4046 Oscillators progressively getting worse over time

fonik · Posted: Fri Dec 07, 2012 1:12 pm Post subject:

it is wicked as he said. all 4 VCOs show the same weird misbehaviour, on two different PSUs...

so IMHO this is not an error n the VCO builds. 4 times the same single error i mean. it has to be something general that applies to all 4 builds, or to both PSUs. wrong parts, wiring...
that it is getting worse over time is irritating. it is not a working/not working situation.

the electrolytics are the 1st promising scent, though.

LektroiD · Posted: Fri Dec 07, 2012 1:21 pm Post subject:

Blue Hell · Posted: Fri Dec 07, 2012 2:02 pm Post subject:

LektroiD · Posted: Mon Dec 17, 2012 1:32 pm Post subject:

Ordered new electro-caps, 10µF/50V. Still have the same problem. Maybe it was luck that it stabilised for a few hours after replacing them last time.

Could the obscure waveshape be related to the fluctuations in pitch?

Could it be the 4046 chip itself? I have no others left to try it with.

I have no idea what else to try.

I'm at a loose end here, so any suggestions are welcome.

fonik · Posted: Mon Dec 17, 2012 3:32 pm Post subject:

this is so awkward!

i have had one bottle of tempranillo tonight, and i have the feeling that the reason is so simple, yet hidden. as soon as it became obvious everyone would say: so clear! what else!
alas, we are still one step before that revalation...

what 4046 are you using? maybe i should mail you a good NXP one?

LektroiD · Posted: Mon Dec 17, 2012 5:36 pm Post subject:

fonik · Posted: Tue Dec 18, 2012 6:23 am Post subject:

check PM

JovianPyx · Posted: Thu Dec 20, 2012 9:26 am Post subject:

Just a thought about electrolytic caps, I had read some years ago about a Chinese manufacturer that had gotten stolen manufacturing process plans from an American mfr, but the stolen plans were missing one page of instructions. The capacitors were made without one of the required chemical processes. These capacitors would work - for a while and then fail. I'm wondering if somehow some old stock of those is still running around or if the plans were sold to someone else? The idea of a different mfr sounds like a good idea. And also the idea of a different stock of 4046. The X-4046 design ought to be quite solid coming from TH. Please let us know what the final answer is.

fonik · Posted: Thu Dec 20, 2012 9:50 am Post subject:

LektroiD · Posted: Thu Dec 20, 2012 10:56 am Post subject:

Blue Hell · Posted: Thu Dec 20, 2012 11:28 am Post subject:

LektroiD · Posted: Sun Dec 30, 2012 8:50 am Post subject:

I've just got back from visiting my family and arrived to find a package with a 4046 waiting for me (many thanks Matthias). I eagerly got to work replacing the chip with the new NXP model, and guess what...

*drum roll*

The problem still persists...

On the bright side, it has eliminated one more possible cause, but I really have no idea what else to try now...

Blue Hell · Posted: Sun Dec 30, 2012 8:54 am Post subject:

This is quite weird all Confused

LektroiD · Posted: Sun Dec 30, 2012 9:07 am Post subject:

I'm now thinking about replacing the hand matched transistors with a SSM2210...

fonik · Posted: Sun Dec 30, 2012 9:09 am Post subject:

maybe four eyes see more than just two? you could mail me one of these troublemakers...

LektroiD · Posted: Mon Dec 31, 2012 12:12 am Post subject:

prgdeltablues · Posted: Wed Jan 02, 2013 6:50 am Post subject:

Am I right in understanding that originally, all 4 oscillators were stable, but have become unstable over time? And that the instability manifests as a minor random variation in pitch(not eg four octave jumps).

I couldn't hear any periodicity in the variation in the sample you posted. Does an FTT analysis show any - could you compare one with that for a stable VCO? (I'm looking to whether the problem is random - which would suggest, though not confirm, a noise issue, or has any periodicity, which might point something thermo-mechanical?

Can you run a module off a battery supply - to eliminate noise either in either of your PSUs or the mains as a cause. (White noise on the ground suggests itself?) Also, take it somewhere else, turn off as much nearby mains equipment as possible. (To check it isn't caused by something in the local electric environment)

I assume you've checked the current draw isn't excessive?

The gradual development of the problem is really puzzling. And it sounds as if it developed at different times, and different rates, in all four modules (so it's not, for example, a deterioration in your mains supply).

I don't see that it can be a problem with soldering, affecting four boards the same way. What about the board material itself - could that have changed over time (baked in the long hot UK summer?)

Sorry, I've no real thoughts to offer, but maybe this brainstorming might trigger something.

Peter

fonik · Posted: Wed Jan 02, 2013 9:59 am Post subject:

emmaker · Joined: Aug 10, 2007 Posts: 12 Location: Portland, OR

A few stupid questions.

Are you using some form of VC to get the freq or are you just using the course and fine tune knobs to set the freq with no VC? What I'm trying to say is if you're using VC to control the oscs are you sure that isn't bad?

Have you added any digital modules that have high freq clocks and signals. Could be some form of EFI or PS noise. Or did you add any modules about the time things went south. Might try disconnection them.

Have you checked the sync? Might be getting some random sync pulses that are resetting the osc.

Good luck.
Jay S.

LektroiD · Posted: Fri Jan 04, 2013 2:26 pm Post subject:

emmaker · Joined: Aug 10, 2007 Posts: 12 Location: Portland, OR

LektroiD · Posted: Sat Jan 05, 2013 6:06 pm Post subject:

My scope is analogue. I'm wondering what component would fail to cause a sync issue. It'll probably be easier just replacing that, see it sorts the problem...

Mongo1 · Joined: Aug 11, 2011 Posts: 411 Location: Raleigh NC

I think you should be able to ground that sync line to get it out of the picture

Gary

LektroiD · Posted: Sun Jan 06, 2013 1:36 pm Post subject: