archive.worldofdragon.org

Hello,

I'm very new to Dragons, but I bought a Dragon 64 recently to continue exploring 8-bit computers. As I kind of expected, untested turned into "not working". The good news is I'm fairly experienced repairing other 8 bit computers (Amstrad CPCs, Spectrum). The bad news is, I don't know much about the Dragon itself.

The symptoms is that I get a green screen with a black border around it. No text or prompt on the screen of any kind.

So I opened it up and it's clean and pristine, including the warranty sticker over one of the screws, so I'm clearly the first one opening it. The CPU is the hottest chip (borderline OK to leave my finger on it) and the second hottest is the SN74LS783N (is that some kind of memory support chip?). None of the RAM ICs are more than just luke warm.

It has correct voltages everywhere. I looked with a scope and I see mostly normal behavior: clock signals, CAS, RAS signals, data bus activity, address activity, etc. The only weird thing I saw was Dout on all the RAM ICs has the normal pulses, but also has some weird 2V curved signals (like a capacitor charging). I'd never seen anything like that in other computers (even ones using the same 4164 DRAM).

So my question is, what would you recommend as a next test I can perform to narrow down where the problem might be coming from? Anything to measure? Some ROMs to burn and use that help me diagnose the system?

Thank you!

Two common failures I've found in Dragons are broken CPUs and some kind of weird crystalized deposits over a vertical variable resistance.

Both usually causes garbage screens but that green with black border doesn't rings a bell...

jltursan wrote:Two common failures I've found in Dragons are broken CPUs and some kind of weird crystalized deposits over a vertical variable resistance.

Both usually causes garbage screens but that green with black border doesn't rings a bell...

The CPU seems lively enough from what I can see from the scope.

The green border is exactly what they found in this video, which was traced back to a bad RAM chip. The Amstrad CPC has a very similar behavior (the border means the CRTC was initialized properly, which means the CPU, ROM, and CRTC are all working correctly (at least on the Amstrad CPC). I was hoping it would be something equally simple in this case.
https://youtu.be/PXZfqn2rnIc?t=25m9s

Getting to a blank green screen means at least some ROM code has been successfully executed. The PIAs and SAM (74LS783) have been set up and the video memory has been written to give you a blank screen. (Otherwise it would show patterns of characters)

Interestingly, clearing the screen would be the first occasion on which the memory is called upon to work properly. i.e. this is the first time a subroutine is called and it requires a working stack to be able to return.

A bad memory IC would seem to be the most likely cause at this point, though there could be other reasons. Next most likely is a poor soldered joint on one of the major ICs. I've seen quite a few of those.

If you end up swapping out the SAM to eliminate it, be extra careful with ESD precautions. The SAM seems to be particularly susceptible.

Oh and check the supply voltages (+5, -5V, 12V) if you haven't already. You never know...

Just a little trick I picked up while repairing my Dragon32 a couple of times was to type the commands MOTOR ON and MOTOR OFF in too see if the relay clicked on and off.
If it does you know that Basic is up and running and responding.

In my case it turned out to be a faulty 6847 VGD chip. It was causing a corrupt display and once replaced the Dragon came back to life for a few months... Then the SAM chip failed.

Hope that helps?

sorchard wrote:A bad memory IC would seem to be the most likely cause at this point, though there could be other reasons. Next most likely is a poor soldered joint on one of the major ICs. I've seen quite a few of those.

That makes sense. I don't imagine there's an easy way to narrow down which RAM IC it might be, is there?

Also, the back of the board looks pristine with all the joints bright and shiny, except for C32, but I doubt that was it. I did reflow the joints there a bit though just to make sure.

sorchard wrote:Oh and check the supply voltages (+5, -5V, 12V) if you haven't already. You never know...

Newbie question: Where's the best way to check that? The +5V looks fine all over the board. I checked the pins coming from the ribbon on what I'm assuming is the power transformer in the back, and I found one that were as high as +17V and one that was about -7V. Is that normal and it gets corrected later on, or are they out of spec already?

By the way, I saw in another thread that people are interested in pictures of different boards, I figured I'll post mine. It's probably common, but maybe someone will be interested.

Interestingly, one of the RAM ICs is different from the others. But I could swear I'm the first person to open this machine since it was manufactured.

Hi,

We have that board scanned here, but thanks anyway!

Best of luck getting your Dragon back in working condition!

The power connection between the PSU & main boards reads something like this from left to right:

video, +5V, sound, video, 0V, 0V, -12V, +5V, +12V

There is a -5V generated somewhere on the main board but that is involved in the video display, so probably not relevant here. The 7V & 17V you saw were possibly a result of measuring between 5V & 12V rails.

A memory fault sometimes leaves a tell-tale on the screen that points to a particular IC, but that doesn't appear to be the case this time.

It would be nice if there were a diagnostic ROM that could be plugged in place of the boot ROM, but I'm not sure if such a thing exists.

Progress update: I started removing RAM ICs, socketing them, and putting on good ones (I can go back later and put back the RAM ICs that were not defective). I did 4 of them and ran out of sockets (oops). So far no changes.

One interesting bit that I suspect you're all aware of, is that with a RAM IC completely removed, I get a screen full of characters. Different characters depending on the IC (and one of them was actually a couple of characters and some color). I guess if an IC is completely dead, that's a pattern to look out for. Also, it confirms that we're reading the ROM correctly and that the screen display is fine.

Also, this is one of the best boards I've had the pleasure to work with. It's so easy to disconnect it, make some changes, plug it back in and see how it works. The case might be really bulky, but it pays off when working inside.

If you're narrowing the fault to the RAM, you can always try the piggyback trick over the other four ICs, Try this only if they're cold (I guess they're, you must have already checked for HOT chips...). Look for changes in patterns, depending the error you can even find that the machine works