RJs Adventures in Mega land III

[rubber_jonnie]

I'm lucky enough to have 3 Megas, a Mega 1 that was in great condition and required little in the way of repairs, other than some minor tinkering, such as a Megabus IDE adapter and hard drive.

Then the was the Mega 4. Boy was that an adventure, what with missing chips, frayed wires, bad floppy plugs etc, it was a labour of love to get it up to speed, but it's now my best Mega, with a V1.1c IDE board, CF drive, Emutos 1.0 and TOS on board, it's a great machine.

So now onto number 3, a Mega 1 with a strange 4MB ram expansion. Initial checks are good, in that there's nothing major missing, but running the diag cart threw up some RAM errors in Bank 0 that need addressing.

RAM_Test.jpg (98.64 KiB) Viewed 901 times

Oh dear, accidental bad memory joke there...

Working through the diag cart tests I could do (I don't yet have any loopback cables, but they are on order), the next problem was a non working floppy. I checked the FDD power and that was fine, but the data cable looked a little ropey, so I switched that out for a socket and new cable, but still no joy.

I started out testing the signals from the WD1772 as suggested by the attached service manual, and started off soldering test wires to the FDD to confirm what I saw at the WD1772 was arriving at the FDD itself:

Testing_1.jpg (142.44 KiB) Viewed 901 times

You can see a nice view of the 'scope showing one of the signals, and how the signal is inverted prior to arriving at the FDD. The scaling of channel 1 (Yellow trace) is a bit off, but it does at least show the inversion. I quickly realised I needed a better way than cables soldered to the FDD though, so fitted a 3rd plug to the cable, and used pin headers to break out the signals, as you can see below:

Testing_2.jpg (143.78 KiB) Viewed 901 times

As you can see, my workbench is really not very big, so things can be a bit of a faff, especially with a large machine like a Mega, a scope, solder/rework station and display!

What is handy though, is having a decent workbench PC that I can use to view the schematics and service manual while I work, and search the internet too.

Unfortunately, the FDD still isn't working after working on it most of the day, but I did get this far:

1. The signals that need to get to the FDD do get there, and are all correctly inverted where necessary.
2. The WD1772 does seem to be getting interrupts, though I want to check this again on my other Mega, make sure it looks right.
3. C/S and R/W signals to the WD1772 look fine, and A0/1 seems to be fine too.
4. Oddly, the FDD, which is known working, seems to be seen as an SS drive, not DS in the diag cart. This may be a clue as the same drive shows up as DS in my good Mega.
5. The busy light comes on and the motor starts, but then testing fails.
6. The line DAL0-DAL7 seems to have very low levels compared to a working machine. I think the resistor packs are going to get replaced next, and mandatory fixes done.
7. The write protect signalling works

So that's where I'm at, I'll keep you all posted on progress and (Hopefully) what the final fix is for the FDD and RAM problems.

[sporniket]

Cool pictures, all this work, it's impressive for the newbie that I am.

As you can see, my workbench is really not very big, so things can be a bit of a faff, especially with a large machine like a Mega, a scope, solder/rework station and display!

Still bigger and more permanent than mine. The soldering station looks huge too ! I had looked about it, fancying to get one, but seeing it in situ, I guess it will require a bigger house first.

[rubber_jonnie]

Cool pictures, all this work, it's impressive for the newbie that I am.

As you can see, my workbench is really not very big, so things can be a bit of a faff, especially with a large machine like a Mega, a scope, solder/rework station and display!

Still bigger and more permanent than mine. The soldering station looks huge too ! I had looked about it, fancying to get one, but seeing it in situ, I guess it will require a bigger house first.

As the solder station also does hot air, it's doubly useful, I certainly couldn't have both on there and have space to work. In fact my desolder station is actually behind me to the left as it's too big. I'm hoping to have something a bit more expansive when we move. With the slope of the roof it is more cramped then it looks, hence the mic boom for the microscope, I can swing it in and out as needed.

I do enjoy the troubleshooting stuff, and am getting more patient and procedural about things, fingers crossed I'll get there with this one too

[rubber_jonnie]

Mandatory fix no 1 applied:

Mandatory_1.jpg (99.81 KiB) Viewed 811 times

Leading to a much improved signal on DAL0-7:

Better_Signal.jpg (83.83 KiB) Viewed 811 times

Prior to this I was seeing under 2.5v on these lines, now it looks like a proper TTL signal.

Sadly, it hasn't fixed anything, however I can at least rule that out. Checking on my Mega 4 it looks like that could do with the same fix applying, though the FDD does work on that.

Things seem to be working on the AY, with the necessary pins toggling for drive A/B selection, but I just need to check the 74LS02 at U21, make sure the output is correct and it's going to pin 17 of the 68901 correctly. Bit of an adventure!

[atari030]

Interesting, I'll be keeping an eye on this one. I haven't touched my Ooom kakala Mega till I can square some ST's away.

[rubber_jonnie]

Well, first things first, I wish the symbols for NOR and NAND were not so similar, chasing down the ins and outs of U21 had me confused for a moment until I realised it was actually a NOR, and as a result I swapped out U21. No bad thing, but I can at least say it is good, and I can see the TTL logic happening as it should when the WD1772 sends out an interrupt, and it definitely gets to the 68901.

Then the strange RAM expansion died on me.......

So I removed it leaving only the 1MB onboard and replaced all of the 33R resistors that were removed for the upgrade. Everything jumped back into life and the onboard 1MB tested flawlessly. I have a Marpet style SIMM based RAM expander that was in my 520ST, and recently replaced with one of exxos' low profile units, much better suited to the confined space in the 520, so that will do nicely in this Mega eventually.

So, RAM is good for now.

Back to the FDD then.

All of the signals coming from the WD1772 seem to be good at the FDD and match the descriptions of operation in the datasheet. The drive light comes on, the head moves, track 0 is detected and the drive spins.

Data line DAL0-7 have pulses on them, and with the 10k pullups, the logic levels are fine. But still it doesn't work. I think I'll finish off the service manual troubleshooting, just to be 100% sure, and then I think I need to break out the logic analyser and compare whats happening on the data bus with my working machine.

Ho hum!

[exxos]

You may need to do the bus pull ups and others as well which can cause all sorts of odd things. I don't get why your having problems with the ram as it tested fine with the full 4MB before shipping.

[rubber_jonnie]

You may need to do the bus pull ups and others as well which can cause all sorts of odd things. I don't get why your having problems with the ram as it tested fine with the full 4MB before shipping.

@exxos I know, the RAM thing is odd as I know you tested it, unless my replacement chips were fake, which is a possibility. The board was been handled a lot over the weekend, which may also have had an effect given the fragility of this stuff these days, so I will probably revisit that later, I'm not decided yet, but for now I need it to work normally for the time being and 1MB will do whilst I'm working on the FDD problems. I'll get it back to 4MB by hook or by crook!

I have a ton of bus pull up SIPs as I want to do the whole fleet , so they are on the cards next, just didn't get to it over the weekend, it's mad how much time gets absorbed when you're peering into a scope.

It was a revelation though just what a difference the 10K pullup made on DAL0-7, absolutely night and day!!

[rubber_jonnie]

Ok, so after a LOT of checking and re-checking of things, not to mention soldering pin headers onto my FDD to break out the signals, it looks like they are fine at the drive.

I kind of went over some old ground again, just with a bit more diligence, and what I see at the floppy drive matches the working one, but the data lines seem weird, so I checked those whilst running the disk exerciser from the diag cart.

They just don't seem quite right, so I did a comparison of the DMA CD0-CD7 lines with my working Mega, and I think the DMA may well be bad. The output is kind of weird in comparison.

Having changed the pull-ups and added the FDD data line pull up resistor pack, I know they are good, so next weekend, I think the DMA is getting pulled and swapped for a working one. It's pretty much all that's left, signals are all getting to the floppy, and everything that needs to be inverted along the way is. Interrupts seem good, as are the signals from the YM.

Fingers crossed that does the trick!

[rubber_jonnie]

Well finally after getting well and truly distracted with work and house moving stuff, I finally got some time with the Mega again, and it is definitely the DMA that is bad.

I removed it, installed a socket and dropped in a known good one, and well what do you know, it's working just fine and dandy now, with the diag cart floppy disk tests running fine, and also running just fine under GEM. I do need to test with my UltraSatan, but I'm confident it'll be fine.

It's been a good week so far, what with getting the ZX81 running again

Next up, I need to revisit the funky RAM board again and see if I can get it working as expected. My confidence is a bit low on this one, as I think the replacement RAM I got for it is possibly faulty or fake. Funny, how the ZX81 also needs the RAM pack sorting!

At least if I can't get it going again, I do have a Marpet SIMM adaptor that would fit nicely and take it to 4MB.

After that, I need to work out the screws I need to put the whole thing back together. I think I'll drop a Megabus IDE adapter in there for Kicks too.

Fingers crossed this Mega won't take me down the same long winding road as my Mega 4 did!!