Uh Oh.Electron inbound

[rubber_jonnie]

Ok. Not the result I was hoping for, having tried all 4 of the new RAM chips, I still get the same on screen pattern. Large white bars across the screen with thin black bars, as per my previous post.

I also tried piggybacking the new RAM on the old, but no change. The clocks are all looking fine according to my frequency counter.

Things are different though, in that I do now see activity on the data out lines of the replaced RAM chip, which I didn't have before, so I'll take that one.

I did wonder at one point if the speaker was working, so I decided to test it, and was able to use my function generator to create a suitable sine wave, verify it on my Fisher Price scope (I think it was something like 2Khz) and when I clipped onto the speaker I got a nice sine wave out. I was pretty pleased I was able to do that, it turns out the function generator is a pretty useful tool

So, I think that the video generation circuits are fine, because the image on screen is rock solid, and hsync/vsync read fine too. If they didn't, I wouldn't get a picture.

CAS and RAS look ok, as best I can tell, so given that so much actually works, I kind of think that the ULA might be ok. Also, given what's on screen I am thinking code is being executed, but it's crap, so I think I might look at the ROM next, but the trouble there is it's soldered in, and I'd rather wait to get my desolder gun on it than risk damaging the mainboard.

It was easy with the Beeb as I was replacing sockets, so removing them destructively wasn't an issue as I could heat the pins to remove them afterwards. I can't do that here as I'd need to snip the pins, destroying the ROM.

I also don't have a spare ROM I can use either or the means to burn one. Given I will be in my new home very soon and will have everything I need, I need to think about waiting to avoid damaging the mobo through impatience.

[stephen_usher]

Got a spare 6502 to try piggybacking? I know doing it with CPUs usually doesn't work and you shouldn't do it for long but it's a non-destructive test,

[rubber_jonnie]

They just arrived, but the Elk is still dead. Worried it's the ULA now, but may have to pause now we have a completion date o the house. Have to think about it.

Oh. So no togglng of the /WR line or anything?

Yes, quite a lot seems ok. /WRITE is toggling at the RAM I see address & data line activity, Din and Dout show activity plus RAS/CAS activity and the sync line, CPU pin 7 has activity too. CS at the ROM seems permanently high though, and I'm sure that should pulse too.

What isn't helping is the difficulty numbering the ULA pins for testing, I've not yet found the glop top ULA numbering yet, and whilst I think I found pin 49 (CLK in), as it had 16Mhz on it, counting every time to find pins is a chore, so I need to make up my own diagram I think.

[rubber_jonnie]

Got a spare 6502 to try piggybacking? I know doing it with CPUs usually doesn't work and you shouldn't do it for long but it's a non-destructive test,

I do have a spare CPU, I could borrow one from my recently repaired Beeb, but I've had a think about it and have decided to set it to one side until we're moved in. Whilst it grates to have to do so, I'm concerned about damaging the mainboard desoldering the ROM, as for the RAM I just snipped the legs off so it was easy and I didn't mind binning a RAM chip. I just don't want to do that with the ROM.

Of course by this time Saturday, I may have changed my mind and decided to try desoldering the ROM!

[stephen_usher]

[rubber_jonnie]

Funnily enough, Hoglet over at *. suggested the CPU and mentioned that the black and white screen is what you see if you remove the CPU from a working Elk.

Looks like I need to look at that then!

[rubber_jonnie]

Well my thoughts of parking it didn't last long, I've just had a long chat with the guys over at *. and we have all sorts going on, since the ROM /CS seems to be stuck high, and the pattern on screen is supposedly the same as what happens when there is no CPU.

Off for another round of buzzing out and will be ordering a replacement CPU, more to follow!

The thread over at *. is: https://stardot.org.uk/forums/viewtopic.php?f=3&t=22837

[stephen_usher]

I've been reading that post. Let's hope it isn't a duff ULA!

[rubber_jonnie]

I've been reading that post. Let's hope it isn't a duff ULA!

Me too, I'm quite worried that pin 61 of the ULA is high, essentially from power on preventing ROM access, so next thing will be to check for shorts between the ULA and ROM, but there isn't much between on the issue 6, so kinda worried now.

I think somebody has made replacement ULAs, but they aren't cheap.

It has a rather weird layout, soI did a diagram to help other folks:

ULA_Numbering.jpg (71.36 KiB) Viewed 3484 times

[stephen_usher]

I s'pose if the only issue is that one pin on the ULA and it's controlled by two inputs to the ULA that you could possibly put a NAND gate on A14 and A15 with the output on the ROM output enable to replicate the functionality. Just a thought.