VOGONS


First post, by suntac

User metadata
Rank Member
Rank
Member

A friend of mine recently bought an old Roland LAPC-I card and asked me to repair it for him. The right channel is much noisier than the left one. I believe I can fix that, but there is another problem. The card usually starts sounding quite weird and distorted (digitally) after a few minutes, 5 or so. Sometimes it may work fine for an hour or even longer.

I got some ideas from the MT-32 Services Notes, but the LAPC-I Services Notes would be much more helpful. Does anyone have it and is willing to share it? We would gladly pay for it, if necessary.

Thanks.

Reply 1 of 18, by suntac

User metadata
Rank Member
Rank
Member

Ok, finally solved it. 😀
The first problem originated in the right low-pass filter. One of its opamps was introducing large amount of noise to the signal — that was easy to fix.
The second problem was caused by IC19 — the reverb mask ROM — which was slightly faulty.
When its temperature rose above a certain point, the output levels of its D1 pin started entering the forbidden zone. Copying its content at room temperature to an EPROM has resolved the problem entirely.

Reply 3 of 18, by root42

User metadata
Rank l33t
Rank
l33t

Do you mind to explain a bit how you found the faults?

YouTube and Bonus
80486DX@33 MHz, 16 MiB RAM, Tseng ET4000 1 MiB, SnarkBarker & GUSar Lite, PC MIDI Card+X2+SC55+MT32, OSSC

Reply 4 of 18, by suntac

User metadata
Rank Member
Rank
Member
root42 wrote:

Do you mind to explain a bit how you found the faults?

Thanks to keropi it was quite easy. I located the first fault just by listening to the signal at different spots in the circuit, by "backtracking" the analog signal path from the output connector to the 6 outputs of sample-and-hold circuits.
From the beginning I suspected the second fault to be caused by the slowly increasing temperature. However, an infrared camera revealed nothing unusual. So I waited for the distortion to emerge, then I cold-sprayed the ICs one-by-one. The only responsive chip was the reverb mask ROM, so I desoldered it and read out its contents. When I moderately heated it up by hot air flow, it started to produce verification errors during reads. The failure of D1 was apparent from the differences among the readouts.
Driven by pure curiosity, I soldered a socket onto the LAPC, placed the ROM back and took a look at the D1 line using an oscilloscope. The shift in logic levels was apparent. 😎

Reply 5 of 18, by suntac

User metadata
Rank Member
Rank
Member

Note that the ROM started failing at around 30°C in the LAPC-I, but it took more than 50°C in the programmer to manifest.
If the board was damaged by an ESD-event, other components might start failing sooner or later...

Reply 6 of 18, by root42

User metadata
Rank l33t
Rank
l33t

I would assume that the data retention period for the ROMs might be nearing its end of life... if so, that would be easy to fix, since ROM images are available, right? I hope that no other severe damage has been done.

YouTube and Bonus
80486DX@33 MHz, 16 MiB RAM, Tseng ET4000 1 MiB, SnarkBarker & GUSar Lite, PC MIDI Card+X2+SC55+MT32, OSSC

Reply 7 of 18, by suntac

User metadata
Rank Member
Rank
Member
root42 wrote:

I would assume that the data retention period for the ROMs might be nearing its end of life... if so, that would be easy to fix, since ROM images are available, right? I hope that no other severe damage has been done.

This is a mask ROM, and thus it has much longer lifespan than EPROM or flash. I assume it was caused by some kind of an electrostatic discharge event.
As far as I know, the reverb ROM of LAPC-I is not publicly available. There are 3 ROMs on the LAPC-I board.

Reply 8 of 18, by keropi

User metadata
Rank l33t++
Rank
l33t++

although I have a LAPC-I for more than 20 years now and the service manual for about 2~3 I had no idea there were more roms other than the firmware and samples... never had to look at the service manual and I hope I never have to! 😁 🤣

🎵 🎧 PCMIDI MPU , OrpheusII , Action Rewind , Megacard and 🎶GoldLib soundcard website

Reply 9 of 18, by suntac

User metadata
Rank Member
Rank
Member
keropi wrote:

although I have a LAPC-I for more than 20 years now and the service manual for about 2~3 I had no idea there were more roms other than the firmware and samples... never had to look at the service manual and I hope I never have to! 😁 🤣

Yesterday, I did some experiments with changing the control ROM in a MT-32, and — behold! — a spark and a magic smoke appeared. 😕 I kid you not. Nevertheless, the MT-32 still works fine.

I was just curious how it would behave if I changed the reverb and control ROMs for LAPC's ones, but this is something I expected the least.

Btw, I sent you an e-mail. Can't PM you yet.

Reply 10 of 18, by keropi

User metadata
Rank l33t++
Rank
l33t++

what? weird why this happened? only reason I see is the mt-32 is using some weird non-jedec roms ?

🎵 🎧 PCMIDI MPU , OrpheusII , Action Rewind , Megacard and 🎶GoldLib soundcard website

Reply 11 of 18, by suntac

User metadata
Rank Member
Rank
Member
keropi wrote:

what? weird why this happened? only reason I see is the mt-32 is using some weird non-jedec roms ?

No, no, it was caused just by the binary content. I flashed the control ROM incorrectly with some garbage (with the upper half of some ROM with the Adjarre demo).
Maybe the adjacent gate array got somehow internally damaged, but in a way that doesn't affect normal function. Who knows?

Reply 12 of 18, by root42

User metadata
Rank l33t
Rank
l33t
suntac wrote:
root42 wrote:

I would assume that the data retention period for the ROMs might be nearing its end of life... if so, that would be easy to fix, since ROM images are available, right? I hope that no other severe damage has been done.

This is a mask ROM, and thus it has much longer lifespan than EPROM or flash. I assume it was caused by some kind of an electrostatic discharge event.
As far as I know, the reverb ROM of LAPC-I is not publicly available. There are 3 ROMs on the LAPC-I board.

Ok, sounds reasonable. I assumed that the ROMs were EPROM. Missed that bit.

YouTube and Bonus
80486DX@33 MHz, 16 MiB RAM, Tseng ET4000 1 MiB, SnarkBarker & GUSar Lite, PC MIDI Card+X2+SC55+MT32, OSSC

Reply 13 of 18, by keropi

User metadata
Rank l33t++
Rank
l33t++
suntac wrote:
keropi wrote:

what? weird why this happened? only reason I see is the mt-32 is using some weird non-jedec roms ?

No, no, it was caused just by the binary content. I flashed the control ROM incorrectly with some garbage (with the upper half of some ROM with the Adjarre demo).
Maybe the adjacent gate array got somehow internally damaged, but in a way that doesn't affect normal function. Who knows?

still weird!

🎵 🎧 PCMIDI MPU , OrpheusII , Action Rewind , Megacard and 🎶GoldLib soundcard website

Reply 15 of 18, by suntac

User metadata
Rank Member
Rank
Member
keropi wrote:

still weird!

That's what I say too.
Btw, did you know that the MPU-401's CPU has an HCF (Halt-and-Catch-Fire) instruction which switches the CPU to a special mode which sequentially reads all possible memory locations?
The HCF opcodes are 9D and DD.

Reply 16 of 18, by keropi

User metadata
Rank l33t++
Rank
l33t++

^ nope this is new to me... I have not really messed with the roland mpu at that level...
do you think they used this to test the cards before shipping? burn-in test of some sorts?

🎵 🎧 PCMIDI MPU , OrpheusII , Action Rewind , Megacard and 🎶GoldLib soundcard website

Reply 17 of 18, by suntac

User metadata
Rank Member
Rank
Member
keropi wrote:

^ nope this is new to me... I have not really messed with the roland mpu at that level...
do you think they used this to test the cards before shipping? burn-in test of some sorts?

No, they didn't. It is an undocumented instruction such as 286's or 386's LOADALL.

Reply 18 of 18, by suntac

User metadata
Rank Member
Rank
Member

I did some experimenting with my MPU-401 and it was quite fun. 😀

_ND38625ex2.jpg
Filename
_ND38625ex2.jpg
File size
249.8 KiB
Views
840 views
File comment
MPU-401 with custom firmware
File license
Fair use/fair dealing exception