Berzerk ZPU repair [1]

Level42

New member
Joined
Dec 22, 2006
Messages
5,895
Reaction score
49
Location
Ridderkerk, ZH, Netherlands
I bought 3 Berzerks last saturday and I want to get them as much running as possible.

I also got one extra ZPU board with the deal. I will call that board [1] so I don't confuse things too much.

Started work on this board tonight. Of course I first removed all the batteries, neutralized the acid damage with vinegar and tried to clean the boards as good as possible. Two have little acid damage, the others are a bit worse.

This board had Q2 damaged (broken pins) so I replaced it. I checked the tracks since there was some damage there. Turned out the Base and Emitter were fine but the track from the collector to R28 was open. I jumpered that. No other parts look/feel physically damaged.

Powered up this board with a AT power supply. Took me some time to realize that the board needs two +5V supplies (I connected the other voltages to the test-points). The 2nd +5V is specced as "unregulated" on the schematics. Anyway I jumpered the +5V to the pin of the unreg. +5V, I hope and suppose that's fine.

When I did this and powered up I had the LED on for the first time. There is a clock signal, but the reset line was constantly low so no "flashes".

When I forced the reset line high (by shortly shorting it to +5V, I don't know if it's healthy but it's a trick I learned from my Centipede board repair) I got what looks like 2 or 3 flashes, not sure.

Is it normal that the LED goes off after the flashes ?

So I traced back the reset line and figured I'd examine the 74LS00 and 7406 (7G and 8G).

I used the old "Randy Fromm" trick and hook up the + of my Fluke to the GND and tested all the pins of those two ICs. And indeed, I measure a short at pin 10 at 7G that isn't supposed to be there. In fact, that pin is directly connected to +5V !

When I measure resistance between +5V and GND connection, it gives about 50 ohms, that can't be good. But this also means it can come from "everywhere" on the PCB.....

To be sure, I measured the resistance on the other boards and it seems that about 2k is normal.

The easiest first step was to hook up the Fluke to +5V and GND with some clamps and remove the socketed chips one by one to see if the short would disappear. I started at the RAM at E1 and E2 and I got lucky because when I removed E2 the resistance jumped to about 2k ! Bingo !

Measured the 2114 RAM and there's a short between pin 9 (gnd) and the other pins, and the short was on the Write Enable pin 10.

So, swapped the RAM with one from the other boards and had high hopes. Start up, LED on, but no flashes......

Reset still low.....

Traced back reset through 7G and 8G and pin 9 of the NAND port in 7G is constantly low.

So, there has to be an issue with the transistor circuit I guess. Right ?
Can anyone explain how that circuit works ?
 
Last edited:
I bought 3 Berzerks last saturday and I want to get them as much running as possible.

I also got one extra ZPU board with the deal. I will call that board [1] so I don't confuse things too much.

Started work on this board tonight. Of course I first removed all the batteries, neutralized the acid damage with vinegar and tried to clean the boards as good as possible. Two have little acid damage, the others are a bit worse.

This board had Q2 damaged (broken pins) so I replaced it. I checked the tracks since there was some damage there. Turned out the Base and Emitter were fine but the track from the collector to R28 was open. I jumpered that. No other parts look/feel physically damaged.

Powered up this board with a AT power supply. Took me some time to realize that the board needs two +5V supplies (I connected the other voltages to the test-points). The 2nd +5V is specced as "unregulated" on the schematics. Anyway I jumpered the +5V to the pin of the unreg. +5V, I hope and suppose that's fine.

When I did this and powered up I had the LED on for the first time. There is a clock signal, but the reset line was constantly low so no "flashes".

When I forced the reset line high (by shortly shorting it to +5V, I don't know if it's healthy but it's a trick I learned from my Centipede board repair) I got what looks like 2 or 3 flashes, not sure.

Is it normal that the LED goes off after the flashes ?

So I traced back the reset line and figured I'd examine the 74LS00 and 7406 (7G and 8G).

I used the old "Randy Fromm" trick and hook up the + of my Fluke to the GND and tested all the pins of those two ICs. And indeed, I measure a short at pin 10 at 7G that isn't supposed to be there. In fact, that pin is directly connected to +5V !

When I measure resistance between +5V and GND connection, it gives about 50 ohms, that can't be good. But this also means it can come from "everywhere" on the PCB.....

To be sure, I measured the resistance on the other boards and it seems that about 2k is normal.

The easiest first step was to hook up the Fluke to +5V and GND with some clamps and remove the socketed chips one by one to see if the short would disappear. I started at the RAM at E1 and E2 and I got lucky because when I removed E2 the resistance jumped to about 2k ! Bingo !

Measured the 2114 RAM and there's a short between pin 9 (gnd) and the other pins, and the short was on the Write Enable pin 10.

So, swapped the RAM with one from the other boards and had high hopes. Start up, LED on, but no flashes......

Reset still low.....

Traced back reset through 7G and 8G and pin 9 of the NAND port in 7G is constantly low.

So, there has to be an issue with the transistor circuit I guess. Right ?
Can anyone explain how that circuit works ?



Sounds like you still have more acid damage to track down. Don't forget...7G and 8G...if there is any acid damage 'close'...they are toast. There are also very fine traces under those chips that easily evaporate...and very hard to tell which legs they go to because they jog over directly under the chips. Check all the transistors under the battery...and check that reset button to make sure it is working and not corroded closed. Once you see corrosion and it is enough you think you need to clean it....assume it is everywhere....

Good luck!
 
7G and 8G are amazingly untouched by the acid on this one. The op that I bought the machines of was smart enough to remove the battery from the board earlier on. Regretfully he wasn't smart enough to do it on the machines though....

So this is the cleanest board.

There are some acid spots though, but I think I got all covered. Guess it will be simply measuring step by step each trace and part.
 
Measured ALL the connection around the reset area with transistors etc. Only one track was broken between R34 and R30. Jumpered it, again hoping for the 3 flashes this time....no go. Still constant LED.

Just for fun I powered up a board that had a broken off pin on Q2 and it gives two flashes (like when I short the +5V to the reset pin on the other board).

Wish I could hear the beeps instead of the flashes. Not sure if it's two or three now...
Led goes on at power up, goes out, flashes one time and then goes out permanently. Doesn't look like three flashes to me right ?
 
Is it coming out of reset?

No, stays low. I have trouble understanding how the circuit actually works. What does it do at power up ?
I also don't understand why the board with the cut-off leg at Q2 DID start-up like it looked normal (although it only gives two flashes).

The worst thing that one of the boards actually gave three flashes when it was still in the machine. I made a video of that, so I now know that it should be three flashes and then the LED has to stay on. AFAIR it didn't give three beeps when I actually powered up the machine, but when removing the power connector and back on, it did give the three beeps.

But let's focus on board [1] which never gave three beeps yet, but does give two beeps when shorting +5v to reset.
 
OK, just to see if I could isolate the problem a bit more I shortly shorted base and emitter of Q5 and get the same result as when I force the reset line of the processor high, two flashes. So I guess the section after Q5 works.

Just to be sure: I can run the board without the battery right ? I assume so, but let's not forget to mention the obvious...

Let me mention here that the schematics incorrectly show R32 as R3 (100 Ohms, 1 W, the biggest resistor) and that zener diode CR26 is actually two diodes in series ( I think the original zener plus a regular diode).

At Q1 I measure about 5V at the collector, almost nothing at the base and emitter. Shouldn't this be high in powered up condition?
 
These ZPU boards will function without the battery installed. I had one non-working acid damaged board that I desoldered/resoldered all connections in the area of the battery and it fired right up (lucky I know). Another ZPU that gave no beeps I swapped out the z80 chip and it worked, shotgun repair at best but thought I would mention it.
 
These ZPU boards will function without the battery installed. I had one non-working acid damaged board that I desoldered/resoldered all connections in the area of the battery and it fired right up (lucky I know). Another ZPU that gave no beeps I swapped out the z80 chip and it worked, shotgun repair at best but thought I would mention it.

Could you (or anyone else with a working Berzerk) measure the voltages on Q1 ? That would really help me.
 
Berzerk ZPU

Hey,

I was just looking at a gorf Battery backup that works very much like this Berzerk. The +5 V Unregulated is specified as 7.75 Volts to 13 Volts on the power supply schematic J3-1. It is also specified as +9 Volts on the interconnect diagram.

There is a 5.6 Volt zener in series with the emitter of Q1 and you have to have at least the combined voltage to turn it on. (ie 5.6 + 0.6 = 6.2 Volts) you also have to be able to supply the 300mA or so that is used to charge the batteries.

If someone has a working Berzerk with an original power supply they can take a measuremnet and let you know what the "+5 V UNREG" actually reads on their unit. (it should be around 10 or 11 Volts by calculation... assuming the schematics are right!)

If it was me... I might try hooking up the +12v on that line but I would be very careful to watch for smoke.

Kenneth
 
Hey that sounds very reasonable ! I am testing it on a AT power supply and have the (regulated) +5V connected to the "unregulated 5V" so indeed it might be too low.....
 

Attachments

  • Schermafbeelding 2009-11-08 om 20.17.39.jpg
    Schermafbeelding 2009-11-08 om 20.17.39.jpg
    9.9 KB · Views: 21
Last edited:
Berzerk #1

I did some more checking...
The +12 Volts is too much for this circuit…. The +5 Volts is too little.

At initial power-on the reset line will be held low by R65. The unregulated +5 Volts will be seen at C8 through R19. C8 will charge until Q1 conducts through CR26 and R31. This is the power-on-reset. The unregulated voltage flowing through Q1/CR26 should be high enough to turn on Q5 through R30/R37/R34. (at least 1 diode drop over the 5.6V, but you say there is an additional diode at CR26, so it would be even higher) Q5 then grounds Q3's base through R39 and turns it on, providing a pseudo-regulated +5 Volts to the reset line pulling it high. The unregulated +5 Volts flows through R3 and is kind-of regulated by the +5 Volt Regulated line and CR35 and is then dropped a little by CR36. (I referred to this as pseudo-regulated above) This voltage should look like a partially regulated +5 volts. (It will probably read around +5VDC on a meter after CR36) The batteries are charged through this pseudo-regulated +5 Volts. The reset pushbutton, momentarily grounds the base of Q1 causing it to lose conduction and everything goes back to the same state as initial power-on, reset is low. When the rest button is released the whole reset process begins again. When power is turned off the unregulated voltage will go to zero first and the regulated +5Volts will be last. This will give a reset during power down to keep the processor from writing junk into the battery backed RAM.

So, a long winded answer to your previous question… there should be whatever your "+5 Volt Unregulated" is at both the collector and base of Q1 about 100-200 milliseconds after power-on. If not, either there is an open at C8/R19/Q1-Base or C8 is bad. Check continuity for all connections associated with C8, R19 and Q1. sounds like Q1 is not turning on, so you are stuck in reset. You could just jumper Q5-Coll to Q5-Emit, after power on has stabilized, for troubleshoot of the rest of the board and come back to this.

Your AT power supply provides only regulated voltages, and all the voltages come on and turn off at almost the same time. This will give unreliable Battery backed RAM results but if you can get the voltage just right for the "+5 Volts unregulated" the rest should work.

If you are so inclined… You could try connecting a 220 Ohm, 1/4 watt (or 1/2 watt) resistor in series with the +12Volts from your AT supply to the +5 Volt unregulated input on the ZPU. The resistor will get warm… maybe even hot, but should allow the pseudo-regulator to give you the correct voltage at Q1 and the reset line will go high. (somewhere around +7.5 volts) You should then be able to troubleshoot the rest of the board. It would be best to leave the batteries unconnected, as this solution would not give you enough current to properly charge them.

Sorry it is sooo lengthy a reply
Hope this helps, Kenneth
 
Really no need to be sorry for the lengthy reply, in fact that's exactly the kind of info I'm looking for !

Printing it out right now and will read it while looking at the schematics.

Big thank you for taking the time to write that !

I'm SO eager to get at least one of the boards working, that will make it easier to get all of them running.....at least, I hope :D
 
I tried all the boards on two machines, non still work. Still get the same result with two boards (2 flashes when I force the reset high).

I measured the +5 unregulated when the boards were connected and they're 10,8V.

Thanks for the write up, I understand how it works now...back to investigating now !

PS One of the machines lacked the unreg 5V but I stupidly forgot to check the fuse....
 
I will try to hook up the AT power supply with the resistor for further testing.

What I do remember is that the collector of Q1 was high, but the emitter low. The base did seem to get about half a volt or so, which seems a bit low ?

I tried replacing the cap C8 but I used 22 uF instead of 4.7 uF figuring it would only make it taking a bit longer to trigger Q1. I had no 4.7 around a couple of days ago, I have now....
 
So, I've replaced C8 with a 4.7 uF again. The use of resisors lowered voltage as expected, but didn't seem to change anything with regards to the reset circuit.

Replaced Q1. The base leg broke off pretty easily even tough I had measured the transistor before and it seemed fine. So I had high hopes, but the new Q1 didn't change anything. Still low at emitter and base less than half a volt.

So for now I decided to let the reset circuit be as it is and focus on getting the board running "forced".

Replaced both RAMs (5114 instead of 6513 so I had to change jumpers.) and 2114.

Now, power up or manual (button) reset still don't only give a lit-up LED. But when I short collector and emitter of Q5, it looks like I get three beeps. The LED goes OUT after that, is that the correct behavior for a working ZPU board without any other boards attached ? IF so I can at least test the other boards in the three cabs.....
 
Last edited:
Berzerk ZPU #1

Hi!

I am probably going over old ground here but...
Check for continuity between all components in the reset circuit, and check that S1 is not stuck open or short.
Since you have no voltage at Q1's base, I suspect that R19 is open or the series diode at CR26 is in backwards or that S1 is short.
With power on you should see +5 V unreg at R19 and, after 100 mSec or less, at least 6 or so Volts at Q1's-base.

I believe that I read 3 flashes is all you should get with just the ZPU board hooked up.

Kenneth
 
Last edited:
Three beeps/flashes means ZPU is good and on to the BSC and checking those interconnect cables. Before moving on make sure your reset line (pin 26 on the Z-80) is going from low to high upon power up...

Alan


So, I've replaced C8 with a 4.7 uF again. The use of resisors lowered voltage as expected, but didn't seem to change anything with regards to the reset circuit.

Replaced Q1. The base leg broke off pretty easily even tough I had measured the transistor before and it seemed fine. So I had high hopes, but the new Q1 didn't change anything. Still low at emitter and base less than half a volt.

So for now I decided to let the reset circuit be as it is and focus on getting the board running "forced".

Replaced both RAMs (5114 instead of 6513 so I had to change jumpers.) and 2114.

Now, power up or manual (button) reset still don't only give a lit-up LED. But when I short collector and emitter of Q5, it looks like I get three beeps. The LED goes OUT after that, is that the correct behavior for a working ZPU board without any other boards attached ? IF so I can at least test the other boards in the three cabs.....
 
So, when it flashes/beeps three times, should the LED go OFF or stay ON ?

Hope to move the cab to my home soon, this will make the next steps a lot easier to test.
 
Back
Top Bottom