Weird fault on A1200, how to find it?

jimbob · 28 November 2016, 21:50

I have an A1200 here which seemed to be working properly after a recap except the keyboard wasn't working. No keys worked and the caps lock light would flash three times every so often. I then discovered that the board read "almost" a dead short between 0v and 5V.

A normal board reads several hundred ohms DCR between 0v and 5v and the continuity /diode test reads no continuity and ~200(mv?) in one polarity and ~400 in the other. This board reads about 30ohms DCR and diode test is ~25 in both polarities.

I disconnected everything and suddenly the board would not boot at all. Reset was being held low so I thought the keyboard controller was dead and probably also the source of the almost short. After replacing the keyboard controller I briefly powered the board on and it boots to kickstart screen. This was a very brief test as the apparent short is still present and I don't want to connect anything else to the board or perform any more tests powered up in case something else breaks. I just wanted to see if the reset was fixed, which it is, for now.

Next I suspected the odd CIA U7 because the KBdata and KBclock pins measured quite different from a good board. Maybe the keyboard controller and CIA had murdered each other. I removed the chip without replacing it yet and the fault remains.

The fact it can boot at all in this condition makes me hopeful it isn't a custom chip or CPU. I also tested as much as I could the discrete diodes and transistors in circuit but couldn't find anything obviously wrong.

Having hopefully eliminated the keyboard controller and odd CIA, I began removing and testing decoupling caps, (ceramics and electros), starting at the power input but going through the whole board this way is a bit of a task so any thoughts on a better way to find the culprit?

hooverphonique · 29 November 2016, 12:54

It's not in any way fool proof, but a method is to look for the smallest resistance, because the further you get away from the short, the larger the resistance will be.

voxel · 29 November 2016, 13:10

question : has the old capacitors having heavily leaked on the board, and did you closely looked at where it had before replacing it all ?

because I've found shorts from under ttls and aops on amiga mobos (especially audio aops ...

jimbob · 30 November 2016, 00:49

Quote:

Originally Posted by hooverphonique

It's not in any way fool proof, but a method is to look for the smallest resistance, because the further you get away from the short, the larger the resistance will be.

Good idea, hoped I would be able to at least isolate an area of the board but I didn't manage to find any clues. I have an LCR meter which is a bit more accurate than my cheapo multimeter but I didn't measure anything out by more than 0.05R, just a difference in probe contact I think. DCR actually seems to be ~18.5 ohms.

Quote:

Originally Posted by voxel

question : has the old capacitors having heavily leaked on the board, and did you closely looked at where it had before replacing it all ?

because I've found shorts from under ttls and aops on amiga mobos (especially audio aops ...

This board seemed in good shape physically when I recapped it, no severe leaks, which means I probably didn't inspect under the caps as thoroughly as I should have

. I will take a closer look at all the standard ICs. That funny continuity reading seems more like a failed diode or transistor junction in a chip or discrete to me than a dodgy cap anyway.

This is annoying, hopefully I will find the bastard!

jimbob · 04 December 2016, 22:59

Had a good hunt for this and still no luck. Before I start lifting components one by one I might try this -

1. Something on the board is broken and gives ~20 ohms between 5v and ground.

2. 5V across 20ohms = 1.25W. Anything apart from maybe the biggest chips will get pretty hot dissipating 1.25W

3. Get some freezer spray, power up the board and look for the part which melts the freeze really fast.

Not delighted about powering up the board like this but I think it is a reasonable plan?

jimbob · 07 December 2016, 02:35

Got the bastard! Turned out to be a ceramic decoupling cap at the video DAC. Freeze spray worked a treat, it was very obvious which component was cooking itself when it got a dose.

Everything seems to be working properly now including the new keyboard controller and the original CIA which is now socketed.

hooverphonique · 08 December 2016, 13:04

JonSick · 09 December 2016, 10:55

What a fantastic idea... freeze spray.

I'm nicking that one!

28 November 2016, 21:50	#1
jimbob Registered User Join Date: May 2006 Location: Kilmacolm Age: 46 Posts: 632	Weird fault on A1200, how to find it? I have an A1200 here which seemed to be working properly after a recap except the keyboard wasn't working. No keys worked and the caps lock light would flash three times every so often. I then discovered that the board read "almost" a dead short between 0v and 5V. A normal board reads several hundred ohms DCR between 0v and 5v and the continuity /diode test reads no continuity and ~200(mv?) in one polarity and ~400 in the other. This board reads about 30ohms DCR and diode test is ~25 in both polarities. I disconnected everything and suddenly the board would not boot at all. Reset was being held low so I thought the keyboard controller was dead and probably also the source of the almost short. After replacing the keyboard controller I briefly powered the board on and it boots to kickstart screen. This was a very brief test as the apparent short is still present and I don't want to connect anything else to the board or perform any more tests powered up in case something else breaks. I just wanted to see if the reset was fixed, which it is, for now. Next I suspected the odd CIA U7 because the KBdata and KBclock pins measured quite different from a good board. Maybe the keyboard controller and CIA had murdered each other. I removed the chip without replacing it yet and the fault remains. The fact it can boot at all in this condition makes me hopeful it isn't a custom chip or CPU. I also tested as much as I could the discrete diodes and transistors in circuit but couldn't find anything obviously wrong. Having hopefully eliminated the keyboard controller and odd CIA, I began removing and testing decoupling caps, (ceramics and electros), starting at the power input but going through the whole board this way is a bit of a task so any thoughts on a better way to find the culprit?

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
A1200 - Apollo 1240 card or memory fault?	necronom	support.Hardware	9	04 April 2016 19:07
Fault A1200 2B board	Anemos	support.Hardware	1	02 October 2013 16:34
A1200 possible CIA fault	edimax	support.Hardware	11	20 May 2012 23:16
A1200 Fault Speculation	CDTV1991	support.Hardware	85	28 February 2012 01:07
Help needed with A1200 fault.	Amigafanboi	support.Hardware	3	10 April 2011 18:26

29 November 2016, 12:54	#2
hooverphonique ex. demoscener "Bigmama" Join Date: Jun 2012 Location: Fyn / Denmark Posts: 1,624	It's not in any way fool proof, but a method is to look for the smallest resistance, because the further you get away from the short, the larger the resistance will be.

29 November 2016, 13:10	#3
voxel Amiga Nuts! Join Date: Sep 2006 Location: Le Mayet d'Ecole, 03800, FRANCE Posts: 176	question : has the old capacitors having heavily leaked on the board, and did you closely looked at where it had before replacing it all ? because I've found shorts from under ttls and aops on amiga mobos (especially audio aops ...

04 December 2016, 22:59	#5
jimbob Registered User Join Date: May 2006 Location: Kilmacolm Age: 46 Posts: 632	Had a good hunt for this and still no luck. Before I start lifting components one by one I might try this - 1. Something on the board is broken and gives ~20 ohms between 5v and ground. 2. 5V across 20ohms = 1.25W. Anything apart from maybe the biggest chips will get pretty hot dissipating 1.25W 3. Get some freezer spray, power up the board and look for the part which melts the freeze really fast. Not delighted about powering up the board like this but I think it is a reasonable plan?

07 December 2016, 02:35	#6
jimbob Registered User Join Date: May 2006 Location: Kilmacolm Age: 46 Posts: 632	Got the bastard! Turned out to be a ceramic decoupling cap at the video DAC. Freeze spray worked a treat, it was very obvious which component was cooking itself when it got a dose. Everything seems to be working properly now including the new keyboard controller and the original CIA which is now socketed.

08 December 2016, 13:04	#7
hooverphonique ex. demoscener "Bigmama" Join Date: Jun 2012 Location: Fyn / Denmark Posts: 1,624

09 December 2016, 10:55	#8
JonSick Registered User Join Date: Oct 2006 Location: London UK Posts: 255	What a fantastic idea... freeze spray. I'm nicking that one!

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)