RAM/MUX problem

[Chris_68]

Hi and Happy New Year!

An update, I used a DS1233 to get better reset and checked "all" solder joints so now I get a random pattern and it passes Test 2 using Mikes Diag rom ((A15-A8 high, A7-A2 will be 000010, A1-A0 pulsing). Will get LOCI when they are available, excellent work with .

When switching and using the FF diagram rom (French) I don't get an audio output reporting which RAM it is. D0-D7+BC1+BDIR is high. I'm I right that Mikes ROM indicates 6502 and 6522 seems ok so should be 8912?

Would you try to fix audio output to get an indication of which RAM IC it is or start replacing RAMs?
Not sure I would trust me building DRAMArduino on breadboard, anyone having PCB or tester up for sale?

[iss]

Will get LOCI when they are available...
Not sure I would trust me building DRAMArduino on breadboard...

Our LOCI's are back in stock tomorrow (02.01.2025).
Build DRAMARDUINO on breadboard - it's easy.

[Chris_68]

Hi great news that it will be available tomorrow !

I will give it a go building the tester using a nano.

/C

[Chris_68]

Hi

An update and request for advice,
I tested with Dramarduino (thanks for sharing the design) and replaced two RAMs, it passed Mikes diagnostic ROM but after a few tests it got more and more unstable and I got garbled screen. With a slight pressure in the corners it was stable for an hour but now it impossible to boot.

Proof that it actually did work :-)

The current status is as below, it's not static and pattern usually changes constantly. Mikes diag test gives that it passes RAM test 1 but not the second RAM test.

Everything indicates a bad connection pad/via/track, likely around 6522 but reflowed all joints and haven't been able to find any open circuit. Is there any part of the PCB prone to fail?

I'm I right that ULA and CPU seems ok but the problem is either intermittent 6522 or bad connection in that area?
Will do another continuity test but running out of ideas, any ideas what to test next?

Thank you for all help so far.
/Chris

[iss]

What's yous power supply (picture,pls) ?
You can try to randomly swap RAM chips. For me it's a rule to use always 8 DRAM's from same brand.
As I wrote in DRAMARDUINO Oric is very sensitive to DRAMs.

[Chris_68]

Hi,

Found a bad track on PCB, repaired and get this, so in the right direction:

After boot

Seconds after reset and then goes back to previous pattern

All RAM is same and verified ok, NEC4164-12. MUX and inverter are ok.

C21 is 10uF (removed the DS1233 for LOCI that arrived today), I have measured most pins related to RAM without finding anything. If I change the ROM slightly it changes behaviour so will replace ROM socket and do a continuity test between CPU/VIA/ULA.

Does the screenshot point in any specific direction?

Chris

[iss]

Hm, you are using 74HCT04 and 74HCT257 - can you try with LS TTL chips.
If you have a good decent 9V 2A PSU use it.

[Chris_68]

Hi,

Thanks,

I have tried with two different PSU 9V, don't know what current they actually deliver though.
I know I have 74LS04, the supplier didn't have 74LS257, I can try to order but it worked earlier with 74HCT257. BTW, well spotted identifying them

Nothing indicating bad VIA/CPU?

Thank you for all help
Chris

[iss]

Nothing indicating bad VIA/CPU?

CPU and VIA should be OK.
BUT! On first picture you have #30 (i.e. '0') instead of #20 (space), #53 'S' instead of #43 'C' ...
This indicates something wrong with bit 4 - IC15 (or 4th DRAM from left just above C16).
Is it good inserted in socket, continuity to CPU....

[Chris_68]

Hi

Thanks a lot, you pointed me in the right direction.
I have continuity on IC15 to CPU, measured on pin instead socket. Can't find anything wrong there but by replacing the NEC D4164-12 with same model I get other pattern so definitely RAM related. When I put back same IC I get same pattern as before so don't suspect socket but something not compatible. All RAM tests ok.

The screenshots when I replace with a mixture of 150ns 4164, as seen you were right, the CAPS is almost visible.

Apart from doing another continuity test and maybe new socket for ROM, any idea what to replace?
74HCT257 should be TTL or shall I try to find LS?
Can I suspect NEC 4164-12 to be the reason and worth ordering M5K4164AP-15 or similar?

/Chris

[Chris_68]

Hi

Finally some good news, beyond grateful to guidance in this forum.

I haven't been able to use the LOCI yet (it doesnt identify SD, most likely my media adapter) but just the reset function is invaluable!!

If someone else stumbles on this topic I did the following:
Reflowed all solder joints related to logics/CPU/RAM/ULA/VIA ( I had intermittently bad connection)
Continuity test for the above
Replaced MUX and inverter with HCT (MUX was bad) socketed
Had some bad RAM I think 4 to begin with, replaced worked a few hours and died again.
Used 8 identical NEC 4164-12 and got the screen posted earlier with all zeros.
As advised I moved around the RAM without success
replaced new RAM with a mixture of 150ns versions, think three different brands and voila
I would prefer to have 8 identical and new RAM but I won't touch it ATM.
I had to sort the spaghetti soldering for the keyboard connector posted earlier

As said by so many, start with inverters and MUX, if they are ok use Mikes diag ROM, and if you suspect RAM I recommend to start using the ROM FF posted in WIKI. It will tell you which RAM is bad (RAM zero is IC19...), a warning though, my RAM was ok and I could still not boot until I moved the RAM around.

[iss]

Congrats, @Chris_68! Well done.
This problem with DRAM is known but it's not easy to be explained.
Definitively I had 4-5 cases when random RAM swapping helped.
I see LOCI is 0.3.0-0.3.0 already - it should work too, so enjoy your Oric!

[Chris_68]

Thank you!!

LOCI is just great, found a small USB-C to A adapter with thumb drive that is identified by LOCI, so will get the hang of it soon

/Chris

[Chema]

I know the Oric has tight timings to accommodate three memory accesses in every cycle (2 ULA and 1 CPU), thus the end of 120ns RAM and versions of 6502 capable of running a 2.0 MHz. It is possible to use slower RAMs (I've seen that) but it might be out of spec.

Could it be that certain CPU & Memory combinations work because they fit in the timings just by chance? In my Oric clone I had stability issues that were solved by changing the CPU (and, by the way, it is using 150ns RAM chips).

[Chris_68]

Yes must be something with timing and the order of RAM.

It took many trial and errors to find this combination of ram and in the correct order. Actually quite weird, usually they recommend same brand and timing and 120ns should be fine.

Well, the problem now is if it dies again….

Once again thank you all !