Topics

Help with strange audio fault #qcx


Ted 2E0THH
 

Blimey Justin!

Really pleased that it sorted your problem. My QCX has been powered up virtually every minute since and works like a charm.
I've gone up a license class since and just waiting to sit my advanced.
However I would never have solved the problem without all the help and guidance from the kind souls on this forum.

73s Ted
2E0THH  


justin.munger@...
 

Ted (and others helped resolve this problem),

I have just built a QCX+ a few weeks ago and after it working for 2 weeks, i experienced the same exact problem you described.  After a few days of troubleshooting on my own, I posted the problem in the QCX group and someone recommended I make the post in QRPLabs group.  I'm glad I searched through all the posts and found yours.  All my troubleshooting steps fell in line with yours--even all the voltage readings on IC10.  I currently have a 1uF electrolytic capacitor in place of the ceramic one until i can get a replacement.

Thanks for leading the way with solving this problem.

Cheers!

--Justin
KN6IPA


Ted 2E0THH
 

Thanks Jerry
I would normally have opted to clip the component up but there was a risk of collateral damage to adjacent components so I opted for my lasso method I have used many times in the past. You just have to be careful to apply gentle persuasion else you risk pad damage. 
Many thanks for the hints, very useful indeed.
73s Ted
M7ECH


Jerry Gaffke
 

Congratulations on fixing it!

Unless you need to save the part, on throughole it's easiest to just clip the part away,
and then pluck the the wires out one by one.  Then wick out the solder, or heat and blow it clean.
Solder is heavy, a bit easier if you use gravity to wick or blow the solder down not up.
Though no real need to clean the old wires and solder out of the holes,
could just solder the new part to the nubs.  (Makes it easier to remove the next time. ;-)
If you want to save the part, use two irons.

For surface mount R's and C's, heat both ends with two irons,
lift straight up off the board by holding the part between the irons.
For a 3 pin surface mount part, first wick away most of the solder, then heat and
lift the single pin on the one side, now use two irons on the remaining 2 pins.

For a nasty pin surface mount SOIC or the like, use a hot air gun of about 250 watts
and ever so gently pluck the part up with tweezers.
A $20 embossing gun from a hobby shop should work fine.

Jerry, KE7ER


On Wed, Jan 16, 2019 at 03:44 PM, <qrp@...> wrote:
It took about an hour to gently tease C22 out of the board. After I sucked away most of the solder, I tied some cotton around the component to apply very light tension as heated each pad in turn until it finally gave up. Note that it is almost impossible to avoid some pad damage but I think I have gotten away lightly.


Ted 2E0THH
 

I am delighted to announce that as 15 mins ago, my QCX40 is alive again!

It took about an hour to gently tease C22 out of the board. After I sucked away most of the solder, I tied some cotton around the component to apply very light tension as heated each pad in turn until it finally gave up. Note that it is almost impossible to avoid some pad damage but I think I have gotten away lightly.

I temporarily replaced it with a 100nF (until the 1uF caps arrive at the weekend) and fired it up. BINGO!

I am deeply indebted to everyone who has offered advice and guidance, but especially Alan G4ZFQ and James W4JED without whom I surely would still be floundering.

My apologies for the ridiculous bandwidth.

73s Ted
M7ECH


Roland Williams
 

Tantalum capacitors can work for a surprisingly long time when installed backwards. Imagine the havoc this causes when you get it on a production line; I’ve experienced this first hand. In a signal path the deterioration may be un-noticed for ages even though it’s measurable easily. If they’re across a power line, though, the results can be spectacular, with noisy operation before they fry. 

By contrast, the standard aluminum electrolytics die very quickly when reversed. 

Roland AE6VL 

On Jan 16, 2019, at 10:23 AM, qrp@... wrote:

Brilliant Alan
I'll try that tomorrow and let you know.
Have a good evening
Ted M7ECH


Ted 2E0THH
 

Brilliant Alan
I'll try that tomorrow and let you know.
Have a good evening
Ted M7ECH


Ted 2E0THH
 





Hi Alan

Many thanks for your input.
Just 105 as far as I can tell. If they were polarised and in the wrong way, I can't wee how the set could work flawlessly for 3 weeks.

Best Wishes
Ted M7ECH



Alan <g8lco1@...>
 

Tantalum bead caps will have a dot, line or + to indicate polarity , the bead is round under the encapsulation. Your parts look like ceramic plates that have been dipped so will be non-polarised. What does the printing say?

 

Alan

G8LCO.

 

Sent from Mail for Windows 10

 

_._,_._,_

 


Alan G4ZFQ
 

is any other value i could use to test we are down the right rabbit hole?
Ted,

For testing anything over say, 0.1uF 100n.
It just has to pass some sort of audio for ever instead of 5 seconds ;- ) An even lower value would most likely produce adequate results.

73 Alan G4ZFQ


Ted 2E0THH
 

The image of my c21 & C22 didn't display for some reason so here it is again.
The 1uF caps I ordered should be here on Saturday so fingers crossed, I will extract C22 tomorrow.
Once I have it out, I wondered if there is any other value i could use to test we are down the right rabbit hole?
I have a few odd caps lying around.
Best Wishes
Ted M7ECH


Chris Wood
 

To answer the questions posed: 

Alan:  yes, they have a + mark as the photo you added

Ted:  yes, they are yellow/buff coloured, see Alan's photo;  fit with lead marked + on inboard side, so adjacent to R38, other lead adjacent to earphone socket.

That's what I've got in mine and it is now working.  Doesn't mean yours is the same, but I hope this might help.

Chris
G4CWS



Alan G4ZFQ
 

Alan may remember that I had a similar fault with my QCX 30m, and described in earlier posts.
Chris,

Maybe I remember vaguely:-)

I think the
ones supplied in my kit are actually tantalum and are polarised.  At least they look, for reference, identical to the part sold by CPC as MCCB1V105M2ACB which is a 1u 35v tantalum.  They do not look like standard 105 ceramics, which are less elongated and marked differently.
The Farnell illustration shows a + mark. Are they like that?

73 Alan G4ZFQ


Ted 2E0THH
 

Hi Chris

I am pretty sure these are ceramics:



Are your tantalum caps actually yellow? I didn't notice any polarisation markings during construction, but I have to confess I wasn't looking for any. Did your set work at all Chris? I just checked the construction manual V4 and Hans lists C21 and C22 as ceramics.

Now I'll have a careful look!

Many thanks for your help.

Best Wishes

Ted M7ECH 


Chris Wood
 

Hello Alan, Ted,

Alan may remember that I had a similar fault with my QCX 30m, and described in earlier posts.

My kit was version 4 and the capacitors supplied for C21and C22 are indeed 1u instead of 10u.  The manual says they are ceramic, but upon further inspection since building, (and checked just now) I think the ones supplied in my kit are actually tantalum and are polarised.  At least they look, for reference, identical to the part sold by CPC as MCCB1V105M2ACB which is a 1u 35v tantalum.  They do not look like standard 105 ceramics, which are less elongated and marked differently.

I assumed they were ceramics when fitting them because they are the same colour as the others, and I was expecting them to be ceramics, but that was my mistake.

So I now think my problem was that I had one or both of these reversed, and I corrected the error unwittingly when I removed and refitted them, producing a working QCX.

Obviously Ted's kit may have had ceramics supplied, not tantalums, but I do again recommend checking.

Best wishes

Chris
G4CWS

 


Alan G4ZFQ
 

Are you implying that C22 might be faulty? If so I better get one ordered.
Ted,

Assuming the PCB is clean it seems by far the most likely fault.
Funny I thought that before I realised it was not an electrolytic.

73 Alan G4ZFQ


Ted 2E0THH
 

Hi Alan

I am out on call right now but was just reading more carefully you last email on my phone.
>> If pin 3 IS connected to pin 7 then R38 +95K=215K confirming a leak to ground at pin 6, possibly through C22

Are you implying that C22 might be faulty? If so I better get one ordered.

Huge thanks

Ted
M7ECH


Ted 2E0THH
 

Good morning Alan & James

I have snipped the top end of R37 as you suggested Alan.
I measured both resistance scenarios, the non snip measurements are somewhat at variance to those of yesterday, I have no idea why.
No R37 snip
8) 1371
7) 273k
6) 154K
5) 5.3K
4) 0
3) 273K
2) 30M
1) 30M

R37 snipped
8) 1371
7) infinite
6) infinite
5) 5.3K
4) 0
3) infinite
2) 30M
1) 30M

I haven't had a chance to try any of the foil bridging yet, I will get a chance to try later this afternoon after I fix the snip and reinsert IC10

73s Ted
M7ECH


Ted 2E0THH
 

Dear James and Alan

I am so sorry. I was on a call out so that last message was rather garbled.
As well as checking for dry joints I also tried bridging in case there was a break in the foil either way off R38.
I only had time to try it with without soldering so that it my next task.
I will also try the same with R39.

I don't have any spares of the op-amp but I already tried swapping IC9 and IC10 with others on the board (I can't remember which) but the fault remained and repeated. I think I will order some in any case.

Alan, I rechecked the resistances on the empty IC10 socket
8) 1371
7) 240K (yes I can't type Alan)
6) 118K (I will check this later with one end of R37 removed)
5) 5.3K 
4) 0
3) 240K
2) 27M (rising)
1) 27M

And voltages in the empty IC10 socket
8) 12
7) 0
6) 0
5) 6
4) 0
3) 0
2) 0.9
1) 0.9

Right, I will try and see if any foil bridges work and then get some sleep. I should get stuck in properly again tomorrow afternoon (GMT)

Again, my sincere thanks gentlemen.

73s Ted
M7ECH


James Daldry W4JED
 

Hi, Ted

Where did I write the word "solder"? My guess is broken foil on the ground end of R39. While you have IC10 out of the socket, check the voltage on pin 5. If the battery is 12 volts, it should be 6. If it is 6, try swapping the IC with one of the others of the same type. Easy way to eliminate the chip as a suspect. After all, if a chip is going to fail without smoke, it will probably happen during the first few months of use.

Jim

On 1/15/19 1:53 PM, qrp@... wrote:
Gentlemen

I had another solder around r39 but to no avail.
I did socket the OP amps when I built the QCX40 so I just removed IC10

These are the values in ohms from the socket holes to ground
8) 1371
7) 216
6) 95k
5) 5.3K
4) 0
3) 217k
2) 15M (climbing as I read it) 
1) 15M connected to 2

I have been called out for a few hours so will jump back in at 2200h.
I am immensely grateful for all help.

73s Ted
M7ECH