electro 4D random note stutter issue

[groover]

.....and i turned off the keyboard for and hour, turned it back on.... and the keys are back...

[groover]

I wonder if it could be a diode used in the addressing of the keys. I am not sure how this is mapped in the Nords, but I guess that there is some multiplexing used. https://en.wikipedia.org/wiki/Keyboard_matrix_circuit
One dodgy diode could be affecting several keys at the same time, perhaps?

[groover]

This last time the notes have disappeared they don't seem to want to come back no matter how many times i press them. Its 4 notes that disappear together, (in this order), A (after middle C), F, C#, and high A. Doesn't matter if i use the octave shift or transpose, its still the same keys gone. I now believe the stuttering may have been when the key kicked back into life (albeit at full velocity) after disappearing, and cutting out again and coming back repeatedly as it struggled to keep working.
My last question before i send it to a tech is:

It can't be a contact issue if 4 keys from 3 different octaves are involved together can it? They all stop working together, and when they come back its together also.

Any thoughts are much appreciated.

My thoughts:
The notes in your failure pattern are separated by 8 notes. This is a typical (byte-wide) scan interval (in the key matrix scanning per Mr.G's post) and suggests a problem with one line going from the keybed to the main circuit board.

The simplest explanation is an oxidized or badly seated connector. Trying to unseat and (correctly!) replace the ribbon cables that run from the keybed to the main board might resolve this.

This doesn't explain the intermittent stuttering note problem...
My best guess for this is a cracked pull-up resistor or cracked solder at the pull-up resistor at key-matrix scanning IC. The purpose of the pull-up is to restore the "inactive" state when no key is pressed.
In a case where the pull-up is missing, the wire to the IC's input pin "floats" making it susceptible to noise and would cause the Nord's processor to interpret the key as still pressed.
All IC inputs leak small amounts of current so the input pin drifts eventually to some constant-voltage state - this would account for the decaying behaviour you reported. The noise-susceptibility means the input wire picks up any nearby electromagnetic noise (probably 60Hz or 50Hz from the power supply). -This would modulate the signal going into the scanning IC and might explain the stuttering.

Circuit board heating expands and contracts components and can open and close cracked solder joints... It just occured to me that cracked solder on one of the scanning IC's input pins would disconnect the input pin from it's pull-up (and from the keybed wiring). This could explain the dead keys too. [Probably the IC --As pablo suggested]

If you note which keys exhibit the stuttering behaviour, it might help aid diagnosis.

Bet you regret asking for "Any thoughts".

Sorry to hear you're having these problems. I hope they get resolved quickly. Thanks for sharing and I hope you find the time to share any findings you our your friendly Nord technologist make.

-pterm
Edited to restore opening "[\quote]" delimiter, add Smilie

[groover]

Thanks for that. I'll make sure i get some information when i get it fixed and share it with you all

[pablomastodon]

+1 to the Mr. G and pterm

something in the data line path of the upper keybed contact PCB. Could be ribbon cable, socket, solder joint, or even scanning IC at the far end of that data line path. You'll need a tech to sort that out for you. Please let us know how it goes.

Bless,

Pablo

[groover]

Thanks. I was all ready to box it up this morning and send it away, but now i have all notes, and have had since last night. I've been playing and playing but it seems to be working now with no problems. I'm pretty sure this isn't going to last, but i don't want to send it away without it displaying a problem to fix. I guess i'll wait for it to happen again then i'll send it away. And i will let you all know what the tech has to say.
Cheers

[groover]

Hi guys,

I pulled apart the electro just to make sure all the ribbon cables were properly seated (as per pterm's suggestion), and i think i found that the little one that goes from under the keybed to the main board was only half way in. There is 3 from the keybed to the main board, 2 larger ones i believe handle half the keyboard each and one small 6 conductor ribbon cable with its own little circuit board screwed onto the bottom of the keybed. (I'll post a pic that i found showing the connector i mean). It was the connector into this little circuit board that i think wasn't fully seated down. Does anyone think that this could cause the issues i had. The keyboard is performing normally now, but this problem was intermittent so i may not have found the culprit and it may reappear at any moment.

Also i have swapped the ribbon cable that controls the top half of the piano with one from an electro 3 thats not working at the moment to see if maybe it was a faulty ribbon cable. Maybe i should have just swapped the 2 ribbon cables around in my piano and see if the problem pops up in the bottom half of the keyboard instead... Then i would know it was that cable.
Let me know if you think that semi loose connection into that little circuit board could be the culprit that could cause such an issue, I'm not sure what that smaller cable is used for, might have nothing to do with it......
Thanks guys

[pablomastodon]

that little one has nothing to do with it -- that's a lookup table for the key calibration which is loaded into memory during boot.

Bless,

Pablo

[groover]

There is 3 from the keybed to the main board, 2 larger ones i believe handle half the keyboard each and one small 6 conductor ribbon cable with its own little circuit board screwed onto the bottom of the keybed. (I'll post a pic that i found showing the connector i mean). It was the connector into this little circuit board that i think wasn't fully seated down. Does anyone think that this could cause the issues i had. The keyboard is performing normally now, but this problem was intermittent so i may not have found the culprit and it may reappear at any moment.

that little one has nothing to do with it -- that's a lookup table for the key calibration which is loaded into memory during boot.

Interesting: - I've never seen that calibration device on this forum before (heard about it, but never saw it).

There is a small possibility that the poor connection you described caused the misbehaviour. I have no insight into the software running on Nords so this is just speculation: Assume that the Nord motherboard loads the lookup table from the keybed and because of the poor connection some calibration factors load a zero instead of their correct factor. The software might behave in an unexpected way in this scenario. -- I would expect to see random notes exhibit the misbehaviour and I'd also expect some notes to sound unusually loud or quiet.

The capacitor (silver cylinder) shown in your picture seems very large for a simple EEPROM memory. This suggests the IC on the tiny board draws more power than the motherboard can reliably provide during a long continuous read. A poorly-connected cable might worsen the power supply to the chip. In this scenario, as the motherboard reading from the keybed's calibration table progresses, the little board's power supply (from the capacitor) drains away faster than it can be replenished. In such a case, the first part of the lookup table would load correctly while the power is okay (assume the "bottom" calibration factors load first). As the power level lowers, errors become more likely, so it might only affect "top" calibration factors (assuming they load last).

Capacitors typically lose their capacitance at temperature extremes (relative to room temperature). If your gig where you originally saw this problem took place under hot conditions (e.g., stage lights) I expect to see higher likelyhood of errors.

The higher-temperature scenario might explain the intermittent nature of the problem. It applies equally to any of the possible hardware-type explanations suggested so far.

Thanks groover for the photo and thanks Pablo for explaining what it showed.

I hope you've solved your problem so you don't have to send your board away for repair.

-pterm

[pablomastodon]

I wonder if it could be a diode used in the addressing of the keys. I am not sure how this is mapped in the Nords, but I guess that there is some multiplexing used. https://en.wikipedia.org/wiki/Keyboard_matrix_circuit
One dodgy diode could be affecting several keys at the same time, perhaps?

The 8x8 multiplexing mentioned is valuable info for many here. When failures occur of 8 consecutive notes (or multiples of 8) or in intervals of 8 semi-tones, the answer will almost certainly be found in that signal path. However, "dodgy diodes" are extremely rare in my experience. Only encountered two such cases in nearly 6 years. More importantly, in that limited experience, diodes do not fail intermittently. When they fail they are dead, done and gone. Thus, if there is any intermittent aspect to the failure symptoms, diodes can be ruled out as a possible cause.

There is a small possibility that the poor connection you described caused the misbehaviour.

This will not be the case here. A bad connection there will simply result in an interruption of the boot process and the generation of an error message. In this event. pressing SHIFT will cause the boot process to pass over this and complete and the instrument will be playable, sans calibration. Lack of boot interruption and error message generation would be an indication that this was not the problem in this case.

Additionally, the kinds of adjustments being made via this calibration process are very, very small. Keep in mind that no one else on the planet goes to the trouble of doing this to the Fatar keybeds installed in their instruments. These are effectively the same Fatar keybeds as are installed in all Kurzweils, along with many other kbd makers, including some models of Korg and Roland. Natural microvariations in the keybed manufacturing process result in microvariations in velocity response from note to note. Nord, in their pursuit of perfection, developed and began implementing this process during Stage Classic Rev. C production and all Stages made since then are calibrated, each instrument one by one, each individual note. All Electro 3 and newer and all Piano/Piano 2 instruments get the same treatment. But absence of calibration would definitely not result in the kinds of wild velocity swing symptoms as noted by OP. Most mere mortals and middling players (such as myself) do not possess the level of skill/touch needed to discern the microvariations existing in the uncalibrated keybeds in the first place.

Bless,

Pablo