A500 1mb mod gone wrong - Keep blowing CIA chips

[hypnoshock]

Hi There,

Hoping to get some advice from a hardware head as to what I've done wrong. I attempted the do the A500 Rev6 motherboard mod where you can get the trapdoor ram be seen as chip ram so that the result is 1MB chip RAM. My result so far has been a blown Agnus as well as a blown odd CIA.

Below are the steps I took to wreck my Amiga. Any advice at all would be most appreciated as I'm looking improve my low level hardware knowledge more than anything.

* JP2 - Cut the trace running from pad 2 and 3 and bridged pads 1 and 2 (top two pads).
* JP7A - Cut trace between pad 1 and 2
* solder 3 wires on JP2 so a jumper can be set to undo the mod. The soldering is where I'm thinking I may have blown the Agnus chip but I'm not sure.

On first boot up after the mod I got the green screen which I read is a chipram issue and the first thing to try is to flip the CIA chips around to see if the behaviour changes. This resulted in the same green screen.

Agnus I noticed was running red hot so guessed she died in the process so I bought a new one. Today I reverted the mod so bottom 2 pins were bridged on JP2 (by jumpering the wires I had soldered) and bodged a wire on JP7A so the bottom 2 pads were connected again (or rather pad 1 and 2 if counted from left to right with the label 'JP7A' the correct way up).

Amiga booted! but disk drive head didn't move. Flipped the CIAs around and grey screen so assumed the CIA that's now in odd was broken. Put even into odd, had odd remain empty and Amiga booted. Put a CIA chip from another Amiga into the empty even slot and then I had a fully working Amiga!

Now if I had just dropped it here, I'd have a fully working Amiga but I had one last go at the mod by jumpering the top two pads via the wires I had soldered onto JP2 and using a voltmeter to check the continuity which checked out and confirm there was no continuity with the bottom most pad. I desoldered my bodge wire, plugged in, turned on and..... Dead CIA in odd socket!! Moved the even chip into another amiga board and it boots so assume that even CIA wasn't fried. I tried booting the other Amiga with the CIA that was in the odd socket in the modded Amiga but it doesn't boot so I believe it's definitely dead.

I won't attempt this mod again as it's expensive but I'm more looking for understanding of what I'm doing wrong as I want to improve my hardware skills. Was it the desoldering of the bodge wire where I zapped the board? Is there another trace I was supposed to cut somewhere but haven't so the CIA is getting shorted? I just can't work out what I did wrong but I'm obviously doing something wrong because I'm killing these CIA chips. Luckily I didn't fry my new Agnus chip which I can confirm is definitely a fat Agnus (required for the mod to work).

What a day, I waited a week for these parts and I blew it again :-(

[DMWCashy]

JP2 is just an address line, so I have no idea how this could have fried both the agnus and cia. Is it possible that there is a fault with the expansion card itself. To fry these chips voltage would have to be pushed somewhere that it is not designed to receive voltage.
I have only blown a CIA once, and this was a long time ago when I built a stepper motor board, and installed a mosfet incorrectly and sent 48 volts down the parrellel port .
Apart from static, the amiga chips are not that easy to destroy by just messing about with address pins.

[DMWCashy]

I would check for dead shorts on the output of the CIA, has some solder got stuck to bottom of the board, what was it resting on when you powered it up. Without a dead short or voltage going into an output pin, I dont see how you can fry a CIA

[solarmon]

Show some clear pictures of your mods and of your motherboard.

[hypnoshock]

Originally Posted by DMWCashy

JP2 is just an address line, so I have no idea how this could have fried both the agnus and cia. Is it possible that there is a fault with the expansion card itself. To fry these chips voltage would have to be pushed somewhere that it is not designed to receive voltage.
I have only blown a CIA once, and this was a long time ago when I built a stepper motor board, and installed a mosfet incorrectly and sent 48 volts down the parrellel port .
Apart from static, the amiga chips are not that easy to destroy by just messing about with address pins.

This is what I thought which is why I'm completely perplexed as to what I'm doing! When I had two working CIA chips before I did the desoldering and reattempted the mod I did check the RAM board with SysInfo and it appeared to be reporting as 512k. I was also able to boot into a 1mb game so although I can't rule out it's not a dodgy board, it seems on the face of it to be working o.k. Since the whole blowing up the CIA mishap, I have tried the board again in the Amiga I originally pinched the CIA chips out of and it still appears to be working.

How on earth is voltage getting to that odd CIA socket? You'd think if i had bridge something by accident I would have instantly fried the chip on turn on at the point I put in the new Agnus and swapped over the chips.

I'm so so lost!

[hypnoshock]

Originally Posted by solarmon

Show some clear pictures of your mods and of your motherboard.

You've got it:
https://docs.google.com/document/d/1Bsf26uj7MY4fzwviOAJ63x3s53avj1MFzm5whDcXqDw/edit?usp=sharing

If too blurry I can upload each of the photos seperately and link them instead of putting them in one Google Doc. Google doc is lowering the resolution of all my photos so they aren't as clear as the originals.

There isn't a lot to this mod so I really don't understand how I have managed to fry

3 CIAs
1 Agnus
1 CPU (I forgot to mention that)

If I had bridged something top side or bottom side the board wouldn't have worked at the point I had fitted the new Agnus chip, two lower JP2 pins bridged with the bodge wires in place to undo the JP7A slash fest.

[hypnoshock]

I have listed out all the states this board has been in in my GDoc (posted above) but I'll paste the list here too. I just don't know what I'm doing wrong to keep killing my chips.

States this board has now been in:

* JP2 top two pads jumpered and JP7A bottom two pads severed. Trapdoor expansion in
- Green screen boot up (Agnus fried at this point but I didn't know that)
* JP2 top two pads jumpered and JP7A all traces cut. Trapdoor expansion in
- Green screen boot up
* JP2 lower two pads jumpered and bodge wires on JP7A to undo the trace cutting. Trapdoor expansion out.
- Green screen boot up
* Switched CIAs around
- Green screen boot up
* New Fat Agnus chip
- Boots into kickstart disk drive doesn't function
* Switched CIAs around
- Boot to grey screen
* Replace odd chip
- Boots up with a functioning disk drive. Boots into game but game needs expansion
* Expansion in. Game boots and SysInfo reports all O.K. Appears to be fully functioning Amiga again
* Mem mod re-attempt: Top two pins jumpered and bodge wires desoldered. Trapoor expansion out.
- Booted to black screen
* Trapdoor expansion in
- Booted to black screen
* Put bodge wires back in and jumpered back to top two pins on JP2
- Booted to black screen
* Switched CPU
- Grey screen
* Switch CIAs over
- Boot no disk support
*Unpopulated the board and gave up :-(

[solarmon]

Thanks for the informative document. Whilst I'm still digesting it, I have a few more questions.

Was the rev 6A board working OK before the mods?

Also, what ram expansion are you using? Has this been tested on another Amiga and proven to be working?

What power supply are you using, have you checked the voltage rails to make sure that they are OK?

[solarmon]

This won't help you, but just for reference, this is how I modded my jumpers. I used right-angled header pins and shaped them appropriately and solder directly on to the jumper pads. You have to be extra careful as the pads can easily be damaged by heat and mechanical stress.





I can then use jumper connectors to easily switch it, or use Dupont jumper wires to connect to them easily.

[hypnoshock]

Originally Posted by solarmon

Thanks for the informative document. Whilst I'm still digesting it, I have a few more questions.

Was the rev 6A board working OK before the mods?

Also, what ram expansion are you using? Has this been tested on another Amiga and proven to be working?

What power supply are you using, have you checked the voltage rails to make sure that they are OK?

Firstly, I'd like to say thanks for taking the time to look at this. This tiny mod is pretty small fry as modding goes so I appreciate you looking at something so tame.

Yes the expansion board was working before the mods and also was working again at the point of undoing the JP7A slash, putting in new Agnus and putting in a working CIA chip. The RAM expansion is an original A501 REV 6C. CAKE 11/15/88 GRR/FISH (I'll add a photo)

Tested in another Amiga and appears to be working. I haven't done a thorough mem test but on the face of it it seems O.K

I'm using an original A500 power supply. I can test the rails but I would have thought if the power supply was faulty I wouldn't be able to power on the board I've just populated with what's left of my good components. Also I have been using that same power supply extensively with my A1200 with no adverse effect i.e the A1200 seemed to run stably for hours and that was with a 030 running Personal Paint along with HippoPlayer playing S3M mods using AHI sound so I was giving it a good thrashing.

[hypnoshock]

Originally Posted by solarmon

This won't help you, but just for reference, this is how I modded my jumpers. I used right-angled header pins and shaped them appropriately and solder directly on to the jumper pads. You have to be extra careful as the pads can easily be damaged by heat and mechanical stress.





I can then use jumper connectors to easily switch it, or use Dupont jumper wires to connect to them easily.

Very tidy and something I would do if I was going to keep the CPU socket populated with just a 68k however I have an expansion board with IDE and 8MB that fits into the CPU socket which I was going to use when I had this thing working and that would have obscured the pins hence why I ran the wires out.

Nice work on your board BTW :-)

[solarmon]

With the board unpopulated, I would test the VCC pins at each of the chip sockets to see if they get a good 5V and it is not under or over voltage. But it could also be spikes in the voltage that might not be picked up by a standard (slow) multimeter.

[hypnoshock]

Originally Posted by solarmon

With the board unpopulated, I would test the VCC pins at each of the chip sockets to see if they get a good 5V and it is not under or over voltage. But it could also be spikes in the voltage that might not be picked up by a standard (slow) multimeter.

I like your style and I need to buy a decent meter - any recommendations? I would like to get some proper equipment e.g my oscilloscope is also a crappy pocket one that needs replacing.
If I wanted to measure spikes, I guess I'd need a decent scope or is that something I can do with a decent voltmeter?

[solarmon]

I'm not the right person to give out advice on what multimeter to get! I've only got an old and basic ISO-Tech IDM63N, but that has served me well so far.

Maybe this will help you better - there are three specs to look out for - accuracy, range and frequency:

https://www.fluke.com/en-gb/learn/bl...racy-precision

[solarmon]

A logic probe is also highly recommended - you can get some cheap ones from eBay and Amazon.

[hypnoshock]

So in summary the next course of action would be:

* Desolder my bodge wires on JP7A so it's in the 'all traces severed' state.
- This allows the trapdoor RAM to be assigned to chip RAM (right?). At least it 'disables' it as an expansion or something?

* Jumper the top two pins of JP2.
- These are 'address lines' and allow Agnus to address the extra ram directly as chip

* Check for dead shorts on the output of the CIA using continuity check on my volt meter

* Power on the board unpopulated and check the voltage on all the IC sockets. Check for under or over voltage - we would expect 5V.
- If unpopulated, would all these sockets be receiving 5V as normal or do I need some chips in there for this to work?

* If readings appear o.k, check for power spikes on power on
- I don't have the equipment for this but I guess I can't skip this step because if it's spiking above 5V on power up this might be the cause of the fried chips. At this point I really don't know what I'm doing to kill the chips and this is the best I have to go on other than the possibility of dead shorts which I can't visibly see.

* If no spikes, repopulate the board and it should in theory work.
- Can I test without the trapdoor expansion or does the mod require that I have the expansion always present now?

* If boot up, retry with expansion in

* If boot up, then populate the empty even CIA with a working one when I manage to fine a replacement.

At this stage I should have an Amiga with 1MB chip RAM. Last thing to do would be put in my 8MB with IDE expansion and have the 1MB / 8MB combo I wanted.

This is haaaaard! Thanks everyone for looking at this! Seems like such an easy mod but how I'm just killing components left right and centre has left me really puzzled.

[solarmon]

I would revert the jumpers back to default- so bottom pads joined on both jumpers. As per my pictures.

A few bits of information to be aware of so help with testing and determining when something may not be working as expected.

JP2 is what tells the system whether expansion ram is Slow (default) or Chip.

JP7A tells the system whether there is any expansion ram enabled- default is to pass the signal through to the expansion ram and let it decide (some expansion ram connect it directly to ground, other expansion ram have an enable/disable switch). The top two pads will connect this directly to ground - so it effectively says expansion ram is enabled/present.

If JP2 is modded to say expansion ram is chip ram, and JP7A is grounded but there is no actual expansion ram board connected, the A500 will not boot up (the power LED will flash and you will get a boot loop). It is expecting chip ram memory in that area and if it is not present then it will go through this boot loop.

[hypnoshock]

Originally Posted by solarmon

I would revert the jumpers back to default- so bottom pads joined on both jumpers. As per my pictures.

A few bits of information to be aware of so help with testing and determining when something may not be working as expected.

JP2 is what tells the system whether expansion ram is Slow (default) or Chip.

JP7A tells the system whether there is any expansion ram enabled- default is to pass the signal through to the expansion ram and let it decide (some expansion ram connect it directly to ground, other expansion ram have an enable/disable switch). The top two pads will connect this directly to ground - so it effectively says expansion ram is enabled/present.

If JP2 is modded to say expansion ram is chip ram, and JP7A is grounded but there is no actual expansion ram board connected, the A500 will not boot up (the power LED will flash and you will get a boot loop). It is expecting chip ram memory in that area and if it is not present then it will go through this boot loop.

Wow, I think we're almost here with this then and that the ram must be present makes 100% sense to me so that makes me feel better. I think the last bit I don't understand is that forum posts mention to sever the bridge between the lower two pins of JP7A but no mention of bridging the top two which I guess would leave this jumper floating and perhaps disable the expansion? I don't really get why I needed to cut this in the first place as I thought I would want to pass that signal onto the expansion which would have defaulted to grounded/enabled or am I supposed to set it to disabled so it's not seen as fast ram but Agnus can address it somehow? (That doesn't make sense to me but maybe the instructions I read were incomplete and they had conflated the solder more RAM onto the MB mod with the one I'm trying to do). So if I had JP2 as top two bridged and JP7A as factory default wouldn't that give me the 1MB chip result I was after?

I understand why one would want to cut the bottom two traces to set it to disabled (if floating means disabled??) if one had soldered more mem chips directly to the motherboard.

I think if I understand this last piece of the puzzle I can let this go working mod or not.

[MiaM]

My guess is that it's either ESD related or related to grounding of the Amiga and the soldering iron.

You should electrically connect the soldering iron with ground on the PCB you are working on and have an ESD wrist strap connected to those grounds. All those grounds should preferable be connected to something that has some kind of true ground (preferable the ground on an electrical socket).

If your soldering iron has a grounded plug and it's connected to an ungrounded socket and it has a switching power supply and/or a mains filter it will likely leak a small current from mains to the tip.That can be enough to fry semiconductors.

[solarmon]

Originally Posted by hypnoshock

Wow, I think we're almost here with this then and that the ram must be present makes 100% sense to me so that makes me feel better. I think the last bit I don't understand is that forum posts mention to sever the bridge between the lower two pins of JP7A but no mention of bridging the top two which I guess would leave this jumper floating and perhaps disable the expansion? I don't really get why I needed to cut this in the first place as I thought I would want to pass that signal onto the expansion which would have defaulted to grounded/enabled or am I supposed to set it to disabled so it's not seen as fast ram but Agnus can address it somehow? (That doesn't make sense to me but maybe the instructions I read were incomplete and they had conflated the solder more RAM onto the MB mod with the one I'm trying to do). So if I had JP2 as top two bridged and JP7A as factory default wouldn't that give me the 1MB chip result I was after?

I understand why one would want to cut the bottom two traces to set it to disabled (if floating means disabled??) if one had soldered more mem chips directly to the motherboard.

I think if I understand this last piece of the puzzle I can let this go working mod or not.

I was in the same boat trying to understand this mod initially. I also didn't get why JP7A needed to be cut or changed. It was only when I looked at the schematics and tried out the various permutations did I understand what it is for. So this information is based on my own experience.

If the middle pad of JP7A is floating then it will be seen as HIGH, as there is a pullup resistor connected to this _EXTICK signal (which is 'active low').