcould be improved on by the user over time.? That is how it has been discussed and marketed.? In My Opinion, If that is not what you wanted you probably bought the wrong radio because you didn't do the proper research.? YMMV

Jens, I'm sorry you had this experience. It was clear to me that it was an in--process, somewhat-experimental transceiver. Those of us in my club who bought them expected to be involved with at least making updates and maybe lighting up the soldering iron. Mine has served me well as my on holiday transceiver.

开云体育

...what if you bought a kit car, that you had to assemble yourself?

YT here:

Petr, OK1RP

Hi all,

I have the Sbitx V2 #005 for sale. Unit is new, almost untouched, original plastic foil on LCD, complete with all accessories from factory in original sealing and boxing.
On arrival the radio has been unboxed, tested shortly for RX/TX > presented hot receiver and full 40W output during few QSOs
and boxed back then it stays untouched until now on shelf.
Unit has no MODs, no updates, no any changes in hw or fw. So, you can modding it according to your preference's own self.

This unit has been dedicated as donation for our local kids radioclub but due to personal changes in the radio-club leadership?
no one is able to be unboxing it and start to play, modding and enjoy it in the club.

Additionally, it comes with BNC adapter, BNC-BNC 1m jumper, additional blue power banana cable, 3D printed sloped stands and the HDMI plate?that allows easy access to the HDMI port > all unused and sealed. Includes original working SD card installed in factory. I am asking $529 shipped worldwide double boxed. Its orig. factory box + additional box to enhance protection of the unit.

73 - Petr, OK1RP

I just wanted to give an update to the group of everything that went wrong i.e. a perfect storm. First of all the original motherboard was defective. Which intern tossed obviously a power spike or something to the RPI which also corrupted the SD card. In initially I had no reason to believe that the last two were defective ?so I was very frustrated upon putting a new motherboard in but it’s still didn’t work ?I knew logically that there were only certain things that would cause this not to work I was confident that the new motherboard was working correctly as I took power measurements. Therefore I eliminated the main board and focused on the RPI. I end up having to purchase a new RPI. Once I plug that in is when the SD card problem came about eventually I got the right image loaded on the SD card and everything worked finally.?

my sincere thanks to all those that responded whether positive or negative.?

--
Jack, W8TEE

I find that amazing you could do that with an iron.... I had a tough time doing it with a hot air station.

To hold the IC in place I took a metal strip from a windshield wiper blade and bent it like a hair pin
to clamp the IC down.?I tinned the leads on the IC... clamped it down and lined it up just right. Then I
could hit it with the hotair gun without it blowing away.
You could clip a ground on to it if your worried about static. I didn't take a picture when I did the raduino
but this is the basic idea.




73's Dave K0DLP

Steve -- surprise, surprise, surprise!
Now we may be getting somewhere.
I reconnected D30 and L201 to bring the switching systems back to where they are supposed tobe.

Then using two 9V batteries in series, I applied +18V to the point marked "HV" on C201.
That should back-bias the low pass filter diodes by 4-6 volts or so....

And surprise!? ?The isolation only improved by a small single digits dB.? ?Like maybe 2 or 3 dB; the entire curve with all of its series resonances (I'm looking from 2MHz all the way to 82 MHz now, with tracking generator and spectrum analyzer).

I was hoping it would have become FLAT at -60db or -70dB.? ?It did not.

(A) So I think we do have some evidence that signals can LEAK? ?*around* the excellent low pass filters.

(B) But if you look at the photos that I shot of the filter performance WITH all the switching diodes in place (see:? ?/g/BITX20/message/105633?) you don't see much evidence of signals getting "around" the low pass filters.

I don't understand how both (a) and (b) can be true.? ? ?I'm still missing something.

Copied from that previous message, here are the measured (IN CIRCUIT) rejections of the various filters:


BAND? ? ? -1db point? ? ? ?2nd H. Freq? ? -24 dB point? ? ?3rd H. freq
80 M? ? ? ? ? 5.4 MHz? ? ? ? ? -22dB? ? ? ? ? ? ~ 7.06MHz? ? ?~-60 dB
40M? ? ? ? ? 11.4 MHz? ? ? ? ? -21 dB? ? ? ? ? ~ 14.6 MHz? ? ?-51 dB
20M? ? ? ? ? ?16.4 MHz? ? ? ? -44dB? ? ? ? ? ? ?21.6 MHz? ? ? ? -57 dB
15M? ? ? ? ? ~3dB@...? ? ?-21 dB? ? ? ? ? ? ?43 MHz? ? ? ? ? -57 dB? ?(I had to change the Siglent to go to much higher frequencies to get this one measured)

How can the 20meter 2nd harmonic rejection be -44dB and my measruements of the 20 meter 2nd harmonic be -30dBc??? ?That suggests the output diode is mixing to creating more of the 2nd harmonic on the spot.? ?

In post?/g/BITX20/message/105656? ?I showed that a modest attempt at shorting out the "output" diode on the filter used for 15 meters resulted in a 6dB reduction in the 3rd harmonic (to almost legal levels) while the 2nd harmonic was not much changed.? ? ?I did NOT test shorting out the "input" diode -- indeed it could be making harmonics and on some bands the filter has only 20dB reduction on 2nd harmonic, so if the switching diodes are making harmonics....that could be again the source of the problem.

I have not yet done the drastic step of disconnecting all of this switching circuitry and just going directly into a filter and out to the antenna, which I discussed in post?/g/BITX20/message/105666? ? ?

Since I still can't really explain what is going on, I may? have to go ahead with THAT experiment....If that succeeds (who knows, it might!)? then the solution would likely be to go back to the DE circuity with relays to switch the LPF's and diodes to do the T/R switching.? ? But there is no proof at this point that this is the solution.?


Gordon KX4Z



But it looks like?

开云体育

Steve, thanks very much for the ideas. ?Inhad not thought of what you just pointed out!! ?I could get an 18 V source and connect it via resistor to the HV point and solve that…. ? ?maybe I can try that tomorrow. ? Thanks!!

Gordon - sorry, I don't have a V2. Maybe some day...

Your tests appear to demonstrate that a reverse bias needs to be applied to the switching diodes to turn them off (or reduce their junction capacitance.)

Note that because of D20 and D21, the reverse "HV" bias will be LESS than the forward bias if one of the low pass filter sections is NOT turned on. Only by turning ON (selecting) one of the filters will there be a voltage drop across R201/R202 and R203/R204, which will allow the reverse bias voltage to be higher than the forward voltage.

Do that, then try bypassing L22 at R203/R204.

73; Steve, N3SB

So I lifted one end of D30 to get rid of any possible transfer via the TR switch.
NO CHANGE.

I'm injecting the tracking generator with the shield of UG-174 at the ground end of C204.? ?The center conductor goes to the connection point of the output of T7 -- having lifted the wire from T7.

I tried touching all of the inductors hoping something would change.? ?Basically there were negligle changes.??

So, I unsoldered my center conduct thinking could I be just conducting through the air, some sort of CAPACITANCE?
Nope.? ?When I did that the signal plain went away.

So there IS some connection from one end of the filter system to the other
-? it depends on some diode or non linear device, because it is bleeding MORE with power off, than with power on.? ?
- It does NOT go through D30? ?(I have it disconnected)
- it does NOT go through L201? (I have it disconnected)

OK, I'm at a loss.? ?I don't know what is happening

Gordon KX4Z

A surprising result tonight.? ?I found a yellow toroid T2-25 I think, put 22 turns of magnet wire on it, and verified with ancient MFJ259B that it was > 650 ohms reactance in the 40-60 MHz range.
Tried putting that in series with L201, figuring it could reduce the "inflow" if the bypass was via L201/L202.

It did NOTHING to help.? ? In fact, even DISCONNECTING L201 did not change the "bypass" around the filters.

What I did notice, however, was that the "bypass" was 10+ dB? ?STRONGER when the unit was UNPOWERED, than when it was powered.? ?The result was striking.
I'm injecting tracking generator at C205/C206 (having disconnected the output of T7 PA XFMR) and looking at what gets to the ANTENNA connector.? ?
There is MORE getting there when the unit is unpowered, than when it is ON and RECEIVING.

This makes? me wonder about the other path --- through the voltage doubler to the "HV" designated point on the schematic, and then from there to R216 and thence through inductor to the output.? ? This just doesn't look terribly likely, but there are not many possible paths left!

Gordon KX4Z

Gary and Tony,
Thanks for the tip about better choice of tape for anchoring static sensitive parts for soldering. I have some aluminum and copper foil tape on hand, but didn't think of it. Hopefully, I'll remember "next time."

Mark,
Your help was essential to resolution of this issue. It is very much appreciated.

We all agree that it is indeed important to get it all right.? ? I'm astonished at the advance that Ashhar has made virtually unaided except for help from Hans.? ? Apparently Hans had some problems with series resonances himself and I think improved his own design, which Ashhar utilized.? ? I think the FT8 code also came from another (I could be wrong) and thus the difficulties were not anticipated.

I can't say if my unit #141 is "representative."? ?It is only unit #141.? ?A? ?*lot* can happen between "prototype" and "delivered manufactured units" -- especially with supply lines being so diffiult in our current world.? ?Has anyone else done the work to measure their own unit?? ? I think a confirmation either way is important.? ?My first measurements were faulty because I hadn't recognized the harmonic/IMD-producing DIODE in the Heathkit cantenna circuitry.? ?Now with a commercial "pad" I think (but cannot be certain) that my measurements are correct.

The unit appears to be at, nearly at, or better than required USA specs on
80 meters (better than required)
40 meters (2nd harmonic is "close" at the low end of the band)
"kinda close" on 10 meters...
"kinda close"? on 30 meters
18 MHz -- right at the spec, so useable there also.

I have not checked all the bands.

So I think you could proceed on most or all of those bands.?
The ones that are most bothersome ON MY UNIT appear to be 14 MHz and 21 MHz.? ?On the trace that I inadvertently caught the bleed-around on, it appears there is a series resonance band from about 32MHz - 40MHz and this MAY the reason that the 2nd Harmonic of 20 meters is a problem, and the 2nd harmonic of 15 meters is a problem? ? That particular trace ended at 52 MHz so I don't know what allows the 63 MHz 3rd harmonic of 15 meters.? ? ?

I need to make up some small chokes (not sure I have the most appropriate cores) and add series impedance of several hundred to thousands of ohms in that series resonant area, and see how much improvement we get.? ? There may be a way to attack the production of IMD in the output switching diodes also.? ?Maybe put two in parallel??? So that each carries less RF current?? ?This is fascinating new stuff for me to learn from -- never tried diode switching before at all, nevermind doing it at 20+ watt level!!!

I think we're going to see solutions.? ? I'm slow and not there yet, but others may solve this before I even get there.? ? ?
I don't have time right now to work on sofftare (and I'm not that good at it) so others can tacckle that.? ?We have 7th graders learning Arduino at my high school where I teach, so there is nothing magic about it, but it does take time.

This is an amazing radio.? ?It deserves for us to make it great!

73
Gordon KX4Z