D550 s/h buffer opamp replacement

[Mooger5]

My d550 has a crackling noise in the left out.
The fault is in the opamp or the cap in in the sample nhold between the dac and the output stages. Can i replace the 5238 with a 5532? The chips are smd tough to swap back and forth. The 5532 is easy to get here. Its not as fast as the original and its bjt instead of fet. But its good for audio...

[jxalex]

TRY! No harm done with that.

Still, FET should be replaced with FET and BJT should be BJT as the BJT has lower input impedance.

TL072 would be next retry. Then AD797 (which is a costy though).
with uA740 would be also worth to try.
NJM4580 too. Not 4570

[Mooger5]

Thanks for your advice jxalex

I´ll study the datasheets and search for those chips in the local market.
I´ll do no more than two maybe three swaps and the last one must be definitive. There are no SMD chip sockets AFAIK...

Mind, am still not certain which component is faulty. The cap(s) would be an easy fix. No opamp rolling needed.

However, there´s this residual noise in the reverb tails and the sounds release. I wonder if the lower specced 5532, which I like a lot for audio, could actually filter the noise out without degrading the high frequency response.

I´ll report my findings.

[jxalex]

Also when the op/amp is removed does it still make crackling noise?
HOw look the power supply lines? noise free?

it is no problem if you just temporarily solder the wires to the SMD areas and so, put the socket. THEN solder the capacitor between the pins 4 and 8 (the power pins). The capacitors still do not swap out if these are ceramic and just in the sample/hold circuit. Neither all electrolyte capacitors will say up.
Still you can try that if the signal path between the s/h input and the main output is working, by just disconnecting the s/h input from that s/h buffer input pins

[Mooger5]

Also when the op/amp is removed does it still make crackling noise?
HOw look the power supply lines? noise free?

Haven´t done any "surgery" yet, just freezed the area to locate the culprits.
I always check the PSU and follow the traces. Everything in the module works fine apart the very dim LCD backlight.

it is no problem if you just temporarily solder the wires to the SMD areas and so, put the socket. THEN solder the capacitor between the pins 4 and 8 (the power pins).

I could solder long jumper wires going to a solderless breadboard, and decouple the ic from psu, but the possibly offending cap is not for bypass, it´s across the non inverting input and ground...

The capacitors still do not swap out if these are ceramic and just in the sample/hold circuit. Neither all electrolyte capacitors will say up.

I´m not sure I understand what you mean here, but don´t ceramics can go wrong too? I think the Korg Delta filter caps are ceramic and they fail. These S/H caps are mylar and they have to constantly charge/discharge at suprasonic rates to fill the gaps due to the multiplexing.

Still you can try that if the signal path between the s/h input and the main output is working, by just disconnecting the s/h input from that s/h buffer input pins

Pretty sure after removing the IC that the crackling will stop

[Mooger5]

SMD TL082 is available locally for less than 1 euro each. Lower slew rate than the stock M5238 and noisier. I´ll try it anyway.

[jxalex]

Try all, Interesting to hear results. Still the most important is the high input impedance of the op-amp in these circuits.
THose noise figures really matter only with circuits where the high gain is in use.
So in such circuits where you have, it is close to unity gain and it really is not audible difference where the input signal is noisier than the op-amp itself and it is more just a question of measuring than hearing.

[Mooger5]

THose noise figures really matter only with circuits where the high gain is in use.
So in such circuits where you have, it is close to unity gain and it really is not audible difference where the input signal is noisier than the op-amp itself and it is more just a question of measuring than hearing.

WOW that´s the kind of information I appreciate. Not just something to take for granted, but one that makes absolute sense! Indeed more gain = more noise. At unity gain "garbage in, garbage out" so to speak, not just MORE garbage. Many thanks for this.

I can´t think of why some opamp will be less noisier than other when adding gain, besides the quality of internal components. Common mode rejection also comes to mind, but that only gets rid of the noise catched in the process I think, like balanced audio lines...

[jxalex]

Well it is similar to question like how come that the HF-VHF-UHF transmitter transistors from 1970's have so modest gain compared to the newest solutions.

technology, manufacturing technology used to make these, their purity of materials. With ICs it is also schematic solutions on that time which affect.

[madtheory]

Jonson noise is the cause.

[Mooger5]

Yes, circuit topology and impurities and thermal variance too.
Well I was about to post some boring good news: reworked the joints and the crackling stopped. So the 5832 is fine. But I´ve decided to try the 5532 and the 082 anyway.

This is the stock chip, before removal.

[Mooger5]

This is the 082. As usual it doesn´t look so bad in real life. I´ll post a better pic tomorrow.



And it works quite well. I think it sounds slightly darker. But it´s all very subjective.

I´ll try the 5532 tomorrow.

[Mooger5]

The 5532 has a more informative midrange. Sounds fuller, with the tonal colours more pronounced. It could be some placebo effect but I´m having fun with this. So I like it and its implementation in a D550 is´justifiable in terms of "design coherence" since this chip was launched in 1985.

I´ll have to rework the solder joints on those caps later.

[madtheory]

Well I was about to post some boring good news: reworked the joints and the crackling stopped.

This is great news, for me I have a D50 here with the same issue. I had suspected the relays on the outputs. Plus broken sockets and sticky buttons. It was not loved.

So you've found that the joints were bad on the op-amp- but is it possible that the heat from the soldering iron gave a temporary fix to an internal open circuit in the chip? Doubtful I know, I can see from the pics that your soldering skill is perfect! Very nice job on SMD

[Mooger5]

Thanks, but SMD is just as easy as THC. With the proper tools and a bit of practice anyone can do it, really.

Wasn´t the D50 one of Roland´s first made in house products to use SMD? It was probably all hand soldered. So I think the crackling is caused by dry solder joints.

I´m going to compare the 5532 with the 5238 now.