Roland Alpha Juno 2 VERY noisy chorus needs repair help

[trondl]

Recently I bought this unit cheap knowing a few things needed repairing (skippy dial fixed by resetting the synth, uneven velocity fixed by cleaning, presumably weak battery but works ok) .
All of them fixed easy but I created a new one:
A noisy chorus circuit!

The noise can be described as digital noise/silence shaped with a square wave 50% / 50% timeframe (drrrrrt / silence / drrrrt / silence...) and the speed of this noise responds with the LFO.
The noise increases with CSR RATE, but goes away if bellow ca 70 value.
If I sett CSR RATE value low, play a note and increase CSR RATE it plays correctly. But if i play a new note with the now high setting it will distort the sound.

VCF and VCA LFO is not touched by this error.

The unit is fairly clean and shows no obvious signs of shorts or defect parts.

Any help here?

I have the 20 page service manual available if requested.

[acreil]

Check the BBD clock, input and output with an oscilloscope if possible.

[trondl]

I don't have access to an oscilloscope, and I'm afraid that it's abit above my experience.
But if it is the clock that's the problem, shouldn't it then also affect VCA and VCF?

A small correction in my problem: CSR RATE affects both speed and intensity of the noise and goes away at ca CSR RATE: 65.

[acreil]

The chorus works like this: There are 2 BBD chips, each with a separate driver chip that produces a high frequency clock. The clock signal is modulated by an LFO that is software generated. So what you need to find out is whether it's the modulation signal, BBD or driver chip that is misbehaving. It may be something simple like a bad solder joint, but an oscilloscope can help immensely and often makes the difference between "what the hell am I doing?" and "oh, that was easy".

I looked at a schematic briefly (MKS50, actually), but my copy is poorly scanned and difficult to follow, so it's probably less than helpful for specific tips.

[trondl]

Don't have access to a scope but this is how it sounds on CSR RATE: 127 and dropping towards 0 value after 5 seconds. Note that this is recorded at very high volume and is not very audible when played at normal levels.

[acreil]

That's not going to be of much diagnostic value beyond "check the chorus circuit". You're still going to have to trace the signal to find the fault.

[GrantB]

This thread interests me greatly because it seems to be similar to the famous MKS-50 chorus problem, which I have in my MKS. You can find some posts about this on the alphajuno-mks50 yahoo group.

alphajuno-mks50 : The Roland Alpha Juno Series email list

I thought this was only a MKS problem though; this is the first time I have heard of it on the Alpha Juno.

I have traced the circuit and it appears to me that the (my) problem comes from a distortion of the LFO CV which sweeps the BBD delay time. This cv is generated from the digital section and then passes through an op amp where +5V is added to the cv from the +5 (digital) regulator. The input of this OPA looks fine, as does the +5V supply, but the output briefly flops from rail to rail before settling on each cycle of the LFO.

One day I got frustrated and replaced that OPA and some electros surrounding it, plus the electros and the 1/4W carbon comps in the chorus power filtering section, but the problem remains the same.

My next guess is the OPA inverter which comes after the voltage rescaling. It inverts the LFO for one side of the chorus, and my problem seems to be much worse on that channel. Unfortunately, it's one of those SIP dual OPA packages, so I don't have one lying around.

Another possibility is that it's something in the +5V ps, but there's some oddball parts in there too that I don't have on hand, and also that section seems to measure fine.

For those who would like to follow along on the MKS-50 service manual common on the net:

Page 16 on my PDF
the LFO comes from IC3(4051) - pin 4 (looks good here)
5V is added in the feedback loop (?oscillation here?) of IC2(TL072) - pins 6 and 7 (looks good at the input, bad at the output).

Next we jump to page 12
the CV is inverted by IC13(4570) - first OPA pins 234. The same messed up CV can be measured here. Not sure if this could be the problem, somehow working it's way back to IC2.

Also check out the power supply voltages in the chorus. Something's not right here either, even after replacing local chorus PS filtering. Could be a side effect of the opamps going rail to rail?

[GrantB]

Just found a better MKS-50 scan on synfo.nl

http://www.synfo.nl/servicemanuals/R...VICE_NOTES.pdf

[acreil]

Originally Posted by GrantB

The input of this OPA looks fine, as does the +5V supply, but the output briefly flops from rail to rail before settling on each cycle of the LFO.

Bit of a wild guess, here: does it resemble this? I wouldn't ordinarily expect this to just appear suddenly, but if it's an already marginal design with aging components...

[trondl]

Thanks for the input and the "good" news that I'm not alone here :D
GrantB: does your chorus sound more or less the same as the audio sample I posted?
It's taken from the first patch available (Poly2 ?)

[GrantB]

Originally Posted by trondl

Thanks for the input and the "good" news that I'm not alone here :D
GrantB: does your chorus sound more or less the same as the audio sample I posted?
It's taken from the first patch available (Poly2 ?)

The sound is similar, but the period of noise is shorter on mine compared to yours (not 50/50).

Re-reading your first post now, I think the behavior is not exactly the same. The MKS-50 problem happens even with no notes playing (only with chorus on), and if I understand you correctly, your problem happens only when a note is playing? The MKS-50 chorus problem is often masked by the sound of notes, so some people just put a noise gate on it to remove the noise in between.

[GrantB]

Originally Posted by acreil

Bit of a wild guess, here: does it resemble this? I wouldn't ordinarily expect this to just appear suddenly, but if it's an already marginal design with aging components...

Thanks acreil. I had considered this which is why I dropped a new TL072 in there for IC3. Now I wonder if it's the next one IC13(?), but that's a 4570 which is bipolar.

[acreil]

Bipolar isn't necessarily immune, it depends on the topology. I've seen it happen to LM324s also. Reverse parallel diodes between the inverting and noninverting inputs should be a reasonable workaround.

But make sure you don't have some offset voltage creeping in there, or sagging power supply rails. If that's really what's happening, I wouldn't expect it to be some op amp's weird failure mode, so look for what's causing it.

[trondl]

GrantB: The recording is with no notes played and with full chorus rate, dropping towards no chorus rate.

[GrantB]

Originally Posted by trondl

GrantB: The recording is with no notes played and with full chorus rate, dropping towards no chorus rate.

Hmm. These problems may be related yet. Mine is still there with low rate, but less noticeable.

I wouldn't mind seeing the Alpha service manual if it is easy for you to send. Perhaps something will come from comparing the schematics.

For now, I am out of ideas except to throw more parts at it.... which of course is a poor troubleshooting methodology.

[GrantB]

Originally Posted by acreil

But make sure you don't have some offset voltage creeping in there, or sagging power supply rails. If that's really what's happening, I wouldn't expect it to be some op amp's weird failure mode, so look for what's causing it.

PS voltages are also fluctuating, only in the chorus section. I have replaced the local filtering components.

Can you be more specific about an offset voltage and where I might look for it? There is the intentional offset voltage from the +5. It's hard to measure for offset at the inputs of the opamps in question because of the rail-to-rail bit.

[acreil]

The op amp is probably drawing more current when it's distorting. But if the overall power supply voltage is sagging out of spec, or if there's an unintended offset on the modulation signal, it could be causing the op amp (which may fine) to distort. I don't know anything more specific than that.

Maybe measuring it with the chorus off will give you a better idea.

[trondl]

The Juno 1 service manual can be found here

[trondl]

Still no clues?
Really eager to put this babe into production again

[drifter7508]

Any solution to this? I have the same problem with my newly bought MKS-50. The chorus "vobbles" when it's ON. When no notes are played.
It's very apparent with headphones or at very high volume.

[biam]

Same here with my MKS50!

[bebo]

I think, I have the same problem with my MKS-50. But on mine the right channel has this problem when you look from the rear side (or the headphones). In the schematics it's the chorus block B.

1. I did some research so far. I crisscrossed the signal where R88 is located and the noise went to the left (A) channel.

2. Then I crisscrossed the pins 5 and 7 of both channels. And somehow the noise went also to the left (A) channel.

Before I swapped IC21 and IC16 with the ones from the other channel but the noise remained on the right channel.

I also swapped D6, D5, C42, Q7, R36 and R34 with their counterparts from the other channel, but the noise was still on the right channel.

C33 and C32 are also new, although they are the used by both channels.

I also crisscrossed the channels where R35 is located. But the noise stayed on the right channel. But since I nearly swapped every component above IC16 I will try it again. Maybe I did something wrong. Then I will have a look at the couterpart of Q5 and the IC13.

Is there a list with the component names of block B? Would save a lot of time compared to following the traces from one component to the next.

Btw measuring using my old analog oscilloscop was no help. It seems that the level of this sweeping noise is in the range of the normal noise level of the circuit.
The only difference I could measure between the two channels was that pins 3 and 7/8 of the 3009 IC have some impulses that have different level. The frequence of the impulses is related to the LFO. And on the bad channel they were 3 times as high.

[bebo]

Hi again,

after a few more days of investigation I finally replaced both MN3101 in the chorus part.

Now I can move up gain to the limit without any of that sweeping noise. Thus the MN3101 was the problem in my case.

For those who have the same problem and live in Germany: You can get brandnew MN3101 for 1,50 Euro at Reichelt.


I just saw that there is no grounding wire a the power socket. Do your power connectors have a ground connection to the chassis of your MKS-50s?

Have a nice day,
Bernd

[bil_g]

Nice job, bebo! I'll have to try this on mine. I've been hoping someone more motivated and smarter than me could get a lead on this issue.

[Mandaloo]

Hi Bebo,

wow, what a great job!! I have the same problem with my MKS-50 since years, now there's hope to get it solved and my wonderful synth will run including the chorus soon! :-)

As I suppose (because of the hint "Reichelt.de") you are from Germany? May I ask two questions (via Email if you want) conserning the MKS-50 repair and the mounting of the MN3101 in German language? Would be a little bit easier for me, thanks!

Nice greetings from Munich
Mandaloo