JP8000 Ribbon "relative" and "hold" buttons not working

Hi udream,

Exactly the same issue here.

I replaced the battery, did do the necessary calibration of pitch bend and ribbon according to the service manual, and also upgraded to latest firmware version v1.05. Bend and Ribbon work absolutely fine over the full scale, BUT:
When I push "Relative" or "Hold" buttons, corresponding LEDs light up correctly, but no function. Means, the two-pole switches are working fine. Either a deeper electronics- or a software issue.

Addendum : According to my understanding,
The ribbon sensor is a variable resistor. This function could be also solved with a capacitive sensor but appears to me here as an unnecessary and too complex approach. The ribbon value is constantly (or upon interrupt?) Analogue/ Digital converted, we may also say sampled, and the resulting value is been interpreted by the microcontroller according to its assigned meaning.
Means, if this value shall only represent cutoff, this value only influences cutoff, if also resonance shall be represented, the value also influences this parameter and so on. Interpretation is to be done by the microcontroller. Which parameters shall be influenced is subject to predefined assignment of the sound patches / performance. This can be changed via the "Ribbon Assign" button / function.
The "Hold" button / function is doing the following: it takes current ribbon value and stores it temporary into volatile memory (RAM). The firmware tells the microcontroller where to find this value in memory and how to interpret it (see above).

As after calibration my ribbon sensor (so as the bender) show correct values over the full range in accordance with the service manual, both ribbon sensor and A/D conversion (sampling) should work properly.
In addition, when I release the sensor, value switches back to lowest value "0". The service manual says, when you push "Hold" , current value is been read at the time when you *release* the ribbon sensor.

Remark : when I enter test mode and select Ribbon [4] , it never says Ribbon "OFF" at the beginning as shown in the manual but shows two flickering hex numbers.

Furthermore, the "Hold" button LED lights up upon activation means the switch itself is working fine.
Activation of "Hold" should trigger two independent actions: light up the LED (ok) AND temporary store the sampled value in volatile memory (we don't know if this works correctly). Same story with the "Relative" button / function.
Finally, this sampled/stored value needs to be correctly interpreted by the microcontroller. No change tells me: the microcontroller simply doesn’t know that he requires to take action OR storage of the value didn’t work out correctly. I'd suspect this path should be checked first.There are e.g. digitally controlled analogue switches in this path which are subject to failure in a lot of dated electronics equipment including vintage synths so could be one candidate. In the end it could be also a weak connection / lose cable.

Would be great to get comments on the above from someone technically skilled who has better insight in / experience with this synth and this issue.

Anyone with same issues, anyone who could get this sorted?

Finally, in my opinion the JP-8000 is a fantastic but underrated synthesizer which is sadly often only limited to its supersaw. There is soooo much more it is been capable of, a real great piece of Japanese electronics design.

Kind regards,
Pete66

OK, I detached the ribbon and made some measurements. It's a 22K Ohm linear resistive type of sensor. When touched, full scale between a few Ohms up to 22K can be achieved. I temporary replaced the sensor with a potentiometer (and a resistor at its middle position), ran the calibration routine and achieved identical results vs. the ribbon. I turned the poti into any position, pushed the "Hold" button and disconnected the wire from the middle (= release finger from ribbon). Result: same as with ribbon, current value is not been captured.
At the end of today I'd say, I'm pretty much certain the ribbon is not the root cause. Ah, just to add: the ribbon sensing line goes to the basis of an NPN transistor Q101 which I checked with a simple diode test (B-C, B-E; plus at B, result around 0.7V each as expected). This transistor seems to be fine as well.

To be continued...

Hi folks,

Anyone with a fully functioning JP-8000 *and firmware version 1.05 installed* who can confirm the initial message displayed when entering test mode and then selecting [4] , Ribbon Set & Check? The service manual says "Ribbon OFF", I always get a hex value around 8C...8F displayed, corresponding to decimal value 140...143.

Would be very nice to get some feedback - MANY thanks in advance!

Kind regards,
Pete66

Really no-one?

Test mode can be entered easily and will not cause any harm or change of your system:

1) Turn on power while holding [OSC2 SYNC], [-12dB/-24dB] and [FILTER TYPE] to enter Test Mode. Wait until display indicates "[1] MIDI Test."

2) Press the [4] button to display "[4] Ribbon."

What is been displayed here, with a properly functioning JP-8000?

Don't be afraid, nothing will be changed / damaged when you are in this mode. Just leave it with...

3) Turn off the power of the JP-8000.

Thanks in advance.

Kind regards,
Pete66

Hi Piet66,

I just got to your message. I will do this for you in the morning. It is 8:50PM here.
A question though, how do I check my firmware version to see if it is 1.05?

Cheers,
mockba_the_borg

Hi mockba_the_borg,

This is great!
To check the firmware version :
1. Turn on the power while holding down [LFO1 Waveform] [1] and [3].
2. The version number appears on the display.
3. After checking the number, turn off the power.

Kind regards,
Pete66

Hi there,

Without touching the ribbon it shows only "OFF" below the "[4] Ribbon".
By touching the ribbon it shows an hex value which goes from around 0CE to 3EE as you move your finger.
Hopefully it helps.

Let me know if you need any more tests or info I could provide.

Cheers,
Mockba the Borg

Hello Mockba the Borg,

Thanks very much, this is indeed very helpful information.
Can you please confirm your firmware version.

The fact that "OFF" is been initially and correctly displayed (without any touch) shows to me that my synth seems to have somewhere a sort of "unwanted trigger". Also, something seems to "automatically" pull the ribbon resistor to a certain (low) value.
My headaches here, this also occurs when I fully disconnect the ribbon sensor. Looks as if there is some electronics behind the sensor responsible for this behavior.

Will do some dedicated measurements / testing and report any results here in the forum.

Thanks again for your kind help!

Best regards,
Piet66

Yeah, interesting.

I do have firmware 1.05.
Looks like you will have to follow traces and look for some stray resistance maybe?

But wait, isn't the ribbon capacitive? I would expect it to be a variable capacitor.

Cheers,
Mockba.

[Update] Never mind, I just saw your 22K pot test above.

OK cheers, our synths then have same OS version, that's a good basis for further investigation and comparison.
And yes, ribbon should be of resistive type.

Interesting new finding:
I suspected that Ribbon Hold function may be *always on* (independent on switch state) and then pulled to a certain value.
For testing this out, I selected preset PERFORM P48 (Back to the 60's) and increased LFO1 Depth to maximum. Both ribbon controller switches, Relative and Hold, are been set to "off".

Now, when I tip the ribbon sensor at any position *from the middle up to the far right position*, current Envelope value is been sampled and hold (remember, hold switch is sett to off!) When I tip the sensor *from the middle to the left position* there is no recognizeable change. When I set Hold switch to "on" no change/ no difference. When I set Relative switch to "on", a value around the middle position is been fixed. This can be released when I put Relative to "off" and tip any position right from the middle.

Keep on investigating.

Best regards,
Piet66

Hi,

Today I finally found some time for further investigation.
I took my oscilloscope and did do some measurements around the "hottest" candidate for the malefunction, IC3, TC74HC4051 on Panel Board B. It looks as if there is a bleeding through of the ADC2 control signal (fed to pin9) to the output signal at output pin 3 (com). There is always a pulse of 0.7V at 3ms to be measured on the output.
I turned all other knobs connected to IC3 to zero (Porta time, LFO2 Depth, LFO2 Rate). The Ribbon delivers an accurate DC signal up to approx 4V. Voltage goes to 0V when I release the sensor. All fine here.

To cross check I measured the output of IC2, another 4051, and as expected no signal here when all related knobs are turned down to zero.

Next is to replace IC3...

After thinking twice, the abovementioned output level of 0.7V corresponds with the voltage measured at RBN2 (PNLC_AD4, pin1 of IC3) which is by default not zero, and it is a constant value (the resistor of the ribbon doesn’t change, only at the tap position RBN1, signal PNLC_AD5, pin 5 of IC3, when touched). Due to the multiplexer this value turns into a pulse as the input pins are scanned sequentially, and the value is been output to pin 3. IC3 might not be the evil one, will neet to check further.

Cheers,
Piet66

Hi,

Just to add that I also checked all related signals which are responsible for Ribbon Assign / Hold / Relative push button scanning. All seems to be working fine.

Not an easy one, unfortunately.

If a Roland technician should read this post: have mercy, any hint would be highly appreciated.

Kind regards,
Piet66

Hi All,

After a longer pause I recently resumed with my investigation. With success!
Today I finally fixed the issue: Root cause actually was ...the ribbon sensor (sic!). As I couldn't get hold of the original part I used a non-compatible but similar sensor (which needs to be implemented slightly differently) but which works quite well (usual calibration procedure in test mode required).
Both Hold and Relative back to life, and the factory patches and performances also sound authentic again.

Special thanks to mockba_the_borg and udream who kindly supported me on my journey.

Piet66

Piet66 wrote:

Hi All,

After a longer pause I recently resumed with my investigation. With success!
Today I finally fixed the issue: Root cause actually was ...the ribbon sensor (sic!). As I couldn't get hold of the original part I used a non-compatible but similar sensor (which needs to be implemented slightly differently) but which works quite well (usual calibration procedure in test mode required).
Both Hold and Relative back to life, and the factory patches and performances also sound authentic again.

Special thanks to mockba_the_borg and udream who kindly supported me on my journey.

Piet66

Some more details can be found here:

https://forums.rolandclan.com/viewtopic.php?f=51&t=53446

Best,
Piet66