Diagnosing temperature related noise?

[ms660]

Originally Posted by Atmosferit

Another thing I will have to identify, is the correct values (and what they are for) of VR1/VR2, VR3/VR4 and VR11/VR12, because I would have expected them to be in the same positions, or mirrored positions, but they are not actually perfectly mirrored

VR1 & VR2 adjust the high frequency playback level relative to the lower frequency playback levels, the value of those controls are 20k each, in the schematic they are in the feedback network that's between the collector of TR3/TR4 and the emitter of TR1/2, that network is switched in on playback by contacts 7 & 8 of the Rec/Playback switches.

VR3 & VR4 adjust the output level on playback, the value of those controls are 50k each, in the schematic their wipers are connected to contact 13 of the Rec/Playback switches.

VR11 & VR12 are for setting the meter levels relative to the line out levels, the value of those controls are 5k each, in the schematic they are connected to the meter +ve via D3/D4.

Lawrence.

[DMcMahon]

The six VR pots you list appear to be all covered in the set-up procedures plus their values are shown on the schematic. I have not checked thoroughly, but a quick glance at the channel pairs, they look fairly symmetrical circuit wise.

The VR1/VR2 on the Servo board, I cannot easily see their values or set-ups.

Response aimed at Atmosferit not Lawrence.

[Atmosferit]

Originally Posted by DMcMahon

Changing the subject, do you like the scope ?
I have been thinking of upgrading my old Tektroinix for something more modern but am unsure about the all singing/all dancing modern scopes, though certainly would be very good to get the storage capability and nice to get 4 channels .

I do like it, but disclaimer, I'm just going back on the saddle, I've not done anything electronic seriously since I was a teenager, and when I had fun last year trying to fix the Brionvega TV with my girlfriend, we decided to both do some more stuff this year, so we bought the Siglent as a common christmas present.

We picked up that one because it has four channels, split over two separate circuits, so it stays at full speed when you use channels 1 and 3, and only drop to half speed when you use the two channels that are on the same circuit.

Also the display was very nice, it can be connected on the ethernet to be remotely used through a webserver which would be handy for saving results, etc...

tldr; it's the best we found in the price range, but it may be overkill depending of what you are doing.

VR1 & VR2 adjust the high frequency playback level relative to the lower frequency playback levels, the value of those controls are 20k each, in the schematic they are in the feedback network that's between the collector of TR3/TR4 and the emitter of TR1/2, that network is switched in on playback by contacts 7 & 8 of the Rec/Playback switches

Ok, I'm going to use the rec/playback switch pins as the reference to find where the signal is going bad (or possibly the switch itself could be having problems?)


One thing we noticed: After leaving the tape deck on for about one hour, there was a start of a smell, kind of in between ammonia and roten fish, not very strong, but definitely coming from the deck: Is that the typical smell of a dying capacitor, and can they emit this smell even if they are not visibly bulging or leaking?

[DMcMahon]

Personally I do not know. I have only experienced split/leaking electrolytics a few times, no obvious smell but these were probably old leaks in my case.

I have occasionally read people saying that they had a leaking capacitor that smelled (not sure what type of smell).

[DMcMahon]

Coming back to the possibility of the noise being generated by a bad transistor. I have a lot of old Akai R2Rs. There is quite a lot of press in the Akai community about old Akai's potentially suffering with noise due to 2 main components.

One being transistor 2SC458LG and the other an Op Amp LD3141. The 458 issue being the same as Lawrence previously referred to (the oxide coating on leadouts reference).

I only experienced transistor noise in a couple of my Akai's, one a GX-286DB suffered severe noise (ran the whole gamut of noise symptoms) mainly random and short duration, never actually proven to be temperature related though.

By time consuming scope monitoring able to pinpoint the individual various noise symptoms and I think in the end changed sequentially at least 12 transistors before all the noise symptoms were fully eliminated and am pretty sure that each transistor replaced, fixed/improved a noise symptom, i.e. none were changed in vain.

I think there were 3 different transistor types involved including the 458.

In at least 3 of the bad transistors their measured DC voltages at their 3 pins were also bad (which I find is not common), i.e. voltages were way out compared to the values shown on the schematic, of course schematic transistor voltage values (if/where shown), are only typical/ballpark figures. Replacing those transistors restored more typical voltage readings.

My only other noise experience with the Akai's was on a 4000DS Mk. 1, where a couple of 2SC458s were generating hiss & crackle (fairly constant) & clicks (random) noise.

There remained a strange echo type noise, where in desperation I replaced all four LD3141 Op Amps (they had such bad press) with transistor replacement circuits (the 3141s being long obsolete and difficult to find) but to no avail.

This echo type noise is on a long list of things to get back to !

[TIMTAPE]

Originally Posted by Atmosferit

...Ok, I'm going to use the rec/playback switch pins as the reference to find where the signal is going bad (or possibly the switch itself could be having problems?)...

I suggested in post #15 you check the switch. Dirty rec/play switch contacts are a very common fault in tape machines. I would do this first. Even just working the switch back and forward say 20 times can be enough to temporarily improve the switch for diagnostic/elimination purposes.

[Atmosferit]

Originally Posted by TIMTAPE

I suggested in post #15 you check the switch. Dirty rec/play switch contacts are a very common fault in tape machines. I would do this first. Even just working the switch back and forward say 20 times can be enough to temporarily improve the switch for diagnostic/elimination purposes.

I did that yesterday: I used a non flammable air spray can on all the board to make sure there was no dust, then I used some contact spray in the switch enclosure, and I switch it in and out a large number of time, but that did not seem to have any impact on the noises.

Coming back to the possibility of the noise being generated by a bad transistor.

Regarding transistors, there was one close from the input stage that was almost bent on itself so the legs were almost touching. I straightened it, but that did not change anything.

Guess I need to go through them all and check the references.

Assuming I go through and replace components, I guess I should probably not just order a random batch of cheap transistors and capacitors from Ali Express at $1 for 100, but ordering "audiophile quality" at $100 for 1 is probably not going to cut it either.

Most big (Farnell, ...) shops have dozens of references for the various types of chips, do you have any guidelines on which type of components to chose when doing replacements?

[ms660]

I would troubleshoot the fault down to component level first, you will learn more that way, you should be able to pinpoint the source of the trouble with your oscilloscope.

Pot's, especially the miniature preset ones can be another source of noise, usually gently probing their wipers with a plastic knitting needle can aggravate the symptom.

Lawrence.

[Atmosferit]

Originally Posted by ms660

I would troubleshoot the fault down to component level first, you will learn more that way, you should be able to pinpoint the source of the trouble with your oscilloscope.

Makes sense.
I spent a few hours today with some big photos I took, the schematics, pcb diagrams, block diagrams, and tried to document/relate what was supposed to go there.

Based on the schematics, the pin 14 of the big switch is supposed to connect to pin 5 of the NE545 through C22, but when I look on the board, I'm at loss to see how C22 connects to the pin 5, there's the big filter assembly on its own small board, there may be a wire under, but I don't see anything Oo

I'm also not sure about the pinout of the TA7066P (is it 100% AN5515 compatible?) and in particular I'm not sure where the pin 1 is

Pot's, especially the miniature preset ones can be another source of noise, usually gently probing their wipers with a plastic knitting needle can aggravate the symptom.

I tried touching them, did not do any difference in the sound.

Anyway, probably going to take me a long time to figure it all out, but I'm going to find what's wrong.

[DMcMahon]

I would second Lawrence's suggestions.

At this stage I would not think about chip replacements, it is not impossible that it could be an IC. but much more likely I would say to be a bad contact (switch/pot etc), bad transistor, bad capacitor etc.

Buying transistors is maybe a personal preference, from component distributors such as RS & Farnell (to name just 2) I would hope one would get a quality product with some traceable quality standards.

I have also brought transistors from sites such as eBay without any problem, personally I would not buy large quantities of cheap transistors, that is maybe asking for potential trouble.

Hopefully with some intensive scope etc monitoring you will be able to home in on the problem area/component (s) etc, once again I know from bitter experience that this is not always easy.

[DMcMahon]

Did you mean C21 ?


Sorry I now see that C21 is the left channel and C22 is the right.

[ms660]

Originally Posted by Atmosferit

Based on the schematics, the pin 14 of the big switch is supposed to connect to pin 5 of the NE545 through C22, but when I look on the board, I'm at loss to see how C22 connects to the pin 5, there's the big filter assembly on its own small board, there may be a wire under, but I don't see anything Oo

Blue jumper wire on the component side of the board?

Lawrence.

[Atmosferit]

Oh crap... this "pseudo" mirroring made me look on the wrong side, and of course there are the two blue cables crossed over the filter, but on only on the left channel, not the right one.

These "filter" blocs, are they standard ways of building things, with known components, or is that custom? I can't find any reference in the service manual other than "it's the filter block and it has 3 pins"

[ms660]

Originally Posted by Atmosferit

Oh crap... this "pseudo" mirroring made me look on the wrong side, and of course there are the two blue cables crossed over the filter, but on only on the left channel, not the right one.

These "filter" blocs, are they standard ways of building things, with known components, or is that custom? I can't find any reference in the service manual other than "it's the filter block and it has 3 pins"

The blue jumper wire is in the RH signal channel, there will be another wire jumper in the LH signal channel.

The "filter block" looks to contain an inductor, probably a capacitor or some capacitors associated with it as well, it's probably there as a multiplex or bias signal filter.

Lawrence.

[Atmosferit]

There are indeed mistakes on the schematics: The service manual page with all the components on the PCB has two references for C48 and C47 around the TA7066P chip, but both the actual schematic blocs and actual PCB identifies the ones connected to the pin 3 of the TA as C50 and C49.

I also checked the record/play switch for contact issues, get a clear contact for each of the 9 sub switches in both position.

I found out that coloring the areas of interest (switches, adjustable resistors) on the schematics makes it a heck of a lot easier to find what is where on the actual board, finding that 23 connects to the pin 1 of the TA7066P confirmed me where the pin 1 actually is (was not sure looking at the chip, I guessed it was on the left side when facing the side with markings), but it's kind of disturbing to follow the tracks and reach components that the exact same service manual disagrees about what it should be

[DMcMahon]

Yes documentation errors/omissions can be very frustrating, it is often difficult enough without the documentation confusing further/not be very helpful.

[Atmosferit]

Damn, this is difficult, did not have time this week end, tried again now with the oscilloscope and the schematics, using the working channel as a reference for what type of signal I should get.

From what I can see, I do have a valid (but weak) waveform on pins 13 et 14 of the big switch, and by the time we reached the ping 23 (that connects to the amp circuit), the sine curve has been replaced by some noise, so the issue happens somewhere in between.

I tried to see what happens in between, I still have a valid sine at the input of C22, but on the other side (which is connected to the pin 5 of the Dolby circuit) it gets all flat, I thought it was maybe the reason, but if I do the same thing on the other channel, I also get the signal disappear.

I tried poking around the various pins of the dolby chip on both side and could not get any signal, so I guess the oscilloscope itself is interfering with the device.

How do you diagnose this type of thing?

[DMcMahon]

Sorry have not had time to look at the Dolby circuitry area. Are your original problem symptoms still the same, i.e. intermittent noise maybe on 1 channel only. Possibly interpreting your latest post I could read it as 1 channel is now not working properly all the time.

[Atmosferit]

The same issue has not changed:
- Both channels working properly for about 20-30 minutes
- Then one channel degrades with more chaotic shape, noise spikes, and then after a while (5-10 minutes) it gets almost totally flat
- The other channel works fine at all time, which is cool because I can use it as a symmetrical reference of what I should expect on the other side

to get the defective channel back working, I need to let the machine rest for quite a while

[DMcMahon]

Originally Posted by Atmosferit

Damn, this is difficult, did not have time this week end, tried again now with the oscilloscope and the schematics, using the working channel as a reference for what type of signal I should get.

From what I can see, I do have a valid (but weak) waveform on pins 13 et 14 of the big switch, and by the time we reached the ping 23 (that connects to the amp circuit), the sine curve has been replaced by some noise, so the issue happens somewhere in between.

I tried to see what happens in between, I still have a valid sine at the input of C22, but on the other side (which is connected to the pin 5 of the Dolby circuit) it gets all flat, I thought it was maybe the reason, but if I do the same thing on the other channel, I also get the signal disappear.

I tried poking around the various pins of the dolby chip on both side and could not get any signal, so I guess the oscilloscope itself is interfering with the device.

How do you diagnose this type of thing?

Does not appear to make much sense why no real signal other side of C22 ?

When you say the signal also disappears on the other channel, I assume this must be the good channel at C21, when you loose the good channel at C21 on the scope, do you also loose the audio output ?

From your reference to sine/sine curve I assume you must be playing the 3 kHz test tape, not just looking at any standing signal noise levels.

Am not familiar with these Dolby NE545B chips, but I would have thought that the scope probe connection should not have any undue effect on the signals.

Are the DC voltage levels good at the pin 5 inputs, in all conditions, including when you connect scope ?

So at this stage I cannot really offer any constructive helpful advice.