Roland Jupiter 8 repair

Hello ,

I know this may be a bit OT but is there anybody familiar with the Roland Jupiter 8 on a repair level ....???

I got one with a couple problems and could use some tips to get it back to shape .....

As far as I can notice it has the following issues :

On POLY 1 mode some keys don't seem to work ... the inoperative keys are spread along the scale and they seem to follow some patern ....
All the keys work on POLY 2 , SOLO and UNISON modes ...

There is also a high pitched sound along with some notes or chords , this either in POLY 1 or POLY 2 modes .....

I have the service manual but quite frankly I don't have a clue where to look for this malfunction .....

Regards and :guinness:

Luis

Luis

The best place to ask is over at the Synth DIY list which has many parallels with this bunch of people. Friendly, knowledgeable, interesting topics.

See here :
http://dropmix.xs4all.nl/synth-diy/

Peter

Thanks Peter ...... !

I will give it a try there ......

Best

Luis

I have 2 of them on my workbench at the moment, and I have a service manual here as well, so if you like we can try and fix it by remote control if you are brave enough and have a scope available. It might be an interesting thread for some of the synth guys. :wink:

The first thing you could try is to open it up and throw the 2 dip switches on the front RHS PCB, and then power up. Make sure that you aren't in unison mode, hit the "panel" button, and the unit will then be in test mode. You will then see that the 8 LED's near the switches light up as you play keys. Also the front panel led's will light as well. This allows you to tell which voices are faulty as you play the keyboard. If the VCO pan pot on the front panel is centered, and a voice is dead, then it is not faulty oscillators, it is control voltages, VCF, VCA or envelope generators. If a few voices are dead, it may just be a failed CV demultiplexer.

Anyway, use the test mode (read about it at the back of the service manual) and start by determining which voice numbers aren't sounding, and we will go from there.

Hello Steve ,

Thanks so much for your help on this ........ :grin:

So as per your instructions it looks like voice 3 and 4 are definetly not working and on voice 5 there is this high pitched sound along with the note ...

Another strange thing is that when starting the test the led's where not sequencing normaly from 1 to 8 .... They would go 1 , 2 , 5 , 3 , 4 , 6 , 7 , 8 ... or sometimes 1 , 2 , 8 , 3, 4 , 7 , 5, 6 ......

If I go to POLY 2 mode and them back to POLY 1 they sort of reset and start sequencing normaly again from 1 to 8 .....

Should I look for problems on the module board corresponding to those voices or on one of the controler boards .... ????

:guinness:

Luis

Hi again Steve ,

I just noticed the high pitch sound on voice 5 relates only to VCO 1 as when I pan to VCO 2 that voice sounds Ok .......

Luis

Well, normally the LED's will light in sequence if you tap an individual note over and over. If you play a chord then they will get out of sequence. Do you have an oscilloscope available? If you do we can get down to tracing the voices in question, checking to see if the VCO/VCF/VCA/EG control voltages are present. If not, we go back towards the DAC via the demultiplexers and their associated op amps.

Another thing we can try is to drive the VCF's into self oscillation, or even better, to turn on the noise generator and see if the same voices are still dead. They probably will be as it is unlikely for both VCO's to die on the same voice, but it is worth checking.

So if you can, get a scope, a digital multimeter, an iron, and some desoldering braid or a gun together. The fault may well be on one of the module controller boards, as you have said, as both the dead voices are on the same module board. I will look at the schematics tomorrow and suggest which chips are the likely suspects, then when we get the dead voices running, we can get onto the high piched niose from the other VCO. There is also an individual VCO power supply on each module PCB, so check this as well, as this would kill the 4 VCO's on that board if it were down. The test point to measure this voltage on each board is one of the first things mentioned in the calibration procedure at the back of the service manual. I am at home so I don't have it to hand, but it is easy to find in the back of the manual, so put your meter on it and see. At least it sounds like the dead voices are on the top boards, so you shouldn't have to hinge all the boards up.

All the top boards are on hinges, so don't accidentally remove the hinge screws! There are plenty of earth points on all the boards in the machine, the little yellow studs with wire loops are test points, and the earth ones are marked with a circle with a diagonal slash through it.

Ok , this is what I could gather so far .....

On voices 3 and 4 the VCO 1 works in Triangular , Sawtooth and Square wave but not on Variable Pulse and the VCO 2 is working on Sinewave , Sawtooth and Noise but also not on Variable Pulse .....

I have checked the voltages coming from the controller board (+13 and -13 ) and they are Ok .... The +15 , -15 supply is Ok as well along with the +5 and +10 voltages ....

With the scope I checked the outputs of the VCO's of voices 3 and 4 on D20A and D20B ....they are Ok ....
On test points 2A and 2B I have -15 Volts and according to the manual and comparing to the other board I should have a square wave of +10V amplitude ....

Could the problem be on the waveform select or the waveform converter circuits them ....???
Or maybe I'm missing the control voltage for the Variable Pulse circuit .... ??

:guinness: :guinness: :guinness:

Luis

Good one Jaakko ..... :wink:

Well at least after this I guess I will have a better idea of how this Jupiter 8 works ........ :green:

One thing to be noticed is the built quality .... It's a shame that most modern equipment does not get built like this anymore ....
The service manual could be a bit less criptycal though .... :?

I sure hope to get it back to shape ......

:guinness: :guinness:

Luis

Try looking at what is coming from the PW sample and hold line from the module controller board, and the PWM level line from the DMPX sample and hold line, and compare them with those going to a working board. You should see them on connector 30 and 32 of the module board. It does seem likely that the issue is with the controller rather than the module board. see if you get a changing control voltage as you move the PW slider, or an LFO signal if you set the PW mod source to LFO.

Hi Steve ,

I do have the PWM control voltage on pin 32 of the LM connector but I have nothing on pin 30 (PW/PWM) ..... maybe I'm using a wrong ground reference ......??

Anyway it looks as these control lines are used by both module boards A/B (voice 1/2) and C/D (voice 3/4) .....
So if everything is working on voice 1 and 2 could the problem is on the RWM or Waveform Converter sections of the C/D module ....
On IC 21A and B I have this -15V offset on the output pins (1 and 7) although when I move the PW mod slider I do get a control voltage there in the form of a variable amplitude square wave ...
When I go to Test Points 2A and 2B all I get is the -15V offset ....

Does this make any sense to you ...????

Thanks very much for your help on this .... !

Luis

The 3 lines you need to look at are 28, 30 & 32. Line 32 should be a varying control voltage, and 28 and 30 are switching logic to IC 17 and 23. You should see activity on them as you throw the PW toggle switch on the front panel. From the looks of the diagram, IC 23 is used in both voices, so put your scope on lines 30 and 32 while throwing the PWM toggle switch, and see what happens. Then scope pins 4 and 14 of IC 23 and see if the logic is switching the outputs as they should, relative to the chips inputs on pins 3,5,12 & 13 as it is an analog demultiplexer being used as an electronic switch, the layout is clear in the diagram. A faulty IC 23 could give you the problem you describe (it is a 4053). There were a couple of versions of the JP8, I hope our diagrams match, mine is 1st edition, Oct.10 1981.

Yes , we are using the same road map ... Mine is also OCT.10,1981 ....

I have checked the input and output pins of IC23 and aparently it's working Ok .... all the control voltages are present on pins 9 and 11 along with the correct outputs on pins 14 and 4 .....

So what else is comon to both voices on Variable Pulse mode ...??? :?

Luis

BTW , just to make sure the problem is on the Module Board I have swapped the faulty C/D one with board G/H and the problem is the same but now on voices 7/8 ......

So it looks like the problem is on the Module Board and the controller boards are Ok ......

Luis

How is the input and outputs of IC 21 looking compared to the working boards? It is common to both voices. Try working along the PWM path and see where things change relative to a working channel.

Also, check IC 14, which is the waveform selector IC, as it is also common to both channels. This perhaps should be your next port of call on the suspect list.

Looks like IC21 has some problems ..... :grin:

Comparing with a working board I have modulation on the inputs 6 and 2 but with a -15V offset ...... Same thing on the outputs 7 and 1 ....
It seems to be on locked on the -V rail ...

Now the inputs 6 and 2 of this opamp are virtual ground so I should not be able to measure a voltage swing there .... There is nothing there on the working board so I guess I'm correct on this ....

I'm going to replace this IC21 opamp and see if it fixes the problem :roll:

Luis

YEEESSS .... :grin: :grin:

Had a new opamp in and now everything looks Ok on Pulse mode ...
I have used a TL072 instead of the original 4558 .... I guess I will not have any future problem with that ...

So now it's just that high pitch sound on VCO 1 of voice 5 .......
I have just checked and the VCO gets all the modulation signals from the panel sliders .... The range control is also working Ok ....

Any suggestions for that one .....????

Luis

Well done! Now for the ugly one.... Is the noise there on all waveforms? First thing to do is see if it is the VCO, or the waveform convertor. Probe around the VCO sawtooth output on diode 20A cathode and see if the noise is there. If you set the VCO range as low as it will go, and put the unison button on, and play the bottom note, then it may make it easier to see the high frequency tone modulating the sawtooth. Is it at D20A? If not, check out the outputs of the waveshapr converter at IC5 pin 7, and the anodes of diodes 2A and 3A.

If it is the VCO, suspect transistor 1A, there transistors always go leaky in Roland gear, or if the main pitch of the VCO is off, perhaps the RC3046 is failing. Look out for suspect caps as well.

Is the oscillator responding correctly and in pitch with a HF sound superimposed? or
is the vco just way above pitch? Have you checked for cruddy trimpots on the VCO range and scale and high trims?

Anyway, congratulations on fixing the other 2 VCO's.

Thanks Steve .... :grin:

I checked the VCO output and it looks to be just way above pitch ....
The output waveform is Ok , but with a higher frequency (about 4 x ) comparing to other VCO's .....
All the functions are working normally and the VCO is receiving all the correct modulation voltages ....

I will dig into it a bit more and let you know .....

Regards and :guinness:

Luis

YESS AGAIN ..... :grin: :grin: :grin:

I have found the problem on the VCO 1 .....

The opamp IC1A was not good and it was causing an offset on the VCO 1 control voltage ...... !

Another good old TL072 and the VCO 1 is working on the correct range again ...

Now just to get the Jupiter as new again I would need to find a Pitch Bender lever since mine is broken ....

Any idea where I could find one ...???? Or a complete assembly ...???
Or maybe some part from other keyboard that would fit ... :green:

:grin: :grin:

Luis