JX3P Index ] [ Pictures ] [ Tips-N-Tricks ] [ Modifications ] [ Schematics ] [ Links ] [ Troubshooting FAQ ]

We are here: [ home ] -> [ roland_jx3p ] -> [ jx3p_troubleshooting_faq ]

Roland JX3P - Trouble Shooting F.A.Q.

This is not really a list of frequenty asked questions, but a list of some questions which people asked me (and the answers I gave):

Q: A friends Korg Polysix has been destroyed by battery leakage. Should I replace the battery in my JX-3P to avoid this?
A: The JX-3P uses a lithium battery which will not leak that much as the rechargeable NiAh-battery in the Polysix and the leakage will not be as agressive. So you do not have to fear that a dying battery night destroy your synth immediately. But nevertheless any 20 years old battery should be replaced. The JX-3P uses CR1/3N Lithium batteries, which are still available (even in the same pinout as the original one).

Q: When I turn the JX-3P on, it makes a buzzing noise like an old fluorescent light bulb. You cannot hear the noise in the headphones or through the keyboard outputs, but you can hear it outside the keyboard. The noise is coming from the right side.
A: This is a mechanical noise coming from the transformer of the powersupply. Those transformers have some damping material inside, which shall avoid mechanical hum and noise. This material starts drying out after 20 or 30 years. What you hear is normal for old transformers. There is no danger. If the sound annoys you, you might replace the transformer, but it is not really necessary.

Q: The keyboard is out of tune. When I play a C, it is off by about 1/2 to a whole note. I tried the tuning knob on the back and it didn't help. The sequencer notes played back are out of tune with the keyboard notes!
Also: the pitch bender doesn't bend down the full amount. it will go up a fifth, but not that far down. when you hold a note and switch the pitch bender switch (wide/medium/narrow) the note changes
A: Both symptom hint for the same cause: The bender sends continuously a "detune" "to the synth. I assume, the bender is trimmed wrongly or one of the trimmers is gone bad. In worst case the bender itself is broken. But first you should try to correct the trimming: There are two the trimmers VR6 and VR7 on the left half of the panel pcb.

Q: My Speed/Edit/Brilliance potentiometer is crackling/broken
A: The first attempt should be to clean the fader. For this remove the frontpanel pcb from the panel, unsolder the fader, open up the housing (by lifting four metal clamps with flat pliers); then clean the graphite track and the wiper track with "caig faderlube 5" "using a q-tip. Then remount housing and solder it back to the pcb. Do not simply spray caig faderlube in the potentiometer; it will leak on the pcb, become sticky and dust will collect. This sticky dust pad will shortcut other circuits.

If cleaning The exact spare part would be an ALPS fader, type S3018P405B15 (100k lin). But I assume, you won't find it in a shop. So the best trick should be to search someone who bought a SH-101 repair set from ebay. The SH-101 uses the same faders. May be you can get one.

Q: I'm fixing up a JX-3P right now and having some headache regarding a single voice which has zero resonance (all the other voices are ok). The cutoff is working but changing the resonance parameter doesnt have any effect on this voice (ch 6). The Resonance CV is correct all the way up to the TR103/TR104 transistor pair. So I thought the problem would be the IC 103 (or 503 in this case for channel 6) OP amp so I replaced it but unfortunately the problem remains :(
A: It cannot be the IC503, because you would not hear anything. My guess are all parts around TR103, TR104 R128, R117, R125, R126. Check for cold solder joints. And check whether the potentiometer really makes contact everywhere.
Also I would measure in comparison. Means: press hold and then six notes. Then check the signal with oscilloscope on R425 then on R525. Then at TR404, then 504. And so on.
You definitely have to control the audio signal in this resonance path. Either simply using an audiocable and a mixer, or (better) with an oscilloscope. Do this again comparing with an working channel.
(final response back then: You solved my problem! Just like you predicted, it was a broken solder joint.)

Q: My friend is moving to europe so he is going to sell me his jx-3p super chepa. Only issue is that one of the voices on it does not work. I have to assume this is a dead voice chip. What is the best/cheapest  source for these?
A;As first: There is no dedicated voice chip in the JX3P. So something more discrete is gone

Do the following:
* switch the synth off
* press and hold PrgNo.2 and switch on the synth. The LEDs will flicker a little bit from now on.
* watch the LEDs 11-16 the show which voice is played.
* select program A1
* play one note six times repeating with a pause between each note. This will play one voice after the other. Check wich voice is the dead one.

If you can identify a dedicated voice as broken:
* switch the synth off
* open the front panel (two screw inside each sidepanel, four screws at the rear bottom)
* switch on the synth and select program A1
* press hold and press six notes
* use an oscilloscope or some audio testing (open cable to mixer...) and check at the corresponding voice. Voice 1 is most right on the voiceboard, Voice 6 is the most left.
* check the divider-outs at the following pins:
voice1  DCO1 IC-50 pin-10
voice2  DCO1 IC-50 pin-13
voice3  DCO1 IC-50 pin-17
voice4  DCO1 IC-49 pin-10
voice5  DCO1 IC-49 pin-13
voice6  DCO1 IC-49 pin-17

voice1  DCO2 IC-48 pin-10
voice2  DCO2 IC-48 pin-13
voice3  DCO2 IC-48 pin-17
voice4  DCO2 IC-47 pin-10
voice5  DCO2 IC-47 pin-13
voice6  DCO2 IC-47 pin-17

These should provide pulsewaves.

If these provide all good signals:
* check the DCO-signals from a working voice and the dead voice: IC n04 pin 1 for DCO1 and IC n04 pin 7 for DCO2. "n" "is the count of the voice.

If there is no good signals, then check the Control voltage coming into pin2 of IC n04 and pin6 of IC n04.

If the DCOs show good signals at the dead voice:
* check the signal before and after the VCF-Chip IR3109 - pin 2 is before, pin 15 is output.

If there is no output at the VCF
* check the control voltage for this VCF on the dead voice and on a working voice with a voltmeter. Check it at pin 9 of the filter.

If there is output:
* check at the Testpoint TP-n of the corresponding voice (its VCA-out).

If there is no signal
* check the control voltage for the VCA at the resistor, which connects the S&H-stage with the VCA-transistor (unfortunately the number is unreadable at the schematic).

If all the control voltages are wrong then assumingly something on the controlling addressbus P0,P1,P2 is wrong

Q: I have recently tried to adjust the resonance of my JX-3P. I also tried to tune to VCF.
When i play the organ patch it is out of tune and sounds very bad.
When i turn the resonance down it fix some of the tunning up.
How do i go about returning the  VCF and Resonance so it is in tune?

A:You have to distiguish the "filter tuning" "from the "filter octave tracking".
* Tuning means: all six filters resonate on the same frequency, if the keyboard tracking is set to zero.
* Octave Tracking means (or would mean): If I set keyboard tracking to a certain value (typically 13), then each filter plays clean scales with the keyboard.

The tuning can be trimmed as I described on my page.
The octave tracking can not(!) be trimmed. It is set by an fixed resistor R109,209,....,609. This resistor has a precision of 10%. So the exact octave tracking may vary between the individual voices.
With the standard resonance setting the difference usually cannot be heard, but if you increased the resonance trimming very much, then the octave tracking veriations between the voices may become more appearent.

You now have three possibilites:
* Either you reduce the resonance again
* or you do not use the organ preset
* or you replace R109,209,....,609 by a resistor of 39k in series with an variable resistor of 10k. Now you can trim the  octave tracking amount.

See also...

to top of document

© florian anwander Feb. 2000