MV1FZS - Click of death, SP1 always disabled, suspect address decoding failure

hkz

Kuroko's Training Dummy
Joined
Jul 26, 2013
Posts
70
Hello everyone!
It's been a few years since I posted here, but I've recently recovered a small lot of defective MV1 boards and started dabbling with them again, for fun!

WARNING: Wall of text incoming!

First of all, I know these boards are relatively cheap by now, and it's economically more viable to just get another if one proves to be too stubborn to be repaired, but this is not the issue: I already have multiple MVS board (some of them you helped me repair!) that are going strong, so this is mostly a hobby/learning exercise.

So, let me present the patient:
A MV1FZS board suffering the dreaded click-of-death, it was pretty dirty when it got to me: some kind of oily substance mixed with dust was covering part of it.

Before working attempting any repair I cut away the battery (little to no leakage traces, but more on that later), removed the sponge below the pcb, desoldered the dipswitches, volume potentiometer, the two crystals and threw the board into my ultrasonic cleaner filled with distilled water.
The board came out in decent shape. I then used an hair drier at low temperature to displace most of the liquid residue and then left the board to dry completely for two days. Then I resoldered everything I removed before the cleaning.

Before attempting to power it on, I checked for shorts, and found one on the 5v line. After careful examination of the board I found out two damaged ceramic caps on the bottom side: one of the caps below the YM2610 was missing, the other was chipped, no other damage was visible.
Removed the damaged cap and replaced both with two 100nF ceramic caps (through hole with legs attached to the pads, couldn't be arsed to search for SMD caps). This made the short on 5v disappear.

At this point I tried to power on the board for the first time: click of death, the "fast" variant, with multiple resets per second.

I disconnected the board and started checking for bad traces. A visual inspection of the battery circuit showed little leakage traces: very light blackening in some zones, nothing on ICs.

The following were tested for continuity:
* CPU1 (A1-A15) -> RAM1 (A0-A14)
* CPU1 (A1-A15) -> RAM2 (A0-A14)
* CPU1 (A1-A16) -> NEO-E0 (A1-A16)
* CPU1 (A1-A16) -> SP1 (A0-A15)
* CPU1 (A17-A21) -> NEO-C1 (A17-A21)
* CPU1 (D0-D15) -> SP1 (D0-D15)
* CPU1 (D0-D7) -> RAM2 (D0-D7)
* CPU1 (D8-D15) -> RAM1(D0-D7)
* CPU1 (D8-D15) -> NEO-C1 (D8-D15)
* RAM1 (OE,WE) -> NEO-C1 (RAOU, RAWU)
* RAM2 (OE,WE) -> NEO-C1 (RAOL, RAWL)
* CPU1 (VPA, DTACK, RW, UDS, LDS, AS, CLK) -> NEO-C1 (VPA, DTCK, RW, UDS, LDS, AS, 68KCLCK)
* SP1 (OE) -> NEO-E0 (AND0)
* NEO-E0 (ANI0) -> NEO-C1 (SROMOEU)
* NEO-E0 (ANI1) -> NEO-C1 (SROMOEL)
* RAM2 (A0-A14) -> RAM3 (A0-A14)
* RAM2 (D0-D7) -> RAM3 (D0-D7)
* RAM1 (A0-A14) -> RAM4 (A0-A14)
* RAM1 (D0-D7) -> RAM4 (D0-D7)

At this point, I took out my logic probe and started checking some ICs:
NEO-C1
* A17-A21: Pulsing
* A22I, A23I: H
* 68LCLK: Pulsing
* AS, LDS, UDS, RW: Pulsing
* CRDW, CRD0, SDW, CRDC: H
* DTACK: L
* PORTOEU, PORTWEU, PORTOEL, PORTWEL: H
* SROMOEL, SRAMOEL, SRAMWEL, SROMOEU, SRAMOEU: H
* SRAMWEU: L

NEO-E0
• AND0, ANI0, ANI1: H

Work and Backup RAM
* CE: L
* OE: H
* RW: H

Checked all the I/O on the 74HC32 piloting the BRAMs, and saw nothing out of the ordinary (no missing signals, and all the outputs of the OR gates were correct).

The clock chip was pulsing on TP1.

The most interesting part is that I also scoped SP1, and it never gets enabled, not even a small blimp on pin 20: always high. In fact, the system behaves the same even without the ROM installed.

At this point, i started replacing ICs:
• Replaced the SP1 rom with a diag (but this was obviously useless, given the issue described above)
• Replaced RAM2, RAM1, RAM4 (not RAM3, not yet at least: i was out of good spares, will get one this week). Yes, I did pay a lot of attention to avoid static charges :)
• Replaced the CPU
None of this brought any change to the behaviour of the machine.

At this point, I begin to fear something is not good with the initial address decoding when the CPU gets reset, I suspect something in NEO-C1, but I don't have spares to test this theory.

Any opinion? Ideas? Something else I could try or test?
If you need/want any photos of the board, just ask.

And if you read up to here, my thanks and congratulations :)

TL;DR
Got a broken MV1FZS to play with. SP1 never gets enabled, traces seems ok. Probably dead custom IC. Ideas?
 

hkz

Kuroko's Training Dummy
Joined
Jul 26, 2013
Posts
70
Replaced RAM3. No changes, as expected.
 
Last edited:

furrtek

Super Sidekicks 2
Joined
Feb 6, 2011
Posts
106
Hello

On NEO-C1, when A23I A22I A21 A20 = 1100 and RW = 1, either SROMOEU and SROMOEL should be low depending on UDS and LDS.
If not, NEO-C1 probably has dead pins.
 

hkz

Kuroko's Training Dummy
Joined
Jul 26, 2013
Posts
70
Wonderful! Thanks furrtek!

I guess this confirms my suspect, but tomorrow I'll recheck the pins to be sure.

Now I wonder if it is worth trying to redo that logic with some TTL ICs or maybe a GAL. Hmmm :)
 

furrtek

Super Sidekicks 2
Joined
Feb 6, 2011
Posts
106
If you dare, the logic should be:

SROMOEU = !RW or UDS or !A23I or !A22I or A21 or A20
SROMOEL = !RW or LDS or !A23I or !A22I or A21 or A20

Not sure if AS is thrown in the mix or not :spock:
 
Last edited:

hkz

Kuroko's Training Dummy
Joined
Jul 26, 2013
Posts
70
Probably yes, given that it indicates a valid address on the bus, but it's not a problem to verify: if it turns out it's not needed, I'll just tie the line to GND on my gal and keep the rest of the signals.

To be fair, I'm not expecting this to bring the mv1 back to life, as probably the dead line is one of the inputs (both sromoeu and sromoel dead would be fishy) but it can surely confirm whether there is something wrong with the C1 if the enable line of SP1 starts pulsing again :)
 

hkz

Kuroko's Training Dummy
Joined
Jul 26, 2013
Posts
70
Implemented the logic functions (including AS) in a GAL, and used it to generate SROMOEU and SROMOEL.
The SP1 gets enabled now, but I still get a click of death, not that it was unexpected: probably C1 has other issues.

This was still very good though: now I have certainty that the NEO-C1 is damaged, and I should search for a replacement. Thanks again furrtek!
I wonder if someone here has a scrap board I can buy...
 

hkz

Kuroko's Training Dummy
Joined
Jul 26, 2013
Posts
70
Heh, I imagine: they're not certainly easy to come by, and probably not worthy for a MV1FZS.
I'll put this in the pile of things to check at a later time. Meanwhile I started fixing a MV1B: burnt ram, broken trace and at least one dead neo-buf... Now it almost works, almost. But I'll make a post on this if I get stuck :)
 

RyuX

Crossed Swords Squire
Joined
Sep 19, 2018
Posts
179
Keeping the only spare NEO-C1 I have very preciously :/

I gladly have two donor boards - however not sure if the NEO-C1 is working there. But funny that on my MV-1FZ it also seems to be a NEO-C1 problem similar to hkz's problem (just different symptoms)
 

hkz

Kuroko's Training Dummy
Joined
Jul 26, 2013
Posts
70
I gladly have two donor boards - however not sure if the NEO-C1 is working there. But funny that on my MV-1FZ it also seems to be a NEO-C1 problem similar to hkz's problem (just different symptoms)

One can always try, worst case, I won't solve the problem.
If you're willing to take the hassle of shipping the boards to Italy, I'm certainly willing to send you a Paypal payment. Let me know in a private message, maybe. Thanks :)
 
Top