Atari 520 STFM faulty - Project

[zChris]

I'm not sure if this question is suitable here or in the hardware section. But since i'm still a noob on these things i post it here.

I have retrieved my faulty Atari 520STFM for a fun project for me to learn how it works and see if i will be able to diagnose and maybe fix it. So i have been reading up on pullup-resistors. So my understanding is that pullup resistors are used to get a default value when it is not used so the line does'nt float and are'nt either logical one or zero.

If i look at MC68000 in the schematics for Atari 520stmf i see that the adresslines have pullup resistors to VCC. Does that mean that the adresslines default to 1? If you can put it in that way.

What does this mean on tristate outputs from 68000? If i'm correct those outputs are either logical 1 , 0 or floating and does not have a pullup/down resistor on the line?

And as a last question. When pressing the reset button should all datalines / adresslines go to logical 1 (high) ? If thats not the case. That some are 1 and some goes to 0 can the CPU then be faulty?

/zChris

[stephen_usher]

(Most) TTL chips and memory can "short" lines to bring them down but can't supply voltage to drive them to high so you have relatively high value resistors to supply the +5V voltage to allow the lines to "float" up to a high value. Because the traces on the board have a very low capacitance the time taken to "float" up is very short.

"Tristate" lines mean that the chip can effectively isolate the pins from the rest of the chip so it's neither trying to force them high or low.

The reset switch merely puts the reset lines on the chips to low, what they do then is up to them.

Your first steps in your diagnosis should be (on the CPU):

• Check the voltage.
• Check that there's a clock signal.
• Check to see if the /RESET line goes high after a short interval after switching on.
• Check to see if the /HALT line goes high.

Next check if the output enable on the ROMs is going low. If not the the MMU isn't turning the ROMs on for the CPU to read and the machine won't be able to get any further.

If this is all OK you can look to see if there's activity on the address and data lines. If the CPU is not halted and is working it should be doing something, even if it's crashing.

At this point a Diag ROM may start to be useful with something listening on the serial port, though is low RAM has an issue it may not be able to run.

Anyway, good luck.

[zChris]

Ive measured what you suggested @stephen_usher and here is the results:
Pin 49 VCC 5.2V
Pin 14 VCC 5.2V
Pin 15 CLK 7.813 -8.065 Mhz
Pin 17 HALT Logical Low (240mV)
Pin 18 RESET Logical High (4.08V)

Here is an image on the CLK and it doesnt look good? Never got a stable curve.

clk.png

/zChris

[stephen_usher]

Does /HALT go high at all after reset or power on?

The image of the oscilloscope screen looks as if the 'scope isn't sampling fast enough rather than it being a problem.

Does the ROM enable go low on any of the ROMs after reset?

[zChris]

The oscilioscope im using is DS1052E 2channel 50Mhz 1GSa/s.

/HALT goes high then low on a power on. It react on RESET.

OE (Pin 22 on the roms L0-L2, H0-H2) is always low.

/zChris

[PhilC]

@zChris tey changing your time base to say 40 or 80ns instead of 200. You’ll get a better look at the waveform then. Then also set the grid to say 2V per interval. That should give you a pretty good view of what’s really happening.

Also check the clock signals on the shifter, should get 32mhz from the oscillator and then the shifter will send out 16mhz to the rest of the system.

[stephen_usher]

If the ROM select is never going high this is suspicious as it would have to go low to allow the CPU to access RAM.

[zChris]

Here are the CLOCKs measurements.


EDIT: Made a misstake and measured the wrong Pin on GLUE 8MHZ. The correct pin measures correctly




M68000K
Pin 15 (DEFAULT 8Mhz)

M68000Pin15.png

SHIFTER
Pin 2 (DEFAULT 32Mhz)

ShifterPin2.png

Pin 39 (DEFAULT 16Mhz)

ShifterPin39.png

MMU
Pin 5 (DEFAULT 16Mhz)

MMUPin5.png

Pin 20 (DEFAULT 8Mhz)

MMUPin20.png

Pin 19 (DEFAULT 4Mhz)

MMUPin19.png

GLUE
Pin 15 (DEFAULT 8Mhz) <--- MADE A MISTAKE. THIS IS THE WRONG PING
Pin 34 (DEFAULT 8Mhz) <--- THIS IS THE CORRECT PIN AND IT LOOKS LIKE THE OTHER 8 MHZ


Pin 54 (DEFAULT 2Mhz)

GlUEPin54.png

Pin 43 (DEFAULT 0.5Mhz)

GLUEPin43.png

DMA
Pin 39 (DEFAULT 8Mhz)

DMAPin39.png

WDA1772
Pin 18 (DEFAULT 8Mhz)

WD1772Pin18.png

ACIA
Pin 3 & 4 (DEFAULT 0.5Mhz)

ACIAPin3.png

[PhilC]

I know I always say this but remove the DMA, it’s quite likely then that the CPU will now no longer be halted, ie /halt us 5v. Then test things like the mad lines from the mmu to ram, there should be activity on all of them. Then do the same for the address and data us lines, then whilst holding reset, all address and data lines should pull up to 5v, if they don’t do any of the above then look in that direction.

Lately I’ve had more than my fair share of 74LS244 & 373 fail, so using the data sheets for those chips and scope, check to see if they are operating as they should be.

Typing this as you’ve just posted the latest pics, if pin 15 of the glue should be 8mhz then it looks like either a corroded track or the socket and chip need the pins adjusting.

[stephen_usher]

Before doing any soldering I’d definitely reseat all the socketed chips, especially the PLCC ones such as the MMU and use a good contact cleaner such as Electrolube EML