Author Topic: Roland VA-3 boot issue (Read 6251 times)

frisc0 · « **on:** August 17, 2021, 10:11:23 pm »

Hi,

My nephews Roland VA-3 keyboard has the issue at power on, it won't boot; during play suddenly it crashed and refused to reboot again. I want to have a go at repairing it, if possible. It's a keyboard with few buttons but a touch screen lcd to operate it. At power on the lcd lights up white/grey color and 3 buttons with red leds light up. A few seconds later the color of the lcd changes to yellow/orange and one of the leds of the 3 buttons switches of, leaving 2 lit.

I've checked the service manual (https://www.synthxl.com/wp-content/uploads/2020/01/Roland-VA-3-Service-Manual.pdf) for the factory reset procedure, but the keyboard does not seem to respond to (any of) these.

I've checked the PSU, that is combined with the Power Amplifier on 1 PCB. The PSU delivers a steady 5V and 12V, and also powers the LCD inverter board, which works. Also, the analogue circuit still works correctly. There is a RCA input (to play along with an external music source) and this input gets amplified and played correctly through the speakers and headphone jack(s).

I suspect it must be something with the boot procedure failing, possibly a flash rom or something (system program can be written/updated in flash rom from a floppy), but I cannot figure out how to proceed or what to check / do next. I've inspected all pcb's in the keyboard, detached and reattached all connectors, but to no effect. No visible damage to pcb's or components, no caps blown or leaking as far as I can see (or smell).

Also tried to detach components (button pcb, keyboard, floppy drive, D-beam pcb, jack/midi board and booted up to see what happens, but same behaviour at all time: lcd lights up, but it won't boot further, never anything is shown on the lcd. Also checked the lcd contrast pot at the rear of the keyboard, makes no difference varying this.

What would be the best way to proceed this analysis?

All help greatly appreciated! Thanks.

frisc0 · « **Reply #1 on:** August 17, 2021, 10:39:39 pm »

The behaviour the VA-3 shows is very similar (yellow blank lcd screen) to what Juno-D users report after a failed system update:

https://www.keyboardforums.com/threads/roland-juno-d-system-update-1-04-trouble-synth-doesnt-work-after-that.3585/

LateLesley · « **Reply #2 on:** August 17, 2021, 11:39:36 pm »

Well, If you have verified the Power lines (+5V, +12V, -12V), I think the next step would be to verify power at the Microcontroller and RAM/ROM/Flash on the MAIN PCB (page 17,19), these will be the basics for booting i'd suspect. If the power all looks good there, then I'd try getting a test disk and system disk, and see if you can get it into the system update mode, or test mode (page 8,9).

It might also be worth, if you have the means, to check the clock signal for the ICs is there and functioning.

But I suspect like frisc0 says it could be bad software, i've noticed cases of flash going bad seem to be increasing, and maybe if you can reload the system, it might work. You might need to try and find someone with a VA-3 to write you a system disk, as I can't seem to find them online.

frisc0 · « **Reply #3 on:** August 18, 2021, 12:44:38 pm »

Thanks for your reply LateLesley! All voltages look good, so it must be something in the digital area I expect.

I also thought about reloading the system to Flash, but the procedure to do so according to the service manual is to hold 3 buttons (piano + one touch + intro) while powering on. This then should display a menu (from static ROM I expect) from which you can choose to save or load an updated system or the test program. However, this first menu doesn't even show, I have never seen any characters or images on the lcd, just blank.

Now there is a Flash ROM initialization procedure (holding melody intell pressed while powering on) but this doesn't do anything either where it should display a message.

I think the boot software with basic operations (BIOS like) is in static ROM and the system program runs on top of that and is stored in Flash ROM.

Since the boot ROM doesn't even come up, maybe it's not a Flash ROM issue, but something around the components and software involved in the boot process. I don't have a scope to check signals, only a multimeter. Are there any suspect components I could check?

I can also imagine the system (BIOS) does some checks/diagnostics prior to actual startup to see it receives the right signals from components like the lcd ('is it switched on and working?') and maybe for other components as well and maybe from these checks it halts because a component is failing. You think the PSU should give some signal as well like 'I now can produce a stable 5V..' for example?

Btw: I do note the 5V voltage regulator gets very hot (incl the heat dispenser) and also R1 and R11 get very hot (cannot touch with bare finger). Would this be normal?

Audiorepair · « **Reply #4 on:** August 18, 2021, 02:29:53 pm »

Check the reset circuit, XRST on page 22. There is an electrolytic cap there.

It is surprising how often the reset signal is the culprit in non boots.

LateLesley · « **Reply #5 on:** August 18, 2021, 04:43:27 pm »

I no doubt think getting as hot as to not be able to touch is indicative of a fault, sounds like something is drawing too much current. (I take it this is R1/R11 on the power board?). They are series resistors in a power line, paralleled, so I'd expect some heat, but if they are getting toasty fast, and the IC1 regulator is too, i'd say something is faulty. You can narrow that part down relatively easily by disconnecting the FDD, Jack board, and Inverter connectors, and then connect one at a time, and see which one makes IC1 hot. That should identify which one is drawing too much power, and help you know where to concentrate your efforts.

Audiorepair makes a good shout too, worth checking the reset signal for the digital side of things, and if IC1 is getting hot quick, and the inverter is causing it, it could just be from the LCD being permanently blank. It may be it's highest state of draw. I do suspect you have a good grip of the bootup process, I'd go for MCU, ROM, RAM, and LCD being the basics needed. But you do have to be aware of other components pulling down the busses (Control, Data, Address) and stopping comms if something is faulty. You might be able to check this with your multimeter, measuring a digital line, if it's hard high or low, it'll read 0V or 5V, if there's comms, you should see it dance around a bit. So any data line with steady readings could be suspect. Though they would all be pretty steady if the reset signal is enforced. So check that the reset is operating correctly before you dig into the data lines n stuff.

If the reset is ok, I'd suspect chips in this order : Flash, RAM, MCU, ROM. The reason for this, is Flash does wear out, RAM is sensitive and can fail, Mask ROM tends to be quite hardy, so i'd suspect the MCU being faulty next. I take it you have done the finger test to see if any of these get hot?

frisc0 · « **Reply #6 on:** August 18, 2021, 08:07:04 pm »

Wow, good suggestions both, thanks, I'll try later this evening!

Meanwhile, two photo's of the keyboard to get an understanding of how the boards are placed/connected. The photo of the top of the keyboard is to show what happens if connector CN4 is detached; with everything connected at first boot it's intro+fill+variation that light up and after a few seconds (when the lcd changes from white to yellow) intro shuts down and only fill+variation keep being lit. I think this shows there is some logic still working the first seconds, to come to a halt in the process.

Audiorepair · « **Reply #7 on:** August 18, 2021, 08:23:22 pm »

From that position, you could try a warm boot by briefly grounding the CPU reset pin (shorting C71 should do it).
There should be 5v across C71 normally.

Might show something new.

frisc0 · « **Reply #8 on:** August 18, 2021, 10:29:53 pm »

Interesting stuff: I detached multiple power consumers and checked the heat production on the voltage regulator and resistors R1 and R11. As soon as the jack board is connected, heat rises to untouchable levels. Testing temperature of the components on this board I found that IC6 (NJM2360M, a DC/DC control IC: https://datasheetspdf.com/pdf/455002/NewJapanRadio/NJM2360/1) gets very hot!

So I decided to measure the voltage on the pins and found (against analogue ground pin 9 on CN3):
pin1: 5V
pin2: nearly 0V
pin3: -12.3V
pin4: (which according to the datasheet should by GND) measures -14.69V
pin5: -13V
pin6: 5V
pin7: 5V
pin8: 5V

Can we conclude this IC is broken?

It's the only hot component on the board, the big resistor next to it (R39) does not get hot.
Also, the IC5 (79L12 negative voltage regulator, https://datasheetspdf.com/pdf/973358/UnisonicTechnologies/79L12/1) does not get hot. On the datasheet I read it has a thermal shutdown protection...?

I totally overlooked the -12V which is not produced on the PSU board, but on the jack board.

Your advice please

LateLesley · « **Reply #9 on:** August 19, 2021, 01:07:26 am »

Actually, it all seems fine. Because it's generating a negative voltage, The IC ground isn't connected to the system ground, otherwise it would short out and destroy it. It's a type of "floating ground circuit"when the IC ground will actually be connected to the negative output voltage, so that all the voltages look positive to the IC with respect to the IC ground pin. But in reality the IC ground is negative, below the actual system ground.

This article might explain it better. https://www.electronicspecifier.com/products/power/creating-a-negative-output-voltage-using-a-buck-converter

So I would check the TP8 and make sure you are getting -12V, and check C43,C44,C45,C46 and D6, to see if they get hot, which might indicate a fault. It seems to supply various ICs on the jack board (Page 25, look across from the letters K and M on the left, you'll see various op amp symbols withe the +Vcc and -Vcc supply lines) is ICs 2,9,10,14,and18, and their associated caps could all be suspect, and worth checking for heat, and in the case of the caps, with the power off, shorts.

Audiorepair · « **Reply #10 on:** August 19, 2021, 12:46:09 pm »

The fact that the unit passes audio from the CD inputs to the power amp suggest that both 12v rails are present, and I believe your measurements confirm this.

If IC6 is getting really hot, and the 7912 regulator isn't, that would indeed suggest a problem with the circuit around IC6.

Since IC6 does appear to be working, or you wouldn't get -12v from the 7912, it kind of points more towards IC6 working too hard.
My first suspect would be C46, as this going bad would force poor old IC6 to pump out loads of power to get enough voltage to run the 7912.

Audiorepair · « **Reply #11 on:** August 19, 2021, 01:35:21 pm »

Also, I doubt this issue has anything to do with the non booting issue, as removal of this jack board should remove the boot problem, but you say it doesn't.
It's probably just another sign of a piece of equipment growing old, like we all do eventually.

Most of the electrolytics will probably not be the same spec they were born with.

frisc0 · « **Reply #12 on:** August 19, 2021, 10:08:53 pm »

Thanks for explaining LateLesley, learnt something new again

!

Strange thing is, none of the components get hot on the jack board, only IC6... whereas the big resistor R39 next to it, which is a 1W Oxide Resistor of 0.22 Ohms and I think was designed to dissipate the heat, but it doesn't since it does not get hot. I unsoldered it and measured resistance, which was floating between 0.3 and 0.4 Ohms. Is this an acceptable variance?

frisc0 · « **Reply #13 on:** August 19, 2021, 10:12:35 pm »

Quote from: Audiorepair on August 19, 2021, 12:46:09 pm

The fact that the unit passes audio from the CD inputs to the power amp suggest that both 12v rails are present, and I believe your measurements confirm this.

If IC6 is getting really hot, and the 7912 regulator isn't, that would indeed suggest a problem with the circuit around IC6.

Since IC6 does appear to be working, or you wouldn't get -12v from the 7912, it kind of points more towards IC6 working too hard.
My first suspect would be C46, as this going bad would force poor old IC6 to pump out loads of power to get enough voltage to run the 7912.

By chance I had a replacement for C46 (1000u-25V) lying around, so I changed it, but with no effect whatsoever unfortunately. Thanks for the advice though, it sounds logical...

TP8 still measures -12V, so it does still deliver the right voltage(s).

frisc0 · « **Reply #14 on:** August 19, 2021, 10:14:18 pm »

Quote from: Audiorepair on August 19, 2021, 01:35:21 pm

Also, I doubt this issue has anything to do with the non booting issue, as removal of this jack board should remove the boot problem, but you say it doesn't.
It's probably just another sign of a piece of equipment growing old, like we all do eventually.

Most of the electrolytics will probably not be the same spec they were born with.

Thing is, the jack board cannot be removed for a boot-test, since the jackboard passes on the required power for the mainboard. The mainboard itself does not have power lines of its own.

LateLesley · « **Reply #15 on:** August 20, 2021, 03:18:25 am »

Yeah, with an ordinary meter 0.3 $\$\Omega\$$ to 0.4 $\$\Omega\$$ Will be close enough. It's hard for meters to read low ohm resistors, you usually have specialised meters for that. It's at the level of measuring the resistance of the meter cables themselves. So the reading you got seems reasonable.

It'll also be the reason it doesn't get really hot. As it's a very low resistance, there won't be much voltage drop across it, so it won't actually dissipate that much power. It would need a lot of current to start it heating.

As that circuit is producing the right voltage, and it doesn't seem to be disappearing with the power on, I think you can rule out that particular circuit from having a fault. Sometimes some components do run hot.

I think you're now into checking the Digital side of things, the power, reset, and data of the digital stuff.

frisc0 · « **Reply #16 on:** August 20, 2021, 09:54:17 am »

It does deliver the right voltages, but I still find it strange this IC6 is heating up that much without a proper heatsink, one would think it's not designed to get this hot. Also, why use such a big resistor designed to handle high temps/amps but it does not heat up at all...

But I guess you are right that the failure to boot is not because of this issue (might be something that will be causing trouble later... :| ).

Looking at the reset lines: Audiorepair, I guess you mean shorting C76 on the mainboard, on the main CPU, right? The C71 you mention is on the synth board (XPGS4) and has it's own CPU. I think the CPU on the mainboard will be responsible for the boot program, but I'll try both

.

Audiorepair · « **Reply #17 on:** August 20, 2021, 10:31:12 am »

Quote from: frisc0 on August 20, 2021, 09:54:17 am

Looking at the reset lines: Audiorepair, I guess you mean shorting C76 on the mainboard, on the main CPU, right? The C71 you mention is on the synth board (XPGS4) and has it's own CPU. I think the CPU on the mainboard will be responsible for the boot program, but I'll try both .

Oops, yes, my mistake.
There are other resets scattered around too, one on the floppy controller pg 20 for instance.

It's just worth a go measuring then forcing resets, just in case something happens, maybe characters on the LCD that at least tell you the LCD is working.

And short the cap rather than the tiny reset pin, so you don't risk shorting adjacent pins, but I'm sure you've got that already

LateLesley · « **Reply #18 on:** August 20, 2021, 04:04:43 pm »

Quote from: frisc0 on August 20, 2021, 09:54:17 am

It does deliver the right voltages, but I still find it strange this IC6 is heating up that much without a proper heatsink, one would think it's not designed to get this hot. Also, why use such a big resistor designed to handle high temps/amps but it does not heat up at all...

That may actually be the reason, you pick big resistors so they don't heat up. especially with audio circuits. heat can A) change the resistance, and B) they'll probably be noisier.

https://eepower.com/resistor-guide/resistor-fundamentals/resistor-noise/

Audiorepair · « **Reply #19 on:** August 20, 2021, 07:55:16 pm »

Quote from: frisc0 on August 20, 2021, 09:54:17 am

It does deliver the right voltages, but I still find it strange this IC6 is heating up that much without a proper heatsink, one would think it's not designed to get this hot. Also, why use such a big resistor designed to handle high temps/amps but it does not heat up at all...

I really don't think also that IC6 should be getting hot, it is only powering a few op-amps.

IC6 is run from the 5v supply.
If it is drawing excess current, this may be putting ripple on the 5v supply, which could cause all sorts of mayhem with the computer stuff.

If there is a way to easily disable the 5v supply to the IC6 circuit to rule it out as a culprit, then I would try that, you don't need the audio to test the boot.

Audiorepair · « **Reply #20 on:** August 20, 2021, 08:04:01 pm »

But it could also be the case that the 5v supply is crippled, in that the extra current IC6 needs is the straw on the camels back that makes the whole system fall over.

frisc0 · « **Reply #21 on:** August 22, 2021, 10:39:27 pm »

Like the analysis...

C49 can easily be removed from the board, as well as R39. That should do the trick to shut down IC6, right? And in the process I could replace C49 right away.

Added some photo's of the 'suspect area' that might help. Still find it strange they fitted this big a resistor for R39 even with the legs protected without it ever getting hot...

frisc0 · « **Reply #22 on:** August 22, 2021, 10:44:17 pm »

Also added a photo of C88 on the back of the mainboard which looks melted; it seems to be a bridge between analogue and digital ground; not sure this could harm? Could replace it with a normal ceramic cap...

Also added a photo of the PSU for reference.

frisc0 · « **Reply #23 on:** August 23, 2021, 08:30:33 am »

Quote from: LateLesley on August 20, 2021, 04:04:43 pm

Quote from: frisc0 on August 20, 2021, 09:54:17 am
It does deliver the right voltages, but I still find it strange this IC6 is heating up that much without a proper heatsink, one would think it's not designed to get this hot. Also, why use such a big resistor designed to handle high temps/amps but it does not heat up at all...

That may actually be the reason, you pick big resistors so they don't heat up. especially with audio circuits. heat can A) change the resistance, and B) they'll probably be noisier.

https://eepower.com/resistor-guide/resistor-fundamentals/resistor-noise/

Makes sense in an audio circuit, again learned something new, thanks

frisc0 · « **Reply #24 on:** August 23, 2021, 10:18:30 pm »

Okay, quick update: removed C49 and R39 which leaves IC6 without power, but no boot...

Also replaced C88 (low hopes) and replaced C49 with a fresh cap with higher specs (100u 63V) and resoldered R39 again, still no boot...

Then I measured voltage over C70 (around -9.7V) and C69 (expected +12V but found 100mV varying). So I replaced C69 for a fresh cap again 63V, but to no effect. I expect this missing +12V is the issue here, but where in the chain could this go wrong? Or is the low voltage measured on this C69 cap correct?


EEVblog Main Site	EEVblog on Youtube	EEVblog on Twitter	EEVblog on Facebook	EEVblog on Odysee

Author Topic: Roland VA-3 boot issue (Read 6251 times)

Share me