Another Quad CD66 with distortion

[PWH]

I got this with described faults as 'will not play discs and if it does is distorted'. On first test, I found it very accurately described!

I have re-capped the machine apart from the two 100uF/16V in the output as these 'green ones' seem to be 'special' maybe audio grade? and the 10uF bi polar as I don't have one.

The two settings for the laser were checked, laser cleaned it and now the machine spins up and plays every time- but the sound is distorted. It sounds roughly like an FM station off tune. I presume the fault must be something common to both channels as the distortion is the same on each. Display is fine and the remote works.

Referring to my other closed thread about a different specimin I have checked the voltages around the op amps and they are + and - 15 where the data sheet says.

I have the manual available here:
https://www.dadaelectronics.eu/uploa...ice-manual.pdf
but am unsure what to do and I don't want to wreck this through ignorance.

What would be the effect of the 56k resistor(3705) which feeds the 4 transistors after the op amps being duff?
Any help would be appreciated.

[wave solder]

I would be checking the output from the DAC first before condemning anything on the analogue side, but I'm not a specialist on this side of things.

[PWH]

Thank you. Before I can do that,
Can someone please advise me what the components are that are marked as a rectangle with 2 lines either side? P/N 3657 3658 for instance. Fusible resistors??

( I have checked the 56k resistor it is OK and there is continuity on the circuit)

[dave cox]

I would guess fusible or fire-proof resistors, given the [!] symbol

An oscilloscope on the output may be a good diagnostic. Especially if you can create a disk with some single tones.

dc

[PWH]

By popular request, a pic of a very shabby sine wave.
1 kHz @-14dB FS (what is FS?) from Alan Parsons/Stephen Court Sound check CD

[PWH]

Well I found out that "FS" is full scale.
Does anyone have any idea whether the graph is pointing to the cause of the distortion and what to do next?
Thanks..

[dave cox]

UGLY! Same on both channels ?

It's not a stuck bit, for sure!

I think you'll need to probe the output of the DAC, if that looks the same I would then isolate the DAC output and verify the error is still there. The DAC is claimed to be 'conventional' so the only mechanism to create that signal, I can think of, is the oversampling screwing up somehow

My guess would be something in the linear side, after the DAC.

The linear side is 2 op-amps, both have some filtering plus there is muting on the output. The first op-amp has a JFET switching some extra filtering in/out -no idea what that is for - pre-emphasis perhaps?

dc

[Radio Wrangler]

What sort of DAC does it have?

Delta-Sigma or R-2R network?

David

[PWH]

Thank you for your replies.
Both channels are equally horrible

Originally Posted by Radio Wrangler

What sort of DAC does it have?
Delta-Sigma or R-2R network?
David

Its a TDA1541a single crown in a socket. I have re-seated it..

http://pdf.datasheetcatalog.com/data...s/TDA1541A.pdf

I do have another of these players so I can put the suspect DAC into the good machine and test it but i am wary of ruining that machine/DAC. Is there anything else I can do before I try this?.

[Radio Wrangler]

There is the small bank of filter capacitors for the dynamically shuffled current sources for the most significant bits. With the DAC out you could probe and measure these.

David

[Red to black]

This is a Philips made player with a CDM4 mech.

First check the supply lines very carefully a this is common to both channels and everything else, I have had the regulators fail on these type of machines before, also check for dry joints especially around the regulators.

The Left and Right data from the Disc is in a single stream (interleaved) right until it leaves the DAC chip where it is then split into L and R at the DAC output. I have had both the DAC and the preceding decoder chip cause this fault before, usually one or the other I don't think I ever had them both fail together, another possibility is also the Ram chip, I have had this fail before as well, again it has only been one chip faulty, some of the Philips machines also had a digital filter chip as well, this could go faulty too but this usually only affected one channel or other as it was after the DAC.

I would start at the supply lines first though before starting the chip guessing games, and it is literally a guessing game too, as I said any one of the three (or four) can fail and give similar/the same symptom with voltage checks here being inconclusive and no way of realistically checking the data stream.

[Radio Wrangler]

Originally Posted by Red to black

and no way of realistically checking the data stream.

This is very true. Oversamplers work by creating a pseudo-random data flow whose average, taken over N samples, is equal to the wanted higher resolution value. This DAC chip runs with N=8

So you want a test CD with a simple waveform like a sinewave or a triangle wave on it. Your computer can be used to write one, that's easy.

To inspect the data to the DAC and to see if it's damaged, you'd need to catch 44100 samples per second times 16 bits per word times 8 oversampling ratio which is 5.6448 Megabits/second serial data flow per channel.

Then you'd have to collect 16 bit words, by monitoring load lines, then take averages of 8 words at a time, lowpass filter to remove noise > 100kHz or so (and there will be LOTS of noise up here, that oversampler is still a noise-shaper) Then you plot the result. Do you get the waveform from the test CD you made? Is the triangle or sine undistorted?

You'd need a digitising oscilloscope and then to do a lot of post-processing on the data it captures. This isn't beyond the capabilities of instruments on the market, but I don't think the software for this is available.

Plan B would be to steal the data flow and feed it to a known good DAC. Do the appropriate analogue lowpass filter, and check the waveform.

Plan C would be to just replace chips until you get a good result.

ALL of these would be after checking power supplies, looking for twitches on a scope just as much as looking for the right average voltage with a meter.

Plan A is R&D territory, but it does check the data being fed to the DAC is OK.

Plan B is factory test jig territory, but you could make one.

Plan C is a field repair strategy. Most probably a proportion of chips that get replaced are innocent, which wastes money and time, but the costs of plans A and B in equipment will equal that of multiple new CD players. They are only worthwhile if you have a significant flow of faulty sets to diagnose.

There is a spectrum of approaches from 'Don't test anything, just start changing things' through to 'Amazing amounts of testing (and test equipment) can ensure that the only part you replace is the faulty one'. Sanity lies in a band somewhere between the extremes. Does a new part and the time swapping it cost more or less than the labour and equipment needed to test it?

You could put a scope on the data flow to the DAC and as a minimum check that the logic levels are good and that the edges are clean and not slow or jittered. Check that load pulses and chip select pulses etc. look OK.

If it was on my bench, I'd poke around, checking supplies and that logic voltages and edges were reasonable. Then I might just shotgun decoupling capacitors because they're cheap and quick to swap. Then it would be chip swapping time.

David

[PWH]

Thank you everyone for your advice. I'm not sure I understand it all but what I can do is take the DAC out and measure the supply voltages to it. I will of course check the regulator voltages too and report back. I have one of the Hantek PC scopes at my disposal along with a DVM and a CD test disc. However this is just a hobby so my knowledge is limited..

[dave cox]

Also check the supply lines to the rails of the output op-amps with an ac coupled scope.

Swapping the DAC with a known working one would be simple - its just a pity they cost so much!

dc

[PWH]

Originally Posted by dave cox

Also check the supply lines to the rails of the output op-amps with an ac coupled scope.
dc

This is where my understanding is limited. I have checked the +/- 15v to the op amps with a meter- looks OK.


How do I ac couple the meter probe from the scope? Usually I connect one to the chassis and the other on the point under test, or is there a setting on the scope (software) to do this. What am I looking for?
Thanks

[Red to black]

Connect up the way you usually do but Switch the 'scope to AC input mode on a Y input, you can then set it to say 0.5V per centimeter when measuring say a 5V DC supply, in AC input mode it will ignore the standing DC offset allowing you to see/measure any AC ripple/hash sitting on top of that DC rail easily.

Don't forget though you need to take the measurement in relative percentage terms of the supply rail here, by this I mean it is no good measuring at 0.1V per cm on say a 100V rail, the ripple would look enormous here even on a normally good working supply, on a 1.8V supply though it would be a suitable setting, in general anything under say 5 to 10% of ripple on the nominal rail is usually ok.

For example scoping a 15V DC rail in AC mode something like 0.5V, 1V or maybe up to 2V per cm would be a sensible setting to show if you had excess ripple on that rail, as I say it is the ratio of ripple to DC that is important here, this is a good way to see duff decoupling caps so long as sensible settings are used.
Touching on the X amps settings here, the speed of the horizontal axis of your scope should be also set to something approaching the expected frequency of that supply, by that I mean a 100Hz supply derived from a bridge rectifier and smoothing cap is far slower than that needed than scoping across a smoothing cap that is decoupling a much faster SMPSU. within reason, scoping the much faster supply with a duff cap at the slower speed gives an envelope type display rather than the actual ripple waveform, this in itself is also ok once you get to know what it is you are looking at, trial and error at first is not a bad way to learn here.

Have a play, experience gained playing about here will soon give you of an idea and a 'feel' of what is normal and expected ripple voltages on a given supply rail voltage and what is not.

[PWH]

The good news is I swapped out the suspect DAC and put it in the known good machine and it sounds exactly the same as the known good DAC so the DAC is fine.
The bad news is the PC Oscilloscope does not do AC coupling!!. I wonder if I put a capacitor in line with the probe that would do it? In the meantime I will start measuring the DC voltages around the DAC socket.

[Red to black]

Ah ok, I didn't realise you were using a PC scope, not that that matters as actually I didn't know what scope you were using anyway I was brought up so to speak on old analogue scopes, a valve jobby which was ancient even then, I think I got a "crash course" in half a morning at the shop I started work in, I did go to college and picked up better techniques however in the day job at the time I still had "my Engineer" if I got stuck lol

[Red to black]

Still on that CD player there is really only three chips, and only two left now!, one the decoder and two the RAM, unless you have some obscure component causing your fault.
The next one I would try (actually would have been my first) but it is a coin toss would be the Ram chip, sods law it will be the decoder now I have said that.
Anyway the faulty chip/s never caused any other damage to any other of the same chipset or machine so that is some good news for you.

[jamesperrett]

Looks like it has a digital audio output. Do you have an external DAC or computer audio interface that you could feed it to?