Troubleshooting a MV-1FZ with the infamous "Z80 ERROR"

hkz

Kuroko's Training Dummy
Joined
Jul 26, 2013
Posts
70
Before trying to swap out the SM1 (and perform desoldering on the working 1FZ), I'll have a look if i have a compatible EPROM around. Maybe I can just use a big and fast enough one, put it in a breadboard and connect the wires.
As for the Z80, I don't think it's an overclocked one: the MV-1FZ has a smd Z80 (84C00AM-6), which is rated for 6Mhz (at least that's what the datasheet says).
As for the crystal, I think I have a frequency meter somewhere, and it's surely worth a check. But given that the games run fine (without audio) I doubt the crystal is at fault.

Thanks for the ideas.
 

Xian Xi

JammaNationX,
Joined
Dec 1, 2005
Posts
27,694
I doubt it's the 24mhz crystal because if the clock isn't working then nothing would display at all.
 

hkz

Kuroko's Training Dummy
Joined
Jul 26, 2013
Posts
70
Yes, it was a long shot.
Within next week i should receive some spare smd Z80s, but given the probe measurements I doubt swapping the current processor will produce any result. Still, as I have no idea where to get another NEO-D0 chip, it's probably the last thing i can try after swapping the SM1 and SRAM (and i have little to no hope on those).

My situation looks an awful lot like the one in this thread...
 

mjs256

Kula's Candy
Joined
Apr 1, 2010
Posts
292
Yeah it sucks when you have to replace one of those custom ICs - the only place to get another one is a parts board. I did not realize that MV-1FZ's have a different Z80 CPU. My thought was that it was the same overclocked Z80 as most other NEO GEO boards, so I guess scratch the bad crystal theory (I thought the crystal may be a bit fast - not enough to keep the 68k from working but enough to mess up the already overclocked Z80). You may be onto something with the NEO-D0. I am working on a bad board and I think the NEO-D0 may be to blame but I also do not have a spare NEO-D0 lying around :(.

Maybe smkdan can suggest something based on the specifics of what the diagnostics BIOS is saying.
 

hkz

Kuroko's Training Dummy
Joined
Jul 26, 2013
Posts
70
My money is on the tiny Z80. You can find 'em on older Sega Game Gear systems.

Good to know you bet on that: gives me some hope!
I have a small number of those SMD Z80 stuck at customs, which should get released in a few days. I'll wait for those before butchering my Game Gear (which I painfully restored from the dead capacitor problem ;) ).

In the meantime, I took the time to probe the SM1 and the NEO-D0
SM1
1 - H
2 to 15 - P
16 - L
17 to 23 - P
24 - L
25 to 29 - P
30 to 32 - H

NEO-D0
1 to 10 - L
11 to 19 - P
20 - H
21 to 24 - P
25 - L
26 - H
27 - P
28 - P
29 - H
30 - P
31 - P
32 - P
33 - H
34 - H
35 to 40 - P
41 - H
42 - L
43 - H
44 - P
45 to 49 - H
50 to 53 - Not connected
54 - L
55 - H
56 - L
57 - L
58 - H
59 to 64 - L
 

smkdan

Galford's Armourer
Joined
Dec 30, 2009
Posts
452
That test result is similar to what causes a regular Z80 ERROR. 68k asking the Z80 (which is still running from SM1 ROM) to do a slot switch so it can use M1 on the game cart, but the Z80 can't respond to it for whatever reason because it's either faulty, crashed at some point or can't talk to 68k.

It's a lot to check for but you covered most of the possible causes already. Sometimes a Z80 ERROR can be worked around with the test cart but sometimes it gets stuck and shows that error. It's more likely to catch stuff that doesn't cause a Z80 ERROR but does cause sound problems in gameplay like this. AES and boards without SM1 don't have to deal with this though.

Here's some other stuff to try checking if you haven't already. I'm guessing the ones that don't run straight to the Z80 are overlooked a lot of the time:

* NEO-C1 (SDW) -> NEO-D0. Will cause comm. issues if this is bad.
* Z80 (SDAxx, NMI, IORQ etc.) -> NEO-D0. Did you try those or just the stuff on the bullet points at the article?
* NEO-D0 (6116CS) -> 6116 /CS
* NEO-D0 (SDROM) -> (some circuit with LS32? involved, not sure on 1FZ) -> SM1 /OE. You already posted activity on both enables so it's a slim chance.
 

hkz

Kuroko's Training Dummy
Joined
Jul 26, 2013
Posts
70
* Checked SDW from NEO-C1. Fine
* Z80 SDA, NMI, ..., those were included in this table I think. Checked out fine.
* 6116CS checked out fine
* As for this, the connections are as follows: SDROM from NEO-D0 (10) connects to pins 2 and 5 of LS32, pin 3 of LS32 connects to /CE (pin 22) of SM1. Nothing of LS32 connects to /OE.

What a nasty bitch of an error :-)
The spare Z80s got released from customs just today: I should get them by tomorrow or the next day.
 

smkdan

Galford's Armourer
Joined
Dec 30, 2009
Posts
452
It was either /CE or /OE so that sounds normal. SDMRD on NEO-D0 should go to /OE on SM1 in that case. It's already showing activity on your list but there's nothing else to check on SM1 if data+address+power are already covered.
 

hkz

Kuroko's Training Dummy
Joined
Jul 26, 2013
Posts
70
Checked that one too, fine.
I guess there isn't much left to do except waiting for the replacement Z80 and install it!
 

hkz

Kuroko's Training Dummy
Joined
Jul 26, 2013
Posts
70
This morning I received my pack of chips, Z80 included. I replaced the processor and...


channelmaniac wins! IT WORKS!!! I had no time to test with diagnostic bios, but the Z80 ERROR is gone. Pics and updates in the evening when I get back home.
Thank you all for advice and help!
 

hkz

Kuroko's Training Dummy
Joined
Jul 26, 2013
Posts
70
Damn, I might have spoken too soon...
The Z80 ERROR is indeed gone, but the diagnostic cart shows another problem. Now I get...


Z80 DEAD / ERRORED / COMM. ISSUE
EXPECTED: 3C
ACTUAL: 50
Z80 REPORTED ERROR CODE: 10
YM2610 ERROR

Any idea on what to check before I replace the YM2610?
 

mjs256

Kula's Candy
Joined
Apr 1, 2010
Posts
292
The saga continues... Z80 error always makes my shudder. I would say recheck the Z80 with your logic probe - particularly paying attention to the control lines INT, NMI, HALT, WAIT, RESET, etc...

If you end up deciding to replace the YM2610 I have a spare lying around here somewhere...
 

Xian Xi

JammaNationX,
Joined
Dec 1, 2005
Posts
27,694
It's the same error I've been stuck on with an MV1F, I'm almost certain it's a bad Neo-D0.
 

hkz

Kuroko's Training Dummy
Joined
Jul 26, 2013
Posts
70
mjs256, thanks for the offer, but I also had a YM2610 laying around... so I decided to remove and socket the one on the 1FZ.

I think i got lucky: as soon as i replaced the chip, the error went away. Now the diagnostic bios shows

ALL TESTS PASSED

I've yet to try with a game, but this screen is a lot more encouraging than the YM2610 error :)
IMG_20130830_225512541.jpg
 
Last edited:

hkz

Kuroko's Training Dummy
Joined
Jul 26, 2013
Posts
70
I will test the board with the original bios and a game tomorrow, when I get back home. Hopefully, if there are no problems in the analog audio section, the game should play fine and with sound: UNIBIOS already allowed me to play with the damaged YM2610 and Z80.


Now I'd like to know what could have killed both the Z80 and the YM on this board...
 

channelmaniac

Mr Neo Fix-it
Joined
Nov 29, 2005
Posts
4,230
Glad I could help!

Z80 chips just die... I don't think anything in particular killed it... The only thing I could think that might would be inserting/removing a cart with the power on. I killed my TRS-80 Color Computer doing that when I was a kid. Killed the CPU by inserting a cart with the power on. OOPS.
 

hkz

Kuroko's Training Dummy
Joined
Jul 26, 2013
Posts
70
Tested the board and everything is working fine :)

Alas, now I have one problem with the previously working other 1FZ board: stuck on green screen, AKA calendar error.

Already checked the obvious: the D4990A is powered, there is no corrosion around it (the board was coin battery modded by me, by removing D1, D2, R3, and worked fine before and after the mod), it is connected to the NEO-F0 and the HC32 nearby. I also tried replacing the crystal with another one (32.768 KHZ), but no dice, the capacitors between the crystal and GND are still well connected.

I also probed the D4990A:
1 (C2) - H
2 (C1) - H
3 (C0) - H
4 (NC)
5 (STB) - L
6 (CS) - H
7 (DATA IN) - L
8 (GND)
9 (CLK) - L
10 (DATA OUT) - L
11 (TP) - L
12 (OUTPUT ENABLE) - H
13 (NC)
14 (/XTAL) - Nothing shows, maybe the probe can't pick the signal
15 (XTAL) - Same as above...
16 (Vdd)

As TP is not pulsing, I'm suspecting the D4990A died...
 

hkz

Kuroko's Training Dummy
Joined
Jul 26, 2013
Posts
70
Just forgot: when the D4990A is powered by the PSU, it receives 4.3v, not 5.0

Also, STB and DATA IN go high for a brief moment one second after the board is turned on. I guess that is the command to check the clock
 
Last edited:

channelmaniac

Mr Neo Fix-it
Joined
Nov 29, 2005
Posts
4,230
Just forgot: when the D4990A is powered by the PSU, it receives 4.3v, not 5.0

Also, STB and DATA IN go high for a brief moment one second after the board is turned on. I guess that is the command to check the clock

Sounds like you have a corroded plated thru hole between the backup battery section of the board and the power pin on the D4990 chip.
 

hkz

Kuroko's Training Dummy
Joined
Jul 26, 2013
Posts
70
I'll surely check all of them and see if one has a resistance above the standard.
I also had a look at the working (the one that was repaired in this thread) 1FZ, and it receives 4.3v on the D4990A also, but the clock in that one works fine. The datasheet for the uPD4990AG states that the operating voltage for the chip is between 2 and 5.5V, so I thought that 4.3 was still fine for normal operation.
 

Xian Xi

JammaNationX,
Joined
Dec 1, 2005
Posts
27,694
On the FZ, 4.3v is normal for everything in the backup section. The +5v line goes through a resistor or diode and drops to 4.3v. Should only affect the LS32, D4990 and Backup RAM.
 

hkz

Kuroko's Training Dummy
Joined
Jul 26, 2013
Posts
70
Ah, "good" to know. Good in the sense that I won't have to check all the platethru holes, but the D4990 is probably really dead.
The last thing I can do before replacing it is replacing the two caps on the crystal.
 

Xian Xi

JammaNationX,
Joined
Dec 1, 2005
Posts
27,694
Ah, "good" to know. Good in the sense that I won't have to check all the platethru holes, but the D4990 is probably really dead.
The last thing I can do before replacing it is replacing the two caps on the crystal.

Put the D-bios back in and check it. Hold A+B+C+D and press start+select and you will bring up the test menu. Select the one for the calendar. It will test it in real time and you can see if the chip is bad or whatnot.
 
Top