Help needed diagnosing newbuild QCX+


Leo
 

Visible symptom: line of black characters at top of display. There is the expected seashell whisper in earphones.
Checks performed: ATmega seated okay. Solder joints checked visually and by continuity from lead to pad and pad to next circuit point. Coils are good. Checked all ground solder joints with respect to JP6 ground (and others where convenient.
Voltages (where they differ significantly from expected I have coloured them red):

+ve supply terminal: 11.26 (11.98) (my battery needs charging)
Clk0: 0.9 (~1.7)
Clk1: 0.9 (~1.7)
lhs C2: 3.40 (3.40)
rhs D2: 4.94 (4.96)
T1 pins 1,2,3 & 4: 0 (0)
T1 pins 5,6,7 & 8: 2.47 (2.36)
C43: 2.47 (2.28)
C44: 2.47 (2.44)
C45 1.3 (2.42)
C46: 0.87 (2.30)

IC5 pin 1: 2.47 (2.64)
IC5 pin 2: 2.47 (2.46)
IC5 pin 3: 2.47 (2.26)
IC5 pin 4: 0 (0)
IC5 pin 5: 0.89 (2.28)
IC5 pin 6: 1.33 (2.46)
IC5 pin 7: 1.38 (2.44)
IC5 pin 8: 11.23 (11.67)

I haven't recorded the voltages for the subsequent opamps as the discrepancies carry through.

It seems that the output of the FST3253 is throwing the voltages out in the receiver section.

IC2 appears to be non-functional and this may be why the SI351A is not outputting the correct Clk on 0 & 1.

I measured the voltages on each of the IC2 pins without and with the chip in place and they would seem to indicate something is going on, but it is not communicating with the display.:
pin 1:   0        4.6
pin 2:   4.93   4.92
pin 3:   4.93   4.92 
pin 4:   4.93   4.92
pin 5:   4.93   4.92
pin 6:   0        0
pin 7:   4.94   4.93
pin 8:   0        0
pin 9:   0        2.42
pin 10: 0        2.45
pin 11: 1.08   1.08 
pin 12: 0        0
pin 13: 0        0
pin 14: 4.94   4.94
pin 15: 1.93   1.93
pin 16: 0.01   0.02
pin 17: 0        0
pin 18: 0        0
pin 19: 0        0
pin 20: 0        4.15
pin 21: 0.1     0.01
pin 22: 0        0
pin 23: 2.45   2.45
pin 24: 1.93   1.93
pin 25: 0        0
pin 26: 0        0
pin 27: 3.47   3.47
pin 28: 3.40   3.40

The display voltages are:

LCD pin 1:   0      (0)
LCD pin 2:   4.93 (4.94)
LCD pin 3:   0.87 (0.57)
LCD pin 4:   3.48 (4.92)
LCD pin 5:   0      (0)
LCD pin 6:   0.04 (0)
LCD pin 7:   4.92 (1.08)
LCD pin 8:   4.92 (1.08) 
LCD pin 9:   4.92 (1.06) 
LCD pin 10: 4.92 (1.04) 
LCD pin 11: 4.92 (0.00) 
LCD pin 12: 4.92 (4.94) 
LCD pin 13: 4.92 (4.92) 
LCD pin 14: 4.92 (0.00) 
LCD pin 15: 4.92 (4.41) 
LCD pin 16: 0      (0) 

There's a conclusion to be reached here but it is eluding me.
I'm about to embark on another component by component check of value, orientation and solder joint. It's probably worth saying that I measured every single resistor before I placed it.
I haven't had the scope out yet as there are clearly at least a couple of fundamental issues to chase.

Hopefully this will all ring a bell with somebody and it will boil down to a bad joint or misplaced component or another face-palm moment. That's okay - I'm a programmer so debugging has taught me that no matter how many times you look at something, you still miss the obvious. 

Help would be much appreciated.


Alan G4ZFQ
 

Leo,
Visible symptom: line of black characters at top of display.
Unfortunately you start here. http://qrp-labs.com/qcxmini/trouble.html#blank Most of the rest of your tests need to be done when the LCD is reading correctly.

73 Alan G4ZFQ


There is
the expected seashell whisper in earphones.
Checks performed: ATmega seated okay. Solder joints checked visually and by continuity from lead to pad and pad to next circuit point.


Leo
 

Thanks.
Naturally I have read the troubleshooting sections including the page you linked, which is why I included a reasonably exhaustive list of actions and measurements taken. Unfortunately the remedial actions had no effect which is why I am asking here.


Alan G4ZFQ
 

Naturally I have read the troubleshooting sections including the page you linked,
Leo,

OK but you did not say you looked there.
It is essential that that stage is reached. Nothing will work until you have the correct display. There must be something wrong in that little section.
Have you checked the 20MHz crystal?

I do not know how good you are but check the solder joints, one of the most frequent causes of problems.

73 Alan G4ZFQ


which is why I included a reasonably exhaustive list of
actions and measurements taken. Unfortunately the remedial actions had no effect which is why I am asking here.


N3MNT
 


Check to make sure the connections between the front panel and the main board are correct  Looks like the connector may be misaligned.  Also check IC4 orientation.
Photos  may help.


Leo
 

Sorry Alan, I thought the nature of the testing and that I'd mentioned the seating of the ATmega inferred that I'd read around the problem. Not to worry - the link is there for others.

Thanks N3MNT - that was the second thing I checked!

Okay, it was the face palm I was looking to precipitate by posting in here.
After a mega (pun intended) magnified inspection of all of the solder joints, I noticed that one of the IC2 joints didn't appear to have a pin poking out of the top. A quick continuity check top to bottom followed by a careful inspection looking between the socket and the board revealed that one of the pins had bent over.
I pondered this for an hour or so, started to desolder then decided to try some micro surgery on the socket instead. After clearing the hole, I managed to feed a bit of offcut signal diode lead up behind the clip where I soldered it at the top of the socket. I needed to bend that pin on the IC a little to clear the now displaced clip, but it all works!

QCX+ @ 30m is functional. Now I just have to wait for the band to open which it never seems to do in the north of England.

It is a little microphonic so I'll have to open it up to check there are no overlong leads contacting the case and perhaps hot glue the coils once I've tuned the power.


Alan G4ZFQ
 

Leo
Sorry Alan, I thought the nature of the testing and that I'd mentioned the seating of the ATmega inferred that I'd read around the problem.
Sorry, I think I got close referencing you to " it usually turns out to be OUR MISTAKE, "

73 Alan G4ZFQ


Rod Smith
 

Hi Leo, I'm glad you've sorted your QCX issue.  Do you happen to know which pin on your IC2 was a bad joint please?

I'm struggling a bit with my QCX mini, where I plucked disaster from the jaws of success when trying to find the source of intermittent crackles after replacing a faulty IC7, and would be grateful for any advice or guidance. 

I'm seeing remarkably similar voltages to yours, around IC5 and C45, C46.  (Clk0 and Clk 1 are fine, though, at ~1.7V) and I've been working to decide whether I should try replacing the FST3253 or whether perhaps the IC5B (and onwards) biasing is being messed up by something else.

I don't have any problems with my display though or, so far as I can tell, the menu.  Also the Si5351 seems to be providing good output - definitely at the right frequencies.  I get a definite but poor (4) amplitude bandpass peak and,of course, no IQ balance.  The bias from T1 is OK at the input to the FST3253 but then seems not to be making it to the other side for IC5B.

My guess is that your IC2 poor connection meant you had temporarily no output from the Si5351 - and that I may have permanently low output.  But I hope I'm wrong!!

Best wishes & good DX on 30

Rod
G0VKX


Leo
 

I think it was pin 20, but any pin would probably cause similar problems. 

It sounds like your QCX mini problem is with the opamps - well covered in this forum. Check all of the pin voltages. 

The FST3253 is, apparently, hard to obtain, but I bought two at the weekend from a UK eBay dude for a couple of pounds each. Opamps are easier to get hold of - I bought 10x original spec from China, took about a week and also 5x NE5532 from UK on eBay. I’m planning a scratch rebuild of my mini after the regulator went bad and fried the board. 

Whatever anyone says this has been a very educational journey and I’ve enjoyed (almost) every minute. I even ordered a Hantek DSO2D10 scope from AliExpress at a 50% saving. At least it will be if it ever arrives. Now I need to improve my 30m antenna. WSPR doesn’t tell a great story even when the QCX is pumping out 5W. 


geoff M0ORE
 

If you are having no success with WSPR on 30m running 5 watts, you have a problem. Check you are on the correct frequency, check your clock time, check you have the transmit cycle starting at the beginning of a even number of minutes. Any errors in these parameters will result in no WSPR decodes.

At the time of writing, 14:50GMT, the WSPR map is covered with decodes.

On 25/01/2021 14:40, Leo wrote:

I think it was pin 20, but any pin would probably cause similar problems. 

It sounds like your QCX mini problem is with the opamps - well covered in this forum. Check all of the pin voltages. 

The FST3253 is, apparently, hard to obtain, but I bought two at the weekend from a UK eBay dude for a couple of pounds each. Opamps are easier to get hold of - I bought 10x original spec from China, took about a week and also 5x NE5532 from UK on eBay. I’m planning a scratch rebuild of my mini after the regulator went bad and fried the board. 

Whatever anyone says this has been a very educational journey and I’ve enjoyed (almost) every minute. I even ordered a Hantek DSO2D10 scope from AliExpress at a 50% saving. At least it will be if it ever arrives. Now I need to improve my 30m antenna. WSPR doesn’t tell a great story even when the QCX is pumping out 5W. 


Leo
 

Thanks Geoff, I had a bunch of hits this morning, but all near Europe except for on eastern USA. Then I altered that params - I may well have set an odd start time. Anyway this evening’s project is to scope the power output and experiment with the CAT. 

Leo M0NNQ


Rod Smith
 

Thanks very much Leo -

I've checked everything I could think of around the SMD ICs.  They were fine before I "fried" the voltage regulator (which was after I replaced a bad IC7); after,though, the bias to IC5, 6, 8 (but not 7) were v close to what you had.  I only feed 7.5V ish to the regulator when first testing and aligning - so that's what arrived on the 5V rail after fry-time; it's hard to believe that would give the audio ICs a problem, 'though I do feel lucky the processor and ACT00 came though it OK.  And surprised the 5351a and 3253 are still functioning at least to some extent.  The IC5 and onwards bias voltages derive from the 5V at the T1 CT winding so I'm still suspecting the FT3253 or 5351a.

A couple of the FST3253 arrived from (excellent) RFplus today so .. I''ll have one more go around the existing wisdom on the forum about audio ICs, poke about the mini a bit more and then contemplate whether to (a) replace IC6 (with my one remainingTS9221D) even though it was working fine before or (b) replace the FT3253.  I'm suffering from one-step-forwards-two-steps-back-itis lately so I'm tending towards look very carefully before  leaping!

Absolutely these journeys are educational and enjoyable.  They can also be frustrating and time consuming but that's just how things are sometimes.  No complaints from me!

Good luck with the antenna.

Rod
G0VKX




Leo
 

Your sequence of events sounds remarkably similar to mine. 

After checking Farnell I discovered they carry chips such as the SI5351 for a fraction of the China prices, so worth checking there before hitting eBay. They do next day delivery but I’ll have to find another £28.04 of goodies before the shipping is free. 

I have a comatose FT-101 on the bench which will live and breathe again, so this exercise in checking, swapping and measuring at microscopic level will pay dividends when I return to working on what is a tractor by comparison. 


After checking the power on the QCX+ I couldn’t get it above 1 Watt. As a sanity check I scoped my Icom 718 and it seems that on 10W it scopes the same as the QCX, so a factor of 10 error somewhere it would seem. I’m confident in the 718’s output as it was serviced by Icom last autumn, which means at 12v supply the QCX is pumping out 10W. I’m confused. Back to WSPR!

Leo M0NNQ


Alan G4ZFQ
 

autumn, which means at 12v supply the QCX is pumping out 10W. I’m confused.
Leo,

Measure the extra current taken by the QCX on TX.
Calculate the wattage. VxA
The maximum TX power can not be more than 90% of the increase from RX to TX.
10 watts at 12 volts seems very unlikely, 1 watt is not unknown..

73 Alan G4ZFQ


Leo
 

I hadn’t realised but my 3253s came from RFplus too. And he threw in a spare!


Rod Smith
 

He threw me a spare too!  Great touch (and hope I don;t need it!)