ZX Spectrum stuck before Sinclair copyright message.

[Muttley Black]

I did that test yesterday with new SN74LS157N. Same results. I order some new lower and upper ram IC's to give a try to. When i got some free time i am gonna desolder ALL sockets to check the traces under them. After that i put new round pin sockets and do the same tests.

I have to say that painful and weird repairs like this one, are extremely interesting!

[Ast A. Moore]

Indeed. Well, keep us posted!

[1024MAK]

When i got some free time i am gonna desolder ALL sockets to check the traces under them. After that i put new round pin sockets and do the same tests.

Before you do that, can you try these tests:

With the upper RAM chips and the associated 74LSxxx chips not fitted, try running with only one of the 4116 (or equivalent) DRAM chips fitted at a time. First with a DRAM chip fitted in position IC6. Then with a DRAM chip fitted in IC7, etc... After fitting each chip, power up, note the colour of the border. Power down and back up, again note the colour of the border. Do this at least four times, before trying with the next DRAM chip.

We are looking for the border to be white EVERY time. As this indicates that the Z80 has managed to execute the the code in the ROM before it gets to the point that working RAM is needed. If at any stage you find the border is not white. Note which socket it is that you are trying DRAM in. Try a different DRAM chip and see if the result is the same.

If a DRAM chip works in every socket, repeat the above process with two DRAM chips, keep one fitted in IC6. Continue until the border is not white...

Mark

[1024MAK]

One other thing. Did you say earlier that you had some or all of the 4116 (or equivalent) DRAM chips in double sockets? That is, DRAM chip in a turned pin socket, which in turn was in the stamped pin socket mounted on the board?

I can’t remember if the /RAS, /CAS and /WE pins on 4116 (or equivalent) DRAM chips float high if not connected, or if instead are high impedance and hence get influenced by any nearby electric field from other circuits. But if there is/was poor connections on any of these pins, it would seriously mess up the operation of the DRAM, including possibly outputting data on the data bus when it should not be. So check the continuity of these chip pins to the source of these control signals, which is the relevant pins of the ULA chip.

Mark

[Muttley Black]

I did the test and i get always white border. Yes i fit round sockets on slide sockets to be sure about the connectivity, cause slide sockets are not in good shape. In some sockets i can see big gap that for sure an ic leg can't touch the connection point. With double socket fitted, i check the connectivity and it is good. All this made for the testing progress. After that, i put the new round pin sockets sodered on the pcb for sure.

[Muttley Black]

Early in the morning i started to remove the old sockets to be able to inspect the traces under them. In the lower ram all traces are ok, but in upper ram i discover 3 traces till now, that have already repaired from someone with thin wire. Connectivity looks ok, but i am not 100% sure if wire was just connect only the broken trace path and didn't short near paths too.

I remove every socket and see what else i am gonna see.

[Muttley Black]

I remove all sockets and i clean the board isopropyl alcohol. I get a fresh inspect on traces on every corner and the results are as you already know the damage in upper ram...



... and pin 10 on IC23. That's for sure not happens in my desoldering progress, but i can't say for sure that the connectivity was bad before.



Then i started to solder the new round pin sockets. That's mean better connection for IC's and fresh solder to every pin.



Before i finish the job, i decide to run the ZX as 16k unit, so i leave the upper ram sockets to wait a bit and did the same for IC'S 23,24,25,26. That also mean that i didn't fix the traces too.



Fresh good clean with isopropyl alcohol and 10 minutes after, i was ready to perform a test as is.

AND........




The same results! I thing i need a cigarette!!!

[Ast A. Moore]

Aside from the passives and physical open circuits/shorts, we’ve pretty much ruled out any other components. You’ve got to find the culprit by reexamining the traces once more, I’m afraid.

[Muttley Black]

I see... The only good news is that i get out of my head that maybe there was a trace problem under some sockets or a connectivity problem. Now with all sockets new i know that i need to look elsewhere... It's something.

[1024MAK]

... and pin 10 on IC23. That's for sure not happens in my desoldering progress, but i can't say for sure that the connectivity was bad before.

This fault alone would have disabled the upper RAM. But is not responsible for messing up operation using the lower RAM.

However, although you have not yet fixed it, you have done good work so far, as you have eliminated a lot of things. Keep at it

Mark

[Muttley Black]

Yes! As already said, issues like this one I enjoyed very much. It is great training according my skills level. I am already a winner and he know it!

[Muttley Black]

Don’t worry about A14 or A15 at the moment.
Mark

I start with ULA, to other components continuity test and I found no continuity from A15 (cpu) to pin 27 (ula). Is that normal? The truth is that none of my spectrums has continuity in this line. Just asking to understand the schematic and the way that this line works.

[Ast A. Moore]

I found no continuity from A15 (cpu) to pin 27 (ula). Is that normal? The truth is that none of my spectrums has continuity in this line.

There should be continuity, definitely, but even if there isn’t, this shouldn’t affect the Spectrum running in 16K mode. Keep checking, though.

[1024MAK]

I start with ULA, to other components continuity test and I found no continuity from A15 (cpu) to pin 27 (ula). Is that normal? The truth is that none of my spectrums has continuity in this line. Just asking to understand the schematic and the way that this line works.

A15 should be connected. A15 is used by the ULA in various internal functions including the address decoding.
Do however watch out for errors in the various schematics for the ZX Spectrum. Apart from the ULA, cross check the function and the pin with a pin-out or datasheet. Also, on some boards, the data lines and address lines to the DRAM chips / multiplexer chips, you may find that some of the PCB tracks/connections have been swapped (compared to the schematic) so that it was easier to route the PCB tracks.

So of the known errors may be listed somewhere in the World of Spectrum forums. But finding them is, err, ‘fun’

Mark

[Muttley Black]

There is an error on issue 3b schematic. ULA pin "27" is for sure the pin "37", like you see it on issue 2 schematic. And i have continuity.

[1024MAK]

I do keep meaning to collate details of the errors on the schematics, but I never manage to find the time...

Mark

[Muttley Black]

• I resolder every joint on board

• All ULA databus to lower ram (pins 2,14) have continuity

• All ULA databus to R1-R8 have continuity. Also R1-R8 are in working order

• All R1-R8 to R9-R16 have continuity. Also R9-R16 are in working order

• Both resistor group "points" with ROM databus have continuity

• All ULA address bus +(dramwe) to Lower ram pins have continuity.

• Lower ram lines IC6 to IC3 have continuity

... nothing new!

666 views on this thread. Is it a coincidence? i don't thing so!



How useful would be if i had an oscilloscope in this specific situation?

if I know how to use it, is another question?

[Ast A. Moore]

How useful would be if i had an oscilloscope in this specific situation?

You could check the power rails and determine if there’s too much ripple on each. You could also examine digital signals and see if there are any irregularities. A data analyzer would be even more useful.

[speccyplus]

Wow this thread sounds like mine, gonna try the RAM test, did the OP ever find the issue?

[Muttley Black]

Not yet. Still faulty in my closet. I give it a try when I have time.

[HappyLittleDiodes]

How useful would be if i had an oscilloscope in this specific situation?

You could check the power rails and determine if there’s too much ripple on each. You could also examine digital signals and see if there are any irregularities. A data analyzer would be even more useful.

This was my thought after reading the thread. Is that voltage regulator doing its job?
Also, have you checked your crysal/clock frequencies?

[1024MAK]

The most useful tools when fault finding a ZX Spectrum are:

• A good quality multimeter
• Another ZX Spectrum of the same type to test any suspect chips
• A expansion interface containing a test ROM (the code in the ROM can run without any working RAM, but can still tell you what RAM faults it finds)
• A logic probe (preferably with a pulse detector)
• A logic analyser (an inexpensive 24 channel channel type is most useful, but you can also do a lot with 16 channel versions, but don’t bother with 8 channel versions)

An oscilloscope is useful if you need to see the actual wave shape (wave forms) of signals, but due to the high cost, it is only worthwhile if you expect to use it a lot. However, it can greatly help with some difficult faults that the above listed test gear may not pick up. An oscilloscope is the most useful tool for dealing with problems with any of the analogue circuitry (such as audio and video circuitry).

In terms of test ROMs, there are various types available, two of the best are this one and this one.

Mark