Ensoniq SD-1 stuck at test mode (no boot)

Pulling out your hair? Don't know what to do or where to go? Ask in here.
Forum rules
READ: VSE Board-Wide Rules and Guidelines

If your Help request has been solved, please edit your first post in order to select the Image Topic Icon to let others know your topic has been solved.
Post Reply
belzrebuth
Newbie
Newbie
Posts: 30
Joined: Tue Jul 15, 2014 12:43 pm

Ensoniq SD-1 stuck at test mode (no boot)

Post by belzrebuth » Sat Apr 11, 2020 5:01 pm

Hi,
I've got an Ensoniq SD-1 which is in a very bad cosmetic state and (of course) stuck in the test mode.

All keys are in a box cause the previous owner tried to clean them and never bothered to put them back so I have only the keyboard board, mainboard, psu and panel on my bench in order to try to repair it.

The insides exhibit no corrosion, no signs of excessive moisture and no previous attempts of repair.

The exterior has probably seen a lot of moving around hence the bad cosmetic condition but since I never had one I thought it would be fun to make it work again.

Pressing most buttons does nothing but some of them just display random characters on the screen.
From what I've seen on the service manual the test mode characters are mostly okay.
Some buttons may not behave correctly but it's probably because the panel is not screwed correctly.

In any case I believe that regardless of being stuck in test mode the panel board is probably okay .
All I want is to be able to boot normally and then slowly address other issues.
There a lot of work to be done anyways given its condition.

Okay so here's what happening:

Checked all voltages which seem okay.

I replaced the EPROMS just in case (it had an older version so I would do that anyway).
The low EPROM was replaced by a 27c1001 instead of a 27c010 but I believe those are compatible.

A new UART IC is on its way since pin35 on mine is stuck high so per Rasputin's instructions here:
viewtopic.php?t=96381
that is worth checking. (this is the chip select pin (active low) and should be low most of the time)

Made a new 20pin ribbon for the keyboard (the old ribbon cable was bad anyway).

Checked for CLK and reset signals on the CPU and these seem fine as well.
VRES is also okay (sits at 5V).

I'm open to any other suggestions on what I should check while I'm waiting for the UART chip.

Thanks:)

belzrebuth
Newbie
Newbie
Posts: 30
Joined: Tue Jul 15, 2014 12:43 pm

Re: Ensoniq SD-1 stuck at test mode (no boot)

Post by belzrebuth » Sat Apr 11, 2020 9:25 pm

After some more messing with it, resoldering some bad joints etc it booted up.
Passes keyboard calibration and can trigger it via midi, sounds very nice too.
BUT,
It doesn't trigger via the keyboard and no button on the panel actually registers anything.
Can't even perform a hard reset :shock:
Any ideas are more than welcome since I feel I'm too close to make this working!
Maybe that UART chip is partially defective?
Is it possible to have the unit booting up and accepting midi data (and playing too) with a faulty DUART chip?
I find it very weird that both the keyboard and the panel are not working..

User avatar
madtheory
Supporting Member!
Supporting Member!
Posts: 5330
Joined: Mon May 01, 2006 12:45 pm
Real name: Tomas Mulcahy
Gear: Wurlitzer Opus 1536, Model F, Morovdis Arpeggiator, Maplin My First EQ, Jeff Wayne Thunderchild rack, Thermostat, Buck Owens' Moog.
Location: Cork, Ireland
Contact:

Re: Ensoniq SD-1 stuck at test mode (no boot)

Post by madtheory » Sun Apr 12, 2020 3:07 pm

Great that you're close. Subscribe to Rainer Buchty's mailing list, it's very nice and a lot fo folk there with experience fixing Ensoniqs.

Rasputin
Senior Member
Senior Member
Posts: 803
Joined: Thu Jul 21, 2016 4:52 am

Re: Ensoniq SD-1 stuck at test mode (no boot)

Post by Rasputin » Sun Apr 12, 2020 7:35 pm

belzrebuth wrote:
Sat Apr 11, 2020 9:25 pm
I find it very weird that both the keyboard and the panel are not working..
If it was the DUART then it would not pass calibration. I suspect what you have done is bypass the keyboard processor, and now the serial comms from the display are going directly to the DUART. This automatically takes it out of calibration mode, but the DUART will not receive pressed panel buttons correctly because the keyboard processor "shifts" the panel commands around some before forwarding them to the DUART.

Since the keyboard processor is out of the loop, the music keyboard will not register keys either. The DUART handles MIDI separately from they keyboard/panel buttons.

You've effectively got it stuck in "rack" mode. It is possible to simulate both the keyboard (obviously) and the panel buttons over MIDI though.

I will have to think about this more and examine my notes to give you better advice at this point.

belzrebuth
Newbie
Newbie
Posts: 30
Joined: Tue Jul 15, 2014 12:43 pm

Re: Ensoniq SD-1 stuck at test mode (no boot)

Post by belzrebuth » Sun Apr 12, 2020 8:12 pm

It is indeed possible to change presets and get notes to play from midi.
I can also see midi out data when I'm messing with the mod or the pitch wheel.
So I think the DUART chip is indeed okay.
I can't request SYSEX data or a bulk dump from the machine but sysex messages may be disabled.
Of course, no midi out messages when I press any of the keyboard keys.

The bad thing is that pins 42 and 43 of the 68hc11a1p (keyboard board MCU) stay high regardless of any key pressing.
Clock on the KPC board is okay. (8MHz)
Pin 2 of the 74HC000 is 2MHz not 8MHz though.
This may be normal, idk.

Pin 39 (reset) of the MCU is high so good as well.

5V on the keyboard board is okay. (from the 7805)
EPROM (KPCv2 v2.33) has been replaced just in case.

7556 IC which was a bit corroded was swapped with a NE556 just to test if anything changes, nothing did so I put the original back.

Clock reaches the MCU and sits at 8MHz and the 68ch11 own clock(?) output (pin 39 I think) sits at around 75 or 80KHz.
I'm also seeing bits around the 74hc373 and various bit activity on the large pin header which connect the KPC board to the keyboard pcb.
I haven't checked the 74hc138 decoders yet but I doubt the actual keyboard or its minimal circuitry is at fault here.

So all the basic checks have been performed.
I haven't yet seen one single key change anything though.
It's like the whole keyboard is dead :(
Haven't checked the other motorola keyboard scanning IC yet but I don't know how it should behave anyways.

I think we now need a more systematic approach.

How one ends up in "rack" mode?!
Since everything else seem to be working and serial data (from the panel) is passing thru the keyboard board that means that the fault is probably at the KPC board?

Rasputin
Senior Member
Senior Member
Posts: 803
Joined: Thu Jul 21, 2016 4:52 am

Re: Ensoniq SD-1 stuck at test mode (no boot)

Post by Rasputin » Mon Apr 13, 2020 12:54 am

belzrebuth wrote:
Sun Apr 12, 2020 8:12 pm
How one ends up in "rack" mode?! Since everything else seem to be working and serial data (from the panel) is passing thru the keyboard board that means that the fault is probably at the KPC board?
On the SD-1, this normally would be done deliberately by bridging J11 (keyboard connector) pins 17 and 19. I'm wondering if you've somehow connected those two pins together when touching up soldering on that header, or when repairing the keyboard ribbon... something like that.

The serial traffic from the display panel usually goes into the keyboard processor, and then it is merged with the keyboard specific data, and then sent out again by the keyboard processor to the DUART.

The keyboard processor can be bypassed by sending the serial traffic directly from display to DUART, but this also disables the proper function of the display buttons, so it's only good for temporarily testing that the mainboard is working and/or MIDI and audio.

If the display and its buttons work under self-test, and if the mainboard is confirmed working when bypassing the keyboard processor then the keyboard processor needs to be put back into the communications loop.

In other words: If you have an SD-1 that boots, but has non-responsive panel buttons and keyboard then you're actually moving in the wrong direction!

You need to figure out why the SD-1 is automatically skipping calibration, and then prevent it from doing that. This seems like a step-backward because the SD-1 will no longer boot, and will fail into self-test again, but you'll have to take one step back to take two steps forward at this point.

So, yes, the problem lies somewhere in the keyboard section, but it seems that somehow another problem has been introduced which has cut the keyboard processor out of the loop. You'll have to put the keyboard processor back in the loop, and then fix any remaining problem(s) with the keyboard.

blueknob
Active Member
Active Member
Posts: 286
Joined: Sun Jul 06, 2008 12:30 pm
Gear: ESi32/E6400U, Novation Supernova, Ensoniq, vintage Macs and the like.
Location: England, UK

Re: Ensoniq SD-1 stuck at test mode (no boot)

Post by blueknob » Mon Apr 13, 2020 7:17 am

Is there not some magic button presses that call up an internal diagnostic mode like there is with the EPS16+?

belzrebuth
Newbie
Newbie
Posts: 30
Joined: Tue Jul 15, 2014 12:43 pm

Re: Ensoniq SD-1 stuck at test mode (no boot)

Post by belzrebuth » Mon Apr 13, 2020 8:45 am

Rasputin wrote:
Mon Apr 13, 2020 12:54 am

On the SD-1, this normally would be done deliberately by bridging J11 (keyboard connector) pins 17 and 19. I'm wondering if you've somehow connected those two pins together when touching up soldering on that header, or when repairing the keyboard ribbon... something like that.
But if pins 17 and 19 were somehow shorted I would get a black screen (test mode).
I've shorted them and this gets me straight in test mode.

On my case when the keyboard is attached the synth boots normally, passes kbd calibration and boots straight to display with the 6 preset sounds.
I can also see the presets change (underscore highlight) when I send a midi program change.
So the display data in and out is working.

It's just that the no panel key is registering anything.

As I've said before I can also see bleeps when I press a key on the 4pin display connector pin1.
So I'm definitely not in "rack mode" per se.

It's like the synth "thinks" it's okay and boots normally so it must get some kind of response from the keyboard on boot-up but there's obviously a serious breach/bad data in its communication loop.
The weird thing is that when on "calibrating keyboard, do not touch" if I press a key on the keyboard I get an error.
So it must sense the keyboard on power up and then after booting up it somehow loses it.
I would even suspect a "local off" thing but that scenario would have the panel working normally.

Rasputin wrote:
Mon Apr 13, 2020 12:54 am

If the display and its buttons work under self-test, and if the mainboard is confirmed working when bypassing the keyboard processor then the keyboard processor needs to be put back into the communications loop.

In other words: If you have an SD-1 that boots, but has non-responsive panel buttons and keyboard then you're actually moving in the wrong direction!
I'm not sure I bypass anything, I probably have a working motherboard as you say but I think if I bypassed the keyboard the synth wouldn't boot , display its welcome message etc.
I get all kinds of messages only when the keyboard is connected, otherwise the screen is black.
Is this wrong?
I mean without the keyboard attached I should get anything else than a black screen?
When I short pins 17 and 19 I get a total black screen and I'm in test mode and when pins 17 and 19 are shorted a I get an underscore on the leftmost top character "slot" and it's in test mode again.
Rasputin wrote:
Mon Apr 13, 2020 12:54 am

You need to figure out why the SD-1 is automatically skipping calibration, and then prevent it from doing that. This seems like a step-backward because the SD-1 will no longer boot, and will fail into self-test again, but you'll have to take one step back to take two steps forward at this point.
I don't follow.
It's not skipping calibration.
It seems to perfoming kbd calibration and then proceeds to boot seemingly normal it's just that I have no panel buttons and no keyboard.
Rasputin wrote:
Mon Apr 13, 2020 12:54 am

So, yes, the problem lies somewhere in the keyboard section, but it seems that somehow another problem has been introduced which has cut the keyboard processor out of the loop. You'll have to put the keyboard processor back in the loop, and then fix any remaining problem(s) with the keyboard.
Pressing a key while in calibration mode prevets the synth from calibrating the keyboard so I'm not sure if it's completely out of the loop or just malfunctioning somehow.
Passing the calibration must at least require a "ping" response from the keyboard I think.
So I don't know what it's actually wrong at this point.
I would bet that keyboard calibration is a very basic test and a succesfull caliration does not tell anything about the actual keyboard functionality from a user's point of view but it should somehow request a "green light" from the kpc processor in order to move on to boot sequence.

I'm really clueless on how I should proceed at this point.

I mean how can I verify that the keyboard is out of the loop and how should I restore that connection?

Would there be a way to see the merged data that returns to the mainboard?

Meaning can I reroute the panel serial traffic as-is back to the mainboard and confirm that the panel works or there's some data processing taking place in the keyboard processor?

In simple terms, is it possible to take a hypothetical serial out of the display/panel board and place it in a hypothetical serial in to the mainboard and check if actively bypassing the keyboard processor will restore panel functionality?

If we exclude the panel that only leaves us with a keyboard that's at fault and that narrows it down to 3-4 easily sourced chips and two very difficult to find :?

belzrebuth
Newbie
Newbie
Posts: 30
Joined: Tue Jul 15, 2014 12:43 pm

Re: Ensoniq SD-1 stuck at test mode (no boot)

Post by belzrebuth » Mon Apr 13, 2020 10:46 am

Okay some more updates:

I resoldered almost every joint in the KPC board.
I then connected it to the mainboard without the keyboard assembly attached.
The synth reported a keyboard calibration error as expected BUT:
I could now press "IGNORE" and guess what...
It actually ignored the error and booted straight up.
Then I could navigate to the menus and everything!

This sadly only lasted until I reboot it.
On the next power cycle I couldn't boot again.
The synth reports a system error 144 (this is MIDI related AFAIK) or a system error 33 (this is the FX chip).
The FX chip is okay since I've heard a very nice reverb on my previous testing.

So something is preventing the synth from booting normally even without the keyboard attached.
This is probably an intermittent fault which is to be expected with most ensoniq keyboard synths in my experience (never had a rack randomly malfunctioning)

Anyway,
I can now say that there is something really bad happening when I plug in the keyboard assembly to the KPC board.
I should probably remove all the keys and start messing with that board.

I believe there is not much there at all, apart from a few decoders, some transistors and passives so it probably a shotgun approach would probably work in this case.

Worst case scenario is that the keyboard is fine and I have an intermittent fault somewhere else which manifests 99.9% of the time except for once.

Back to work I guess..

belzrebuth
Newbie
Newbie
Posts: 30
Joined: Tue Jul 15, 2014 12:43 pm

Re: Ensoniq SD-1 stuck at test mode (no boot)

Post by belzrebuth » Mon Apr 13, 2020 2:11 pm

Dismantled the keyboard and desoldered all chips from the board.
There were a couple of 74hc04 inverters and I think 2 or 3 74hc138 decoders.
All checked okay in my TL866 eprom burner which has a logic chip tester mode.
I resoldered those back and re-connected the keyboard back to the KPC board.
Still the same problem.

Now at least it's consistent.
The current status is that if the unit boots without the keyboard attached the panel works.
If it boots with the keyboard panel does not respond.
The previous system errors (144, 33) were random in nature so I will not bother with those until the repair is complete.

The question now is what is happening to the KPC board when I plug it to keyboard PCB?
Apparently the keyboard PCB is okay since there is not much happening there anyway.

I fear that it may have to do with that rare motorola chip that perfoms the key scanning but It may be a simpler issue.
Really hope it's something easy to source..
KPC board 74hc00, 74ch373 and SN7406 are good BTW because I've tested them.
EPROM is also new so that's out of the equation too.
So that pretty much only leaves the MCU and the KPC chip..except if without them the KPC board wouldn't operate properly even in my keyboard-less configuration.

edit: Another very strange thing happening is that every time I try to format a floppy disk after the format is complete the instruments plays a drone on itself..!
Happened every time I've tried to format a disk.
This sound doesn't remind me any of the presets and stops only if I try to format the disk again.
Returning to any other menu does not stop the drone from playing..
Last edited by belzrebuth on Mon Apr 13, 2020 8:58 pm, edited 1 time in total.

belzrebuth
Newbie
Newbie
Posts: 30
Joined: Tue Jul 15, 2014 12:43 pm

Re: Ensoniq SD-1 stuck at test mode (no boot)

Post by belzrebuth » Mon Apr 13, 2020 6:02 pm

Another update:

The synth boots and behaves normally (floppy/drone problem still exists though) even without the KPC chip installed.. :shock:
So that chip is the only (?) probable candidate for my problem..
I suspect it won't be cheap/easy to replace this.
Unfortunately I don't have any other ensoniq keyboards at the moment to test if a working chip would resolve this issue.
I'd be much happier to put a working chip in there only to find out that the problem remains of course!

In any case I can't troubleshoot this further without a replacement IC.

Any suggestions would be much appreciated (Rasputin?) since in my experience custom ICs almost never fail.
It almost always is something else so I really hope it's the same case now..

I'm awaiting for any further info/help..

belzrebuth
Newbie
Newbie
Posts: 30
Joined: Tue Jul 15, 2014 12:43 pm

Re: Ensoniq SD-1 stuck at test mode (no boot)

Post by belzrebuth » Tue Apr 14, 2020 2:24 pm

I managed to get a hold of a working S38BC010PS01 IC from a friend's EPS keyboard.
Unfortunately my friend lives far from me so transport is not very easy/quick.
Nothing changes with the known working KPC chip so I'm not sure what is going on here.
It's like the synth does not like the keyboard at all and bypasses the calibration test when it is attached to the KPC board.
I don't know where to look next.
I may try to get a known working 68hc11a1p mpu as well (the EPS has one since it has a very similar board).
I wonder if there's something I'm missing here.
I still don't understand in which condition the synth/processor/whatever decides that the keyboard should be taken out of the serial communication loop.

belzrebuth
Newbie
Newbie
Posts: 30
Joined: Tue Jul 15, 2014 12:43 pm

Re: Ensoniq SD-1 stuck at test mode (no boot)

Post by belzrebuth » Wed Apr 15, 2020 2:46 pm

Finally fixed :D

It was very tricky to hunt down why the calibration process was bypassed each time the keybed was attached to the KPC board.
Since I've tried both another MPU and a KPC chip there was only the keybed PCB left to blame.

After comparing with the SQ-80 schematics my problem came down to broken traces on the keyboard PCB which prevented the synth to complete the calibration procedure.

I don't exactly know what this procedure does (I guess a "ping request" to all key contacts) so even the tiniest cut on any of the traces can result in a dead keyboard.
This is not difficult to happen since oxidation can render traces very fragile.
Since this is a keyboard it's not uncommon for drinks or simply moisture to creep in and destroy those delicate traces.

Repairing those traces is tricky because they are very narrow (about 1mm) and you have to follow the pattern on the PCB to retain their inductive and resistive properties.

I used a single strand of a stranded 28awg cable which fits exactly the width of the trace.
I've carefully scraped the "healthy" part of the trace and apply generous amount of flux and solder to bridge both healthy parts of the damaged trace keeping the strand as straight as possible in order to bridge the gaps in the same exact way the trace indicated.
You may have multiple cuts (I know I had ) so rechecking is key here.
It's a tedius procedure but sure enough when I was done the keyboard passed calibration and the synth is 100% functional again.

Post Reply