I have actually.
It happened so often that I made it a habit to label every part I ever tested.

If it worked, I'll write something like "tested" with a curly mark (dunno what it's called in English) so it's obvious that it worked the last time I checked.

If it doesn't work properly, I'll mark the card with what it did do (example: If a VGA card didn't display anything, I'll write only that it didn't seem to give any display, NOT that it is broken!) and put it in my "tested once but so far it hasn't worked yet"-box.

More then once (actually many many times!) I would go through this box of parts again, give everything an extra clean and visual check and use another test setup and I guess that about half the parts that didn't want to cooperate the first time, would actually make progress in newer attempts. I actually will write down how many (failed) attempts I have done. If it didn't work correctly 3 times or more, then it's probably broken (probably!). This method is also handy when it comes to parts that are flaky (like testing memory modules in ASUS A7V133 boards, this particular board seemed to be pretty critical in mixing memory modules).

Important though is to label all tested parts thoroughly, so don't write down "it didn't work so it MUST be broken!!!", keep calm and diagnose the problem for what it really seems to be.

Tetrium wrote:

I have actually. It happened so often that I made it a habit to label every part I ever tested. […]
Show full quote

I have actually.
It happened so often that I made it a habit to label every part I ever tested.

If it worked, I'll write something like "tested" with a curly mark (dunno what it's called in English) so it's obvious that it worked the last time I checked.

If it doesn't work properly, I'll mark the card with what it did do (example: If a VGA card didn't display anything, I'll write only that it didn't seem to give any display, NOT that it is broken!) and put it in my "tested once but so far it hasn't worked yet"-box.

Important though is to label all tested parts thoroughly, so don't write down "it didn't work so it MUST be broken!!!", keep calm and diagnose the problem for what it really seems to be.

Hey Tetrium,

Broken is just another word for non-functional... whatever the reason or failure may be. I made this distinction as many people just dump / throw away parts just because they can't get them to work... when the fault might not actually be that part!

My repair box is really there for when I feel like fault finding and fixing something as I do from time to time. Not always as some things do take time especially when checking the traces / tracks etc for finding any "open connections" and to close them back up with a wire bridge.

Today not everything was working again from the "for repair" box, but a few things certainly! 😁 The rest will have to wait until I have time to do the continuity test between points and heat the soldering iron up to solder connections. I have already repaired a few of my parts this way!

I think I will take your advice though and start to label things more and get better organized... you're right, it will definitely help ID parts and their status more quickly in future.

The curvy mark you're talking about... is possibly a "tick".

It'd be interesting to check your PSU and see whether it's working correctly or not. It's possible that the parts you have that don't work assume a high quality PSU with good filtering on the output, but your PSU doesn't have good filtering to reduce cost. Newer parts might have additional filtering on the boards themselves, so they work better with cheap PSUs, but the old parts don't so they misbehave.

It could also be something as simple as one of the voltages being a bit low. Some motherboards seem to have a lot of voltage regulators on them to get exactly the voltages they need, whereas older ones were more likely to use the PSU's power directly. So this could also explain it, depending on the age of the boards you are testing (since you say "AT case" this may not be the issue.)

Malvineous wrote:

It'd be interesting to check your PSU and see whether it's working correctly or not. It's possible that the parts you have that don't work assume a high quality PSU with good filtering on the output, but your PSU doesn't have good filtering to reduce cost. Newer parts might have additional filtering on the boards themselves, so they work better with cheap PSUs, but the old parts don't so they misbehave.

It could also be something as simple as one of the voltages being a bit low. Some motherboards seem to have a lot of voltage regulators on them to get exactly the voltages they need, whereas older ones were more likely to use the PSU's power directly. So this could also explain it, depending on the age of the boards you are testing (since you say "AT case" this may not be the issue.)

The PSU does work... only it decides which motherboard it wants to work with. All the motherboards i'm using are from late 80's / early 90's... 384 / 486 and are pretty much all built the same utilizing 5v - 12v DC running through various tantalum caps. Some of the boards have the odd electrolytic so you could be right... I will need to invesigate this further.

Glad a few things are not so dead after all and i can sort an unbais rig for testing.

Think we all have had parts that work in 1 system that don't in another for no good reason.
Depending on what the part is how long I will keep it around. A nothing special video card may only get tested in a 2nd system and thrown out if it fail, Even if a don't really think its the cards fault.
Something I put value on will probably keep going back into storage no matter how many systems it fails on.

386_junkie wrote:

Malvineous wrote:

The PSU does work... only it decides which motherboard it wants to work with.

This sounds like the caps are on their way out.

PCBONEZ wrote:

This sounds like the caps are on their way out.

Possibly, it would not surprize me, I believe the PSU to be the original suplied with the desktop... though I have not yet been inside this particular PSU so I can't comment, as yet. I don't even think it would be worthwile to replace them either even if they are on their way out... a decent PSU new in the UK will go for about £10 - £20.

I may do this soon then and prevent any potential failures to the current AT system via power sag / surges.

386_junkie wrote:

Tetrium wrote:

with a curly mark (dunno what it's called in English)

The curvy mark you're talking about... is possibly a "tick".

Off-topic: I've been trying to figure this out, can you type the punctuation mark Tetrium?

Is it:
a. '
b. "
c. { or }
d. ( or )
e. [ or ]
f. < or >
g. ~
h. `
i. ^
j. @
k. something else?

😀

Stiletto wrote:

Off-topic: I've been trying to figure this out, can you type the punctuation mark Tetrium? […]
Show full quote

386_junkie wrote:

Tetrium wrote:

with a curly mark (dunno what it's called in English)

The curvy mark you're talking about... is possibly a "tick".

Off-topic: I've been trying to figure this out, can you type the punctuation mark Tetrium?

Is it:
a. '
b. "
c. { or }
d. ( or )
e. [ or ]
f. < or >
g. ~
h. `
i. ^
j. @
k. something else?

😀

It was actually pretty hard to find, but this is a pic of what I meant
http://bin.snmmd.nl/m/m1fymkpwhutl.jpg
In Dutch it's called a "krul", which stands for "curl".

It's the symbol that my teachers used when I did a school test and answered the question correctly and as it's more distinct from the checkmark or any other symbols and can be written down very quickly and already had the "this is correct" meaning to me, I just kept on using it for identifying correctly working parts.

And I forgot to mention that I usually also mark all parts I get with the date I got the parts and sometimes how I got it or where it came from, but this also depends on if it's actually worth it (I typically don't bother for the more tiny parts like memory modules as they are really small, I just toss those in a "didn't post" or "lots of BSoD" box without adding any further marks).

Ah, I think that odd mark is unknown in the U.S., we would probably use a checkmark instead. 😀

Interesting reading:
http://www.shadycharacters.co.uk/2012/04/name … -approval-curl/
http://graphicdesign.stackexchange.com/questi … tch-called-krul

Stiletto wrote:

Ah, I think that odd mark is unknown in the U.S., we would probably use a checkmark instead. :) […]
Show full quote

Ah, I think that odd mark is unknown in the U.S., we would probably use a checkmark instead. 😀

Interesting reading:
http://www.shadycharacters.co.uk/2012/04/name … -approval-curl/
http://graphicdesign.stackexchange.com/questi … tch-called-krul

I considered the checkmark (and I actually used it for a little), but the problem is that the checkmark is basically just 2 stripes and the curly mark is just easier on the eye (I can spot it with more ease, making working with my parts more fluent) as the 2 stripes can more easily be confused with other stripy things (like a pen that has it's ink stuck so I have to scratch the pen on some surface to make it write again, which leave ...yup..stripes 🤣 ).

The curly mark is just one of those subtle things that make my life just a little bit easier.

PSUs require a certain load on the output side. If it is too low they wont switch on. The load threshold can be different.

elianda wrote:

PSUs require a certain load on the output side. If it is too low they wont switch on. The load threshold can be different.

or you would ground the PS_on wire

saturn wrote:

elianda wrote:

PSUs require a certain load on the output side. If it is too low they wont switch on. The load threshold can be different.

or you would ground the PS_on wire

I think you missed his point.
Not all.. but many.. PSUs (least modern ones) will not start without a load on them. (Even if you jump PS_On.)
Usually connecting a hard drive is enough to get them to go.
The drive does not need I/O connected, just power.
.
I've run across some that needed two HDDs to have enough load to start.
Those were high power server PSUs though. Not typical home PC stuff.
.

PCBONEZ wrote:

Usually connecting a hard drive is enough to get them to go. The drive does not need I/O connected, just power. . I've run acros […]
Show full quote

Usually connecting a hard drive is enough to get them to go.
The drive does not need I/O connected, just power.
.
I've run across some that needed two HDDs to have enough load to start.
Those were high power server PSUs though. Not typical home PC stuff.
.

You've made an interesting point there.

Ussually when I initially test boards, I do so with minimal load... or at least, with the fewest possible points of failure so as just to test the individual component itself with only the necessary parts to POST.

It never occured to me that PSU's could detect and require a load threshold to be crossed before powering on... most I own or have seen can be switched on with nothing attached at all, hence why I've taken it for granted.

The first time I ran into it I was really stumped.
I pulled a PSU out of a working system and decided to check it's voltages, but it would not start on the workbench.
Then someone suggested adding a HDD. I did and it worked.
I felt pretty dumb. At my tech level I should have already known that by then.
.

When things no longer work I get annoyed and throw them out. 😀

PeterLI wrote:

When things no longer work I get annoyed and throw them out. 😀

That doesn't work out well when your profession is to repair things.
It would be like putting your in-box (and your paycheck) into the trash can.
.

PCBONEZ wrote:

That doesn't work out well when your profession is to repair things. It would be like putting your in-box (and your paycheck) in […]
Show full quote

PeterLI wrote:

When things no longer work I get annoyed and throw them out. 😀

That doesn't work out well when your profession is to repair things.
It would be like putting your in-box (and your paycheck) into the trash can.
.

No, I can totally understand what Peter is saying... and in a display of dissatisfaction from it's failure, you have to take a hammer to it just so you can be totally sure its broke, it's a form of closure.

The user below didn't have it in him to go all the way...