The No.1 Website for Pro Audio
Jupiter 6: Replaced out of tune oscillator..STILL WON'T TUNE..any thoughts?
Old 27th February 2012
  #1
Here for the gear
 
🎧 5 years
Jupiter 6: Replaced out of tune oscillator..STILL WON'T TUNE..any thoughts?

I have a Roland Jupiter 6 that has a tuning issue. Oscillator 6 VCO 1 will not tune to the rest of the synth. It has been to a reputable service shop and after 2 months on the bench my guy gave up and returned the JP6 to me.

While in the shop he replaced both of the B oscillator (aka osc 6) cem3340 chips on the 2 voice board. The problem specifically is the B osc on vco1. It is always about a 1/4 tone sharp no matter how many times you hit the tune button.

He also replaced most caps and resistor stacks associated with that oscillator without any change. This leads me to believe that the problem lies with another chip....maybe the computune IC or the slave cpu chip 8051. Not sure if that would cause the same problem on the same oscillator consistently?

Has anybody had any experience with repairing a similar problem. I have seen a couple of posts where replacing the cem3340 fixed the issue. Unfortunately that did not work here. Any thoughts of where to go next with this would be appreciated.
Old 27th February 2012
  #2
Lives for gear
 
🎧 10 years
First thing you need to do is seek out another tech. One who is competent enough to understand that if the 3340 did not fix the issue then the problem lies in the S&H section as you have a bad autotune CV, detune CV, or keyboard pitch CV coming from somewhere in there. Preferably one who understands the concept of hybrid analogs and demultiplexed S&H CV circuits. Good luck.
Old 28th February 2012
  #3
Here for the gear
 
🎧 5 years
Finding a tech with intimate knowledge of the JP6 would be a dream come true. Until I do I am going to try replacing some parts and hope I get lucky.
I have heard that the 4051 chip has been known to fail and is used throughout the S/H system.
There are 3 that seem to be involved in the auto tune and cv.
IC 20 uses the 4051 in tandem with the autotune itself.
IC 15 uses the 4051 to communicate width to the vco's. Width is the term used in Roland's service manual for the autotune function.
IC 16 uses the 4051 to communicate frequency and cv in the vco's.
There are also resistors that link the s/h and vco's.

Any experienced guess at which one to start with would be greatly appreciated.
Schematic is found here:
http://www.wiseguysynth.com/larry/sc...jp6-page12.pdf
Old 28th February 2012
  #4
Lives for gear
 
LeeYoo's Avatar
 
🎧 10 years
No experience in this Roland, but looking at the schematic, the first thing I would replace is the 40xx "in and out" chips for that VCO.
Relatively easy to MARK and SWAP 2 IC's. One set at the time. Compare surrounding resistor values when IC's are out.
With every swap, you can remove two IC's off the suspect list.
I doubt it is CPU related.
Leo..
Old 28th February 2012 | Show parent
  #5
Lives for gear
 
acreil's Avatar
 
🎧 10 years
Quote:
Originally Posted by The Real MC ➑️
First thing you need to do is seek out another tech. One who is competent enough to understand that if the 3340 did not fix the issue then the problem lies in the S&H section as you have a bad autotune CV, detune CV, or keyboard pitch CV coming from somewhere in there. Preferably one who understands the concept of hybrid analogs and demultiplexed S&H CV circuits. Good luck.
Yeah. Aren't the 3340s in sockets? He should have swapped two of them to see if it changed anything. Sounds like it's a "throw parts at it" kind of guy. Sometimes that works, but these are very expensive parts!
Old 28th February 2012
  #6
Here for the gear
 
🎧 5 years
Swapping out a pair of 4051's at a time sounds like a good place to start. Unfortunately nothing on the board is in sockets.
Forgive my lack of experience here ... I don't want to damage the board. Is there a soldering device that can heat all the pins on each side of the chip at once, or do they get heated and lifted off the board one at a time? I've been warned to be careful not to damage the trace.
Thanks for sharing the voice of experience here.
Old 28th February 2012
  #7
Here for the gear
 
🎧 5 years
Okay... I may be a novice at desoldering chips. I found some helpful video on youtube. It appears all that is necessary is a good soldering iron and a desoldering pump.
I have a good iron and I'll get a pump. Mainly have used desoldering wick in the past, but the pump looks like the ticket.
Also it looks like the best way to do this is to install sockets as the chips are removed and snap the replacement chips into the socket. Saves wear and tear on the board for future work.
Now all I need is a bit of luck to find the right piece that is skewing my tuning.
Old 28th February 2012
  #8
Lives for gear
 
🎧 10 years
This is why "swapping parts out until it works" is always a bad approach.

Being a novice you WILL damage the PC board trying the swap soldered parts around. If unsoldering 4xxx CMOS the risk is even higher for damaging the IC. Trust me I was there.

The correct approach is to probe the flaky CV from the 3340 up stream until you find where it is out of whack. Oscilloscope is highly recommended. You may find noise on the CV and have to trace your way back, or you may have a bad power rail. Only a 'scope will tell you that. This approach is quicker and reduces the risk of damage.
Old 29th February 2012
  #9
Lives for gear
 
LeeYoo's Avatar
 
🎧 10 years
Swapping IC's should not be a problem, but you obviously haven't got the experience.
Get some old circuit boards from the dump, and try to get chips and other parts out.
You will soon find the best and quickest way to do it.
I usually use a pump, a medium size one, but fresh wick also works.
A quick resolder with fresh solder before you unsolder makes the solder less "sticky".
Read up about ESD.
Make sure the gear, AND you AND the soldering iron are (and stay) on the same potential.
Put parts on paper, or antistatic foam. Not on plastic.
If unsure, use a bracelet, antistatic mat, and ground your iron to the same potential as the mat/gear.
Don't use a cheap iron. Use one with a transformer. Or if you have to, attach a wire to the (coldest) metal part of the iron, and clip the other end to the chassis/ground of your gear.
Personally, I don't like using sockets.
And... be very carefull with those 4051s. They are at least $0.50 each....
I doubt if you can find the problem with a sope yourself. The VCO chip seems to be in a CLOSED LOOP with the CPU.
Anything in that loop will try to correct/hide the faulty part.
Leo..
Old 1st March 2012
  #10
Here for the gear
 
🎧 5 years
Thanks for the desoldering advice as well as your read on the closed loop with the cpu. That would explain why my tech had difficulty tracing the problem. He was referred by a respected pro and just couldn't get the support from Roland on this old hybrid. The synth is older than the support people.
I am gathering some good tools to help insure the best outcome. I already have good soldering iron / station. I ordered a high end desoldering pump rated esd safe and will use a bracelet to keep the work and myself on the same plane. I think a practice session on a junk pcb is a good idea too. I have a piece of gear in mind for this. I do have experience pulling caps and pots. resistors and such.
I am also getting advice from Doug at synthparts who provides parts for these old Rolands. He is a contributor to this and other forums. Swapping parts was also his advice.
I am going to swap between the 2 voice and the 4 voice boards to see if the problem moves after each swap and identify the corrupt part in this way. The simplest swap is the comparator. After the comparator (which is the easiest..only 4 pins) I'll go for the DAC's. Then the 4051's and so on.
Probably will have the tools by Monday and then....I'll post the outcome.
Although I don't want to wreck the Jupiter 6 I have nothing to loose. Its just not useable in its present state.
Old 1st March 2012
  #11
Here for the gear
 
🎧 5 years
Is some sort of implement necessary for lifting the pins? If so what works best?
Old 1st March 2012
  #12
Lives for gear
 
LeeYoo's Avatar
 
🎧 10 years
Hi pianohead.
Once you have pumped out all the solder, you can just lift the IC out.
I hope the board is not double sided. This makes it a bit more difficult.
As I said, a quick reflow with fresh solder adds more resin around the spot.
Easier and faster to heat up the pins.
Heat up the IC pin 2-3 seconds longer after the solder melts.
Pump in your other hand. Put over the pin, and suck within 1 second.
Let the chip cool down for 10 seconds before your next attempt.
I would start with the 405x chips. Replacements are very cheap and easy to get.
Leo..
Old 3rd March 2012
  #13
Here for the gear
 
🎧 5 years
Swapped a couple chips today. The 8051 cpu's and the 2732a's from IC5. These were socketed so no problems. Unfortunately it did not change the tuning issue.
I did notice an interesting phenomena though. When the keys are struck with much greater velocity the B voice vco 1 (the problem) oscillator is not nearly as out of tune. Is this a clue?
Still waiting for desoldering tools to arrive. I think my first targets will be the the 4051 on IC20 then the comparator. I'll leave the DAC (if necessary) until I've experienced a couple of desolderings.
Old 6th March 2012
  #14
Here for the gear
 
🎧 5 years
Received tools today and got to work on the Jupiter 6. Swapping parts between the 2 voice and 4 voice boards to try and see which part causes the problem to move. First I swapped the 4051 multiplexer on IC20. Then swapped the comparator. Next swapped the 4051 on IC15. All these swaps made no change in the tuning problem.

Next I swapped the 4051 multiplexer on IC16. The tuning issue did change. It did not move to the other board as expected. It is still on the 2 voice board...but instead of vco 1 voice b being 1/4 tone sharp all the time it is now about 2 octaves flat all the time. Also vco2 is now being affected.
On the 2 voice board:
Vco2 voice Γ€" is always about 2 octaves flat.
Vco1 voice "b"is always about 2 octaves flat.

Not sure if I damaged the 4051 chip or if is related to the original problem. I did have to resolder a couple pins several times... which I did not have to do on the first couple of swaps. Ordered some replacement 4051's today.
Just trying to get a bit of experience before I swap the larger DAC chips.

If anyone has a minute to look at this and share a clue... please let me know.
The service manual for the jupiter 6 can be found here.

http://fa.utfs.org/diy/rolandjp6/Rol...-6_Service.pdf
Old 6th March 2012
  #15
Lives for gear
 
LeeYoo's Avatar
 
🎧 10 years
Hi pianohead.
I guess you should be happy that you have found something that changes the problem.
Next step is to try to understand how and why.

I will try to walk you through the workings of a 4051 in case you don't know yet.
It basically is an electronic version of an 8 position switch.
The common pin is pin 3.
The 8 positions are pin 13, 14, 15, 12, 1, 5, 2, and 4. In the same order as on the schematic.
The switch is operated with a 3bit code on pin 11, 10 and 9.
A bunch of 4051's is driven by that 3bit code, so I don't expect problems there.
Same goes for the common pin, pin 3.
Supply rail pin 16 VDD and ground/VEE pin 8/7 is common with other IC's too.
No harm in measuring pin 16 (+15v) and pin 7/8 (-0.6v).
I would first concentrate on the sample and hold buffers on the outputs of your 4015.
A voltage from the D/A (IC12/TP1) is selectively (the 4051) put onto the 8 output caps, and buffered by op amps e.g. IC23 and 29.
Something migh be wrong there. Time to swap TL064's.
Another not shared pin is 4051/pin 6. IC9 enables one 4051 at the time.
Haven't read the full manual yet (pfew), but I thought to give you already a heads up.
Leo..
Old 6th March 2012
  #16
Lives for gear
 
acreil's Avatar
 
🎧 10 years
You should just replace the 4051s rather than swapping them. They're not worth reusing. I wouldn't expect the CPU or RAM chip to make a difference, similarly it's probably not the DAC. It's more likely either the sample and hold buffer (TL062) or a passive component somewhere. It could also be a problem with the auto-tune circuit (IC20), or maybe there's something happening where the CVs are mixed...
Old 6th March 2012
  #17
Lives for gear
 
LeeYoo's Avatar
 
🎧 10 years
+1
IC20 seems to be the return to the CPU, closing the loop.
But I would expect the problem to lie around the IC that made a big difference.
So you swapped IC 16 with ??
Don't think chips are the only parts that can go bad.
Leo..
Old 6th March 2012
  #18
Here for the gear
 
🎧 5 years
Looks like the TL064 op-amp at IC23 or IC20 is a likely suspect. I've ordered some TL064's along with some 4051's from ebay. They both are relatively inexpensive...so rather than swap I will try replacing some of these chips around IC16...where I caused the tuning problem to change.
Keeping fingers crossed. It would be so nice to have the Jupiter 6 back.

I tried to measure the capacitors at IC16 (and other locations) but get different values every time I test them. Guessing that oxidation on the solder could cause this. Never have had much luck getting readings on capacitance. Hopefully its a chip and not one of the capacitors here. All resistors have been in spec so far.

Will continue to post when parts arrive. Thanks for the input!!!
Old 6th March 2012
  #19
Here for the gear
 
🎧 5 years
Hi Leo,
IC16 on the 2 voice was swapped with IC16 on the 4 voice board.
The problem remained on the 2 voice which makes me suspect the TL064 at IC23 on the 2 voice board (that feeds the IC16 multiplexer). I think you may be right when you suggested its time to swap some TL064's. Ordered a couple of replacements today. I'm going to replace the TL064's at IC23 and 29 as well as the 4051 at IC16.

Any suggestions on how to get a reliable capacitance reading here? My multi meter has settings for capacitance... but gets different readings almost every time I test it.

Thanks for the tips.
Old 6th March 2012
  #20
Lives for gear
 
LeeYoo's Avatar
 
🎧 10 years
Those caps are 100n/10% film?
Do you measure the caps in circuit? Might give false readings.
I think low leakage is more important than absolute value here.
I just realised, the boards are double sided. Inspect carefully after soldering.
Leo..
Old 8th March 2012
  #21
Lives for gear
 
3 Reviews written
🎧 15 years
Did your tech or you did any calibration / adjustments procedures before troubleshooting?
Old 11th March 2012
  #22
Here for the gear
 
🎧 5 years
I did try calibrating the synth with a good multimeter and an oscilloscope. Could not get the linearity on the troublesome voice to line up.

Traveling for a few days. Will get back to working on synth when I get back home.
Old 9th September 2015
  #23
Here for the gear
 
🎧 5 years
Hello pianohead

Now it's a long time ago, but any news about this matter?

I'm very curious to know, since I have a Jupiter 6 here.
The 4 voice module board is perfectly in tune, but the 2 voice one does not line up correctly after pressing tune.

In test mode, the tune button LED stays on, so it should be something with the D/A Converter that is not working.

As Leo mentioned earlier I was going to change the 4051, because they are easy to get.

Anyway reading through this makes me more sceptical that this will work out.

I have not tried yet swapping the CPUs, maybe I should?

What else could be the problem?

Did you succeed in the end and got so excited that you forgot to post here?
Old 29th September 2015
  #24
Here for the gear
 
🎧 5 years
Update: I was able to fix the Jupiter 6 in one go by replacing the IR9022 chip.
Old 20th November 2015
  #25
Here for the gear
 
🎧 5 years
Thanks for your update. I was not successful with my attempts to repair jp-6 in 2012. The synth has been shelved until yesterday. Decided to try again. Your input may prove valuable.
Old 26th January 2016
  #26
Here for the gear
 
🎧 5 years
Hi pianohead.

Not sure if it will work out in your case as it seems not to be 100% the same issue.

Anyway good to see you back here and good luck with the repair, if you have any questions you can always post them here.

I was lucky with my repair job at first, but was also a bit unlucky as we noticed later on, that on the other board one of the CEM3340s was not working... that turned out to be hmmm a bit expensive, but now the JP-6 works like a charm!
Worth also to go through the whole adjustment procedure, although it takes a while to get into
Old 16th November 2017
  #27
Gear Nut
 
egloss's Avatar
 
🎧 5 years
Hello, so many year after...
was you Jupiter 6 ever fixed?
im curious about it
Old 10th April 2021
  #28
Here for the gear
 
droveloder's Avatar
 
I have a similar issue with a JP-6, but with a twist, that may shed some light on the problem.

I can best describe it by using a simple example: Switch panel mode to LOWER. Make a basic patch, mixed with both oscillators 50/50, and VCO-2 tuned to perfectly match VCO-1. Now save the patch. Now switch to UPPER mode. Recall the patch you just saved and play it on the upper keys. On my JP, VCO-2 is WAY too sharp (like 1/3 semitone), and I have to turn the TUNE knob down to about 9 o'clock to restore perfect tuning. If I were then to save this modified patch and recall it in LOWER mode, it would be too flat, and to correct it I would turn the TUNE knob back to 12 o'clock.

Interestingly, in LOWER mode (which has correct tuning), whether you're in 2-4 split or 4-2 split, the tuning is always correct. Likewise in UPPER mode, regardless of which voice board is handling the voices, it always has this sharp tuning. It's like it's being sent out from the main CPU with incorrect values. And I know it's not the TUNE pot or anything on the panel board because this problem surfaces even when just recalling a patch (before touching any knobs).

This is also a problem in WHOLE mode, where all 6 notes of a chord will be too sharp.

I wonder what it is about UPPER and WHOLE that is taking the same patch parameters from memory and sending out different values than when in LOWER mode?
Old 12th April 2021 | Show parent
  #29
Here for the gear
 
droveloder's Avatar
 
Quote:
Originally Posted by droveloder ➑️
I have a similar issue with a JP-6, but with a twist, that may shed some light on the problem.

I can best describe it by using a simple example: Switch panel mode to LOWER. Make a basic patch, mixed with both oscillators 50/50, and VCO-2 tuned to perfectly match VCO-1. Now save the patch. Now switch to UPPER mode. Recall the patch you just saved and play it on the upper keys. On my JP, VCO-2 is WAY too sharp (like 1/3 semitone), and I have to turn the TUNE knob down to about 9 o'clock to restore perfect tuning. If I were then to save this modified patch and recall it in LOWER mode, it would be too flat, and to correct it I would turn the TUNE knob back to 12 o'clock.

Interestingly, in LOWER mode (which has correct tuning), whether you're in 2-4 split or 4-2 split, the tuning is always correct. Likewise in UPPER mode, regardless of which voice board is handling the voices, it always has this sharp tuning. It's like it's being sent out from the main CPU with incorrect values. And I know it's not the TUNE pot or anything on the panel board because this problem surfaces even when just recalling a patch (before touching any knobs).

This is also a problem in WHOLE mode, where all 6 notes of a chord will be too sharp.

I wonder what it is about UPPER and WHOLE that is taking the same patch parameters from memory and sending out different values than when in LOWER mode?
I fixed it. It turns out it was the two trimpots on the bender board. The one for VCO-2 was fairly far out of calibration, and a simple adjustment fixed the issue described above.

I know that the service manual prescribes a procedure to calibrate them using an oscilloscope, but you can just do it by ear--like tuning a guitar. Just undo the screws for the top lid of the synth and swing it back. On the bender board there will be two trimpots that you can access without having to remove that board (I believe they are labelled VR1 and VR2). Make sure that both VCO1 and VCO2 selector switches on the bender board are deselected. Put the synth into one of the split modes, so that you can toggle BENDER on and off for the keyboard section you're auditioning. Enable HOLD and press a key. Turn the MIXER knob all the way counter clockwise so that just VCO1 is playing. Adjust trimpot VR1 and toggle BENDER on and off so you can hear the difference as the influence of the bender board is applied/unapplied to the voice. When it is perfectly adjusted you will not be able to detect any difference in pitch when toggling BENDER. Then do the same for VCO2 (MIXER all the way clockwise). After that, put the MIXER at 12 o'clock to have both oscillators playing. Also put the VCO-2 TUNE knob at 12 o'clock, so that the two oscillators are so close in tune that there's a very slow phasing effect (beating). If the bender board trimpots are perfectly adjusted then the overall sound of the playing note will be the same as you toggle BENDER on and off.
πŸ“ Reply
Post Reply

Welcome to the Gearspace Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…

Forum Jump
Forum Jump