Amiga.org

Amiga computer related discussion => Amiga Hardware Issues and discussion => Topic started by: orange on June 12, 2005, 10:07:46 PM

Title: Decoupling capacitors
Post by: orange on June 12, 2005, 10:07:46 PM
Are they really needed? I've checked EVERY bloody trace to every DRAM chip on {bleep}ty a501 with batt leakage. Also replaced 3 capacitors and checked resistors. But there are some 16(!) decoupling tantal capacitors that arent checked/replaced. they are expensive to replace and I don't know capacitance, anyway. So I thought about simply disconnecting them if they are not necessary?  :roll:
Title: Re: Decoupling capacitors
Post by: mattabat on June 13, 2005, 01:11:27 AM
Well, are the capacitors blown or not?
Quick examination should tell you.
I'd suggest if a capacitor is there, it's needed.
Can't comment on price, but I think you might be pleasantly surprised...
"I don't know capacitance, anyway." Capacitance is measured in farads (F). Match them and you're set. Get a table if they aren't marked clearly. Sometimes you can measure the capacitance even if it's blown. Remove and test :)
Title: Re: Decoupling capacitors
Post by: Jeff on June 13, 2005, 01:25:25 AM
Decoupling capacitors are indeed necessary :-).  Watch the video of Jeri Ellsworth's talk at Stanford University.  She explains why they are so important. Great Story.

http://8bit.dk/jeri/jeri_ellsworth_talk_stansford_uni.avi

 
Title: Re: Decoupling capacitors
Post by: CodePoet on June 13, 2005, 04:45:29 AM
 
Quote

 Jeff wrote:
 Decoupling capacitors are indeed necessary :-).  Watch the video of Jeri Ellsworth's talk at Stanford University.  She explains why they are so important. Great Story.

http://8bit.dk/jeri/jeri_ellsworth_talk_stansford_uni.avi


Great video! :-D
Title: Re: Decoupling capacitors
Post by: orange on June 13, 2005, 07:55:16 AM
What kind of 'examination'? Capacitors look fine.
Could I use cheap alum ones instead of tantal?
What table? from where? I got several schematics. Besides, value of other capacitors and resistors are DIFFERENT from those written on schematics. There are some small numbers printed on capacitors, but they seem to be different for each one  :-? and its not something understandable like '10uF'
hmm, I probably should try measuring their capacitance..
How do I know which one is blown?
Title: Re: Decoupling capacitors
Post by: CodePoet on June 13, 2005, 11:29:15 AM
Not too sure without seeing a snapshot, from experience most decoupling caps i've seen on amiga hardware are 0.1uF ceramic (usually printed as 104 or 104K on the cap itself) for the ones between vrails and gnd. They can be a major pain to test without a decent multi/capmeter, and generally its worth replacing the lot since they're worh about $0.50AUD a pair.

I prefer to use monolythic caps for decoupling my microcontroller based apps (up to 20MHz), they'll set you back $0.50AUD a piece - I havent found any difference performance wise bewteen monolythic and tantalum when used purely for decoupling; both worked equally well.

Also, caps used between oscillators (crystals) and ground are usually 33pF afaik, 22pF should also be fine; there are generally only two of these per oscillator, one between OSCIN and GND, and one between OSCOUT and GND

Hope this helps
/Cp
Title: Re: Decoupling capacitors
Post by: Framiga on June 13, 2005, 11:34:23 AM
can you take a hires picture of the capacitors area and post the URL here?

Title: Re: Decoupling capacitors
Post by: orange on June 18, 2005, 08:25:22 PM
Here it is, finally:
http://qrange.150m.com/pictures/a501.htm    (900Kb!)

Title: Re: Decoupling capacitors
Post by: Doobrey on June 18, 2005, 09:50:12 PM
Like CodePoet said, just rip them all out and replace them.
Just checked on Maplin, they have them (ceramic disc 0.1uF 16v) for 5p each.

Did this 501 ever work in your A500? You say all the traces are intact, and the 501 is a simple design so I'm just wondering if the 501 could be OK and the A500 at fault.
 Maybe a previous owner tried the hack to get the 501 as chip ram or something?
Title: Re: Decoupling capacitors
Post by: orange on June 18, 2005, 10:07:26 PM
A500 works with other slow_fast 512K RAM expansions. I got A501 broken. Actually it did work only once before I disassembled it to remove dead batt. BTW, could it be that metal case is required for operation?  :-?
Title: Re: Decoupling capacitors
Post by: Framiga on June 18, 2005, 11:08:09 PM
no need to use tantalium (btw you should use not-polarized one).

Like CodePoet already said, 0.1µF (marked 104 or so) maybe plastic instead of ceramic (anyway its the same).

Bye

EDIT- if the caps installed aren't in short (misure it with a multimeter), don't replace it.
Title: Re: Decoupling capacitors
Post by: Zac67 on June 18, 2005, 11:26:51 PM
Yes: capacitors - apart from electrolytic ones - extremely rarely break. If they're not short they're most probably OK.

What is the problem with the board? Does the 500 run at all? Is the RAM recognized? Does Amy crash?
Title: Re: Decoupling capacitors
Post by: orange on June 18, 2005, 11:40:34 PM
none of capcitors are short.
A500 runs but always with fast = 0Kb, strangely CLOCK works! I  checked everything except chips (and dont know how to check them)

Someone said once that chips can be tested if they are on for 10 min or so and then checked if they are warm or not, could that be the answer? :-?
Title: Re: Decoupling capacitors
Post by: Framiga on June 19, 2005, 12:29:56 AM
most probably, a broken trace (due the battery leakage)

I don't like those "gremlins" near C59.

Have you checked if the +5volt, goes to the ICs?

Title: Re: Decoupling capacitors
Post by: orange on June 19, 2005, 12:32:31 AM
arrrr... ALL TRACES ARE GOOD!!!
The voltage is a bit lower, about 4.78 V..
Title: Re: Decoupling capacitors
Post by: minator on June 19, 2005, 03:57:45 AM
Quote
Quote:

Jeff wrote:
Decoupling capacitors are indeed necessary . Watch the video of Jeri Ellsworth's talk at Stanford University. She explains why they are so important. Great Story.

http://8bit.dk/jeri/jeri_ellsworth_talk_stansford_uni.avi


Very good!
Title: Re: Decoupling capacitors
Post by: Zac67 on June 19, 2005, 09:25:37 PM
4.78V on the A501 might be a bit low; if you measured the PS input it's way low - why don't you juice up the PS .2 or .3V?

If it's not voltage I'd guess one of the RAS, CAS or WE lines is dead, so the RAM can either not be addressed or not be written to...

Yes, dead RAMs most often get warm pretty fast.
Title: Re: Decoupling capacitors
Post by: Chain on June 19, 2005, 09:30:01 PM
Nice macrophoto :)
...and this is very oln 150ns chips
Title: Re: Decoupling capacitors
Post by: orange on June 19, 2005, 09:48:19 PM
I've measured again, its 4,84 V on all chips. They don't get hot but I don't know for how long A500 should be on.

BTW, if just one of them is faulty none works?
Title: Re: Decoupling capacitors
Post by: Zac67 on July 06, 2005, 07:17:30 PM
Still not working?

Voltage is not really awsome, but should work.

Well, if one chip is bad, Amy might see the error on memory sizing and doesn't add the memory at all. More probable though is the missing _EXRAM signal on pin 32. Is that grounded when the expansion is plugged in (I assume it's a non-ECS 500 w/ 512 KB chip and 512 KB pseudo fast ('Ranger')).

PS: bad chips get hot VERY fast - IF they get hot