Jupiter 8, Voice E(5) VCO1 fault

[Mr. Mazz]

Hoping someone can advise with a Jupiter 8 I currently have for repair. I have the service manual and have diagnosed through the Test mode

The issue is that I initially found that Voice E(or 5) VCO1 is sounding raspy compared to all others that sound the same.
I've swapped Module boards over and it's still there for E(5) VCO1
I've now swapped over the Module control boards and still the same.

So my next option is the Interface board. Looking at the end chain of this, Each Voice outputs via a TL082 (for VCO1 & 2) and then goes to a shared IR3109 ic. Prior to the TL082 is the 4051 1-8 multiplexer ic. I've heard these can be a bit iffy also?

This seems to be the common denominator for my symptoms, having done all the usual steps. Any advice appreciated.

Many thanks.

 

[CJ]

My heart goes out to anybody trying to fix one of those,


Try wiggling the ribbon cables while playing ,

 

[Mr. Mazz]

My heart goes out to anybody trying to fix one of those,


Try wiggling the ribbon cables while playing ,

Thanks, I've been more indepth than that. I think I'll replace the 4051 1-8 multiplexer (IC26) as these can be funky as they get old

 

[zamproject]

My heart goes out to anybody trying to fix one of those

yep...know this, fully restore one some years ago, recap, change all tact switches, various issue etc...and a painful calibration.
But at the end of the (more than one) day, this beast is just so great to play with, I wish I had one but market is totaly crazy about it...

I think I'll replace the 4051 1-8 multiplexer (IC26)

Maybe try swamping IM ribbon cable between upper and lower row before changing IC
Not sure what -raspy- define here ? like filtered ? IC26 handle CV and gate for VCO, not VCF/VCA

 

[Mr. Mazz]

yep...know this, fully restore one some years ago, recap, change all tact switches, various issue etc...and a painful calibration.
But at the end of the (more than one) day, this beast is just so great to play with, I wish I had one but market is totaly crazy about it...


Maybe try swamping IM ribbon cable between upper and lower row before changing IC
Not sure what -raspy- define here ? like filtered ? IC26 handle CV and gate for VCO, not VCF/VCA

Thanks, good call. Yes, you're right, it's the gate to the module boards, but clutching at straws here.
When I increment through in Testmode, all Voices bar E(5) sound the same. This one sounds like it has a "raspy" element to it and yes, I'd say filtered or more of a square wave sound compared to the others that have a smoother sine wave sound. I'll try and upload a video to explain later.

Possibly the connections to Module board E and/ or supplies to check too

 

[zamproject]

Possibly the connections to Module board E and/ or supplies to check too

Yes, as I understand it the issue don't follow modules board nor controllers boards swap, check all wire as @CJ already say

Do you have other issues ? sometimes it appears that 2 issues seem uncorrelated but finally are, can help to narrow on a faulty source

 

[Mr. Mazz]

Yes, as I understand it the issue don't follow modules board nor controllers boards swap, check all wire as @CJ already say

Do you have other issues ? sometimes it appears that 2 issues seem uncorrelated but finally are, can help to narrow on a faulty source

No other issues, it's just VCO1 for the Module Board E(5)

 

[AnalogPackrat]

Do you have an oscilloscope? Have you probed the signal path between VCO1 and the following circuits? Sounds like maybe something is getting clipped. DC offset issue or maybe some analog switch or MUX only passing half of the waveform. Good luck.

Edit...is this an accurate block diagram?

http://www.audiocentralmagazine.com/roland-boutique-jp-08/roland-jupiter-8-schema/

 

[Mr. Mazz]

Do you have an oscilloscope? Have you probed the signal path between VCO1 and the following circuits? Sounds like maybe something is getting clipped. DC offset issue or maybe some analog switch or MUX only passing half of the waveform. Good luck.

Edit...is this an accurate block diagram?

http://www.audiocentralmagazine.com/roland-boutique-jp-08/roland-jupiter-8-schema/

Thanks. Yes, I've a scope but haven't really looked at the module board analogue waveforms as yet as I thought it was something before then having swapped module boards around and the fault stayed at Voice E(5). As mention previously, could be the connections into that Module board and they could be dragging something down causing the issue.

That diagram is the same as my schematic

 

[AnalogPackrat]

Since it doesn't follow the voice board, but seems associated with a slot, it could be a physical connector issue or a mapped control signal problem. Do any parameters (e.g. cutoff, resonance, waveform select, etc.) behave differently for the problem voice? "Raspy" could be goofed up filter CV. Maybe the HPF is stuck at a high freq.

 

[Mr. Mazz]

Maybe try swamping IM ribbon cable between upper and lower row before changing IC

Is this doable? C/D & A/B swapped with G/H & E/F, how would that affect with the Module control board connections still as they were?

 

[zamproject]

It may help to see if IC26 follow the fault as you suspect it ?
I have no idea how the JP8 will behave with uncorrelated gate and KBCV messing the voices, there will probably be strange things with ADSR/GATE select at VCA, but I have not the guts right know to dive again in the whole schematic...

I don't think you will harm anything, it's just the same type CV signal.

I don't remember if there is a debug mod with monophonic (16 unison voices), with same parameter for all, it should trick the CVs destination mess up, and you'll bee able to see if the issue follow a PCB or keep a slot position, as looking at data lines for a faulty control parameter.

 

[Mr. Mazz]

Maybe try swamping IM ribbon cable between upper and lower row before changing IC
Not sure what -raspy- define here ? like filtered ? IC26 handle CV and gate for VCO, not VCF/VCA

So, I've swapped the IM ribbon between upper and lower and the fault moves to the A/B board. So tracking back to the Interface board and back from IM pin 21 KCV E-1, I have found that IC27c (TL082) the first op-amp in this package has noise on its output pin 1. Compared to the other op-amp in the package and subsequent TL082 outputs I've checked, so this is the culprit. Attached some video's of prior checks and the "raspy" looking and sounding waveform. Have ordered up a stack of TL082's. Thank you all for your insights and help

 

[zamproject]

I have found that IC27c (TL082) the first op-amp in this package has noise on its output pin 1. Compared to the other op-amp in the package and subsequent TL082 outputs I've checked, so this is the culprit.

Did you check input to exclude the 4051 demux ?
Also, there is a feedback from the IR3109 (IC29) output to TL082 + input, you may consider/check for a faulty OTA too...

 

[Mr. Mazz]

Yes, input from 4051 to op-amp is solid. I also did a check on the IR3109 feedback and this has noise, but I'm confident it's the TL082. I'd also had some good info. from someone on another forum which seems to point towards the TL082 too:

Most likely you have a problem with the VCO CV sample and hold circuit for the offended voice. If you look at the left side of the Jupiter 8's interface board, you can see a vertical row of eight TL082s and their related buffer capacitors, one chain for each voice. Those are the VCO CV sample and hold circuits. Each TL082 is in charge of two VCO CVs, so 8 op-amps are needed to feed the Jupiter 8's 16 VCOs, one op-amp per voice as told previously.

Take a look with an oscilloscope at the outputs of those TL082s. There you should see a clean DC, if you see something else (noise, a sawtooth-looking waveform or any other type of AC signals) most probably that op-amp and/or its buffer capacitor are defective and must be replaced. The 4051s dividing the multiplexed CVs coming from the DAC can be the culprit too, but it's most likely that the problem is on the sample and hold circuits.

This is a common issue with Jupiter 8s. I had a Jupiter 8 some years ago which exhibited the same issue of yours, and reading on the Internet I saw several other cases of people talking about the same problem. On my synth, after replacing the first TL082 others died, so I finally decided to replace all of those TL082s in one shot.

So I suggest you to do the same and replace all the op-amps. If one failed, likely others will follow in the near/mid term. And while you're at it, consider to install IC sockets of high quality where to put the new chips, instead of soldering them directly to the board, just to make future maintenance easier.

 

[zamproject]

I also did a check on the IR3109 feedback and this has no noise

Don't get how you have noise at S/H buffer output and issue at KCV line but a clean 3109 output ?!?

But if it's only the TL082 that's good...because 3109 are not cheap

 

[Mr. Mazz]

Don't get how you have noise at S/H buffer output and issue at KCV line but a clean 3109 output ?!?

But if it's only the TL082 that's good...because 3109 are not cheap

Apologies, in my haste I meant there is noise at the 3109 output which then is sent to the IM input of the module board, hence the issue. Experience of others points towards it often being the TL082. I'll replace this first and if still prevalent...IR3109 is next.