|
Hi all...
Wanted to put out the question, to Hans and others, in case it's not already existing somewhere, of what one can do, or would want to do, to "bring forward" the earlier PCB versions?
I have a ver 3, and getting back into using it after a hiatus, am happy to see and read about the updates and improvements in later board revisions. But, I'm wondering what might be done by more intrepid DIY'ers. to add some of those features to my old board.
Particulars...
I see that the DC blocking caps going into the audio preamp chips are no longer there, due to biasing issues and latch ups. Is there any reason one wouldn't want to remove and jumper them out on a v3 or earlier board? Having read a bit, guess this likely is a good one to do, along with the other R's and C's associated with the original bias?
Might be a lot to do, but what's the feasibility of retrofitting a SMPS to an earlier board, for the reduction in PIN diode current draw? How easy (or difficult?) is it to hack the parts onto an older PCB?
There was a change in the clock drive that I'm forgetting the details of, sounds like that might be somewhat easier to do, is that easy to add? Benefits worth the hassle?
Any other improvements or changes to a v3 board (or earlier) that I might not be aware of?
Someone let me know if there's already a list out there, or just the information. I started out on this thread, then did some more reading, and found some of my questions already answered.
I've been very happily surprised by the amazing progress since I got my kit, quite a lot of things done and improved. Also, the stories of the research and development on hw and sw, especially the SMPS efforts and the solving of the TX frequency stepping issue... really great engineering!
|
|
|
Hello Jon
What you could do, is yes, make all the changes. What you would WANT to do, depends on a lot of factors... how mad you are, how badly you want the changes, blah. Some changes are easy, some not. Updating to the latest firmware is the most trivial easy thing that will get you immediately some benefits! I see that the DC blocking caps going into the audio preamp chips are no longer there, due to biasing issues and latch ups. Is there any reason one wouldn't want to remove and jumper them out on a v3 or earlier board? Having read a bit, guess this likely is a good one to do, along with the other R's and C's associated with the original bias?
If you have an op-amp lockup problem, then it's NECESSARY to short out these capacitors. If you don't, then shorting them out and removing the other components makes very little difference to anything. But it's also very easy to do and I find removing these components strangely and unexpectedly therapeutic in some sense. Might be a lot to do, but what's the feasibility of retrofitting a SMPS to an earlier board, for the reduction in PIN diode current draw? How easy (or difficult?) is it to hack the parts onto an older PCB?
There was a change in the clock drive that I'm forgetting the details of, sounds like that might be somewhat easier to do, is that easy to add? Benefits worth the hassle?
Any other improvements or changes to a v3 board (or earlier) that I might not be aware of?
Attached, some photos of my franken-QDX. Which is a Rev 3a board (remember, Rev 3a is only different from Rev 3 in so far as the trifilar transformer was moved fractionally further from the board edge in Rev 3a; which also inflicted the infamous C41 via bug, necessitating a trace cut). This board works great still (miraculously). It has the modifications:
1) Final prototype of my SMPS bias circuit, which was built in the same location as it would be on the final Rev 4 PCB, to make sure there was no issue with signal trace lengths, position of the components near the output transformer etc. As closely as possible, I tried to implement it as it would be done in the Rev 4 board. WITHOUT wishing to actually prototype a real Rev 4 PCB, life being so short and all. Necessary PWM output and ADC feedback connections to the microcontroller, where the SMPS constant current control loop is implemented and controlled in firmware, are via 0.33mm enamelled copper wire, through PCM holes and soldered directly to the microcontroller pins.
2) For the VGA and PS/2 keyboard terminal port: 2x 4-pin headers which are installed on a little piece of PCB, which is actually a chopped up QCX-mini prototype board, using 2x 4-pins of the ATmega328 socket position, to solder in the 4-pin headers securely. Not that you wanted to know. There are then very thin wires, which actually came from an old relay coil (not that you wanted to know that either), and pass through the un-used holes header holes in the PCB (now used for high-band and mono-band band selection), through to the underside where they are soldered directly to the microcontroller pins.
3) C71-74 shorted, and all associated bias capacitors and resistors removed
4) On the bottom side, the tiny single-gate logic chip is installed in place of the earlier NPN/PNP transistor buffers.
5) The PTT transistors are removed and bypassed, to be able to use the 3.5mm PTT connector as Seiral Port 3 (enabled in the Configuration screen in the firmware) so I can now easily plug in my beloved DEC VT320 terminal at 19200 baud and it's hacked RS232 level converter, whose purchase did materially significantly dent the Save-The-Hans coffee fund.
6) Various other trace cuts and soldered on connections had been done and undone, as part of lots of ongoing experimentation on the long-suffering unit.
Is it all worth it?
Oh yeah baby, it's SO TOTALLY WORTH IT!
(Disclaimer: Terms and conditions apply, your mileage may massively vary, no responsibility accepted for damage or fatal inoperability of your unit, blah blah blah)
|
|
|
On 20/01/2023 09:20, Jon Pike wrote: what one can do, or would want to do, to "bring forward" the earlier PCB versions? Jon, Certainly use the latest firmware. I'd say leave the hardware alone unless you actually notice any deficiencies. I'm wondering what might be done by more intrepid DIY'ers. Almost anything:-) to add some of those features to my old board. Old board? I've 3 v1s. I see that the DC blocking caps going into the audio preamp chips are no longer there, due to biasing issues and latch ups. Is there any reason one wouldn't want to remove and jumper them out on a v3 or earlier board? Having read a bit, guess this likely is a good one to do, Again, unless any problems are obvious, leave alone. But that's something I've wondered.. Might be a lot to do, but what's the feasibility of retrofitting a SMPS to an earlier board, for the reduction in PIN diode current draw? How easy (or difficult?) is it to hack the parts onto an older PCB? Looks a little more complex, is it worth it? There have been few problems even though it appears the original circuitry exceeded component ratings. I guess the main thing is power reduction if that is important? There was a change in the clock drive that I'm forgetting the details of, I've completely forgotten, does not ring a bell. Any other improvements or changes to a v3 board (or earlier) that I might not be aware of? Improvements? 2 extra serial ports, Hans' VGA display? I've a few personal mods. Someone let me know if there's already a list out there, As far as I know just deeply embedded in the group posts. Oh, Hans has beat me to it, about 100x faster typist! I'll post anyway. 73 Alan G4ZFQ
|
|
|
Of the modifications on the version 3 board that really improved the performance of MY transceiver, this is the shorting of the C71-C74 capacitors. This was especially felt on the 20 m band, where, before refinement, a huge number of FT8 stations did not allow listening to the air at 24 kHz above the receive frequency due to weak suppression of the false receive channel. After removing the capacitors, these interferences disappeared. Without other improvements, you can live in peace.
--
73, Oleh, UY2ZA
|
|
|
Jon,
For Rev 3a and earlier boards replace R2 & R4 (330 Ohms) with higher wattage resistors.
After reading Hans' description of using SMPS for PIN diode biasing (Rev4), one take away item for me was the existing "603" size resistors used for PIN biasing were undersized relative to the amount of current they handle.
I replaced them with 805 sized 1/4 Watt ones. The larger size will fit on the existing pads without overshoot.
I think it helps long-term reliability.
--Al
WD4AH
|
|
|
I'm glad I've found this post Chaps! I too have a V3 that was put to one side (Life got in the way..) and was wondering what to possibly change before it's complete. I'll drag out the manual and have a re-read before commencement. Cheers, Rick.
DE G6UZT
|
|
|
I'm just gonna jump in here and emphasize that clock buffer.. It need to be done. My hat is off to Hans -- I know from personal experience how hard something like that is to track down.. I'd also add some better heatsink to the output transistors.
If anyone wants some NZ17SZ04DFTs (they are tiny) to do the clock buffer thing let me know -- I've got a reel from a previous lifetime. SASE will get you 5-6 of them.
|
|
|
Gary W9TD
Phil,
I would like some of these devices. What is your call so I can look it up in QRZ, or reply privately with address.
Thanks,
Gary
W9TD
|
|
|
Gary, I just replied to your email --
When I did my 3A I found that it was pretty easy. Remove the transistors, one of the caps and one of the resistors (I changed the other resistor to 470K - don't know if that is really required - let us know) - then the part can be mounted on some of the transistor pads -- two pins match up with transistor pads, don't member which -there was some wirewrap wire for one of the power (or ground - don't remember) and either the input or output - (still don't remember). I added a drop of uv resin after it was working to keep it stable - even though 2 pins were soldered down. (When I say easy - I do have a microscope, hot air, and have worked on tiny things a lot.)
|
|
|
On Fri, Mar 17, 2023 at 05:52 PM, phil kirshtein wrote:
I'm just gonna jump in here and emphasize that clock buffer.. It need to be done
Phil, I cannot see any comment from Hans when he changed this. Did I miss it? Is it any more than just a simplification, one IC instead of discrete components? 73 Alan G4ZFQ
|
|
|
That was a 3A to 4 revision of Hans's -- he used a single gate nand. Might also be referred to a 'clock drive' or 'popping'. It's basically a signal integrity issue between the TCXO and the Synthesizer which caused the synthesizer to occasionally go out of lock for an instant.
|
|
|
On 23/03/2023 15:36, phil kirshtein wrote: It's basically a signal integrity issue between the TCXO and the Synthesizer which caused the synthesizer to occasionally go out of lock for an instant. OK, but I thought that was rectified by a firmware update. I thought I'd remember if it required a hardware update. 73 Alan G4ZFQ
|
|
|
On 23/03/2023 15:36, phil kirshtein wrote: That was a 3A to 4 revision of Hans's -- he used a single gate nand. Yes I found it. "In Rev 4 of the PCB, a single NAND logic gate chip is used to buffer the 1.1Vpp signal from the 25MHz TCXO " The firmware hid the effect, the addition of the IC cured it. Thanks. 73 Alan G4ZFQ
|
|
|
From Hans's list below:
4) On the bottom side, the tiny single-gate logic chip is installed in place of the earlier NPN/PNP transistor buffers.
|
|
|
On 23/03/2023 16:19, phil kirshtein wrote: From Hans's list below:
4) On the bottom side, the tiny single-gate logic chip is installed in place of the earlier NPN/PNP transistor buffers. Phil, Yes, I've got that far. (Except from seeing Hans' list below?) What would be useful is a picture of a modified QDX, would help anyone wishing to make the change. 73 Alan G4ZFQ
|
|
|
