Found my old A500 again, and it is nearly running

[MarkL]

My name is Mark, I have been programming the Amiga in the early 90s in machine code. That's what I still like a lot on other platforms.

My old naked A500 has a strange behaviour when joystick is plugged in. Every time I move the stick left or up the machine resets. Any other operation like loading and using the Workbench with the mouse, watching old demos or playing old games from disks w/o joystick performs alright.
I happened to have a lot ICs left over and successively changed both CIAs, ROM, Gary and Paula. And I inspected all electrical contacts and connection to the IC sockets. I removed and refitted even the Agnus and Denise ICs.

That strange behaviour is still there.
Are there any hints or ideas to enable joystick use again w/o causing reset all the time ?
Changing the joystick model does not make any difference.
Mark

[Jope]

Are the joystick port pins all straight?

[nogginthenog]

Sounds like +5V & GND are being connected.

What kind of joysticks are you using?
Does the joystick in the mouse port do the same thing?

I seem to remember there is one type of controller (NES?) that pulls up the pins to 5V when not pressed instead of leaving them floating like most systems. Not sure that would explain this but just a thought.

[gdunlapsd]

I agree with noggin; sounds like a short between 5v and ground, or perhaps a directional pin (don't know if that would cause a reset, but maybe).

Here's a pinout of the joystick port:

http://eab.abime.net/showpost.php?p=1032393&postcount=2

Make sure there is no debris or corrosion between pins inside the port and behind it as the pins go into the motherboard.

Also, take out the motherboard and check the bottom for corrosion under the joystick port (and elsewhere while you're at it. Since it sounds like you got this out of storage, corrosion is a real possibility depending upon how it was stored.

[MarkL]

That explains one possible source of the problem of course. But the pins are straight and the joysticks are those I am using with my other Commodore computers without problems.
The board looks fairly clean from the upper side. I had already checked all capacitors visually.
Anyhow, I will soon inspect the joyport from the under side of the main board. Did not yet do this.

[gregthecanuck]

Welcome to EAB!

[Shadowfire]

I'm looking at the schematic for the A500 Rev 6.

Up/Left on an Atari style joystick should connect pin 8 (ground) to either pin 1 or pin 3 (depending upon direction).

The Amiga port does not connect pins 1 & 3 directly to a GPIO port. Instead, they are multiplexed by U15, a 74LS157, and the mux outputs are fed to Denise (U4). The 74LS157 has no direct passthrough of the signal from one mux input to the output, it buffers the signal; therefore, it is safe to assume that the problem is not with Denise.

Looking at the input signal conditioning, there are capacitors to snub EMI on these lines. If the capacitors have internally shorted, then your symptoms would be of that joystick direction always being activated, not a system reset.

Additionally, RP401 serves as the pullup to VCC. At the existing value (4.7K) there would be a ~1mA power draw whenever the joystick pulls the I/O line low, which should not cause a reset.

The 74LS157 buffer IC is a CMOS level input buffer, which when functioning normally, shows an extremely high resistance level to VCC on its inputs.

However, you ARE getting system resets, and this almost certainly because your 5V rail is tanking when you activate the joystick. Unless you have a mechanical short on to the motherboard, the most likely reason is the 74LS157, and I'll explain why.

The mouse in an Amiga uses 4x infrared LEDs and phototransistors to "see" the holes drilled in wheels which rotate when you move the mouse. The LEDs are powered by the 5V and Ground pins (7 & 8). The phototransistors are all connected from Ground (pin 8) to their respective pins (1 to 4).

In a joystick, a phototransistor isn't doing the shorting from GND to the directional pin; a mechanical switch with a VERY low resistance (typically <1 ohm) is doing it. The phototransistor in a mouse does NOT get nearly this low of a resistance; typically, it will go down to an effective resistance of ~60 ohms (it's more complicated than this, but this is an approximation). Therefore it is entirely possible that the mouse does not load the power supply down enough to reset the computer, while the joystick does, and the joystick does not reset any other machines you connect it to.

The pullup resistors (RP401) *might* be bad and too low of a value, however it is FAR more likely that the 74LS157 has a defect or ESD damage on the input stages tying it to VCC with a low resistance. Take a DMM (with the computer OFF) and measure the resistance from 5V (pin 7) to pins 1 & 3, then check it against pins 2 & 4. Make sure that on each measurement you use the RED DMM lead to probe the 5V pin and the BLACK DMM lead to the pin under test.

I bet you'll find that pins 2 & 4 have a significantly higher resistance than pins 1 & 3. You should have the IC replaced by someone who can service circuit boards; also have them check the EMI capacitors to make sure that they haven't failed (or you may find yourself doing the same repair again in the future).

[MarkL]

Shadowfire, what you are explaining does sound very elaborate. I have tried to follow your description. And I have found that resistance between pin 5 (+5V) and equally pin 4,3 and 2 is 9 kiloohms each. Whereas resistance to pin 1 is about 1 megaohm. I could not locate RP401 I just see a resistors unit called RP404. I have board revision 5.
I am refering to they joyport pinout linked above. I will possibly have a try and replace U15 using my new desoldering gun. It is not that I am totally inexperienced with soldering. I am just new to the computer electronics. In the past I just have done model railway soldering.
Mark

P.S.
If I swap mouse and joystick the amiga does not come up at all. screen stays gray. if none of mouse or joystick is connected everything is alright again.if then i put mouse into port 2 (the joyport) the amige resets and wont come up again. If only the joystick is connected to the mouseport the 'normal' failure behaviour resetting on up or down occurs...

[MarkL]

Originally Posted by gregthecanuck

Welcome to EAB!

Thank you .

[Shadowfire]

I tried to hunt down a copy of the Rev5 schematic to see if the circuit is any different, but no dice. Good news is, I have a working Rev 5 A500 here to measure and walk you through it.

http://amiga.resource.cx/photos/phot...res=hi&lang=en

shows a picture of the rev 5 PCB. RP401 is on the right hand side (further to the right, and slightly above, RP404). You need to take off the floppy drive to see it. In this picture it is a red component; for what it's worth, on my machine it is blue.

A little less than 10K is the correct measurement. This confirms that R401 (4.7K current limiting on the +5V pin of the joystick port) and RP401 (4.7K pullups to +5V) are good on joystick pins 2-4, however, 1 Meg means that you have at least one fault on the motherboard.

Once you have removed the drive and have RP401 exposed, perform the following checks:

Inspect the rear side of the joystick port, EMI421, RP401, and U15 for foreign bodies, solder bridges etc. which might be shorting pins together.

Measure from joystick pin 1 to EMI421 (side nearest the joystick port). You should get <1 ohm.
Measure across EMI421. You should get <1 ohm.
Measure from EMI421 to RP401 pin 6. (RP401 pin 1 is nearest to the right side of the board). You should get <1 ohm. (* this is the one I think is going to fail)
Measure from RP401 pin 1 to RP401 pin 6. You should get around 4.7K.
Measure from EMI421 to U15 pin 11. You should get <1 ohm. U15 is the 74LS157, not the 74LS38.

Get back with which measurement(s) failed.

[MarkL]

Following your instruction I found all values as expected besides that from joystick pin 1 to EMI421. It is above 1 Mega Ohm. I had no problem identifying all pins as the board layout you have posted is just the one I have here.
BTW I have swapped U15 for a brand new one and the A500 still behaves exactly as before failing on joystick moves up and left.
There is no debris or short cuts to be seen. I have cleaned the board using compressed air just after disassembling the computer. There was only very little dust in the case.

[Shadowfire]

You have a break in the electrical path leading from joystick pin 1 to EMI421. This is on the back side of the PCB, so you need to take it out of the EMI shielding to do this. Check the joystick connector for damage to the solder joints; if you can see any cracks, flux up and reflow the solder on all of the connectors on the machine. If there is no visible damage to the solder joints on the connector, look for damage to the track itself. Check the trace itself from joystick pin 1 to EMI421 on the back side with the meter. If the break is in the trace and not the joystick connector joint, use a 24 to 30 gauge wire to jumper from the connector pin to EMI421.

The damaged pin joint or trace, however, is probably not responsible for the spontaneous resets.

Next, yank Denise (U4) out of the socket and power it up with the joystick in port 2. You will *not* have any video, but you will be able to observe the power LED turn on after a few seconds, and hear the floppy drive click.
Now check the joystick and observe the LED/floppy to see if it still makes the machine reset.

If it no longer resets, Denise needs to be replaced. If it still resets, turn off power and measure RP405 (near Denise) between pins 1 and 4. This should be 120 ohms. If you're stuck with fatty meter probes, you can probe it from the back side, or check from Ground to U4.39.

The rev 5 board seems to be different than rev 6 on this signal. Instead of an EMI capacitor, the signal is routed through RP404 (pins 7 and 8). I can see no markings, but it seems to be a string of four 68-ohm resistors (no common rail). Check for ~68 ohms between U4.39 and U15.9.

If it still resets without Denise and RP405 / RP404 check out OK, then I sure hope that you socketed U15 when you replaced it.

[Pyromania]

Welcome!

[MarkL]

I am soryy. But I was misled by the joystick pin layout linked here. It wrongly showed no.5 as +5V and not no.7.
When repeating the initial measurements to pin 1 thru 4 I get a consistent 4.7KOhm.
Of course you are right that a possibly broken joint from no.1 joystick pin to EMI421 has no effect on my problems. And there is no problem either at a more closely inspection neither with joint or trace on the back side. I also was misled by mixing up no.5 and no.7 pin.
However the resets still occur with Denise removed and all resistance readings around RP405 / RP404 came out as expected and hence are alright. I might pause the issue for now and will put in a socket to try my secondary 74LS157 one day. But there is no socket handy at the moment. Anyhow it this a welcome lesson in understanding electronics a bit. It is all rather new to me. BTW where did you get the A500 rev.5 schematics online ?
I also have another patient namely an Acorn Archimedes to be cured. But the oscilloscope I was promised by a friend is still retarded until midth of December. Maybe that one is helping me too with the A500 then.
Regards
Mark

[Shadowfire]

I don't have the schematics for the R5, I was looking at the R5 A500 on my desk. The next step is to try a new 74LS157.

[MarkL]

I see. Thanks. Then I will startup my desoldering tool again to install a socket.