saifbadri wrote on 2020-02-01, 21:59:

Thanks got another freind who recommended this on discord. I will test that today. Caps look fine so I hope its just the regulator. I tried applying 5v to the logic inputtaping into ground and 5+ but got the same results. Maybe its a bad eprom...

This won't work if the regulator isn't failed open.

Im trying to get a another rom chip. I see a chip with amd 1986 on it. And a bunch of numbers highlighted with red ink. Any idea how we can program the above binary attached to a new chip? Thanks in advance.

I tested today with a multimeter 5v exactly out from the power supply to the logic board. EPROM chip was removed and I notice a tiny stain of oil underneath it and touching one leg, so weird maybe it just died...

saifbadri wrote on 2020-01-31, 21:09:

Hi, my wife used a 12v DC instead of 9v DC to power our SB-55 and now it has black screen.

Reverse voltage or not?

It worked for one day, so polarity was apparently correct.

gdjacobs wrote on 2020-02-01, 22:46:

This won't work if the regulator isn't failed open.

Exactly. Also, it's quite courageous to just feed 5V in after the voltage regulator without knowing what fault there is. You may even cause more damage this way.

But anyway, a failed ROM is possible. You can erase the EPROM using UV light and reprogram it using a cheap device like the TL866. EPROMs rarely fail electrically, they just develop flipped bits and can be used again after erasing. If you have exactly the same ROM version, you don't even need to erase it.

yawetaG wrote on 2020-02-02, 07:59:

saifbadri wrote on 2020-01-31, 21:09:

Hi, my wife used a 12v DC instead of 9v DC to power our SB-55 and now it has black screen.

Reverse voltage or not?

No both adapters have Roland's correct polarity. only voltage was different.

derSammler wrote on 2020-02-03, 10:37:

It worked for one day, so polarity was apparently correct. […]
Show full quote

It worked for one day, so polarity was apparently correct.

gdjacobs wrote on 2020-02-01, 22:46:

This won't work if the regulator isn't failed open.

Exactly. Also, it's quite courageous to just feed 5V in after the voltage regulator without knowing what fault there is. You may even cause more damage this way.

But anyway, a failed ROM is possible. You can erase the EPROM using UV light and reprogram it using a cheap device like the TL866. EPROMs rarely fail electrically, they just develop flipped bits and can be used again after erasing. If you have exactly the same ROM version, you don't even need to erase it.

Thank you, I found the same rom on ebay so I got one just in case the old one is damaged. I also ordered the reader will see how that goes!

I read the rom in my SB55 and compared the file to the attached rom in the begining of this thread. I see a difference 0A string is 15 in the attached bin while its 14 in my bin I read from the eprom. This is consistant with the advice I got which is eprom develop reverse polarity! 1 -> 0 since 15 became 14...
The data also seems shifted in my file comparing the hex values. Will try burning the attached bin. fingers crossed will fix it! Power supply output seems working on all 5v pins...

DaCiRO wrote on 2018-11-08, 15:34:

....I get this on-screen error message: '5ur' or maybe instead of a 5 is a capital S 'Sur'...the manuals do not mention anything about this error message which is rather strange since they mention others, ....this has me puzzled.

Just for the record: I had the same issue on a newly acquired SB-55 and it happens when you try to use a floppy formatted to 1.44 instead of 720k. You have to cover the hole on the lower right hand side of the floppy and format as 720k, that's all .

stalk3r wrote on 2020-06-10, 07:38:

DaCiRO wrote on 2018-11-08, 15:34:

....I get this on-screen error message: '5ur' or maybe instead of a 5 is a capital S 'Sur'...the manuals do not mention anything about this error message which is rather strange since they mention others, ....this has me puzzled.

Just for the record: I had the same issue on a newly acquired SB-55 and it happens when you try to use a floppy formatted to 1.44 instead of 720k. You have to cover the hole on the lower right hand side of the floppy and format as 720k, that's all .

I had the same issue of the SB-55 saying "Sur" on the display, turns out my floppy disk head was jammed. Opened it up, gave the floppy head a nudge and it started working beautifully.

And I would like to reiterate the above, along with, make sure that you transfer the files using a floppy drive designed for 2DD (double density) disks. Some of the newer 1.44 drives do not natively support 720k DD disks, and if you take a 720k disk and try to format it using "format a:" command, the drive will error out saying Track 0 (zero) is bad, disk unusable. I found that out after spending a whole day trying to get my SB-55 to read a disk ... 😀

I hope this helps someone who's in the same situation as me. Made an account on Vogons just to share this experience 😀 ... by the way, love Vogons, I've been reading stuff on it since the past year an a half, lovely community !

Jonas-fr wrote on 2016-07-01, 13:04:

Ok I think the issue resides into bad capacitors. I replaced C7 capacitor (it's an ugly soldering job since I don't have tools for SMT) and I gained back some functionnality! The midi file load and plays for a few seconds but stop after that and I got a "dEr" message meaning that my floppy is bad. Since I also tried with a floppy emulator (which doesn't go bad since floppy image is a file and it's load ok with my 486 PC) I think this is because some capacitors are bad too. I'll have to recap the whole thing. I'll try to find someone to do this for me.

In the meantime I dumped my ROM for conservation purpose only and uploaded it here since it's nowhere to be found and can help people to fix broken SB-55! 😉

I gave this a go on my recently repaired unit (with many thanks to the wonderful Alex who is responsible for Bits und Bolts!), as I was curious as to the version - it's a v1.00 ROM. You can check the firmware version by putting the device into standby mode and pressing PROG+SET+REW+FF simultaneously. The display will show the version number.

V1.03 is the latest I've found online, and is what my unit came with. I've attached it here for reference.

Careful playing around with the v1.00 ROM - I tested it in my unit, and it of course works, but in doing so it has upset something on my unit so it now reads Er2 when booting up - it works by powering off and on again, but it’s corrupted something in the NVRAM I guess which sucks. Haven’t found any way to reset it.

Edit: there’s a BR28C16A chip onboard which is 2K of EEPROM. I’ve tried desoldering and wiping in my chip programmer to no avail - it does still retain settings though. Not sure what check it does which flags up the Er2 message, as it doesn’t appear to impair functionality at all.

If someone with one of these can dump their EEPROM it’d be much appreciated.

Right, so I've taken another look at this.

Er.1 occurs when the HM-65256BLSP-12 (the pseudo SRAM) is damaged - this is a showstopper; without this, the device will not work at all.
Er.2 is (I'm assuming) related to the BR28C16A-150 (a 2K NVRAM) ROM contents being bad. From what I can tell, this chip is compatible with AT28C16. This is not a fatal fault - you can hit standby to 'acknowledge' it, and the device seemingly continues to work as expected.

I've replaced both to no avail. I've played around with the 'test' mode as much as possible, but it doesn't actually seem to do anything useful. Its operation doesn't match the service notes, for either version 1.00 or version 1.03 of the ROM. It seems to be broken, or perhaps requires a special disk. It acts differently with different disks inserted (e.g. blank, with music etc.) but again the actual tests do not match with the service notes.

So I've come to the conclusion that there's something in the NVRAM which is missing, and causing the error. I'm assuming there aren't any other chips in the device that have some hidden storage capability.

Frustratingly, I can't figure out what the fault actually is - once you acknowledge the fault, the device performs perfectly, with no loss of functionality.

I'd like to offer a bounty of $20 via PayPal if someone can dump the NVRAM from their working unit and post it here. I was able to dump mine (after it having become seemingly damaged) with a TL866, so a standard programmer is all that is required. Mods, if this is not allowed, please let me know.

I've resolved it!

I noticed that Xgpro blanks the chips with FF's but noticed that one of the other chips I was testing with had 00s. I zeroed another chip with all zeroes, and it now doesn't error at startup. I also confirmed that the original chip was damaged, as programming it with zeros and verifying showed that it read back as 00 FF 00 FF 00 FF etc.

So the TL;DR of this for anyone who has a unit with an Er.2 problem - replace IC8 with another 28C16 with has been blanked with zeroes instead of FFs.

I had the chance to buy an almost brand-new SB-55. At first, everything worked normally, but after I downloaded the 1.06 ROM to upgrade, the machine started behaving just like the original poster described: very hard to power on, and sometimes only displaying ‘---’.

After analysis, it turned out to be the same problem:

C7 (47 µF): electrolytic capacitor failure. Its function is to provide a delay for the RESET/INIT signal.

D2 (RD3.0L): Zener diode/voltage regulator diode, used to clamp and protect the reset level.

R11 (4.7k): pull-up resistor.

Q3 (DTC124EK): resistor-equipped NPN transistor. Since C7 needs to charge, the reset signal is held low for a certain period of time, serving as reset control/buffering.

When the power supply voltage stabilizes and C7 finishes charging, the reset signal is released, and only then does the CPU start running. Therefore, after replacing the C7 capacitor (measured at 284 nF, Rs 4.4k, D 7.99), everything went back to normal.

Another observation to share: the 1.06 ROM provided by the original LED shows “100” on the display when tested in the actual machine, but the sticker on my original ROM chip is marked “103,” and the LED display also shows “103” when checked on the actual unit. So I believe the highest version currently available online might still be 1.03.

Roland SB-55 Main Board Repair Log (No Display / No Standby LED), fix again.
Symptom
No display after power-on.
Standby LED remained off.
Main board powered by a single +5V supply, which was confirmed to be normal.
Current consumption approximately 106mA.
Troubleshooting
Power Supply
+5V rail was normal.
CPU, EPROM, SRAM and other ICs received proper supply voltage.
Reset Circuit
M51953AL reset supervisor operated correctly.
CPU RESET was released normally (~4.8V).
Lowering VCC to about 4.28V asserted RESET and reduced current from ~106mA to ~44mA, confirming correct reset operation.

CPU Activity

16MHz crystal oscillator was running normally.
EPROM /OE, /CE, address bus and data bus were all active.
SRAM showed normal write activity (/WE pulses).
CPU Memory Strobe (MSTB) and I/O Strobe (IOSTB) remained active.

Conclusion: The CPU was executing code normally. The fault was not caused by the CPU, clock, ROM, or reset circuit.

Peripheral Investigation
HD63265FP (IC18) Pin 3 (RESET) was found to remain permanently LOW.
Tracing the reset signal revealed that RESET never reached IC18.
Root Cause

The PCB trace carrying the RESET signal had been corroded and broken beneath electrolytic capacitor C7, which had leaked in the past.

As a result:

HD63265FP remained permanently in reset.
The CPU continued executing code normally.
However, the peripheral controller never initialized.
The unit showed no display and no standby LED.

After repairing the broken RESET trace:

RESET reached IC18 correctly.
The nearby IC13 logic resumed normal operation.
The Roland SB-55 booted and operated normally.
Conclusion

This failure was not caused by the CPU, ROM, SRAM, or reset supervisor, but by a corroded PCB trace damaged by electrolyte leakage from capacitor C7.

The key diagnostic clue was that the CPU, EPROM, SRAM, and bus activity were all normal, leading to investigation of the peripheral RESET network.

Takeaway: On vintage Roland equipment, replacing leaking electrolytic capacitors alone is often insufficient. PCB traces and vias underneath and around leaked capacitors should always be inspected for hidden corrosion and open circuits.

For anyone else repairing these, you may find that the two smaller caps on the power board are very susceptible to the pads getting trashed by the SMD cap leakage. These caps are in line with and used purely for smoothing power to the floppy drive. I placed a 16v 100uF cap between the GND and 5V pins on the floppy connector with some suitably long leads, then hot glued it to the support metal. Worked a treat. The device will operate without them, but I like to try and keep it as close to the original spec as possible.