Unix True 64 - API CS20 / DS20L power supply flaw

Author

API CS20 / DS20L power supply flaw

SDF Poster

2007-07-23, 7:20 pm

I've contacted Microway to see if they have the schematics for the API
CS20 power supply, however, they do not. It seems they only purchased a
portion or all of API's production stock at their liquidation.

I don't know how compatible the DS20L's power supply is, though I'm going
to guess that its nearly exactly. However, I believe the DS20L would have
RMC which was never implemented (unfortunately!) on the CS20. That alone
would make me think there had to be some modifications done to the API
design prior to Compaq releasing the DS20L.

My question is, has anyone discovered a common component failure in the
CS20 power supply? I think anyone who has dealt with multiple units knows
exactly what I'm talking about.

1. Power is applied, a yellow LED illuminates, a green flashes
2. The power button is pressed
3. A relay clicks
4. The flashing green LED goes steady red.

I'm assuming is most likely a bad cap. But before I go in desoldering
parts to test them, I wanted to see if anyone had identified the culprit
and even better, how to resolve the issue.

API made a change to the PS at one point by removing the upper shielding
and replacing it with a thin hard plastic sheet. Obviously the over
heating a component or components was the primary cause of power supply
failure.

I've got a number of CS20s that have been offline due to PS failure and
I'd like to get them going. I'm confident that if it is something as
simple as a fused cap that I'll be able to bring them back on line. I've
heard that there is an IC component of the PS that can fail .. if that is
the case, I'm screwed... I think. Anyone have the schematics? or even
better, a solution?

David Turner, Island Computers

2007-07-26, 7:20 pm

I can get new Power supplies for the CS20/DS20L for $750 each

David

--
David B Turner
Island Computers US Corp
2700 Gregory St, Suite 180
Savannah GA 31404

T: 877-6364332 x201
Intl: 001 912 447 6622

E: dturner@islandco.com
F: 912 201 0402
W: http://www.islandco.com

"SDF Poster" <martians@sdf.lNoOnSePsAtMar.org> wrote in message
news:f8370v$n07$1@chessie.cirr.com...
> I've contacted Microway to see if they have the schematics for the API
> CS20 power supply, however, they do not. It seems they only purchased a
> portion or all of API's production stock at their liquidation.
>
> I don't know how compatible the DS20L's power supply is, though I'm going
> to guess that its nearly exactly. However, I believe the DS20L would have
> RMC which was never implemented (unfortunately!) on the CS20. That alone
> would make me think there had to be some modifications done to the API
> design prior to Compaq releasing the DS20L.
>
> My question is, has anyone discovered a common component failure in the
> CS20 power supply? I think anyone who has dealt with multiple units knows
> exactly what I'm talking about.
>
> 1. Power is applied, a yellow LED illuminates, a green flashes
> 2. The power button is pressed
> 3. A relay clicks
> 4. The flashing green LED goes steady red.
>
> I'm assuming is most likely a bad cap. But before I go in desoldering
> parts to test them, I wanted to see if anyone had identified the culprit
> and even better, how to resolve the issue.
>
> API made a change to the PS at one point by removing the upper shielding
> and replacing it with a thin hard plastic sheet. Obviously the over
> heating a component or components was the primary cause of power supply
> failure.
>
> I've got a number of CS20s that have been offline due to PS failure and
> I'd like to get them going. I'm confident that if it is something as
> simple as a fused cap that I'll be able to bring them back on line. I've
> heard that there is an IC component of the PS that can fail .. if that is
> the case, I'm screwed... I think. Anyone have the schematics? or even
> better, a solution?
>
>

SDF Poster

2007-07-27, 1:19 am

David Turner, Island Computers <dturner@no-spam-islandco.com> wrote:
>I can get new Power supplies for the CS20/DS20L for $750 each

Thanks David. I know you can do that, but I have a stack of 24 power
supplies that need repair. Can you help with that? I don't see why they
should all be scraped just because they've failed. Fixing them, or
bypassing their 'smart soft power broken feature' would really be the best
solution than just buying new supplies that could possibly end up in the
same state.

The design has a flaw and it needs to be fixed. Do you have the
schematics?

David Turner, Island Computers

2007-07-30, 1:20 pm

Sadly we are not (nor is anyone I know at HP) privy to any of that kind of
information

I think the PS states "no user servicable parts inside" does it not?

We have a guy here that might be able to play with one but the labor cost
alone would require that you sit down and breath deeply before you look at
the quote from us for this kind of reworking.

My bet is I can get these NEW for less than $300 if I push it

David

"SDF Poster" <martians@sdf.lNoOnSePsAtMar.org> wrote in message
news:f8bjt3$5r3$1@chessie.cirr.com...
> David Turner, Island Computers <dturner@no-spam-islandco.com> wrote:
>
> Thanks David. I know you can do that, but I have a stack of 24 power
> supplies that need repair. Can you help with that? I don't see why they
> should all be scraped just because they've failed. Fixing them, or
> bypassing their 'smart soft power broken feature' would really be the best
> solution than just buying new supplies that could possibly end up in the
> same state.
>
> The design has a flaw and it needs to be fixed. Do you have the
> schematics?
>

bob.birch@gmail.com

2007-07-30, 7:19 pm

On Jul 23, 5:36 pm, marti...@sdf.lNoOnSePsAtMar.org (SDF Poster)
wrote:
> I've contacted Microway to see if they have the schematics for the API
> CS20 power supply, however, they do not. It seems they only purchased a
> portion or all of API's production stock at their liquidation.
>
> I don't know how compatible the DS20L's power supply is, though I'm going
> to guess that its nearly exactly. However, I believe the DS20L would have
> RMC which was never implemented (unfortunately!) on the CS20. That alone
> would make me think there had to be some modifications done to the API
> design prior to Compaq releasing the DS20L.
>
> My question is, has anyone discovered a common component failure in the
> CS20 power supply? I think anyone who has dealt with multiple units knows
> exactly what I'm talking about.
>
> 1. Power is applied, a yellow LED illuminates, a green flashes
> 2. The power button is pressed
> 3. A relay clicks
> 4. The flashing green LED goes steady red.
>
> I'm assuming is most likely a bad cap. But before I go in desoldering
> parts to test them, I wanted to see if anyone had identified the culprit
> and even better, how to resolve the issue.
>
> API made a change to the PS at one point by removing the upper shielding
> and replacing it with a thin hard plastic sheet. Obviously the over
> heating a component or components was the primary cause of power supply
> failure.
>
> I've got a number of CS20s that have been offline due to PS failure and
> I'd like to get them going. I'm confident that if it is something as
> simple as a fused cap that I'll be able to bring them back on line. I've
> heard that there is an IC component of the PS that can fail .. if that is
> the case, I'm screwed... I think. Anyone have the schematics? or even
> better, a solution?

A common problem in DEC TPM :
Bad modules, power supply, etc but no prints,
no manuals or parts sources (reasonable).

DEC TPM solutions sometimes, require a little creative techniques:

1.Go over the PS and determine, if possible, where
it was manufactured and who actually made it. Or
ask everyone "who made it and where".

2. Google manufacturer "XXX and power supply engineer".
Review the google results and quite often you'll get
An on line resume from someone from that company
who worked on their power supplies. Call them and
ask them if they can help in diagnosing the problem,
for a fee, or know of someone who could help.
Done this several times, works great.

3. Find the city where the PS was made and using "other
Methods" find the local bar where employees hang out.
Have a friend, associate stop by the bar an quiz
Employees on who could help with your problem.

4. Numerous other legal, "edgy methods". After 30+ years
in TPM, never had a part I couldn't find or problem I
couldn't fix at a reasonable cost ;-)

If you have a stack of dead PS's it's gotta
be an easy fix and someone knows it.

SDF Poster

2007-07-31, 1:22 am

<bob.birch@gmail.com> wrote:
>DEC TPM solutions sometimes, require a little creative techniques:

Bob! That has got to work. ;-) Hopefully old DEC trolls are monitoring
USENET because thankfully this power supply proudly states it was
manufactured in the USA (what a rarity). The part numbers are DECish,
no doubt and no surprise. We all know API was the heart of the alpha
project ripped out of the ribcage of digital when the corpse was dropped
shipped on Compaq's door step ... though I'm pretty sure it wasn't via
Route 62.

Anyway .. I'm thankful with your suggestions, though I need to splash in
the water a bit more to hopefully get more attention on the matter. I'm
thankful that Island has said that they could possibly do $300 a supply
with a bulk order. I think that is very generous of them and definitely
is an option. Still, I feel the same way you do. Here lies a stack of 24
power supplies that *use* to work and that all exhibit the same symptom...
if I can save them from going in the Chinese electronics garbage
container, I'd rather try that first.

And yeah, of course I'd hire a DEC engineer to turn on his/her scope
and burn some solder.

Pictures of the part are here.. I didn't take off the heat sinks, but I
suppose if I did thats as far as a screw driver would take me.

http://sdf1.org/curiosity/api-cs20-ps/

Unfortunately very little evidence to go on.

Stephen

bob.birch@gmail.com

2007-07-31, 1:22 am

On Jul 30, 10:49 pm, marti...@sdf.lNoOnSePsAtMar.org (SDF Poster)
wrote:
> <bob.bi...@gmail.com> wrote:
>
> Bob! That has got to work. ;-) Hopefully old DEC trolls are monitoring
> USENET because thankfully this power supply proudly states it was
> manufactured in the USA (what a rarity). The part numbers are DECish,
> no doubt and no surprise. We all know API was the heart of the alpha
> project ripped out of the ribcage of digital when the corpse was dropped
> shipped on Compaq's door step ... though I'm pretty sure it wasn't via
> Route 62.
>
> Anyway .. I'm thankful with your suggestions, though I need to splash in
> the water a bit more to hopefully get more attention on the matter. I'm
> thankful that Island has said that they could possibly do $300 a supply
> with a bulk order. I think that is very generous of them and definitely
> is an option. Still, I feel the same way you do. Here lies a stack of 24
> power supplies that *use* to work and that all exhibit the same symptom...
> if I can save them from going in the Chinese electronics garbage
> container, I'd rather try that first.
>
> And yeah, of course I'd hire a DEC engineer to turn on his/her scope
> and burn some solder.

>
> Pictures of the part are here.. I didn't take off the heat sinks, but I
> suppose if I did thats as far as a screw driver would take me.
>
> http://sdf1.org/curiosity/api-cs20-ps/
>
> Unfortunately very little evidence to go on.
>
> Stephen

Good pics, shows its 500W DC output with
12 vdc @ 40a, 5vdc @ 1am, made in 2000

Looks like they made 4 different PS:
ATX type 300/400/600W with
+3.3vdc, +5,+12,-5,-12vdc with thermal
sensors, fan and voltage sensors.
Slick video at
http://www.dveemedia.com/Api/index.html

Maybe just a bad sensor

SDF Poster

2007-07-31, 7:17 am

<bob.birch@gmail.com> wrote:
>
>Good pics, shows its 500W DC output with
>12 vdc @ 40a, 5vdc @ 1am, made in 2000
>
>Looks like they made 4 different PS:
>ATX type 300/400/600W with
>+3.3vdc, +5,+12,-5,-12vdc with thermal
>sensors, fan and voltage sensors.
>Slick video at
>http://www.dveemedia.com/Api/index.html
>
>Maybe just a bad sensor

Well this is exactly what I imagine it to be. I'm not exactly sure where
they would be, however near the 1800uF orange cap (with 'OK') on it there
looks to be two wafer components that could be those sensors. They
have 0044 / ZR00J on them.

The grey block is WXP 105K X2 which looks like its an RFI suppressor
The black block is a relay (JW5211 / 901 19F)

Everything else is pretty easy to identify .. except for the mystery under
the heat sinks.