Ugly Tezro power

Elf

Storybook / Retired, ex-staff
Feb 4, 2019
792
252
63
Mountain West (US)
I'll discuss Tezro power supplies in more detail at a later time, but have a look at some nasty Tezro power!

Super ugly. Probably needs new capacitors at the very least. Tezro owners, do you know what your power supply is doing? 😰


SCR02.PNG


SCR03.PNG



Also just a spoiler: there is more signaling going on there than a basic ATX supply provides!
 

dancingfools

New member
Jan 8, 2020
14
2
3
Hi, I have a Tezro 4X1.0Ghz, for the moment no problem, but there is also to say that I turn it on occasionally. I basically believe that everything also depends on where they were placed, in what environments. If left standing for months and maybe in damp and dusty places, the internal components are affected by it over the years.
 
  • Like
Reactions: Geoman

Elf

Storybook / Retired, ex-staff
Feb 4, 2019
792
252
63
Mountain West (US)
I basically believe that everything also depends on where they were placed, in what environments. If left standing for months and maybe in damp and dusty places, the internal components are affected by it over the years.
Indeed, there does seem to be some variability about the condition of the power supplies, or electronics in general across all the SGIs. I have an old early Indigo 1 prototype with a working PSU, but a newer (mass production) Indigo 1 with a dead supply.

It is hard to say exactly what the causes are, although all electrolytic capacitors age and have a lifespan to them. I've attached a good document on the subject called "Capacitors Age and Capacitors Have an End of Life," from Emerson, who produce Uninterruptable (back-up) Power Supplies. In general, temperature is the biggest factor at reducing their life.

However it should be noted that 15 years is considered a long service life for an electrolytic capacitor. Versus our SGI machines:
  • Indigo 1 - up to 28 years old
  • Indigo 2 - up to 27 years old
  • Indy - up to 27 years old
  • O2 - up to 24 years old
  • Octane - up to 22 years old
  • Fuel - up to 18 years old
  • Tezro - up to 17 years old
(Dates as of time of post, just based on rough Wikipedia history timelines rather than real research :) )

It's amazing and a testament to good engineering that these devices are still running! But at this point even the Tezro can fall past the end of the expected service life for some of the electronics so I suppose I am not surprised my PSU isn't doing too well :(
 

Attachments

dancingfools

New member
Jan 8, 2020
14
2
3
Indeed, there does seem to be some variability about the condition of the power supplies, or electronics in general across all the SGIs. I have an old early Indigo 1 prototype with a working PSU, but a newer (mass production) Indigo 1 with a dead supply.

It is hard to say exactly what the causes are, although all electrolytic capacitors age and have a lifespan to them. I've attached a good document on the subject called "Capacitors Age and Capacitors Have an End of Life," from Emerson, who produce Uninterruptable (back-up) Power Supplies. In general, temperature is the biggest factor at reducing their life.

However it should be noted that 15 years is considered a long service life for an electrolytic capacitor. Versus our SGI machines:
  • Indigo 1 - up to 28 years old
  • Indigo 2 - up to 27 years old
  • Indy - up to 27 years old
  • O2 - up to 24 years old
  • Octane - up to 22 years old
  • Fuel - up to 18 years old
  • Tezro - up to 17 years old
(Dates as of time of post, just based on rough Wikipedia history timelines rather than real research :) )

It's amazing and a testament to good engineering that these devices are still running! But at this point even the Tezro can fall past the end of the expected service life for some of the electronics so I suppose I am not surprised my PSU isn't doing too well :(
Yes, but also the quality of the components is a very important factor. To give you an example, I had direct problems with the latest Amiga 4000 capacitors. I have Amiga 2000 and 3000 instead which are still perfect, although older in age.
 

Elf

Storybook / Retired, ex-staff
Feb 4, 2019
792
252
63
Mountain West (US)
Yes, but also the quality of the components is a very important factor.
Indeed! Quality capacitors, and higher temperature / lifespan ratings of those capacitors, are a major factor for sure. Better electrolytes and better construction to avoid evaporation and leakage.

I was happy to see in the Tezro power supply that they had used -- if I remember correctly -- Nippon Chemicon. Definitely a reputable brand, and actually quite a well built supply overall.

Apropos of nothing, but perhaps of interest to someone, I generally find these brands to be on the quality list:
  • Nippon Chemicon / United Chemicon
  • Rubycon
  • Nichicon
  • Panasonic
  • Kemet
  • EPCOS / TDK
  • Sprague (now Vishay)
There can be other good brands for sure, but those are my go-tos, particularly the top 4. But even within those brands, selecting the right series (with higher temp ratings and lifespans) is also important.
 

Elf

Storybook / Retired, ex-staff
Feb 4, 2019
792
252
63
Mountain West (US)
Some Tezro PSU testing pictures, including running the Tezro with the PSU outside of the chassis (and skins removed).

Tezro PSU wide.jpg Tezro PSU normal.jpg Tezro PSU external.jpg

With my new DC clamp meter ( :) ) it seems that the motherboard draws about 110 watts off one of the three (!) 12V rails when starting up.
 

Geoman

Member
Dec 28, 2019
53
22
8
Germany
HP used the same Model for their xw8400 Workstations (though with more output capability). I took it apart for cleaning as well as the one that came with the Tezro. They look quite the same on the inside. It is from Delta electronics.

IMG_4059.JPG


Tezro's PSU:

IMG_4069.JPG
 
Last edited:
  • Like
Reactions: Elf

Elf

Storybook / Retired, ex-staff
Feb 4, 2019
792
252
63
Mountain West (US)
HP used the same Model for their xw8400 Workstations (though with more output capability). I took it apart for cleaning as well as the one that came with the Tezro. They look quite the same on the inside. It is from Delta electronics.
Interesting, even more 12V rails! Does it also have that small 5 pin (0.1" pin header) control cable? I can take a picture of the Tezro one if you need a reference, just curious if even the non-SGI ones have it.
 

Elf

Storybook / Retired, ex-staff
Feb 4, 2019
792
252
63
Mountain West (US)
Monitoring the output of the 1.8V voltage regulator on the Tezro motherboard, which I have been having some problems with. Where does that wire go? 🎃

Tezro 1.8V monitoring.jpg


Of course, as soon as I start watching it, the problems go away...
 

Geoman

Member
Dec 28, 2019
53
22
8
Germany
Interesting, even more 12V rails! Does it also have that small 5 pin (0.1" pin header) control cable? I can take a picture of the Tezro one if you need a reference, just curious if even the non-SGI ones have it.
I checked and there are only "power-out" connectors. Here is a picture of every cable that goes onto the main logic board:
IMG_4064.JPG
 
  • Like
Reactions: Elf

Geoman

Member
Dec 28, 2019
53
22
8
Germany
You've got a nice workbench there!
What I dont like about the Tezro PSU is the fan-speed. It is the loudest fan in the system.
 
  • Like
Reactions: Elf

Elf

Storybook / Retired, ex-staff
Feb 4, 2019
792
252
63
Mountain West (US)
I checked and there are only "power-out" connectors.
Thanks! That is helpful information; it would appear then that the 5 pin control header is custom for SGI.

You've got a nice workbench there!
What I dont like about the Tezro PSU is the fan-speed. It is the loudest fan in the system.
Thank you! :)

I have also been noticing that PSU fan speed; even more now that the PSU is running outside of the Tezro. A voltage on a control pin also appears to have the Tezro ramping up the PSU fan's speed from its default. Interestingly even with the PSU out of the case the area where it would be does not accumulate any appreciable heat. Between that and looking at the air flow of the case, it would appear that the PSU is just cooling itself but does not play much appreciable role in drawing air over other components.
 

Elf

Storybook / Retired, ex-staff
Feb 4, 2019
792
252
63
Mountain West (US)
So while I had been having problems with the 1.8V regulator cutting out, now that I have had it under test for two days (including for almost 9 hours today) it has failed to reproduce the issue. My best guess is an old capacitor at the output of the 1.8V regulator that has livened up after some activity (something that I have commonly seen with old electrolytics). Might replace it anyways, we will see...

Anyways, some thermal shots of the area (with emissivity set for plastic IC packages, but not any surround metals or board):

1.8V switching MOSFETs; not doing too badly at the moment (and the Tezro isn't failing either)
FLIR0006.jpg

1.8V regulator output capacitor, running at 91.9F. Not terrible but probably worth replacing. A square of electrical tape is used for proper emissivity.
FLIR0010.jpg

A toasty little MOSFET near the IO9 socket. Purpose unknown!
FLIR0008.jpg

3.3V AUX regulator, probably not passing too much current
FLIR0009.jpg

Tezro on the test bench. Looks better live with the hot exhaust ports, as if it contains some sort of demonic energy inside!
FLIR0004.jpg

Incidentally, astute observers may observe that I now possess the equipment to put the urban legend of the Fuel's "poor thermal design" to rest one way or another. Don't expect anything soon though, I would like to bust it in style such that we never have to hear about it again. ;)
 
  • Like
Reactions: Geoman

Elf

Storybook / Retired, ex-staff
Feb 4, 2019
792
252
63
Mountain West (US)
Desoldered the controller board from the HV/LV board of the Tezro PSU. Yes, that is a PIC16 microcontroller at the heart of it (under the white sticker), and a 24C02 EEPROM on the back!

As @Geoman pointed out, it looks like these units were customized for more than one application, and given the five mysterious control pins it would appear that some use of that programmability was employed here.

Tezro PSU controller - 1.jpg Tezro PSU controller - 2.jpg Tezro PSU controller - 3.jpg
 

Irinikus

Well-known member
Dec 16, 2019
738
553
93
South Africa
www.youtube.com
The interesting thing is that MrWeedser has fitted a standard ATX power supply into his Tezro and it works!

Quoting his post on IRIX Network: "

Just get a normal PSU for ATX, they work perfectly.
Mine is running with a standard corsair HX520, but you can get whatever psu you want. Go with a super efficient enermax or whatever you prefer.

The problem is the 10 Pin CPU connector besides the 24 Pin, because standard ATX PSUs don't have it, they are 8 pin maximum.
My solution was to put in the PCIe connector in the missing 2 pins, see the attached picture."

 

Elf

Storybook / Retired, ex-staff
Feb 4, 2019
792
252
63
Mountain West (US)
Yes, that has been one of my observations; unlike the Fuel with its L1 environmental monitoring, the Tezro does not strictly require the 5 pin control line to turn on, nor does the PSU appear to report fan speed back to the environmental sensors. One can disconnect it and just leave the "dumb" side of the power rails + Pgood / Run connected, and it will run although with an audible difference in fan speed. (In fact one can disconnect and reconnect that line even while the system is running and hear the fans spool up and down)

At the same time I'm also reluctant to recommend using only an ATX supply replacement (with nothing to handle the control lines) until I get a better grip on what exactly the control lines are doing and what edge cases they may cover, since they are clearly custom to the supply from SGI. If one just wants a Tezro to turn on though, an ATX replacement is certainly a viable option. It may also be that it ends up being the most straightforward and best solution, but it's hard for me to say right now.
 
Last edited:

Elf

Storybook / Retired, ex-staff
Feb 4, 2019
792
252
63
Mountain West (US)
I will compile my findings when complete, but just wanted to give one example of something I found today that makes SGI power supplies a little more unique than the average bear's ATX supply.

One of the 5 pin connector control lines is .. and this took some reverse engineering and then a test setup to confirm .. an AC frequency dependent signal to increase the voltage on the 3.3V rail! I had to make sure I wasn't crazy, but sure enough, one of the pins which floats on the 3.3V rail sense is involved in a R/C network in an op-amp's feedback loop and then gets involved in the same path as the 3.3V adjustment trimmer on the PSU's control board. The presence of the R/C network meant that there was an AC component, and it also lined up with a mis-shapen triangle wave being supplied by the Tezro.

Tezro PSU disconnected Pin5 wave.PNG


A bit of a test rig confirms: by varying the frequency of the wave provided, one can increase the voltage on the 3.3V rail (and only on that rail):

Here a sweep from 0.5 to 4MHz shows the 3.3V rail increasing, until the sweep resets back to 0.5MHz and the voltage drops again.

So, they went through the trouble to build in a (somewhat complex!) feedback from the motherboard to the 3.3V rail voltage. Why? What consumes that power, and how important is it? Does it get especially sad if the voltage is lower than expected because an ATX supply doesn't have the same feedback mechanism? It was at least important enough for them to customize the PSU and add to the motherboard BOM, and does raise a few more questions now. :p

Incidentally, the Indigo 2 PSU also has a similar step up / step down mechanism, but not as fine grained.

This is the kind of information you won't get anywhere else! ;)
 
Last edited:

Elf

Storybook / Retired, ex-staff
Feb 4, 2019
792
252
63
Mountain West (US)
Finally caught the 1.8V rail problem in action, and have confirmed that it is real and not just the environmental sensor playing tricks!

Tezro 1.8 VR mosfet overheat.jpg Tezro 1.8 VR visual.jpg

It happens so rarely I haven't been able to catch it until today, but I guess this means the Tezro is going under the knife!

(Note the switch of the units on the thermal image from Fahrenheit to Celcius; this is much hotter than the 105F seen previously under normal operation)

Update: I think I have a better handle on this 1.8V issue. It seems to be related to restarting the machine (from L1 up but system power down). One can restart the machine a lot and not trigger it, but if it does happen, it usually happens immediately or shortly after a restart. Then the Tezro has to be left to "cool down" for a while completely unplugged, and the issue will go away.

The machine can even be left running for days and not encounter it! Strange.
 
Last edited:
  • Like
Reactions: Geoman

weblacky

Active member
Jan 13, 2020
181
45
28
Seattle, WA
Hi Elf,

Did you ever get more info/time on this Tezro 1.8v issue?

I really forget what happened (I looked at an old email to even remember this happened...it's been that long since I've used this system).

My first Tezro does this a couple moments after startup (log shown below), but did start to the PROM and then shortly shutdown. When I bought it second-hand it was working for nearly a year, upgraded RAM from 4GB to 8GB (complete module replacements), added a bunch of PCI Cards, then had a 5V rail dip issue (temporarily resolved by removing my Adaptec firewire card that pulled heavy on my 5V rail. But I think this started to happen (or maybe it resolved with the 5V issue…not sure).

So I guess I have this too…a lot. I want to concentrate on this stuff in December when I’ll hopefully be able to buy a late model curve tracer to test components. For now I have the voltage dip and the firmware forces a PSU shutdown. Seems to be exactly what you’re describing at the end of your post. Did you replace the filter cap in that 1.8v buck convert section and see if this went away? I assume that’s what’s needed because it starts out great, then dips, and I know a lot more than I did back then about this power electronics stuff. So I feel better about doing work on this issue later this year.

But since you saw the same thing, was wondering if any progress was made on this? Obviously, the regulator IC isn’t blinking out or anything, but low voltage on a buck converter is normally a sign of crap smoothing (DC average) by the final cap (correct me if I’m mistaken).

Any new info would be greatly appreciated!



LOG from like 2011 (Dual 800Mhz Tezro Deskside, I know the firmware is old, never really loaded Irix beyond 6.5.19 on this due to hardware issues that cropped up soon after.):



INFO: Cannot enable VRM: 9

INFO: Cannot enable VRM: 10

INFO: Cannot enable VRM: 11

SGI SN1 L1 Controller

Firmware Image B: Rev. 1.32.6, Built 09/27/2004 14:40:25


001c01-L1>

001c01 ATTN: Cooling system stabilized

001c01 ATTN: 1.8V low fault limit reached @ 1.199V.

001c01 ATTN: brick auto power down in 30 seconds

001c01 ATTN: brick auto power down in 25 seconds

001c01 ATTN: brick auto power down in 20 seconds

001c01 ATTN: brick auto power down in 15 seconds

001c01 ATTN: brick auto power down in 10 seconds

001c01 ATTN: brick auto power down in 5 seconds

001c01 ATTN: brick is powering down now!
 

About us

  • Silicon Graphics User Group (SGUG) is a community for users, developers, and admirers of Silicon Graphics (SGI) products. We aim to be a friendly hobbyist community for discussing all aspects of SGIs, including use, software development, the IRIX Operating System, and troubleshooting, as well as facilitating hardware exchange.

User Menu