Clock reading not as expected

[satjanand]

Hi,

Trying to fix a 520 ST with a piggybacked RAM upgrade, I have started to get some experience in using the oscilloscope I recently bought. I have followed the Atari Field Service Manual and Sams Computerfacts Technical Service Data as well as I could.

The machine powers on, white screen is visible, but nothing more happens. I have switched out TOS ROMS, GLUE, MMU, Shifter, but no avail.

Following advice in the manuals, I have measured:

CPU P15 = 8 MHz
MMU P5 = 2 MHz (Edit: this pin was wrongfully seen as pin 5. It was actually pin 8. MMU Pin 5 turns out to be 16 MHz after all)
Shifter P2 = 32 MHz
Shifter P39 = 16 MHz
GLUE P54 = 2 MHz
CPU P17 (HALT) = 0.5 V always
CPU P22 (BERR) = 0,5 V but shows pic 1 on the oscilloscope. Is this the logic low pulse which is mentioned in the Atari FSM?

I read in Sams TSD that GLUE P54 MMU P5 should be 16 MHz instead of 2 MHz.

Booting to white screen is almost there, right? How can I find the problem from here (without first having to uninstall the RAM upgrade)?


Pic 1:

Knipsel.JPG

(Edit 21 Feb 2023: clocks turned out to be fine, so I changed the title of the thread.)
(Edit 2 21 Feb 2023: changed the title back, found an error in my text. Not GLUE P54 but MMU P5 was reading 2 MHz while expecting 16 KHz)

[rubber_jonnie]

As a general rule of thumb, a white border and black screen is a RAM problem, and a fully white screen is usually ROMs. As for P54 of the GLUE, my Atari ST internals book says it should be 2Mhz not 16Mhz.

The rest of the clocks seem to be fine.

Personally I would leave the RAM alone for now, the lack of a black screen or on screen corruption is suggestive it's ok.

I'd personally look closely at the PLCC sockets chips, so pull them out (I'd suggest a proper PLCC chip puller to avoid chip/socket damage), clean the chips and sockets with IPA and adjust the contacts of the sockets: PLCC Contact adjustment.

I also tend to use some contact cleaner on the sockets, such as DeOxit, Electrolube or Servisol prior to re-inserting the ICs. You can use contact cleaner on the DIP sockets if there are any fitted too, such as the ROMs.

You might also want to take a look at the chip select lines on the ROMS to be sure they are being enabled, usually pin 20 of the ROMs.

Not sure if you have one, but in instances like these the diagnostic cartridge can offer some insight via a terminal on the serial port.

If you have any PSU doubts, all STs will boot with just 5v from a bench PSU.

[olivier.jan]

Hi,
Did you check the RESET circuit which is prone to failure ? An Atari ST stuck in RESET or not resetting properly would produce such a white screen.
About the 0.5V you get on BERR and HALT, I can see on your screenshot that your Channel 2 probe is configured as x1 while Channel 1 is x10. Is the probe itself physically configured as x1? If not, then you have a nice 5v…

If reset is Ok, you can start probing Address and data lines for activity but I agree with @rubber_jonnie , cleaning all sockets, and reseating every chip are first things to do.

Probing with the scope is going to be tedious and if you can get a diagnostic cart it might save you hours of troubleshooting.

[frank.lukas]

... once had the case the "reset" was high and "halt" low. The reason was a short between vcc and on data signal ...

[satjanand]

Thank you all very much for responding.

As a general rule of thumb, a white border and black screen is a RAM problem, and a fully white screen is usually ROMs. As for P54 of the GLUE, my Atari ST internals book says it should be 2Mhz not 16Mhz.

The rest of the clocks seem to be fine.

Personally I would leave the RAM alone for now, the lack of a black screen or on screen corruption is suggestive it's ok.

Sorry, I seem to have made a mistake: not GLUE P54 but MMU P5 reads 2 MHz while 16 MHz is expected.

I'd personally look closely at the PLCC sockets chips, so pull them out (I'd suggest a proper PLCC chip puller to avoid chip/socket damage), clean the chips and sockets with IPA and adjust the contacts of the sockets: PLCC Contact adjustment.

I also tend to use some contact cleaner on the sockets, such as DeOxit, Electrolube or Servisol prior to re-inserting the ICs. You can use contact cleaner on the DIP sockets if there are any fitted too, such as the ROMs.

You might also want to take a look at the chip select lines on the ROMS to be sure they are being enabled, usually pin 20 of the ROMs.

I'l get into cleaning and adjusting the PLCC and DIP contacts
And see what I can find regarding the ROM chip select lines.

Not sure if you have one, but in instances like these the diagnostic cartridge can offer some insight via a terminal on the serial port.

If you have any PSU doubts, all STs will boot with just 5v from a bench PSU.

I do have Exxos' diag cart, but shows next to no activity on the terminal: every so many restarts I discover a few random single characters, but no more than that.

I am using a brand new MeanWell RD-50A PSU.

Hi,
Did you check the RESET circuit which is prone to failure ? An Atari ST stuck in RESET or not resetting properly would produce such a white screen.
About the 0.5V you get on BERR and HALT, I can see on your screenshot that your Channel 2 probe is configured as x1 while Channel 1 is x10. Is the probe itself physically configured as x1? If not, then you have a nice 5v…

If reset is Ok, you can start probing Address and data lines for activity but I agree with @rubber_jonnie , cleaning all sockets, and reseating every chip are first things to do.

Probing with the scope is going to be tedious and if you can get a diagnostic cart it might save you hours of troubleshooting.

Reset circuit was tested and came out fine. I'll get to cleaning and adjusting all sockets and socketed IC's next.

About the probe: yes, you were correct! Shows what I know, right :D
Thank you for this!

But what about the clear dip down in the BERR signal? Is that what the Atari FSM means by logic low pulses?

... once had the case the "reset" was high and "halt" low. The reason was a short between vcc and on data signal ...

How could I test for that?

[olivier.jan]

2Mhz on pin 5 of the MMU is strange. Did you do continuity test with pin 39 of the shifter ?
Could you scope both at the same time and post the results?
Also, pictures of the motherboard, revision, etc would be helpful.

[rubber_jonnie]

Yes, P5 on the MMU should definitely be 16Mhz.

Since you're getting 32Mhz at pin 2 of the Shifter and 16Mhz at pin 39 of the shifter, and pin 5 of the MMU says 2Mhz, then you're going to be looking at something in-between the two, at least that is where I would look.

[satjanand]

2Mhz on pin 5 of the MMU is strange. Did you do continuity test with pin 39 of the shifter ?
Could you scope both at the same time and post the results?

No direct continuity between Shifter P34 and MMU P5, but I did measure resistance between MMU P5 and Shifter P39 (120+ Mega Ohms). I couldn't measure direct continuity in a working machine either. If I can find out how Shifter P39 and MMU P5 are connected together, I can check point-to-point.

Pic of 2 channel scope (Probe 1 = MMU P5, probe 2 = Shifter P39):

knipsel 2.JPG

Yes, P5 on the MMU should definitely be 16Mhz.

Since you're getting 32Mhz at pin 2 of the Shifter and 16Mhz at pin 39 of the shifter, and pin 5 of the MMU says 2Mhz, then you're going to be looking at something in-between the two, at least that is where I would look.

I found this capacitor inductor connected to P39 on the bottom side of the motherboard. The close side (to P39) showed 16 MHz, the far side nothing. But not sure if that's relevant.

How do I trace the complete path from Shifter P39 to MMU P5?

IMG_20230222_105928.jpg

Also, pictures of the motherboard, revision, etc would be helpful.

Here are the pictures of my motherboard:

IMG_20230222_110027.jpg

IMG_20230222_110329_edit_21378883639445.jpg

IMG_20230222_110337_edit_21369675907675.jpg

IMG_20230222_105956_edit_21226841620197.jpg

IMG_20230222_110251_edit_21503555767030.jpg

IMG_20230222_110258_edit_21495694629531.jpg

[rubber_jonnie]

That's an inductor not a capacitor, though I'm not sure why it is there, but looks like probably a factory bodge. It looks normal, see the picture of another board here: ST Mainboard

At least 2 mainboards in this section of the forum have the same bodge as yours. Later revisions may not have it, or have it integrated into the PCB.

You can do continuity tests without worrying about damage. You need to go point to point and see if there is anything odd going on.

[satjanand]

That's an inductor not a capacitor, though I'm not sure why it is there, but looks like probably a factory bodge. It looks normal, see the picture of another board here: ST Mainboard

At least 2 mainboards in this section of the forum have the same bodge as yours. Later revisions may not have it, or have it integrated into the PCB.

Thanks, I've corrected my post.

You can do continuity tests without worrying about damage. You need to go point to point and see if there is anything odd going on.

When I look at the schematics, I see P39 go into an arrow shape saying 16MHz, but no traces after that. I see the same kind of arrow shape saying 16 MHz near MMU P5, connecting to it. But I can't see how it routes in between, what components it passes. How do I find that?

I could measure no direct continuity between MMU P5 and Shifter P39, but neither could I in a working ST.