rx2620 power supplies / VRM voltage low

[George Cornelius]

Anyone else had problems with rx2620 power supplies?

We have been helping someone in-house who has a laboratory
instrument with attached rx2620 and no hardware contract.

We are not hardware support people but got him around his
last two hardware problems - bad scsi cable, (soft?) ecc
errors that appeared to be interfering with boots - but
he now has chronic power issues.

The iLo MP shows a message like "VRM non-recoverable voltage
low - check all boards at this address". VRM's - voltage
regulator modules - are on the motherboard and generate
regulated voltages from the output of the power supplies.

Swapping out the power supply resolves this. Over and over
again. We get about a month out of a supply.

Just recently tried inserting two supplies (these are used/
refurbed, with 90 day warranty). One of these died within
a day or so - different error, indicating defunct supply.

Soon after - replacement supply not yet in - VRM voltage
low error returns.

"Check cards with this voltage" is troubling. I would
be tempted to pull out the storage array card - replaced
by a new one from the vendor when the scsi issue arose -
but hard to run without the application's storage and
it would appear the supplies really have gone bad.

Decoding the message to determine which voltage it is
is problematic. An HP person on one of their forums
seemed to state that it is HP proprietary and such
problems _must_ be turned over to HP Support.

Ideas?

Don't ask why they won't pay for support. Something
about the instrument being due for replacement. Research
funding goes through cycles and sometimes they're
pinching their pennies.

Oh - and did I tell you we have no Itanium experience?

George

[George Cornelius]

Post by George Cornelius
The iLo MP shows a message like "VRM non-recoverable voltage
low - check all boards at this address".

Sorry - it's all boards at this voltage.

This seems to mean PCI cards.

[abrsvc]

I may have a few options for you. Please contact me directly at the address below.

Dan

dansabrservicesATyahooDOTcom

[Hans Vlems]

So if you swap a power supply then the message goes away for a certain time?
Which seems to indicate that they deteriorate, possibly owing to a high load? Are there devices connected to the rx2620 that draw power, like an external USB disk.
If you could hook up volt meters to the power supply and monitor them daily you might be able to figure out which power rail is failing and measure the current drawn. Compare that with the system and parts specifications.
It's not that there are very fast disk drives installed that run too hot and draw too much power?
If so check your fans, you might want to do that anyway. If they're dirty and run less efficient they might use more current. I'm guessing that too much current is drawn somewhere which leads to a voltage drop.
Hans

[George Cornelius]

Post by Hans Vlems
So if you swap a power supply then the message goes away for a certain time?

Yes. This is basically a fatal message (actually level 5 - critical, I think).
Sometimes you can resume from it if you wait awhile, but it will degenerate
from there.

Post by Hans Vlems
Which seems to indicate that they deteriorate, possibly owing to a high load?
Are there devices connected to the rx2620 that draw power, like an external
USB disk. If you could hook up volt meters to the power supply and monitor
them daily you might be able to figure out which power rail is failing and
measure the current drawn. Compare that with the system and parts specifications.

Great idea. I don't know if the manual goes into that kind of detail, but I
suppose there are test points somewhere.

Post by Hans Vlems
It's not that there are very fast disk drives installed that run too hot
and draw too much power?

Just the standard internals, which are probably 15K rpm. I think
a boot disk and a spare and an empty slot. Everything else is in
an external Storageworks pedestal.

Post by Hans Vlems
If so check your fans, you might want to do that anyway. If they're
dirty and run less efficient they might use more current. I'm
guessing that too much current is drawn somewhere which leads to
a voltage drop.

Have thought about that from the point of view of power supply
cooling. If they have failed the system can detect that. If
they are inefficient or failing, or if there are airflow
problems, maybe not.

Worth a try.

Post by Hans Vlems
Hans

[Hans Vlems]

I' ve never seen an rx2620 from the inside, I only have an rx2600. Are there unused disk power connectors inside the box? That would provide access to +5V and +12V to begin with.

[Scott Dorsey]

Post by Hans Vlems
I' ve never seen an rx2620 from the inside, I only have an rx2600. Are there unused disk power connectors inside the box? That would provide access to +5V and +12V to begin with.

There's one big connector that provides power to the cpu board, and all of
the supply rails are accessible through that. It's a Molex so you can check
them without pulling the connector out, and it's color coded so you should
be able to figure out what the nominal voltages are without much work.
--scott

[David Froble]

Post by George Cornelius
Anyone else had problems with rx2620 power supplies?
We have been helping someone in-house who has a laboratory
instrument with attached rx2620 and no hardware contract.
We are not hardware support people but got him around his
last two hardware problems - bad scsi cable, (soft?) ecc
errors that appeared to be interfering with boots - but
he now has chronic power issues.
The iLo MP shows a message like "VRM non-recoverable voltage
low - check all boards at this address". VRM's - voltage
regulator modules - are on the motherboard and generate
regulated voltages from the output of the power supplies.
Swapping out the power supply resolves this. Over and over
again. We get about a month out of a supply.

Now, that really does not surprise me. Keep in mind that power is where
everything starts. The commodity stuff coming from China is a crap shoot. You
never know when you turn on the power switch of a new power supply whether
you'll get power, or sparks, smoke, and time for a new power supply. I've had
brand new stuff spark and smoke. No QC in China, that's left to the users. Of
course, the new stuff is much cheaper than things in the past. QC costs money.

Post by George Cornelius
Just recently tried inserting two supplies (these are used/
refurbed, with 90 day warranty). One of these died within
a day or so - different error, indicating defunct supply.

Refirbished - someone returned it as defective, stick it back on the shelf and
send it to the next customer who orders one ....

Post by George Cornelius
Soon after - replacement supply not yet in - VRM voltage
low error returns.
"Check cards with this voltage" is troubling. I would
be tempted to pull out the storage array card - replaced
by a new one from the vendor when the scsi issue arose -
but hard to run without the application's storage and
it would appear the supplies really have gone bad.
Decoding the message to determine which voltage it is
is problematic. An HP person on one of their forums
seemed to state that it is HP proprietary and such
problems _must_ be turned over to HP Support.
Ideas?

Ok, it may not always be the power supply. Perhaps some component is putting
too much of a load on a valid power supply, causing it to go bad. Having really
good stuff to test power supplies would be a good thing. Anybody know where to
get such?

HP wanting their hands in your wallet sure doesn't surprise me ....

[Scott Dorsey]

Post by George Cornelius
Anyone else had problems with rx2620 power supplies?
We have been helping someone in-house who has a laboratory
instrument with attached rx2620 and no hardware contract.
We are not hardware support people but got him around his
last two hardware problems - bad scsi cable, (soft?) ecc
errors that appeared to be interfering with boots - but
he now has chronic power issues.
The iLo MP shows a message like "VRM non-recoverable voltage
low - check all boards at this address". VRM's - voltage
regulator modules - are on the motherboard and generate
regulated voltages from the output of the power supplies.
Swapping out the power supply resolves this. Over and over
again. We get about a month out of a supply.

Where are you getting these supplies?

Post by George Cornelius
Just recently tried inserting two supplies (these are used/
refurbed, with 90 day warranty). One of these died within
a day or so - different error, indicating defunct supply.

Who is refurbing these?
Are they actually refurbing them, or are they just fixing whatever is
broken? Are all the electrolytic capacitors in them new?

Post by George Cornelius
Soon after - replacement supply not yet in - VRM voltage
low error returns.
Ideas?

Check ALL the power supply rails. See if one of them actually is low.
Swap the "bad" power supply into another machine and see if they give the
same error.

When supply rails are pulled down, it's almost always because of bad
electrolytic decoupling capacitors. If you buy a "refurbed" supply, it
should have new electrolytics throughout, and they should be high grade
high temperature ones.

If you have a supply that "tests good" but has 20-year old capacitors
throughout, it will not last very long before another cap fails in it.

Now, you might have bad electrolytics on the CPU board as well... and I
would certainly inspect the board for bulging or leaking ones even if
you don't have an ESR tester to check them in-circuit.

Does the customer have an electronics tech on staff who can check caps?

Post by George Cornelius
Don't ask why they won't pay for support. Something
about the instrument being due for replacement. Research
funding goes through cycles and sometimes they're
pinching their pennies.

That's how it goes.

Post by George Cornelius
Oh - and did I tell you we have no Itanium experience?

This is a good way to get some!
--scott

[George Cornelius]

Post by Scott Dorsey

Post by George Cornelius
Anyone else had problems with rx2620 power supplies?
We have been helping someone in-house who has a laboratory
instrument with attached rx2620 and no hardware contract.
We are not hardware support people but got him around his
last two hardware problems - bad scsi cable, (soft?) ecc
errors that appeared to be interfering with boots - but
he now has chronic power issues.
The iLo MP shows a message like "VRM non-recoverable voltage
low - check all boards at this address". VRM's - voltage
regulator modules - are on the motherboard and generate
regulated voltages from the output of the power supplies.
Swapping out the power supply resolves this. Over and over
again. We get about a month out of a supply.

Where are you getting these supplies?

Post by George Cornelius
Just recently tried inserting two supplies (these are used/
refurbed, with 90 day warranty). One of these died within
a day or so - different error, indicating defunct supply.

Who is refurbing these?
Are they actually refurbing them, or are they just fixing whatever is
broken? Are all the electrolytic capacitors in them new?

Thanks for the response. These are all great comments and
help us focus a bit. Yes, one of the electronics techs
is the one who referred the original scsi issue to us
after replacing array controller did not help. He also
helped with the purchase of the original replacement
supply, from an "approved" vendor. Lab supervisor thought
he was buying new but after it eventually failed, I
noticed corrosion on the case due to something having
dripped onto it from above - no doubt electrolytics
leaking from the supply above it!

So we lined up a local vendor and he has been buying from
them. Their supplies are significantly less expensive
than the first one he bought, and come with warranty,
but are failing as well. Looks like it's time for
a long discussion with our suppliers.

No time to write more, but this is very helpful. Thank
you.

Post by Scott Dorsey

Post by George Cornelius
Soon after - replacement supply not yet in - VRM voltage
low error returns.
Ideas?

Check ALL the power supply rails. See if one of them actually is low.
Swap the "bad" power supply into another machine and see if they give the
same error.
When supply rails are pulled down, it's almost always because of bad
electrolytic decoupling capacitors. If you buy a "refurbed" supply, it
should have new electrolytics throughout, and they should be high grade
high temperature ones.
If you have a supply that "tests good" but has 20-year old capacitors
throughout, it will not last very long before another cap fails in it.
Now, you might have bad electrolytics on the CPU board as well... and I
would certainly inspect the board for bulging or leaking ones even if
you don't have an ESR tester to check them in-circuit.
Does the customer have an electronics tech on staff who can check caps?

See above.

And I do have a Fluke meter that can go down to 10 or maybe even 1 nF
caps - I'll check if it is specified for in-circuit testing.

[...]

Post by Scott Dorsey

Post by George Cornelius
Oh - and did I tell you we have no Itanium experience?

This is a good way to get some!

Yes, several years' worth in just a few months.

George

[Stephen Hoffman]

So we lined up a local vendor and he has been buying from them. Their
supplies are significantly less expensive than the first one he bought,
and come with warranty, but are failing as well. Looks like it's time
for a long discussion with our suppliers.

Swap the parts that do work over into...

http://www.ebay.com/itm/HP-Itanium-rx2620-Rackmount-Server-AB332A-AB331-60101-/361473636652

http://www.ebay.com/itm/HP-AB332A-Integrity-rx2620-Rack-Mount-Server-w-RAM-Upgrade-4-x-1GB-NO-DRIVES-/151480810489

http://www.ebay.com/itm/HP-AD152A-RX2620-SERVER-Dual-1-4Ghz-AD122A-12GB-RAM-3x73GB-HDD-dvd-rom-/251562182532


do the license-transfer with HPE, and move on.

[George Cornelius]

X-News: eisner.encompasserve.org comp.os.vms:81443

Subject: Re: rx2620 power supplies / VRM voltage low
Date: Wed, 10 Feb 2016 11:29:50 -0500

So we lined up a local vendor and he has been buying from them. Their
supplies are significantly less expensive than the first one he bought,
and come with warranty, but are failing as well. Looks like it's time
for a long discussion with our suppliers.

Swap the parts that do work over into...
do the license-transfer with HPE, and move on.

Thanks. My recommendation from day 1, when there was one
available in Minneapolis and I could have made a trip up
there and gotten it same day.

Someone has already invested in a spare motherboard, though,
so if we think that will do it we may decide to give that
a try first.

George

--
Pure Personal Opinion | HoffmanLabs LLC

[Scott Dorsey]

Post by George Cornelius
Thanks for the response. These are all great comments and
help us focus a bit. Yes, one of the electronics techs
is the one who referred the original scsi issue to us
after replacing array controller did not help. He also
helped with the purchase of the original replacement
supply, from an "approved" vendor. Lab supervisor thought
he was buying new but after it eventually failed, I
noticed corrosion on the case due to something having
dripped onto it from above - no doubt electrolytics
leaking from the supply above it!

That is not a refurbed supply, that is a "junkyard" supply. When you buy
junkyard parts, you get what you pay for.

Post by George Cornelius
So we lined up a local vendor and he has been buying from
them. Their supplies are significantly less expensive
than the first one he bought, and come with warranty,
but are failing as well. Looks like it's time for
a long discussion with our suppliers.

Did you not ask them what they did when they "refurbed" them before you bought
them?

Just take one of the supplies and have it worked over by a legitimate power
supply tech. These things are not rocket science.
--scott

[Carl Friedberg]

HP service is reasonably priced. I believe we pay under $1,000/year for
Next Business Day hardware support for an RX2600 (not 2620) on a 3-year
prepaid contract. There was about a $1,300 charge for return to maintenance
(no inspection). If you need a referral, as usual I recommend David Witt at
Flagstaff Technology (contact me off list)

I know that research institutions can be very cheap, and they have graduate
students as slave labor, but surely having an instrument on maintenance is
not an unreasonable expense for a lab?

Carl

On Wed, Feb 10, 2016 at 11:43 AM, Scott Dorsey via Info-vax <

Post by Scott Dorsey

Post by George Cornelius
Thanks for the response. These are all great comments and
help us focus a bit. Yes, one of the electronics techs
is the one who referred the original scsi issue to us
after replacing array controller did not help. He also
helped with the purchase of the original replacement
supply, from an "approved" vendor. Lab supervisor thought
he was buying new but after it eventually failed, I
noticed corrosion on the case due to something having
dripped onto it from above - no doubt electrolytics
leaking from the supply above it!

That is not a refurbed supply, that is a "junkyard" supply. When you buy
junkyard parts, you get what you pay for.

Post by George Cornelius
So we lined up a local vendor and he has been buying from
them. Their supplies are significantly less expensive
than the first one he bought, and come with warranty,
but are failing as well. Looks like it's time for
a long discussion with our suppliers.

Did you not ask them what they did when they "refurbed" them before you bought
them?
Just take one of the supplies and have it worked over by a legitimate power
supply tech. These things are not rocket science.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."
_______________________________________________
Info-vax mailing list
http://rbnsn.com/mailman/listinfo/info-vax_rbnsn.com

[Stephen Hoffman]

Post by Carl Friedberg
HP service is reasonably priced. I believe we pay under $1,000/year for
Next Business Day hardware support for an RX2600 (not 2620) on a 3-year
prepaid contract. There was about a $1,300 charge for return to
maintenance (no inspection).

So ~US$4300, for three years? For comparison purposes and if you have
your own staff, that's a spare server off of eBay up front, plus a new
spare server each year, plus more than a little cash left over for
license transfers and lattés— if it comes to that. Now hauling
somebody with the necessary skills out to the site to swap parts isn't
cheap, but I'd wager that HPE is betting that those boxes won't fail
very often.

[Carl Friedberg]

Yes, but after the first year ($1,300 back-on-service + ~900) it is under
1,000 / year. It sounds like these folks are having problems both getting
parts and diagnosing the problem. IMO it's worth a grand to have someone
else take over these problems. I have more important things to do, and I
want the RX2600's to just work. Of course, in our case, HP has made no
calls on these 2 machines for almost 3 years, so yes, a big win for them.
But also for us, since if something happens, it's a call and HPE will fix
it. End of story. That's why people waste money on service contracts -- it
becomes someone else's problem.

I'm certainly capable of fixing most of the things that would likely go
wrong with one of these; but my time is better spent on what I do best.

Thanks for your answer. HPE is indeed making money on these contracts. This
field service contract business is so profitable, you have to wonder why
HP/HPE worked so hard to cut out VMS.

On Wed, Feb 10, 2016 at 4:19 PM, Stephen Hoffman via Info-vax <

Post by Carl Friedberg
HP service is reasonably priced. I believe we pay under $1,000/year for

Post by Carl Friedberg
Next Business Day hardware support for an RX2600 (not 2620) on a 3-year
prepaid contract. There was about a $1,300 charge for return to maintenance
(no inspection).

So ~US$4300, for three years? For comparison purposes and if you have
your own staff, that's a spare server off of eBay up front, plus a new
spare server each year, plus more than a little cash left over for license
transfers and lattés— if it comes to that. Now hauling somebody with the
necessary skills out to the site to swap parts isn't cheap, but I'd wager
that HPE is betting that those boxes won't fail very often.
--
Pure Personal Opinion | HoffmanLabs LLC
_______________________________________________
Info-vax mailing list
http://rbnsn.com/mailman/listinfo/info-vax_rbnsn.com

[Kerry Main]

-----Original Message-----
Carl Friedberg via Info-vax
Sent: 10-Feb-16 11:45 PM
Subject: Re: [New Info-vax] rx2620 power supplies / VRM voltage low
Yes, but after the first year ($1,300 back-on-service + ~900) it is under
1,000 / year. It sounds like these folks are having problems both getting
parts and diagnosing the problem. IMO it's worth a grand to have someone
else take over these problems. I have more important things to do, and I
want the RX2600's to just work. Of course, in our case, HP has made no
calls on these 2 machines for almost 3 years, so yes, a big win for them.
But also for us, since if something happens, it's a call and HPE will fix
it. End of story. That's why people waste money on service contracts -- it
becomes someone else's problem.
I'm certainly capable of fixing most of the things that would likely go
wrong with one of these; but my time is better spent on what I do best.
Thanks for your answer. HPE is indeed making money on these contracts. This
field service contract business is so profitable, you have to wonder why
HP/HPE worked so hard to cut out VMS.

One needs to remember that HP is a big company with depts. not always
having the same priorities. SW Engineering makes money off licenses,
not HW maintenance, so all Eng decisions are based on prod X license $'s
and not total $'s that come from prod X into the company.

HW support revenue drops because of poor decisions made by SW Eng
and Marketing ignoring the base Custs, but the HW support groups are
simply told to find other sources of revenue.

Gotta love big companies .. :-(

[snip..]


Regards,

Kerry Main
Kerry dot main at st

[Steven Schweda]

Post by George Cornelius
Anyone else had problems with rx2620 power supplies?

Not I. rx2600 is my newest, and it's too noisy for more
than occasional use.

Post by George Cornelius
The iLo MP shows a message like "VRM non-recoverable voltage
low - check all boards at this address". VRM's - voltage
regulator modules - are on the motherboard and generate
regulated voltages from the output of the power supplies.
Swapping out the power supply resolves this. Over and over
again. We get about a month out of a supply.

Has anyone considered the possibility that the VRMs "on
the motherboard" are bad, and not every main power supply in
the world? I know nothing, but I can imagine a case where a
fresh power supply might be just barely able to overcome a
faulty VRM, but not for long. Swapping main power supplies
may be much easier/cheaper than swapping VRMs, but how many
bad main power supplies do you think are out there?

How much would a whole spare system cost?

[George Cornelius]

Post by Steven Schweda
Has anyone considered the possibility that the VRMs "on
the motherboard" are bad, and not every main power supply in
the world? I know nothing, but I can imagine a case where a
fresh power supply might be just barely able to overcome a
faulty VRM, but not for long. Swapping main power supplies
may be much easier/cheaper than swapping VRMs, but how many
bad main power supplies do you think are out there?
How much would a whole spare system cost?

OK, well, thanks to the group.

The answer to my question: this is anomalous.

If this person has been getting supplies from people who
are not being up front with him, then that's the first
problem. The second is that he is asking people who are
not equipped to solve his problem to do his troubleshooting.

Yes, I have component level repair experience, but that
was a million years ago. I don't think I have ever taken
a soldering iron to a motherboard except once in the old
days when a soldered-in rechargeable CMOS battery failed,
I suppose a NiCad. So I probably won't be trying my hand
at such things today.

We do have a number of techs who could check the
capacitors and attach meters to test points. That
will likely not happen either except maybe to convince
the "approved" vendor who sent us the first failing
replacement that he needs to be considered for removal
from approved status.

By the way, it is apparently well known that there
was an era not so long ago when there were large
numbers of faulty electrolytics from China and
Taiwan making their way into PC's and servers,
ultimately resulting in recalls. Certain
manufacturers were affected more than others,
HP being perhaps one of those less affected.
I do have sitting in the hall an HP PC, one that
was shipped as a console for a GS1280, I am told,
that has video display issues along with bulging
capacitors on the motherboard.

I am going to discount that, though, and just
say that electrolytics are known to have finite
lifetimes and there are people who will just
turn around used parts with minimal effort
invested; and those who don't understand that
deserve what they get.

We are perfectly capable of buying a working
system and moving everything over to it, but
I wanted to make sure that there was not, say,
a known issue with power supplies in certain
classes of servers first.

So thank you all very much for your help.

George Cornelius

[abrsvc]

Post by George Cornelius

Post by Steven Schweda
Has anyone considered the possibility that the VRMs "on
the motherboard" are bad, and not every main power supply in
the world? I know nothing, but I can imagine a case where a
fresh power supply might be just barely able to overcome a
faulty VRM, but not for long. Swapping main power supplies
may be much easier/cheaper than swapping VRMs, but how many
bad main power supplies do you think are out there?
How much would a whole spare system cost?

OK, well, thanks to the group.
The answer to my question: this is anomalous.
If this person has been getting supplies from people who
are not being up front with him, then that's the first
problem. The second is that he is asking people who are
not equipped to solve his problem to do his troubleshooting.
Yes, I have component level repair experience, but that
was a million years ago. I don't think I have ever taken
a soldering iron to a motherboard except once in the old
days when a soldered-in rechargeable CMOS battery failed,
I suppose a NiCad. So I probably won't be trying my hand
at such things today.
We do have a number of techs who could check the
capacitors and attach meters to test points. That
will likely not happen either except maybe to convince
the "approved" vendor who sent us the first failing
replacement that he needs to be considered for removal
from approved status.
By the way, it is apparently well known that there
was an era not so long ago when there were large
numbers of faulty electrolytics from China and
Taiwan making their way into PC's and servers,
ultimately resulting in recalls. Certain
manufacturers were affected more than others,
HP being perhaps one of those less affected.
I do have sitting in the hall an HP PC, one that
was shipped as a console for a GS1280, I am told,
that has video display issues along with bulging
capacitors on the motherboard.
I am going to discount that, though, and just
say that electrolytics are known to have finite
lifetimes and there are people who will just
turn around used parts with minimal effort
invested; and those who don't understand that
deserve what they get.
We are perfectly capable of buying a working
system and moving everything over to it, but
I wanted to make sure that there was not, say,
a known issue with power supplies in certain
classes of servers first.
So thank you all very much for your help.
George Cornelius

I can say that the 3 RX2620s I have don't show this problem and never have. I will admit however, that I don't use them 24x7 either. These machines usually run about 10hr per day with some extended runs now and again. I do keep spare parts around just in case. I have seen only 1 PS fail and that was after a power surge from a defective UPS setup. Once I replaced the smoked parts, the unit worked fine. I have not yet seen any issues with capacitors in these supplies. And yes, I repair to the component level (even surface mount devices) when necessary.

Dan

[Steven Schweda]

I still know nothing, but the "HP Integrity rx2600 Server
and HP zx6000 Workstation Operations and Maintenance Guide"
(Document Part Number: 5991-5988) says that there are
distinctive patterns in the diagnostic LEDs for various
faults, including power supply and VRM over- and
under-voltage ("View the SEL for additional information.
For further assistance, contact your HP Support Engineer.")

If the rx2620 and its corresponding manual are similar,
then there may be some useful info/advice there. I'm still
not convinced that every power supply in the world is
defective, and if the system is complaining about the VRMs,
not the power supplies, then it may know more than any of us
about what's wrong. But what do I know? (And I bought two
rx2600 systems at the same time, so I have a spare one in
storage, just in case. Except for the CPU which was donated
to my (quieter) zx2000 to replace its slower one.)

[George Cornelius]

Post by Steven Schweda
I still know nothing, but the "HP Integrity rx2600 Server
and HP zx6000 Workstation Operations and Maintenance Guide"
(Document Part Number: 5991-5988) says that there are
distinctive patterns in the diagnostic LEDs for various
faults, including power supply and VRM over- and
under-voltage ("View the SEL for additional information.
For further assistance, contact your HP Support Engineer.")

We have been through all that, and, yes, the codes are there
at the front panel - sometimes. The manual states that
if you have an iLo MP the front panel status LEDs are
disabled (all the information is in the error logs
you can get to via the ILo) but if your system hits one
of these fatal errors you can eventually get it to show
up on the lights as well (pressing the halt button
or something).

I suppose there have been four failures since last
June, and if he had been sending back the old supply
each time he would have been getting the replacements
for free - until the vendor got tired of sending more
supplies. But it's time to swap in the spare motherboard,
or to do what Hoff (and I) suggested and buy a 2nd
system to swap the parts into.

The manual is somewhat ambiguous but the error does
show up under motherboard failures, so we suggested
either a power supply or a motherboard, and he bought
both. We thought of a motherboard replacement as
being a more invasive fix, so have resisted putting
it in, but now may be the time.

Regarding the online forums, we found some tantalizing
but in the end ambiguous items:

o A thread in the HP forums with the same error
codes, which someone decoded with an, apparently,
HP internal tool called IPF2. This second person
was corrected and told the system in question was
not IPF2 so the decoding would not apply. Thread
was resurrected maybe a year later - maybe 2008 -
with someone claiming to have an _rp_ 2620, and
that is when the same responder was asked to _not_
supply information using the IPF2 tool and the
poster told to call field service.

What we would have been able to obtain if the
thread continued was the decoding of the field
that specified _which_ voltage sensor was
returned the bad reading.

One remark, maybe in response to the first
problem, was that the message indicated the
sensor itself was bad, thus muddyng the
waters even further.

o MG, on another forum, as eMGee I think,
responded to someone's question having to do
with a problem which might be solved by a
motherboard replacement and pointed out that
there are UUID's involved somehow, and the
new motherboard's UUID would not match what
was expected, in which case everything would
work except it would not go naturally into EFI
mode and you would have to do that by playing
games with the iLo MP. But it might have had
to do with swapping, say a ??6200 motherboard
into an rx2600, and might not have applied
to my case. Still, it's confirmation that a
motherboard swap might not be as simple as
it might appear on the surface.

Post by Steven Schweda
If the rx2620 and its corresponding manual are similar,

What you are saying sounds exactly like what's in
the rx2620 manual.

Post by Steven Schweda
[...] I bought two
rx2600 systems at the same time, so I have a spare one in
storage, just in case.

Spares are easy to come by. My experience is, however, that
bureaucracies and eBay just don't play well together, and
that makes it all more difficult and expensive.

George

[Scott Dorsey]

Post by George Cornelius
We have been through all that, and, yes, the codes are there
at the front panel - sometimes. The manual states that
if you have an iLo MP the front panel status LEDs are
disabled (all the information is in the error logs
you can get to via the ILo) but if your system hits one
of these fatal errors you can eventually get it to show
up on the lights as well (pressing the halt button
or something).

The codes are interesting.... but use the meter, and the meter will tell
you what is really going on.

Post by George Cornelius
I suppose there have been four failures since last
June, and if he had been sending back the old supply
each time he would have been getting the replacements
for free - until the vendor got tired of sending more
supplies. But it's time to swap in the spare motherboard,
or to do what Hoff (and I) suggested and buy a 2nd
system to swap the parts into.

Might be, but the meter will tell you if that's useful or not. If the
supply rail values all look good, by all means swap in a new motherboard.
If the supply rails are sagging, it's time to find out why.

Post by George Cornelius
o A thread in the HP forums with the same error
codes, which someone decoded with an, apparently,
HP internal tool called IPF2. This second person
was corrected and told the system in question was
not IPF2 so the decoding would not apply. Thread
was resurrected maybe a year later - maybe 2008 -
with someone claiming to have an _rp_ 2620, and
that is when the same responder was asked to _not_
supply information using the IPF2 tool and the
poster told to call field service.
What we would have been able to obtain if the
thread continued was the decoding of the field
that specified _which_ voltage sensor was
returned the bad reading.
One remark, maybe in response to the first
problem, was that the message indicated the
sensor itself was bad, thus muddyng the
waters even further.

This may well be the problem. THIS is why the first thing you need to do
is measure the actual rail voltages. If the voltages are good, the sensor
is bad and the motherboard should go (or at least the shunt resistors on
the motherboard should be checked).

[random blathering removed here]

Post by George Cornelius
Spares are easy to come by. My experience is, however, that
bureaucracies and eBay just don't play well together, and
that makes it all more difficult and expensive.

Spares that have been sitting on the shelf for a decade are apt to have
problems resulting from years of sitting on the shelf. That includes
just about any rubber parts failing, thermal grease hardening up, and
electrolytic capacitors going bad. You don't want spares from random
sources if you can help it, you want spares that have been gone over and
refurbished by competent technicians.
--scott

[Hans Vlems]

The rx2600 mainboard swap issues msy have been my post. On an rx2600 I had to swap the motherboard to use the pci expansion cage. That fits onto a large connector with many pins soldered on said motherboard. I did just that and the motherboard did not recognize the cpu's id's. The replacement board was not new and had been configured before, obviously those cpu's had different id's. IIRC there is an on board tool to fix the issue but it's only accessible to a sufficiently privileged user. Not the box owner but an HP employee...
So i wouldn't recommend that solution though an rx2660 may behave differently.
Even then only do this with the correct cpu tool.
Hans

[George Cornelius]

The rx2600 mainboard swap issues may have been my post. On an rx2600 I had
to swap the motherboard to use the pci expansion cage. That fits onto a large
connector with many pins soldered on said motherboard. I did just that and
the motherboard did not recognize the cpu's id's. The replacement board was
not new and had been configured before, obviously those cpu's had different
id's. IIRC there is an on board tool to fix the issue but it's only access-
ible to a sufficiently privileged user. Not the box owner but an HP employee...

So if you replace an rx2600 motherboard with an rx2600 motherboard that has
been used before you will have this issue?

Doesn't that also imply that CPU's cannot be changed on a motherboard without
involving HP?

Sounds awfully restrictive.

And, depending on how you want to frame it, it's either a way to track
everything you do or a way to guarantee an ongoing income stream for HP.

So i wouldn't recommend that solution though an rx2660 may behave differently.

rx2620. I will try to check.

Even then only do this with the correct cpu tool.

Hardware tool? Like a mini torque wrench?

George

[Hans Vlems]

The uuid issue is a great tool to control the second hand market. I won't buy spare mainboards nor cpu upgrades for IA64 again, that's for sure whatever the model. A beautiful strategy, why don't eBay sue HP for restricting their business?

The rx2600 cpu's are locked in place with a screw mechanism. The screw has a very specific head and only the tool fits in it. Brute force works for removal but misaligns a center pin in the screw. Which makes locking very difficult. I did curse one of the previous owners of my rx2600 for this reason...
The phillips screwdriver needed to remove the cpu assembly is ver long, 4 inches.
A mechanical variant of the uuid approach.
No wonder customers love HP.

[Steven Schweda]

[...] The screw has a very specific head and only the tool fits in it.
Brute force works for removal but misaligns a center pin in the screw.
[...]

It's been years since I did it, and my memory is worth little, so I
know nothing, but I did transplant a 1.6gHz CPU from an rx2600 into a
zx2000 (replacing a 900MHz CPU there), and I don't remember using more
than simple tools to do it. That is, if I own a "special processor tool
(HP Part Number A7231-67046)" as specified in the manual, I've forgotten
about it, and it's not in the box with the old CPU and power module.

From what I can see (on the old CPU), the captive heatsink screws are
plain T-15 (with a slot which might allow a plain flat-blade screwdriver
to work). I don't see a "center pin" in mine. Do you have
tamper-resistant screws (with the center pin)? Tools for those are
cheap at the local cheap-Chinese-tool store. It's possible that the
screws on the newer/faster CPUs were more exotic, but I don't remember
any serious problems with any of them.

The phillips screwdriver needed to remove the cpu assembly is ver long, 4 inches.

Phillips? Not Torx? Four inches is long? If you want _long_, then
you want the foot-long T-15 tool needed to reach the case screws on an
early Macintosh. _That_'s long. ("That's not a knife...")

[David Froble]

Post by Steven Schweda
("That's not a knife...")

Oh, yeah! Arguably the best scene in that movie.

[Hans Vlems]

I know you can do work on an rx2600 without that tool. Saw the results in the box I own. Hence my earlier remark.
Hans
(thinking American screwdriver :-)

[George Cornelius]

Post by Hans Vlems
I know you can do work on an rx2600 without that tool.
Saw the results in the box I own. Hence my earlier remark.
Hans
(thinking American screwdriver :-)

So is it a simple modification to, say an
Allen wrench (a length of hexagonal rod)?
Like maybe a hole drilled in the center?

Specs say it is 2.5 mm - actually
quite small. 3.5 mm is a miniature
headphone plug, 2.5 mm its little brother,
often used with cell phones.

American screwdriver sounds like a
Phillips - tapered to almost a point
at the tip, roughly a plus sign in cross
section.

George

[Hans Vlems]

Ahem, an American screwdriver is UK English. It translates as a hammer :-)

[Steven Schweda]

Post by George Cornelius
So is it a simple modification to, say an
Allen wrench (a length of hexagonal rod)?
Like maybe a hole drilled in the center?

I assume that he's talking about the Torx screws, not the
hex-hole things. A Google search for, say:
tamper-resistant torx
offers many pictures.

Drilling your own hole in a hardened tool (even a cheap
Chinese one) is more trouble than buying one which was
drilled before hardening.

[Carl Friedberg]

While I don't relish a refernece to Amazon, here is one set for US 5.59

http://www.amazon.com/SE-7519SD-33-Piece-Security-Extension/dp/B002E0S2SQ/ref=sr_1_1?ie=UTF8&qid=1455596156&sr=8-1&keywords=torx+security+bits


On Mon, Feb 15, 2016 at 10:25 PM, Steven Schweda via Info-vax <

Post by Steven Schweda

Post by George Cornelius
So is it a simple modification to, say an
Allen wrench (a length of hexagonal rod)?
Like maybe a hole drilled in the center?

I assume that he's talking about the Torx screws, not the
tamper-resistant torx
offers many pictures.
Drilling your own hole in a hardened tool (even a cheap
Chinese one) is more trouble than buying one which was
drilled before hardening.
_______________________________________________
Info-vax mailing list
http://rbnsn.com/mailman/listinfo/info-vax_rbnsn.com

[George Cornelius]

Post by Carl Friedberg
While I don't relish a refernece to Amazon, here is one set for US 5.59
http://www.amazon.com/SE-7519SD-33-Piece-Security-Extension/dp/B002E0S2SQ/ref=sr_1_1?ie=UTF8&qid=1455596156&sr=8-1&keywords=torx+security+bits

Thanks. I've seen a lot of those screws so it's pretty
obvious that there are people making the tools that
drive them.

It's amaizing how many knowledgeable people there are in this
group, and what a diversity of experience is represented.

When you have a long thread like this one you - or at
least I - begin to wonder if such minor problems are worth
bothering the group with. But everyone's collective experience
in the end can be so valuable. So if some of us are a bit
on the wordy side, well, maybe you can skip past some of
that to get to the really useful information that comes
along in the responses.

Thanks again to everyone.

George

[George Cornelius]

Post by Steven Schweda

Post by George Cornelius
So is it a simple modification to, say an
Allen wrench (a length of hexagonal rod)?
Like maybe a hole drilled in the center?

I assume that he's talking about the Torx screws,
not the hex-hole things.

I'm not sure there was any goal in mind in his
last post other than getting to the punch line.

Post by Steven Schweda
tamper-resistant torx offers many pictures.
Drilling your own hole in a hardened tool (even a cheap
Chinese one) is more trouble than buying one which was
drilled before hardening.

Yep. 50+ years since I learned to heat treat tool
steel. You anneal it, work on it in that form,
then temper it afterwards.

My toolmaking project was a centerpunch. Others
made screwdrivers.

Project for an earlier phase of the class, sheet
metal fabrication, was to take some really hefty
1/8" aluminum sheet stock and form a heavy duty
chassis for a tube based 300 volt power supply
I was building in my home workshop (think amateur
radio).

Now Torx is often chrome plated or otherwise super
hardened, so likely a bit tougher than the tool
steel I had available.

So I won't do it. My girlfriend - gone now - had
a brother, not to mention an ex-husband, trained
in tool and die making at a school in your neck of
the woods (Dunwoodie), so there's no lack of access
to people who _can_ do it.

That brother had a small business here in town at
one time whose sign was based on a silhouette
of a man with a hammer fashioning, presumably,
a piece of red hot metal metal into something
of value.

I see an employee of his now has a welding shop
down the street from me with a sign that is
clearly based on the one his employer once had.

A hammer in the right hands creates works of
art. Or maybe just a little thing like a
screwdriver.

Thanks for the info. Last time you gave me
advice it had to do with using a 3100/30 in
diagnostic mode to change the auto-spin-up
settings on scrap RX23's.

George

[George Cornelius]

Post by George Cornelius
auto-spin-up
settings on scrap RX23's.

^^
RZ23's (and the like)
George

[Stephen Hoffman]

Post by George Cornelius
Spares are easy to come by. My experience is, however, that
bureaucracies and eBay just don't play well together, and that makes it
all more difficult and expensive.

Okay, if the purchasing department has "Best Practices" in play that
exclude eBay or Craigslist or related, then they can and usually also
should pay HPE for service, or pay one of the reputable used-equipment
vendors that'll offer systems with warranties. In general, these
folks might want to review Chapter 55612.3.5 in the applicable "Best
Practices for Continued Employment via Blame Shifting" where it states
that all critical servers must maintain support contracts from the list
of approved providers (see section 55612.3.476 and updates in BPCEvBS
addendum 2016.02 or current) at all times, or digitally-signed copies
of the applicable Business Practices Exception (C-level exception or
trap to interrupt vector badge 00001) must be maintained among the
server manager, the department, the purchasing department and internal
IT and other applicable groups, per BPCEvBS section
55612.3.5772.x4.375. Or just order the eBay spare server out of the
Division Shadow IT^W^WGroup Latté Fund.

[David Froble]

Post by Stephen Hoffman

Post by George Cornelius
Spares are easy to come by. My experience is, however, that
bureaucracies and eBay just don't play well together, and that makes
it all more difficult and expensive.

Okay, if the purchasing department has "Best Practices" in play that
exclude eBay or Craigslist or related, then they can and usually also
should pay HPE for service, or pay one of the reputable used-equipment
vendors that'll offer systems with warranties. In general, these folks
might want to review Chapter 55612.3.5 in the applicable "Best Practices
for Continued Employment via Blame Shifting" where it states that all
critical servers must maintain support contracts from the list of
approved providers (see section 55612.3.476 and updates in BPCEvBS
addendum 2016.02 or current) at all times, or digitally-signed copies of
the applicable Business Practices Exception (C-level exception or trap
to interrupt vector badge 00001) must be maintained among the server
manager, the department, the purchasing department and internal IT and
other applicable groups, per BPCEvBS section 55612.3.5772.x4.375. Or
just order the eBay spare server out of the Division Shadow IT^W^WGroup
Latté Fund.

Buy a spare on EBAY, with guarantee, then sell to company for double what you
paid. Or triple.

Steve is having way too much fun with this ....

:-)

[Paul Sture]

Post by David Froble

Post by Stephen Hoffman

Post by George Cornelius
Spares are easy to come by. My experience is, however, that
bureaucracies and eBay just don't play well together, and that makes
it all more difficult and expensive.

Okay, if the purchasing department has "Best Practices" in play that
exclude eBay or Craigslist or related, then they can and usually also
should pay HPE for service, or pay one of the reputable used-equipment
vendors that'll offer systems with warranties. In general, these folks
might want to review Chapter 55612.3.5 in the applicable "Best Practices
for Continued Employment via Blame Shifting" where it states that all
critical servers must maintain support contracts from the list of
approved providers (see section 55612.3.476 and updates in BPCEvBS
addendum 2016.02 or current) at all times, or digitally-signed copies of
the applicable Business Practices Exception (C-level exception or trap
to interrupt vector badge 00001) must be maintained among the server
manager, the department, the purchasing department and internal IT and
other applicable groups, per BPCEvBS section 55612.3.5772.x4.375. Or
just order the eBay spare server out of the Division Shadow IT^W^WGroup
Latté Fund.

Buy a spare on EBAY, with guarantee, then sell to company for double what you
paid. Or triple.
Steve is having way too much fun with this ....
:-)

You could have fun too. *You* buy a spare on EBAY, with guarantee, then
sell to company for quadruple or more. Remember that you have lots and
lots of "administration costs" and don't forget a "handling charge".

Then *you* split the profits with the OP.

:-)

[Scott Dorsey]

Post by Paul Sture
You could have fun too. *You* buy a spare on EBAY, with guarantee, then
sell to company for quadruple or more. Remember that you have lots and
lots of "administration costs" and don't forget a "handling charge".
Then *you* split the profits with the OP.

Likely this is what is causing the original poster's problem in the first
place.
--scott