Diagnosing temperature related noise?

[Atmosferit]

A few months ago I posted about a Yamaha TC-800D I was trying to fix, and there was some electronic noises ("crunches" and "clicks") I thought were caused by bad solders, problems with connectors, etc... without success.

I tried to diagnose the issue a number of time, but then finally I realized that there was no issue during the first 10 to 20 minutes trying to reproduce the issue.

So I did some tests over multiple evenings: Playing music from "cold start", and indeed everything plays perfectly without any noise for a while, and then after 20 to 30 minutes depending of the ambient temperature, the noises start and get progressively worse until it sounds like some semi loud "pops" in the speakers.

I'm thinking that could be capacitors going the way of the dodo, but before proceeding to do that, I wanted to have some opinions about it, and possibly some suggestions on how to isolate specific components that could be causing that.

Thanks!

Freezer spray is good, a cheaper way is to warm parts up at the start.

[DMcMahon]

Gentle hair drier application to suspected areas/components at the start might induce the symptoms, if it does then freezer spray application as suggested by merlin to areas warmed up may help to isolate to a particular area.

If you have access to an oscilloscope then would be good to monitor the DC power supply and playback amplifier to see if any obvious untoward noise signals show when you hear the noises.

[Radio Wrangler]

Better than a hair dryer, if you have access to a hot air soldering station, you can use various sized nozzles and focus the heat into a smaller area. There aren't many hot air stations around, but the number is increasing to handle SMD.

Freezer spray is very focused, too.

You need to give it some time between tests for the temperature to average out again.

Freezer spray can create thermal shock induced failures.

DAvid

[Atmosferit]

Originally Posted by DMcMahon

If you have access to an oscilloscope then would be good to monitor the DC power supply and playback amplifier to see if any obvious untoward noise signals show when you hear the noises.

I guess the oscilloscope could also be used to diagnose backward:

- Connect the first channel of the oscilloscope to the output
- Play one of these 3khz test tone tapes
- Wait until the machine starts doing noises and that it is visible on the scope signal
- Then use the second channel to measure down the line until the moment when there are no more noise on the display

When the area is detected, I guess the freeze spray can be used to find the faulty component

Would that work?

[DMcMahon]

Yes you can certainly do that.

Personally I consider that a test tape is not actually necessary (lucky if you have one/some). Depending upon the type of noise signal generated it may be easier to see it on a scope while playing a fixed frequency test tone as against a more complex music waveform.

I would assume (tell me if I am incorrect) that you may well get the same noise symptoms if you just play an empty tape with no recordings on.

When you get the noise symptoms do they continue if you press "Pause".

First thing I would do with the scope is to monitor the DC power supply when the noise occurs, because if a component etc is breaking down in the power supply generating a noise signal, then that will reflect through most of the electronics that use that particular power supply rail/line.

[TIMTAPE]

Have you narrowed the noise down to one or both channels? Is the noise pre or post the volume control?

[Silicon]

Apparently the TC-800D was made in 1975 to 1978.

Could it contain Lockfit transistors? These have a reputation for going noisy.

Does it have Polystyrene capacitors?
Polystyrene capacitors are generally good but some can be affected by heat and solvents.

[kalee20]

Originally Posted by Silicon

Does it have Polystyrene capacitors?
Polystyrene capacitors are generally good but some can be affected by heat and solvents.

Good point - spraying freezer spray around these may do them no favours at all.

I've always been very liberal with freezer, but it's been in the design environment, when the aim has been to provoke any sort of temperature- or thermal shock-related design weakness, on a breadboard which may have a working life of a couple of weeks anyway. But when fault-finding, you don't want to create a new fault which you then proceed to fix, not realising the original fault is still there.

So - suggest using freezer spray in very short, 1/4 second bursts, to cool a component down in 30 seconds or so, rather than cover it in a thick layer of frost.

[Atmosferit]

Originally Posted by DMcMahon

Depending upon the type of noise signal generated it may be easier to see it on a scope while playing a fixed frequency test tone as against a more complex music waveform.

It's what I meant: I have a tape that plays a 3khz sinus for about 15 minutes, or do you mean something else by "test tone" ?

I would assume (tell me if I am incorrect) that you may well get the same noise symptoms if you just play an empty tape with no recordings on.

Correct

When you get the noise symptoms do they continue if you press "Pause".

I've not actually tried that, I'll test this week end

First thing I would do with the scope is to monitor the DC power supply when the noise occurs, because if a component etc is breaking down in the power supply generating a noise signal, then that will reflect through most of the electronics that use that particular power supply rail/line.

I guess that makes sense, and would probably be the easiest part to fix

Have you narrowed the noise down to one or both channels? Is the noise pre or post the volume control?

As far as I noticed (by ear so far), it happens on both channels, and changing the volume changes the volume of the noise as well. it's like it was added to the actual audio signal and normally amplified with the rest.

Apparently the TC-800D was made in 1975 to 1978.
Could it contain Lockfit transistors? These have a reputation for going noisy.
Does it have Polystyrene capacitors?

I've no idea, how can I recognize these?

[Atmosferit]

So yesterday evening I did some more testing, still on pure audio listening, and I found out a number of strange things:
- I was able to listen tapes for almost one hour before the sound started degrading
- When it happened, it was on only one side
- I can confirm that the noise persists when in Pause mode

I let it "cool" for the night, and this morning I tried again, and this time the left channel was noisy pretty much immediately, and after a minute or so completely cut off, came back briefly and then was gone again.

I did a recording of that:
- http://defence-force.org/download/Ya...SidedNoise.wav

(the actual music starts about at 40 seconds, all the noise before that was just generated by the tape deck, the smooth volume change was when I checked the volume sliders to see if it was not a contact issue)

Two of the recordings I did yesterday by comparison:
- http://defence-force.org/download/Ya...n-BeastDis.wav
- http://defence-force.org/download/Ya...-Megablast.wav

[DMcMahon]

Originally Posted by Atmosferit

It's what I meant: I have a tape that plays a 3khz sinus for about 15 minutes, or do you mean something else by "test tone" ?

Yes I understood what you meant. No, my test tone reference was meant to be the same as your 3kHz test tape.

I've no idea, how can I recognize these?

I am not familiar with the Lockfit issue, so will leave others to comment further on this.

I am a little familiar with noisy transistors on old Akai's.

[DMcMahon]

I am sure you will have done it already but make sure you do a good visual check of all the components, checking for anything that looks overheated/damaged, any split/bulging/leaking capacitors etc.

[DMcMahon]

Originally Posted by Atmosferit

So yesterday evening I did some more testing, still on pure audio listening, and I found out a number of strange things:
- I was able to listen tapes for almost one hour before the sound started degrading
- When it happened, it was on only one side
- I can confirm that the noise persists when in Pause mode

I let it "cool" for the night, and this morning I tried again, and this time the left channel was noisy pretty much immediately, and after a minute or so completely cut off, came back briefly and then was gone again.

I did a recording of that:
- http://defence-force.org/download/Ya...SidedNoise.wav

(the actual music starts about at 40 seconds, all the noise before that was just generated by the tape deck, the smooth volume change was when I checked the volume sliders to see if it was not a contact issue)

Two of the recordings I did yesterday by comparison:
- http://defence-force.org/download/Ya...n-BeastDis.wav
- http://defence-force.org/download/Ya...-Megablast.wav

That is the trouble with these type of noise problems, they can be somewhat random and not always repeatable.

The noise burst on the first recording certainly sounds electrical. Even before the short noise burst it sounds a bit noisy possibly like transistor noise (although at low volume).

I assume the 2 music recordings are good recordings and that the alarm clock type sounds and radio tuning type sounds are part of the music.

[TIMTAPE]

Is there a record/play changeover switch? If so, it may be worth spray cleaning it while working the switch a number of times (with the machine switched off of course) They often fail due to lack of use.

[ms660]

Japanese small signal transistors that develop a dark or black oxide coating on their leadouts were also notorious for causing noise back in the day, often aggravated by wiggling the top of the transistor(s)

Lawrence.

[DMcMahon]

Had a quick read up on Lockfit transistors. They were UK (Mullard) and came out in the late 60s and were used extensively in the 70s in various equipment maybe could have got into Japanese Hi-FI.

Looking at the Yamaha Service manual the listed transistors I think are all Japanese, but that is not to say that Lockfit are not fitted, because it is not that uncommon to find different components fitted, compared to what is listed in manufacturers parts listing.

Lockfit have a quite distinctive body shape.

[Atmosferit]

Originally Posted by TIMTAPE

Is there a record/play changeover switch? If so, it may be worth spray cleaning it while working the switch a number of times (with the machine switched off of course) They often fail due to lack of use.

I'm not sure if it's what you mean, but there are two long metalic sleeves attached to the main audio board, with plastic pushers inside that are spring loaded and designed to be pushed on by a small metal plate that moves when pressing the various control buttons, one of these feels smooth to the touch, while the other feels likes "broken ball bearings" (or sand or dust).

I am sure you will have done it already but make sure you do a good visual check of all the components, checking for anything that looks overheated/damaged, any split/bulging/leaking capacitors etc.

Yeah, the whole thing is super messy and cramped, I had to remove a dozen of small rubber ties to just be able to look on the board itself.

As far as I can see, there's nothing obviously bulging or corroded, but it looks like the board was somewhat forced into submission by somebody: Some of the components are bent, one of the two fuses connectors is half a centimeter out of alignment, etc...

One very nice thing, is that the output circuitry is neatly symmetrical, so it should be possible to compare the signal where it enters and where it leaves... assuming I can find where in the mess of wires!

[DMcMahon]

Yes the main board certainly is a mess of wires and will be difficult to fault find on, access wise by the looks of it.

Good that the noise symptoms continue when Pause is selected, this basically removes the possibility of the tape/heads interface being suspect. i.e. the noise appears to be inherent in the electronics.

If the noise symptoms continue only to be on the left channel then that is good, as then you can monitor left and right signals (assuming you have a 2 channel scope) and that may make it easier to see the noise signal.

For clicks, bangs. crackle etc would expect the noise to be short duration spiky pulses, for hiss, waterfall etc would expect more wide band constant noise like white noise etc.

As others have stated be careful with any cold/heat applications, I would first try to see the noise and attempt to trace it back to area of origin.

[DMcMahon]

I assume the noise symptoms also exist on the headphones output. Do you get the noise symptoms both with Dolby on and Dolby off ?