30m QCX working – but not for long....


Chris Wood
 

Good morning

I’ve recently finished building my 30m QCX, board rev 4.  Completed alignment ok (after fixing poor joint on T1).  4w output at 13v and had a brief ‘599TU’ style qso with a Spanish station.


So the good news is that it has worked ok, apart from very loud clicks on power on, power off, first press of buttons and first key closure on tx/rx changeover. 

Then the rx died, leaving just a faint ticking and the sidetone faint and distorted.   In ‘rx dead’ state, tx works fine, clean keyed output, menu selection also works and ticking disappears in menu mode.

Switch off, leave for a minute, switch on, no change.  Left overnight, switch on, worked for 3 or 4 minutes then died.  This time ticking, but no sidetone audible at all.  Left overnight again, switch on again, no resurrection this time.  Nothing.

Injecting audio signal at gain pot wiper has no effect, nor does injecting signal from ‘Peak BPF’ alignment process.

DC voltage checks in ‘rx dead’ state with supply voltage set to 12v (per manual section 6.4, but using external DVM for measurements): 

Rx current 120 mA working and 100 mA with dead rx.

Clk0 and Clk1 both show ~1.7, but no nasty noise audible.  T1 pins 5, 6 and 8 also C46 ok.

IC5 to 10 pins 3 and 5 at same voltage as pins 1/2 and 6/7.  Perhaps not surprising if rx is showing no sign of life. 


Probably entirely unrelated, but LCD pins 7-10 show ~4.96v not ~1.08.

I’ve examined the board with a good magnifying glass and by taking photos enlarged on my pc screen.  Soldering looks ok, no shorts evident.

So something is amiss in the rx world, possibly a poor joint somewhere.  I’ve reached the end of my meagre investigative skills, and would very much appreciate the advice of others. 

Thanks

73 Chris G4CWS


Alan G4ZFQ
 

Injecting audio signal at gain pot wiper has no effect,
Chris

This seems to show that something is wrong after the pot.

Q7 failed? Shorting to ground.

And if not give us the actual voltages at Q7 and IC10

Left overnight, switch on, worked for 3 or 4 minutes
This makes me think of bad connection somewhere, but maybe not.

Probably entirely unrelated, but LCD pins 7-10 show ~4.96v not ~1.08.
Not sure, possibly due to good meter, not the integral one that is low impedance, that's what the troubleshooting gives.

73 Alan GZFQ


Miroslav Sperlin
 

Check polarity of electrolytic capacitors.
73 Mirek OK2BUH


Glen Leinweber
 

Can you get CW decoded on the LCD while transmitting?
If so, then that leaves one op-amp (and stuff connected to)....IC10A.
If CW doesn't decode (and the menu item is ENABLED), then
everything past the gain pot (Q7, IC10B, IC10A) is suspect. I'm pointing
out the active devices, not to point suspicion to them, but to broadly
point out the stages - all the parts associated with these stages are suspect.

You have an intermittent - a nasty thing to debug.
I might suspect the earphone jack area. If this thing is overheated during
soldering, it can distort and its spring pressure relaxed...the clearances
inside are tiny.


Chris Wood
 

Thank you all for your suggestions.  To respond to each:
- Mirek - yes, electrolytics are fine

- Glen - the decoder does decode on the LCD when transmitting.  Earphone jack looks ok, and it did work at first.  I did hear a power on click today, so something is getting as far as the earphones.

- Alan - actual voltages IC10 pins 1/2/3 10.8v, pin 4 0v, pin 5 9.6v, pin 6 9.7v, pin 7 11.0v, pin 8 11.6v;    More interesting perhaps, Q7 D 0.4v, G 0v, S 0v.    S is grounded anyway.

So could the problem be a permanently muted rx? Looking at the circuit, there doesn't appear to be much around Q7 which could be a more likely suspect.  Could Q7 fail gradually?  I had assumed, maybe wrongly, that it either works or it doesn't.

Conveniently, I do have some spare BS170s from another project.  Sod's Law applies of course in that the solder pads for Q7 are under the gain pot, making resoldering or removal a bit of a game!  They are not shorting to the pot above nor do they appear to have any short between them. I did look quite carefully before soldering the pot over the top, for exactly that reason.

Any further checks I can do before pulling the board apart?  Of course, that may not be the only issue, but at least it's a start.  Thank you for that.

73 Chris G4CWS


Chris Wood
 

Further to my last message, I've just found a trail of comments concerning CLK0 /CLK1, where Hans confirms that CLK1 is not active if the receiver is muted.  My voltage measurements seem to show both CLK0 and CLK1 are active in my 'rx dead' state, which would seem to suggest that the issue isn't to do with muting.  Or maybe I have misunderstood. 

Not sure what to think...

Chris


Alan G4ZFQ
 

- Glen - the decoder does decode on the LCD when transmitting.
I do not know the significance of that, maybe Glen can expand his comments?

Chris,

First a reply to your following post, I think the state of the clock outputs is correct, I do not see how they would affect your stated symptoms anyway.
- Alan - actual voltages IC10 pins 1/2/3 10.8v, pin 4 0v, pin 5 9.6v, pin 6 9.7v, pin 7 11.0v, pin 8 11.6v;
Are you looking at the manual? These voltages do not match. IC10 pin 5 should be fixed by a potential divider, something around 6V if measured with a DVM, Hans measured 4.2 with the inbuilt meter.
Check R39, 40 and joints around here.

More interesting perhaps, Q7 D 0.4v, G 0v, S 0v.
Shows it is not actually turned on. I'd expect the drain to be 0V. Have you checked the resistance to ground here? And you are positive C21, 22 are the right way round?

Could Q7 fail gradually?  I had assumed, maybe wrongly, that it either works or it doesn't.
Usually that is correct, but we are looking for a fault, just as likely to be a bad joint as actual component failure. The intermittency could be a bad joint but intermittent components are not unknown.

73 Alan G4ZFQ


Glen Leinweber
 

I have made a guess regarding QCX Morse decoding.....
Am guessing that Hans does receiver decoding and transmitter decoding the same way:
Audio is picked off at IC10B pin 7 and goes into analog-to-digital converter IC2 pin23.
From there, it is digitally processed in a 700 Hz bandpass filter, and its amplitude detected.
Am guessing that if transmit 700 Hz sidetone is Morse-decoded, then audio is present at IC10B
output - this is very near the receiver's output.

I may be wrong: Hans may have pumped the key signal into the Morse decoder, and
in that case, audio digital signal would be ignored.
If my guess is right, then that leaves one stage to troubleshoot: IC10A

Alan's troubleshooting suggestions (added below)  are right-on. Those 10V measurements on
IC10B seem high. A resistor value wrong? Those 10K resistors look similar to 1K resistors.
---------------------
Alan G4ZFQ wrote:
Are you looking at the manual? These voltages do not match. IC10 pin 5 should be fixed by a potential divider, something around 6V if measured with a DVM, Hans measured 4.2 with the inbuilt meter.
Check R39, 40 and joints around here.
[…]
About Q7:
Shows it is not actually turned on. I'd expect the drain to be 0V. Have you checked the resistance to ground here? And you are positive C21, 22 are the right way round?


Hans Summers
 

Hi Glen

On transmit the decode input signal is obtained from the key. Not from the 700Hz tone. There is no need to go through the whole detection stage, with amplitude variation tracking etc., when the keying state is already known... 

If the iambic keyer mode is being used, the decoder also skips the stage which tracks keying speed variations, since the keying speed is already known. 

73 Hans G0UPL

On Sat, Sep 22, 2018 at 11:49 PM Glen Leinweber <leinwebe@...> wrote:
I have made a guess regarding QCX Morse decoding.....
Am guessing that Hans does receiver decoding and transmitter decoding the same way:
Audio is picked off at IC10B pin 7 and goes into analog-to-digital converter IC2 pin23.
From there, it is digitally processed in a 700 Hz bandpass filter, and its amplitude detected.
Am guessing that if transmit 700 Hz sidetone is Morse-decoded, then audio is present at IC10B
output - this is very near the receiver's output.

I may be wrong: Hans may have pumped the key signal into the Morse decoder, and
in that case, audio digital signal would be ignored.
If my guess is right, then that leaves one stage to troubleshoot: IC10A

Alan's troubleshooting suggestions (added below)  are right-on. Those 10V measurements on
IC10B seem high. A resistor value wrong? Those 10K resistors look similar to 1K resistors.
---------------------
Alan G4ZFQ wrote:
Are you looking at the manual? These voltages do not match. IC10 pin 5 should be fixed by a potential divider, something around 6V if measured with a DVM, Hans measured 4.2 with the inbuilt meter.
Check R39, 40 and joints around here.
[…]
About Q7:
Shows it is not actually turned on. I'd expect the drain to be 0V. Have you checked the resistance to ground here? And you are positive C21, 22 are the right way round?


Chris Wood
 

Good morning all,

Success!  Alan - if your shack window was open you may have heard the shout of joy from Lymington.

Checked C21 / 22, which are different components on v4 pcb, 1uF blob capacitors, but still polarised - they were the right way round. 
Checked R39/40, which were the correct values, resoldered just in case, and then got more sensible voltages at IC10 pins 5 (5.2v) and pin 6 (4.6v) but still no output. 
So went round the whole board and resoldered everything in sight. 

Bingo.  So I guess I need more practice at soldering.  Thank you to all for your help.  Much appreciated. 

Looking forward to getting my hands on a QSX, whatever colour it may turn out to be. 

Thanks once again.

73  Chris G4CWS


Alan G4ZFQ
 

if your shack window was open you may have heard the shout of joy from Lymington.
Chris,

No, a little too cool and rainy.

Bingo.  So I guess I need more practice at soldering.
Yes, me too. Must be over 60 years since I first soldered and still I get bad joints...

73 Alan G4ZFQ