UK Vintage Radio Repair and Restoration Discussion Forum

UK Vintage Radio Repair and Restoration Discussion Forum (https://www.vintage-radio.net/forum/index.php)
-   Vintage Computers (https://www.vintage-radio.net/forum/forumdisplay.php?f=16)
-   -   Sharp MZ-80K help needed on repair (https://www.vintage-radio.net/forum/showthread.php?t=183455)

pudwink 3rd Oct 2021 6:48 pm

Re: Sharp MZ-80K help needed on repair
 
Sorry ment to say reprogrammed both EPROMs again.

SiriusHardware 3rd Oct 2021 7:39 pm

Re: Sharp MZ-80K help needed on repair
 
Pudwink, thanks for trying it with at least one more EPROM. I am assuming that when you programme these devices your programmer reads back the code just programmed to see if it matches the code you meant to program into the device, usually called 'Verification' in the programmer's software.

The fact that the RAM test code starts up every now and again is screaming 'Half dead buffer IC' to me, but let's see what everyone else says. Unless you are prepared to sit with the diagram and meter out every circuit node on the circuit diagram (costs nothing, except your time and possibly your sanity) then we need some other way to easily detect a problem on the system buses.

I get the impression that you are quite capable and confident about removing and replacing ICs at least by the 'cut pins and remove separately' method, so I think you would be unlucky to have caused any damage while replacing the IC sockets and ICs that you have already replaced.

My suggestion is: Remove, socket, fit new replacement databus buffers IC36 and IC37 (74LS241) only for now. If that does not conveniently fix the problem, having those buffers in sockets will make it easier to rig a NOP test, because we can then disconnect the Z80 data bus from the system data bus by removing IC36 and IC37.

To run a NOP test we could fit 2 * ordinary ICs sockets into the IC36 / IC37 sockets each with 4 * low value resistors fitted from the Z80 data bus pins to 0V. This would be easier than trying to make a NOP rig to go into the CPU socket with the CPU plugged in on top of it.

Any thoughts on this?

Mark1960 3rd Oct 2021 8:08 pm

Re: Sharp MZ-80K help needed on repair
 
Try pressing down on each socketed chip first, maybe also try gently pushing sideways on each socketed chip, while operating the reset switch. See if any of the chips are sensitive to pressure. This might show if the intermittent power up is due to a bad connection in one of the sockets, include any sockets that you replaced just in case.

pudwink 4th Oct 2021 6:03 am

Re: Sharp MZ-80K help needed on repair
 
Hi
I will order some 74LS241if that fails then we can try the test because I really do want to try and get this going.

SiriusHardware 4th Oct 2021 8:21 am

Re: Sharp MZ-80K help needed on repair
 
Please try Mark's 'crush test' first though, as I said earlier, always try the things which are simple and don't cost anything to try first.

pudwink 4th Oct 2021 2:50 pm

Re: Sharp MZ-80K help needed on repair
 
HI
8255 turned up today pugged it in and still 1v on pin 14 so as previously suggested I snipped pin 1 on IC 24 and now 0v on 8255 pin 14.

SiriusHardware 4th Oct 2021 4:10 pm

Re: Sharp MZ-80K help needed on repair
 
Not to worry - I think you felt you needed to replace the 8255 on the grounds of its physical condition alone in any case.

Can you either repair IC24 pin 1 or replace IC24 and leave the new 8255 in place, then, as Mark suggested, put pressure on each socketed IC in turn while turning on the machine to see if putting pressure on any of them makes it start up reliably every time.

If no different, unless Mark has any other suggestions can you remove, socket and replace IC36 / IC37 in sockets and we'll proceed from there. It might just be that those buffers are the actual cause of your problem and replacing them will fix it, but if not, let's explain what we will try to do next.

The idea is that instead of the Z80 running code from the MROM or RAM test ROM we feed it directly with the opcode for a NOP instruction, hence the name 'NOP test'. A NOP instruction is an instruction which does nothing, other than to waste a very small amount of time. The effect of the NOP test is to make the CPU think the whole memory range is full of NOP instructions, so it runs smoothly through the whole address range over and over again executing NOP instructions as it goes. As it's doing this, it generates nice steady waveforms on all of the address lines and exercises the chip select circuits and a few other sections besides, so if you then look through all of those lines even on an old analogue scope like yours, you can quite quickly spot any problems in those areas.

Mark1960 4th Oct 2021 4:18 pm

Re: Sharp MZ-80K help needed on repair
 
Try a 4k7 pull up resistor on the output of the 8255 as it could be programmed to operate as an open collector output.

It might also be an issue with control inputs to the 8255 not allowing the z80 to write to the control registers of the 8255.

SiriusHardware 4th Oct 2021 6:51 pm

Re: Sharp MZ-80K help needed on repair
 
On a quick look through an 8255 data sheet I don't see any modes other than input and output - no open collector mode - for pin 14 which is port C, bit 0. (PC0). That pin can also be configured as an interrupt input but the MZ80K diagram clearly indicates that it is being used as an output and is connected to an input on IC21.

All three ports are initialised as general inputs on reset. In the datasheet for the 'original' 8255 as used here I see no mention of internal pullup or pull down resistors, the 82C55A appears to add that feature to the feature set, but that is not the version which is under consideration here.

SiriusHardware 4th Oct 2021 7:05 pm

Re: Sharp MZ-80K help needed on repair
 
pudwink, things don't quite add up with regard to the voltages you are seeing on the 8255 V-GATE pin so could you check what voltage you have on the +5V supply pin of the 8255 when the machine is switched on? That would be.. <rustle rustle> pin 26 of the 8255. Also check, with power off and your meter on ohms, that pin 7 of the 8255 is directly connected to system 0V / GND.

pudwink 4th Oct 2021 7:27 pm

Re: Sharp MZ-80K help needed on repair
 
0v / GND is fine and 4.94v on the +5v

SiriusHardware 4th Oct 2021 7:40 pm

Re: Sharp MZ-80K help needed on repair
 
Thanks for making that measurement. As Mark suggested, it sounds possible that the 8255 is not being initialised and its port pins including pin 14 (PC0) are staying in the default input mode. That could explain why you only have whatever is on the 74LS00 input on the 8255 pin 14 when the 8255 is connected to the 74LS00 input, and nothing at all on the 8255 pin 14 when it is not.

Any further non-invasive suggestions, or do we go ahead and replace the data bus buffers IC36, IC37 - partly because they could potentially be the culprit, and partly because even if they are not, putting them in sockets will make running a NOP test easier?

Mark1960 4th Oct 2021 7:58 pm

Re: Sharp MZ-80K help needed on repair
 
I was just thinking that the ram checker might have been written to switch that pin between input and active low, rather than a problem with the MROM driving the 8255. This would simulate an open collector output. This would work with most 74LS inputs as they tend to float high, but might not always reach logic high level on every 74LS. I don’t know if age could have an effect on 74LS inputs floating high.

SiriusHardware 4th Oct 2021 8:09 pm

Re: Sharp MZ-80K help needed on repair
 
OK - possibly the way to find that out is to look through the RAM test code, the initialisation of the 8255 will come fairly early on. If PC0 is initialised as a input during the initial setup that will add some weight to your theory, but to be honest why go to all that trouble - why not just use PC0 in conventional digital output mode when it is only driving one TTL input?

pudwink - I asked you some time ago but maybe you did not see it - do you have a link to where the RAM TEST code came from? If it's somewhere on the net please post a link to its location or alternatively attach the code to a post here - you will have to put it into a .zip file in order to attach it because .bin, .hex are not among the permitted file types. If you have been asked not to post it publicly (perhaps by the person who wrote it?) then could you let me know where it came from via a PM?

In order to know what the RAM TEST code does, especially with respect to the way it uses the 8255, we may need to look through the source code (if available) or disassemble the raw code - (turn it back into readable program code) so we can understand what it tries to do.

In the meantime Mark - can we go ahead and remove the data bus buffers, fit sockets and fit new buffers?

Timbucus 4th Oct 2021 8:29 pm

Re: Sharp MZ-80K help needed on repair
 
I vote for the buffers, in my current H8 restoration that is usually what has got a card running once the shorted tantalums are swapped...

Mark1960 4th Oct 2021 8:30 pm

Re: Sharp MZ-80K help needed on repair
 
I wonder if one of the 8255 control signals is disconnected, resulting in write to data direction instead of data.

I’m not sure about replacing more ICs unless we can see why the ram checker started to give bad results. I still regret advising change of the video data buffer when it was the z80 reset writing over ram due to reading FF.

SiriusHardware 4th Oct 2021 8:53 pm

Re: Sharp MZ-80K help needed on repair
 
My angle on this is that pud's scope is better at seeing repetitive signals than it is at seeing one-shot or infrequent signals so the sooner we can get to running something like a NOP test the better the information we will have. It could be done by using a couple of stacked sockets plugged into the CPU socket, yes, but doing it that way is a lot more fiddly and delicate. Even a NOP test will not really show us how well the databus buffers are working, especially in the 'to the system' direction. Substitution (with known good devices) is really the easiest way to be sure about that.

The idea to replace (one of) the video data buffers was a reasonable call at the time. I want to replace the main data bus buffers because it serves two possible purposes, one, the buffers themselves may be the cause of the problems including the apparent inability of the system to initialise the 8255, two, putting them in sockets will make it easier to run a NOP test which I think we are going to have to do to make progress if replacing the data buffers does not fix it.

Buzby123 4th Oct 2021 9:20 pm

Re: Sharp MZ-80K help needed on repair
 
2 Attachment(s)
Here is my Z80 NOPulator.

I made this way back in the day, and have not used it since then.

It may or may not still work, so I'd prefer to send it to Timbucus or Sirius or Slothie first, who can then test it, and then forward it to Pudwink as a known good device.

One switch forces 'all READ nops' or 'all WRITE nops'.

The other switch forces scanning of 'Mem space' or 'IO space'. ( IIRC )

If you think this may help I'd be pleased to lend it out. I've not needed it for 30 years, so I won't need it back soon. ( He says, but you never know. )

Cheers,

Buzby

SiriusHardware 4th Oct 2021 9:28 pm

Re: Sharp MZ-80K help needed on repair
 
Well, you can see why I thought asking pudwink to build something like that to go into the CPU socket might be a bit ambitious at this stage. ;)

My current thinking is just to have pud fit 2 * empty sockets into the IC36 and IC37 sockets, (once there are sockets fitted), and wire / solder four resistors from four pins on each of the plugged in sockets to 0V. Physically speaking, not too difficult to do, although a quick pencil sketch would probably explain it better.

pudwink 5th Oct 2021 2:43 pm

Re: Sharp MZ-80K help needed on repair
 
1 Attachment(s)
Hi
Here is the memory test program I have been using just waiting on the replacement chips coming then I will have the sockets installed to try a NOP test(unless the chips solve things) and yes a sketch and simple instructions would be most welcome.


All times are GMT +1. The time now is 3:27 pm.

Powered by vBulletin®
Copyright ©2000 - 2024, vBulletin Solutions, Inc.
Copyright ©2002 - 2023, Paul Stenning.