Robotrontechnik-Forum

Registrieren || Einloggen || Hilfe/FAQ || Suche || Mitglieder || Home || Statistik || Kalender || Admins

Willkommen Gast!

Robotrontechnik-Forum » Technische Diskussionen » A7100 repair » Themenansicht

Autor

Thread - Seiten: -1-

000
06.09.2025, 17:13 Uhr
RetroNora

Hello,
I got a set of A7100 and alpha 1 monitor. The monitor got a defective psu board. However when powered from two bench PSUs (+12V and +30V) it seems to work fine.

At first the computer showed no sign of life. It was due to missing 12V rail (the fuses inside the 12V PSU module corroded). After replacing the fuses the computer turned down after couple of seconds with error code LED flashing - as I read online it was a fan failure. After cleaning the optical sensord the A7100 stopped turning off.

Now it turns on, beeps once, sometimes twice (with second beep much quiter then the first one), when reset button on the front panel is pressed it beeps at about 200 Hz. With the monitor connected the screen flashes green, sometimes flashes green/black bars. With reset button on the keyboard pressed it seems to reset fine. Sometimes it is solid black with a quick green flash when the system is turned off.

Any ideas if this can be an issue with monitor or the computer ot both?

Profil || Private Nachricht || Suche

Zitatantwort || Editieren || Löschen

001
06.09.2025, 18:16 Uhr
Ordoban

There should be a beep 8s after reset/switch on, and a shorter beep 2s later.
(if you get nothing here, the CPU is not even running, or the buzzer is broken)
(if you get a pulsing permanent tone here, the CPU cannot access the RAM)
After 10 more seconds there should be on more beep and some outputs starting on the screen.
(If you get more than one beep here, the CPU cannot access the graphics, or the graphics report an error)
Even more 15 seconds later there should be action on the floppy drives.

Since you report that you not always get the fist two beeps, I would think the ZVE (Main-CPU) is not running stable.
I would start by measuring the main (5V) supply, the CPU clock, CPU Reset signal, CPU S0-S2 signals.
--
Gruß
Stefan

Profil || Private Nachricht || Suche

Zitatantwort || Editieren || Löschen

002
06.09.2025, 21:24 Uhr
RetroNora

Hello, thanks for answer.
So I am getting a beep after 6-7 seconds after power on, there is a shorter beep 2s later. And a third beep 9 seconds later. After third beep there is silence. (Waited for 2-3 minutes). Pressing the restart button repeats the pattern. There is no floppy activity.

Greetings
Filip

Dieser Beitrag wurde am 06.09.2025 um 21:24 Uhr von RetroNora editiert.

Profil || Private Nachricht || Suche

Zitatantwort || Editieren || Löschen

003
06.09.2025, 21:45 Uhr
Ordoban

So the system runs and recognizes the graphics. There should be something on the screen.

Can you measure the video signal with an oszilloscope or logic analyzer?
Pin 1+2 GND
Pin 6 Video 2 (Pixel signal)
Pin 7 Video 1 (Pixel signal)
Pin 9 BSYN (Combined HSYNC+VSYNC)
This are open collector TTL outputs, so you need external pullup-resistors and some voltage for measuring.

Edit:
should look like that:

(My A7100 came without monitor, so i wrote a display module for my logic analyzer as an interim solution. Today we also have a VGA adapter device. see
https://www.robotrontechnik.de/html/forum/thwb/showtopic.php?threadid=21315
and
https://www.robotrontechnik.de/html/forum/thwb/showtopic.php?threadid=21380)
--
Gruß
Stefan

Dieser Beitrag wurde am 06.09.2025 um 22:00 Uhr von Ordoban editiert.

Profil || Private Nachricht || Suche

Zitatantwort || Editieren || Löschen

004
07.09.2025, 13:06 Uhr
RetroNora

Well, it does not look TTL-ish at all.

CSync:

Video 2 on top and Video 1 on bottom:

While I was testing it I have heard some bubbling/boiling noise from the 5V PSU so I need to take care of it first

Profil || Private Nachricht || Suche

Zitatantwort || Editieren || Löschen

005
07.09.2025, 14:48 Uhr
Ordoban

Zitat:

RetroNora schrieb
Well, it does not look TTL-ish at all.

Looks somehow like a broken ground connection.

Zitat:

RetroNora schrieb
While I was testing it I have heard some bubbling/boiling noise from the 5V PSU so I need to take care of it first

A nice greeting from the (often failing) golden "ROE" caps, or the useless redundant mains filter. There are also more rotting fuses inside this PSU.
--
Gruß
Stefan

Profil || Private Nachricht || Suche

Zitatantwort || Editieren || Löschen

006
07.09.2025, 17:16 Uhr
RetroNora

Zitat:

Ordoban schrieb

Looks somehow like a broken ground connection.

Well, the ground connection was made on the monitor cage, to this cage are connected the negative clips of two bench PSUs I am using to power the monitor. Will check all the cabling tho.

Zitat:

Ordoban schrieb
A nice greeting from the (often failing) golden "ROE" caps, or the useless redundant mains filter. There are also more rotting fuses inside this PSU.

Somehow the fuses are fine, the primary filter caps (these in metalic cans with rubber cap at the top looks fine, however the output filter caps are somehow lifted. There was no smoke out of the PSU or the blowing caps.

The +5V rail is at 5.03V so I'd say it's fine

Dieser Beitrag wurde am 07.09.2025 um 17:31 Uhr von RetroNora editiert.

Profil || Private Nachricht || Suche

Zitatantwort || Editieren || Löschen

007
07.09.2025, 17:36 Uhr
Ordoban

I would not see the output caps as "liftet". They look normal to me.
I suspect the mains filter (the thing with the green terminals in top left on your picture).

This thing is an LC-Filter element, filled with a kind of wax. If the isolation of this thing fails, it will go quite hot, and boil out some of the wax. In my picture you can still see some drops of wax.
--
Gruß
Stefan

Dieser Beitrag wurde am 07.09.2025 um 17:37 Uhr von Ordoban editiert.

Profil || Private Nachricht || Suche

Zitatantwort || Editieren || Löschen

008
07.09.2025, 18:19 Uhr
RetroNora

Oh it might be the case. You can see some kind of yellow goo leaking out of it (under top red cable).
Will bypass it and see if the boiling stops.

Dieser Beitrag wurde am 07.09.2025 um 18:19 Uhr von RetroNora editiert.

Profil || Private Nachricht || Suche

Zitatantwort || Editieren || Löschen

009
09.09.2025, 14:05 Uhr
RetroNora

Hi,
have bypassed the said filter and the noise stopped, the 5V rail is at 5.02V.

It was not a bad ground just too big resolution on my cheapo oscilloscope.

Vid1, Vid2 and CSync look okay-ish.

I also checked the video signal at the CRT neck and compared it to the oscillograph from monitor schematics. And if I understand the oscillograph they do look similar so I guess the monitor is not the issue.

After all there is no floppy activity so even if the bad video is a result of broken monitor the computer also has an issue.

Do the order of the boards matter? Also, do anyone has a pinout of the backplane?

Dieser Beitrag wurde am 09.09.2025 um 14:08 Uhr von RetroNora editiert.

Profil || Private Nachricht || Suche

Zitatantwort || Editieren || Löschen

010
09.09.2025, 17:08 Uhr
RetroNora

okay, I got a clue. It seems that the video output and the video processing on the monitor side is correct. The issue is the address lines for onboard VRAM are stuck low.

So far I have found that they are MUXed by four 74153 near the backplane connector. The enable pins for these dual 4-to-1 muxes are connected together. The enable signal is 1,67V so I guess I got a bad output of one (or more) ICs somewhere.

My A7100 got the K7072 board, I've been trying to find schematics for it, but got only a theory of operation.

Dieser Beitrag wurde am 09.09.2025 um 17:08 Uhr von RetroNora editiert.

Profil || Private Nachricht || Suche

Zitatantwort || Editieren || Löschen

011
09.09.2025, 17:19 Uhr
Ordoban

Zitat:

RetroNora schrieb
Vid1, Vid2 and CSync look okay-ish.

BSync looks good so far.

Zitat:

High means bright green on the screen. The low pulses are the time between the lines. So this matches the screen output in #0.
There should be only very short high pulses on Vid1 (the very bright cursor), and a somehow chaotic signal on Vid2 (the normal bright text).
This means the video output is broken somehow.

Zitat:

i agree.

Zitat:

After all there is no floppy activity so even if the bad video is a result of broken monitor the computer also has an issue.

Yes, but the video should be the first thing to fix, simply because you can see the selftest then.

Zitat:

Do the order of the boards matter?

Yes, there are just a few allowed combinations. From top to bottom:
ABG (graphics output)
KGS (graphics processor)
KES (Floppy / Harddisk gateway) or OPS (Memory) or ASP (serial / paralell ports)
KES (Floppy / Harddisk gateway) or OPS (Memory) or ASP (serial / paralell ports)
ZVE (processor) or OPS (Memory) or ASP (serial / paralell ports)
OPS (memory) small slot
ZVE (processor) or OPS (Memory) or ASP (serial / paralell ports)

Zitat:

Also, do anyone has a pinout of the backplane?

Not just the backplane, but the full schematics in KiCad on my github: https://github.com/Stefan311/MMS16-KiCad-Umsetzung
There are also some original schematics and other stuff here: https://www.tiffe.de/robotron/MMS16/
--
Gruß
Stefan

Dieser Beitrag wurde am 09.09.2025 um 17:19 Uhr von Ordoban editiert.

Profil || Private Nachricht || Suche

Zitatantwort || Editieren || Löschen

012
09.09.2025, 17:43 Uhr
Ordoban

Zitat:

RetroNora schrieb
okay, I got a clue. It seems that the video output and the video processing on the monitor side is correct. The issue is the address lines for onboard VRAM are stuck low.

So far I have found that they are MUXed by four 74153 near the backplane connector. The enable pins for these dual 4-to-1 muxes are connected together. The enable signal is 1,67V so I guess I got a bad output of one (or more) ICs somewhere.

My A7100 got the K7072 board, I've been trying to find schematics for it, but got only a theory of operation.

The enable inputs (pin 1 and 15) should be pinned to ground. And yes - it is weird that they are connected by a trace and then connected to ground instead of direct connected to the ground plane.
--
Gruß
Stefan

Profil || Private Nachricht || Suche

Zitatantwort || Editieren || Löschen

013
09.09.2025, 18:04 Uhr
RetroNora

Thanks for the informations.
I guess the video output is not broken, but is only displaying first bit of every RAM IC, and I guess they are just blank.
I'll verify the connection for the enable pins of the MUXers. Maybe they are interconnected to ground plane, but there is a trace going to two caps near the connector. Maybe one of them is bad or semi-shorted. For sure there is no muxing of the address lines, all address are low.

The order of the boards fits one of the combinations you have described.
Also, the floppy controller is in the middle one of three front slots, I guess this is also correct.

Dieser Beitrag wurde am 09.09.2025 um 18:08 Uhr von RetroNora editiert.

Profil || Private Nachricht || Suche

Zitatantwort || Editieren || Löschen

014
09.09.2025, 18:40 Uhr
Bert

Zitat:

RetroNora schrieb
The enable signal is 1,67V so I guess I got a bad output of one (or more) ICs somewhere.

Keep in mind, that also a bad input could result in wrong voltage levels...
--
Viele Grüße,
Bert

Profil || Private Nachricht || Suche

Zitatantwort || Editieren || Löschen

015
09.09.2025, 21:35 Uhr
RetroNora

Zitat:

Bert schrieb
Keep in mind, that also a bad input could result in wrong voltage levels...

Oh yes, of course it can, but a voltage of ~1,7V in TTL circuits most likely shows an output that is floating. Of course it can mean a bad input but it is less likely.

Have checked with my multimeter and the enable inputs show open line to ground. So maybe a broken trace somewhere.

Profil || Private Nachricht || Suche

Zitatantwort || Editieren || Löschen

016
09.09.2025, 21:50 Uhr
Ordoban

To be on the same page: you are talking about this inputs?

They are connected to ground, and should have 0,0V and not a single milivolt more.
--
Gruß
Stefan

Profil || Private Nachricht || Suche

Zitatantwort || Editieren || Löschen

017
10.09.2025, 07:14 Uhr
RetroNora

Yes, I am talking about these exact inputs

Profil || Private Nachricht || Suche

Zitatantwort || Editieren || Löschen

018
10.09.2025, 12:38 Uhr
RetroNora

Well, this was not helping for sure.

We got a screen output, but it signals some issues.

There is something wrong with 'PIC' - is it Programmable Interrupt Controller? or some german name?

Also it shows the NOGO for KES

Profil || Private Nachricht || Suche

Zitatantwort || Editieren || Löschen

019
10.09.2025, 12:46 Uhr
Rüdiger
Administrator

PIC (KGS) is OK, you only did not enter a key when expected.
KES is damaged or the board is plugged at wrong slot
--
Kernel panic: Out of swap space.

Profil || Private Nachricht || Suche

Zitatantwort || Editieren || Löschen

020
10.09.2025, 13:15 Uhr
RetroNora

Okay, so the KES is the only issue.
The order of the board is:

With the KES board removed the system beeps and does not boot.
Since the error is at the start of the testing I guess there is some communication error or something like that.

Profil || Private Nachricht || Suche

Zitatantwort || Editieren || Löschen

021
10.09.2025, 14:05 Uhr
Rüdiger
Administrator

Zitat:

RetroNora schrieb
The order of the board is:

...correct.
--
Kernel panic: Out of swap space.

Profil || Private Nachricht || Suche

Zitatantwort || Editieren || Löschen

022
10.09.2025, 17:11 Uhr
Ordoban

Zitat:

RetroNora schrieb
Since the error is at the start of the testing I guess there is some communication error or something like that.

Yes. The KES does not answer at all.
You can check the Clock, Reset, Address and Data on the KES Z80-CPU.
You can also trigger a reset to the KES by typing "o 100,2" and "o 100,0" in the monitor program. ("o 100,1" and "o 100,0" does a wakeup call to the KES)
--
Gruß
Stefan

Dieser Beitrag wurde am 10.09.2025 um 17:12 Uhr von Ordoban editiert.

Profil || Private Nachricht || Suche

Zitatantwort || Editieren || Löschen

023
10.09.2025, 18:19 Uhr
RetroNora

I don't know much about the behave of this computer. But with two top boards removed (KGS and ABG) so the KES board is easy to probe with scope I think the onboard Z80A CPU is faulty.
The clock signal is fine, the reset if fine, there is some activity on the data lines, but all the address lines are low. The /RD output is ar 1,45 V. And after a short while the CPU is burning hot to the finger. My thermometer shows 71 °C after a minute of work.

After trying to power the KES board from my bench PSU - at 5.0V it is consuming 3A. The Z80A is burning hot, no other chip is even slightly warm. I guess it's time to replace the Z80A.

Dieser Beitrag wurde am 10.09.2025 um 18:34 Uhr von RetroNora editiert.

Profil || Private Nachricht || Suche

Zitatantwort || Editieren || Löschen

024
10.09.2025, 18:54 Uhr
Ordoban

It is also possible some of the other Address Bus members are faulty and the Z80 CPU have to fight against a short circuit.
I have an idea: please measure the common high signals "H1" and "H2". This are always-high signals by resistors to +5V. If these are somehow shorted to ground, some of the 8287er bus transmitters will change the direction, and become hostile to the CPU.

Edit: the KES is rated at max 2,8A, so yes, something bad is going on.
--
Gruß
Stefan

Dieser Beitrag wurde am 10.09.2025 um 19:00 Uhr von Ordoban editiert.

Profil || Private Nachricht || Suche

Zitatantwort || Editieren || Löschen

025
11.09.2025, 11:35 Uhr
RetroNora

So it was the Z80A that was bad.

Now, the last thing to do, to check if it is working, is to boot it. I was looking for a boot disk image online but found only for A7150, I guess they are not compatible.

Can I ask for a boot image for A7100?

Profil || Private Nachricht || Suche

Zitatantwort || Editieren || Löschen

026
11.09.2025, 11:44 Uhr
Rüdiger
Administrator

Zitat:

RetroNora schrieb
I was looking for a boot disk image online but found only for A7150, I guess they are not compatible.

Insufficient information.
--
Kernel panic: Out of swap space.

Profil || Private Nachricht || Suche

Zitatantwort || Editieren || Löschen

027
11.09.2025, 11:58 Uhr
RetroNora

I got this

Profil || Private Nachricht || Suche

Zitatantwort || Editieren || Löschen

028
11.09.2025, 12:37 Uhr
Rüdiger
Administrator

Not usable at A7100.
--
Kernel panic: Out of swap space.

Profil || Private Nachricht || Suche

Zitatantwort || Editieren || Löschen

029
11.09.2025, 12:58 Uhr
RetroNora

So could you please provide me with something A7100 usable?

Profil || Private Nachricht || Suche

Zitatantwort || Editieren || Löschen

030
11.09.2025, 16:37 Uhr
Ordoban

Zitat:

RetroNora schrieb
So it was the Z80A that was bad.

Very unusual.

Zitat:

Can I ask for a boot image for A7100?

You got a PM. (I don't want to post the link in the open, there may be still a copyright issue)

You know about this type of drives? They look like standard 360K, but they are 80 tracks instead of 40. So same parameters as 3.5° 720K drives.
--
Gruß
Stefan

Dieser Beitrag wurde am 11.09.2025 um 16:53 Uhr von Ordoban editiert.

Profil || Private Nachricht || Suche

Zitatantwort || Editieren || Löschen

031
11.09.2025, 17:14 Uhr
RetroNora

Thanks,
aren't they the same drive used in 1715 or A5120? So K5601 or their Teac equivalent.

So for making a boot floppy a 1.2M drive and 360K disk?

Profil || Private Nachricht || Suche

Zitatantwort || Editieren || Löschen

032
11.09.2025, 17:20 Uhr
Ordoban

Yes, same drive in 1715. I don't know about the A5120.
1.2M drive does not work, it turns to fast (360rpm instead of 300).
You also should know, the format of this disk is not compatible with modern disk controllers.
(256 byte sectors). So you maybe need a older floppy controller to write this disk image.
--
Gruß
Stefan

Profil || Private Nachricht || Suche

Zitatantwort || Editieren || Löschen

033
11.09.2025, 17:24 Uhr
Rüdiger
Administrator

Zitat:

Ordoban schrieb
1.2M drive does not work,

Wrong.

Zitat:

You also should know, the format of this disk is not compatible with modern disk controllers.
(256 byte sectors).

Wrong again.
--
Kernel panic: Out of swap space.

Dieser Beitrag wurde am 11.09.2025 um 17:25 Uhr von Rüdiger editiert.

Profil || Private Nachricht || Suche

Zitatantwort || Editieren || Löschen

034
11.09.2025, 17:30 Uhr
Ordoban

Zitat:

Rüdiger schrieb

Zitat:

Ordoban schrieb
1.2M drive does not work,

Wrong.

Well, there are 1.2M drives who can change the rotation speed.

Zitat:

You also should know, the format of this disk is not compatible with modern disk controllers.
(256 byte sectors).

Wrong again.

The teledisk manual says so, if i remember right.
--
Gruß
Stefan

Profil || Private Nachricht || Suche

Zitatantwort || Editieren || Löschen

035
11.09.2025, 18:47 Uhr
RetroNora

If the drive is the same as on 1715 - then making a boot disk on 1.2M drive works. At least on my unit.

I am getting no sector found errors for the first track? Is the first track somehow special (like FM/MFM encoding not supported by IBM compatible controllers?).

Once I had to make a boot floppy for Olivetti M20 that required first track to be 128 bit/sector using FM. I have some old WD controller from 1987 that did the job. Is it something similar for the A7100 adn SCP1700?

Thanks

Profil || Private Nachricht || Suche

Zitatantwort || Editieren || Löschen

036
11.09.2025, 18:58 Uhr
Rüdiger
Administrator

Zitat:

RetroNora schrieb
Is it something similar for the A7100 adn SCP1700?

Correct. A7100 uses FM at boot tracks too.
--
Kernel panic: Out of swap space.

Dieser Beitrag wurde am 11.09.2025 um 19:00 Uhr von Rüdiger editiert.

Profil || Private Nachricht || Suche

Zitatantwort || Editieren || Löschen

Seiten: -1- [ Technische Diskussionen ]