Got me an Atom

[SiriusHardware]

You seem to have clocks all right, and if pin 0 (A0) and pin 34 (R/W) are active then the machine isn't as dead as we had the impression it was at first.

So now could you follow up on KeithsTV's suggestion and look on the enable pin on the PROM - pin 20. In normal circumstances we'd expect to see this mainly high, with regular low-going pulses.

By the way, did you look through the PROM as you planned to do? Look for a few plain text words. Words like LOADING or ACORN are likely candidates. If you can see some correctly spelled words in the PROM then it is fairly likely to be OK.

If you can't see ANY plain text words or the PROM reads as all 00, all FF or all some other value, then you probably have a duff PROM. It certainly is possible that a few bits have been corrupted: The easiest way to check that is to get on OS file image from somewhere and programme it into an eprom.

However, do the pin 20 check first. If the machine isn't even accessing the PROM then you'll need to fix that first.

[jay_oldstuff]

This is getting odd, I got the logic probe out and probed around ic20 on switch on pin 20 was low almost all other pins where high as soon as I pressed the brake key all the data line became active including pin 20. it looks like the hart of the computer is running its just not talking to the video decoder ic.
unfortunately I failed in my attempt to read ic20 my reader just doesn’t seem to support the device. its a MM512164GMF/N can't find much data on it on the web.

Jay

[wave solder]

A MM52164 is a 8192x8 bit ROM.

[SiriusHardware]

What did you try to read it as? From the data Wave Solder provided it looks as though you could read it as a generic 2764 eprom.

When you hold the machine in reset you will see most of the lines switch to steady states, mainly high, except for the reset input which should be steady low.

You say you have a steady low on the PROM pin 20 during normal running mode - you're sure about that?

[jay_oldstuff]

Ok thanks for that Info I may stand a bit more chance trying to read it now I have an idea what it is. My reader does have an auto detect function which is usually quite good but in this case just came up with defective or unknown device.

No pin 20 was lo on power up pressing the brake key and it shows date activity rapid hi lo switching. Press the brake key and it goes lo again until the key is released.
Jay

[wave solder]

The ROM does not have the same pinout as a 2764. it is the same as a 2564 and 2364 according to my datasheets.

[jay_oldstuff]

I’ve managed to read ic20, it read as a generic 2732 and I was able so save a copy to my computer as a .bin file. it looks fine on screen much as I would expect it to look. what I could possibly try is getting an Atom emulator to run with the ROM image. I kind of suspect the ROM is not the issue though.

Jay

[jay_oldstuff]

I've just saved the image as both .bin and .hex files zipped them and uploaded them below. see what you guys think.

Jay

[wave solder]

Reading the IC as a 2732 only accesses half of the data. The PROM is a 8192x8 device not 4096x8 device.

[SiriusHardware]

Originally Posted by wave solder

The ROM does not have the same pinout as a 2764. it is the same as a 2564 and 2364 according to my datasheets.

Duh, sorry, what was I thinking?

For a start the 2764 is a 28-pin device whereas this device is a 24-pin chip.

Although the 2732 has 24 pins, reading it as a 2732 won't work as expected because a couple of the pins are placed differently and the 2732 has no A12 pin because it is only half as big memory wise as the chip you are trying to read.

jay_oldstuff, can your programmer handle the 2564 or 2364 as Wave Solder suggests?

[jay_oldstuff]

The ic will not read as a 2764 it just comes up read error. in fact the ic has a different number of pins between the 2764 and 2732 the 64 has an extra pair. the software for my programmer even shows you how to orient the ic in the ziff socket the diagram matches my ic for number of pins when set to read a 2732, when I set the software to a 2764 there are two pins left over compared to my ic.
also the hex file is 4kb which I’m sure is what the acorns ROM file is meant to be.

Jay

[jay_oldstuff]

Ah post crossed. In that case i have a problem, unfortunately my programmer doesn’t support 2564 or 2364 ic's

Jay

[SiriusHardware]

Originally Posted by jay_oldstuff

The ic will not read as a 2764 it just comes up read error.

Yes, sorry, my huge mistake, you can't read it as a 2764 for various reasons, not the least of which is that the +5V power pin is in a different place.

If possible, try to read it as one of the devices Wave Solder suggested?

It could be read as a 2764 if you made a suitable converter / adaptor to reroute some of the pins on the PROM to where they would be on a 2764. One way would be to stack two or more 28-pin turned-pin sockets with certain through connections between the upper and lower socket cut and rerouted to the correct pins on the lower socket with thin link wire (like wire-wrap wire).

Edit: Post crossed again! So you can't read those devices. If you get to the point where you feel you have to eliminate the PROM from enquiries I'll make a suitable adaptor and post it to you.

However, this might be an unnecessarily complex undertaking just now as there's still some basic faultfinding to do on the machine itself.

Did you re-check what's happening on pin 20 of the PROM in normal running mode? Is it low all the time, mostly low, mostly high or high?

[jay_oldstuff]

Originally Posted by SiriusHardware

Did you re-check what's happening on pin 20 of the PROM in normal running mode? Is it low all the time, mostly low, mostly high or high?

Yes see post 45

Jay

[jay_oldstuff]

😄 It's running. I spent a few hours last night prodding round various things with the logic probe and scope, tracing signals and starring at the blank monitor screen with that patterning on it. For some reason I pressed the brake key while watching the screen and something happened, the patterning changed slightly, as I pressed keys on the keyboard I could see slight changes to the patterning with every key pressed. Something must be working, I'd all ready tried swapping the 6847 video IC with one from my dragon 32 with no effect and the 6847 from the atom in the dragon and it worked so I'd ruled it out. I amongst my micro junk I found the board from a TRS80 colour computer which also uses a 6847 (the dragon is basically scoops of the TRS80
coco) I extracted the 6847 plugged it in the Atom and as you can see we have progress. It's rolling at present but I thinks that's to do with my monitor not liking an NTSC 60Hz signal. So need to fix the keyboard as a fair few keys don't work at present but it's a star

Jay

Super! Well on to 99.9% of the way, there are some soundcard tape simulators out 'there' to enable easy programme loading.

[KeithsTV]

If you're in need of a 6847 I've just found one. Can bring it to Golborne.

Keith

[jay_oldstuff]

Yes please Keith, then I can put this one back on the Coco board. I'm just puzzled why the original 6847 will run in my Dragon but not the Atom and the Dragons wont run it the Atom but works fine in the Dragon. Any ideas anyone?
I have software on my PC to play .tap files from my PC's sound-card which will work for now but I'm looking into making an MMC card interface for it (building one for my BBC master first) that way I can have everything I could want software wise and all virtually instantly. and to that end I purchased a 6522 VIA as my Atom didn’t have one installed, I actually ended up with four of them so if there are any other Atom owners out there who need one I have a couple available.

Jay

[KeithsTV]

OK will be with me at Golborne along with all the other stuff I'll be bringing.

Keith

[SiriusHardware]

Originally Posted by jay_oldstuff

I'm just puzzled why the original 6847 will run in my Dragon but not the Atom and the Dragons wont run it the Atom but works fine in the Dragon. Any ideas anyone?

None, but I'm glad you've got it working. It might also be interesting to know whether the Dragon's 6847 works in the Coco and whether the Coco's 6847 works in the Dragon.

If Keith's 6847 proves to be like the one in the Dragon (and won't work in the Atom) I have a loose 6847 here too, you are welcome to try that and keep it if it works for you. Mine is Motorola branded, MC6847P AN48243.

According to my 'Motorola 8-bit Microprocessors Data Manual' (1983) the 'P' on the end of mine just means it has a plastic package, but interestingly there was also an MC6847YP which is the 'interlaced' version whereas the standard MC6847P is the 'non interlaced' version.