PDA

View Full Version : Amiga 500 - blank/black screen



r00li
August 26th, 2016, 06:36 PM
I just got my first Amiga a few days ago. I bought it off of ebay for a decent price (with original packaging actually). It was a bit dirty and it looked like it's been sitting in a garage somewhere. Classic ebay really.

I tested it when I got it (it booted, and read a random floppy disk game just fine) and then it was off to cleaning. So I disassembled everything, cleaned the keyboard, retrobrited the case, checked if anything suspicious was going on on the board (leaked caps and such). So I put everything back together, downloaded some software and put it on floppies and started playing a bit with it.

Then I noticed that the thing sometimes won't work properly. When I turn the power on, the power light comes on and I can hear the floppy clicking, but nothing is displayed on screen. The TV actually gets a video signal (at least it doesn't display the "no signal" message) but everything stays black. If I put in a floppy it will actually start reading it and the disk light will come on. If I just let it sit at the first screen (for a few minutes) the display will usually slowly start displaying stuff. First I get the outline of the floppy and then the rest of the image shows up. It's like a fog lifting from the screen or something :p . Rebooting it doesn't seem to have an effect. Just waiting a bit does.

I tried both the A520 and the internal monochrome composite out - both have the exact same issue. The whole thing looks like it's working properly, but the video signal is a bit "wrong". Because even when it works I can still see an occasional color line (horizontal) on the screen. And for some reason the whole display looks like it's shifted a bit to the right. I also tried making an RGB-Scart cable but if I plug that thing in the the Amiga just won't boot (no power led). But that may be just because I wired something incorrectly.

Any ideas on what could be wrong?

giobbi
August 26th, 2016, 07:31 PM
Ok, I don't know much about Amiga, but here you have my two cents...

If it was working before you opened it, we could suppose something happened while cleaning... maybe (or maybe not, of course).


First attempt, if you didn't try it yet, you should try re-seating the socketed ICs; just pull them out from its socket a bit a re-seat. Take care not to bend any pin! No need to completely extract it.

You could also check solder joints to see if there's some bad. I had a problem with a VIC-20 showing strange lines during programs execution and I discovered it was a broken solder joint.
Particularly, take a look the solder joints of the video connector.

Of course it could be a bad IC (Denise video chip?), but it seems the Amiga begins to work when it warm up a little. Are you sure caps are all ok (including ceramic) ?

Try to remove add-on card too (like RAM expansion board).

Video RAM could be the culprit, but I'm not sure if Amiga has dedicated video RAM or it's sharing system RAM.

r00li
August 27th, 2016, 12:31 PM
I don't know if it was working before. I just did a quick test.

Well I tried reseating the ICs just now. All except fat agnus - that thing needs a special tool to remove (which I don't have, but I did order it). So I just pushed it down a bit until I get that tool. It now takes about 10 seconds to a minute after I turn on the power until the image starts appearing and then about 10 seconds more before the image clears up. First the outline of the disk appears, then the background (it's black before), then the color lines disappear and the image clears.

The image takes longer to appear after the power has been out for a while.

This leads me to suspect the caps, but they really do look fine. No leaking, no bulging, nothing. They still may be bad of course, but there is no obvious way for me to figure that out. Solder joints seem ok (and if those were the problem they probably wouldn't affect both the internal composite out and the RGB out). There are no add-on cards/boards in my model .

EDIT: I also tried inserting a game disk and turning the thing on. The thing starts reading the floppy immediately and I'll get the game sound way before I get the image. So it must be something with the video output.

KC9UDX
August 27th, 2016, 01:20 PM
Don't pull Agnus unless you have to. That socket is super fragile, and even with the tool it may break.

If you pry up the corner of it and push it back down it's as good as pulling it out to reseat it.

giobbi
August 27th, 2016, 03:13 PM
Don't pull Agnus unless you have to. That socket is super fragile, and even with the tool it may break.

If you pry up the corner of it and push it back down it's as good as pulling it out to reseat it.

Yes, I discovered it on my own.
However I don't think the fat agnus can be the culprit.

Solder joints sometimes seem good when they aren't. If I was in your shoes, and if you are confortable with solder iron, I would try to take care of the video section. Just use some solder flux and/or some solder wire to refresh the joints. Can't hurt (if you know what you're doing).

----

Tantalum, polyester and ceramic caps don't give you any clue when they're bad. Electrolytic sometimes.

r00li
August 27th, 2016, 05:48 PM
I was pretty sure that it wasn't the solder joints because the problem seemed to repeatable and not random enough. But now I am not so sure. Basically I picked up the thing while it was hooked up to the screen and no image was showing. Then I shook it and the screen suddenly flashed into life and went all disco on me.
On the other hand, shaking the machine when the video is on just makes a random line or two flash on screen. I would think that a bad solder joint would make the screen randomly disappear as well. But it doesn't. After I get the video on screen the thing stays on.

I certainly don't have any issues with a soldering iron. If I can solder 0402 SMD components I really shouldn't have any issues soldering through hole stuff. But I will give it a try and see how it goes.

As for caps... I don't think I've seen any ceramic caps fail before. But then again... it's possible. Though I really don't want to go replacing the caps if I am not very confident that they are the problem.

inotarobot
August 27th, 2016, 06:37 PM
I was pretty sure that it wasn't the solder joints because the problem seemed to repeatable and not random enough. But now I am not so sure. Basically I picked up the thing while it was hooked up to the screen and no image was showing. Then I shook it and the screen suddenly flashed into life and went all disco on me.


Well my first port of call to look at is the power in connector on the 500 mother board. Either the solder joint to the board, or possibly one of the pins inside the connector has a break in it. As you will see the the pins turn at right angles and I have seen similar connectors fracture inside the shell at the bend point.

To check get yourself a matching plug for the power in. solder some flying leads to it and then connect a multimeter to one lead at a time, and other lead to a flying lead soldered to mating pin on the back of the motherboard. then wiggle the plug.

If no break then look at quality of the socket to see if it is corrosion on it OR one part of the this female connector has broken and fallen out; this not really then making 100% good contact to the pin on the power modules plug. Some sockets use 2 split sides that hold the mating plug pin. I have seen sockets where one 1/2 of this split socket contact is completely missing and its only by good fluke on plug positioning, that it actually mates.

Slim chance also you have a fault in the power module cable or plug that feeds the 500

Then start looking for an intermittent joint in the DB23 socket and plug on video cable.

giobbi
August 28th, 2016, 08:58 AM
I was pretty sure that it wasn't the solder joints because the problem seemed to repeatable and not random enough. But now I am not so sure. Basically I picked up the thing while it was hooked up to the screen and no image was showing. Then I shook it and the screen suddenly flashed into life and went all disco on me.
On the other hand, shaking the machine when the video is on just makes a random line or two flash on screen. I would think that a bad solder joint would make the screen randomly disappear as well. But it doesn't. After I get the video on screen the thing stays on.

It seems to me you have a bad contact / socket / solder there.

Old sockets are known to be unreliable, expecially C= ones (I've had bad experiences with PET sockets, I don't know about Amiga). When I'm in doubt, first I replace sockets. You would at least try to replace the video IC socket. Can't hurt...

In my experience, simply try to clean it usually doesn't make any difference.




As for caps... I don't think I've seen any ceramic caps fail before. But then again... it's possible. Though I really don't want to go replacing the caps if I am not very confident that they are the problem.

Ceramic caps are quite reliable, yes, unless they're broken... But I've already seen many bad polyester caps.
However, if shaking the board makes difference, can't see a cap, IMO. It seems to be a trivial issue.

KC9UDX
August 28th, 2016, 11:08 AM
Amiga sockets are not poor quality. Quite different from PET sockets. In fact, I think by the VIC-20 these were different.

I have very little experience fixing Amiga hardware problems, other than bad CIAs and battery corrosion. I've got a 2000 that stumped me, I had to abandon it last year. I don't know if I'll go back to it.

I don't actually have an idea in this case, but if I were to start swapping custom chips I'd probably start with Denise.

I agree that this seems more like a mechanical problem though. inotarobot may be on to something with the power supply connector. I do seem to recall resoldering a few of those over the years but don't recall the failure mode.

r00li
August 28th, 2016, 11:09 AM
Update...

I re-soldered all pins around the video section and power in. That had absolutely no effect.

So I plugged the machine in and started randomly poking the chips with my finger. As soon as I pushed down on Denise the whole image suddenly cleared up. So I pulled it out, cleaned the socket and the chip with isopropyl alcohol and put everything together. No difference. But again pushing down on the chip with almost ungodly force makes the image clear up/disappear. Especially the middle of the chip. Pushing down around Denise has similar effect, just not as profound. So my guess is that there is either something wrong with the chip (broken internal connections possibly?), the socket or the PCB around it.

EDIT: Replaced the socket. No change. I think Denise has a problem.

EagleTG
August 28th, 2016, 01:23 PM
So my guess is that there is either something wrong with the chip (broken internal connections possibly?), the socket or the PCB around it.

Remove the socket, take a VERY close look at the traces, concentrate on parts that are near board vias and any through holes. I'd also suggest using a multi-meter on continuity to verify the traces aren't showing resistance, etc. I'm thinking possibly broken traces rather than a bad Denise.

Edit: Use a magnifying glass or microscope if you have one to inspect the traces.

giobbi
August 28th, 2016, 05:12 PM
I'm thinking possibly broken traces rather than a bad Denise.

I agree, it sounds like broken traces. Before to remove (again!) the socket, I would check every trace from every denise socket pin to next solder joint (i.e. another IC) to see if you find a broken trace. If I was in your shoes I would remove the IC and check from the socket contact to the next solder joint.

Everything is possible, but a bad contact inside the denise IC seems to me very improbable.

r00li
August 29th, 2016, 02:25 AM
To be honest I suspect Denise more than the PCB. Because a problematic PCB trace wouldn't really explain the fact that the problem disappears after a while by itself. I actually do suspect a bad contact inside the chip. Because when it warms up the material expands/contracts. After the image appears and stabilizes pushing down on the chip and around it seems to have very little effect. It would also explain why the image appears quicker when I just power cycle the machine without any pause in between.

I did a visual inspection of the traces around that area with a magnifying glass and they seem fine. I tried testing continuity on a few of them and that was fine as well. Sadly testing continuity on all of them would be quite a hassle.

I ordered a spare chip from ebay for 13€. We will see if that solves the problem or not.

KC9UDX
August 29th, 2016, 03:23 AM
I've seen bad solder joints do all of the above. And yes, I mean the kind that you look at them with your highest power loupe and can't see a defect unless there's mechanical movement.

But for your sake, I hope you're right and it's the chip.

r00li
August 29th, 2016, 08:21 AM
You may be right. I've rarely seen broken traces myself. It was usually on home-etched PCBs where the etching went a bit wrong or when a sharp object impacted on a PCB. But then again... I have little experience with older computers and larger boards.

But I did a quick experiment just now. I pulled out the chip and heated it up using a hot air gun, inserted it into the socket and turned on the power. Result: image was clear and visible immediately. That does somewhat support my theory. Though it could be pure "luck".

I really hope that I am right as well. Rewiring the traces would be a pain.

KC9UDX
August 29th, 2016, 08:55 AM
I've seen broken traces, but what I almost always see instead is bad solder joints*. They get hairline cracks that you can only see under magnification, and only when the crack is spread apart, which it isn't always. I've run into dozens of those over the years.

I've never seen a bad solder joint on an Amiga, not to say it couldn't happen.

What you did makes me think you just had a bad connection between the chip and the socket. Time will tell.

*The exception is when someone drags something sharp against the board, or, by battery leakage.

r00li
September 7th, 2016, 07:28 AM
Got a new chip today... It works!

Immediately after turning on the power I get a white screen and after that the floppy. Pushing in or around the chip doesn't cause any distortions or anything. So it apparently was the chip. I can now finally spend some time actually testing and using the thing.

Many thanks for the help!

giobbi
September 7th, 2016, 10:15 AM
Glad to know your A500 came back to life!
I have to say I've never seen a bad chip with a behavior like yours; usually they work, or they don't.

Rick Ethridge
September 7th, 2016, 08:16 PM
Good show!