In need of a replacement PSU.

[Windows7ge]

I recently purchased this 2U JBOD system here on a Scratch & Dent deal.

 

 

As far as I've seen so far the chassis itself is completely free of marks but the PSU grill that protects the fan is completely smashed in and the fan is wreaked.

 

 

I didn't really trust the PSU to begin with so I want to replace it.

 

Due to the 2U form factor the PSU must use an 80mm fan on the front/back it cannot use a 120mm top/bottom fan. It'd be completely suffocated.

 

Does anyone know of a well branded PSU that meets this spec? One that's not considerably overpriced for the wattage? I'm looking for ~400W with ~30+A on the 12V rail.

 

Worst comes to worst I'll buy an SFX PSU and just make a custom bracket for it.

[Princess Luna]

@LukeSavenije Don't you know some Delta's that might fit his usercase?

[LukeSavenije]

4 minutes ago, Princess Luna said:

@LukeSavenije Don't you know some Delta's that might fit his usercase?

you'd get into the server parts of delta at that point... that'll cost a lot more than that sfx plan

[Windows7ge]

16 minutes ago, LukeSavenije said:

you'd get into the server parts of delta at that point... that'll cost a lot more than that sfx plan

I'm not particularly worried about redundancy. The pool is only ever going to exist on the JBOD box if it goes offline because it needs a PSU swap no biggie. Plus backups are always going to exist else where on the network, to further state it's not a huge issue.

 

Thinking more and more about it I do have one SFX PSU around, I might just be able to swap it for an ATX then use it in the server. But if we can find a more appropriate ATX unit for ≤$75 that'd be great.

[LukeSavenije]

well... from taking a bit closer look... i think it's a flex atx... that's not going to be easy

 

I'm thinking about the few units fsp has for this so far

[Windows7ge]

15 minutes ago, LukeSavenije said:

well... from taking a bit closer look... i think it's a flex atx... that's not going to be easy

 

I'm thinking about the few units fsp has for this so far

In the photos? It's standard ATX. Server is a 2U. However I forgot about the existence of FlexATX. That'd get the job done. Some don't appear all that expensive. I can make a bracket to convert ATX to FlexATX.

[LukeSavenije]

54 minutes ago, Windows7ge said:

In the photos? It's standard ATX. Server is a 2U. However I forgot about the existence of FlexATX. That'd get the job done. Some don't appear all that expensive. I can make a bracket to convert ATX to FlexATX.

well, good luck finding whatever psu you need... at least now you know a bit where to look for, and I'll just be here whenever you want to show some units you found

[Guest]

First off:  That's a Zippy PSU.  One of the best PSU OEMs out there.  Might even consider just swapping out the fan.

 

Second:  It's a standard ATX PSU.  Not 1U, 2U or Flex.  Just good ol' PS/2 / ATX.  Not even the connectors are proprietary. 

 

Thirdly:  That's not a 2U chassis.  That's a 3U chassis.  Almost anything will fit in there.

 

[iamdarkyoshi]

27 minutes ago, jonnyGURU said:

First off:  That's a Zippy PSU.  One of the best PSU OEMs out there.  Might even consider just swapping out the fan.

 

Second:  It's a standard ATX PSU.  Not 1U, 2U or Flex.  Just good ol' PS/2 / ATX.  Not even the connectors are proprietary. 

 

Thirdly:  That's not a 2U chassis.  That's a 3U chassis.  Almost anything will fit in there.

Not 100% sure on the last one. If a 1U server is a little thicker than a 40mm fan, then a 2U should be a little thicker than an 80mm fan, which looks to be the case here.

 

I'd definitely agree with the fan replacement route though, Looks to be an NMB 12v 300mA fan, which should be super easy to find a replacement for if OP is comfortable doing a fan swap.

[Windows7ge]

32 minutes ago, jonnyGURU said:

First off:  That's a Zippy PSU.  One of the best PSU OEMs out there.  Might even consider just swapping out the fan.

 

Second:  It's a standard ATX PSU.  Not 1U, 2U or Flex.  Just good ol' PS/2 / ATX.  Not even the connectors are proprietary. 

 

Thirdly:  That's not a 2U chassis.  That's a 3U chassis.  Almost anything will fit in there.

 

1) I don't know if it was pressure or blunt force trama. I'm not against getting a hammer and banging out the metal to replace the fan but if the PSU was dropped hard to cause this I don't think it matters if zippy made it it shouldn't be used.

 

2) When I say 2U I'm only referring to the enclosure. I'm aware the PSU itself is standard ATX.

 

3) This goes against my understanding of U's. It's 1/2 the hight of my 4Us and is marketed as a 2U. If you say it's a 3U I'm willing to hear your argument as to why. And yes I'm aware any ATX/PS2 PSU will fit the issue is there's no room for PSUs with 120mm intake fans to get air. There's no gap top nor bottom the PSU would suffocate.

[Windows7ge]

1 hour ago, iamdarkyoshi said:

I'd definitely agree with the fan replacement route though, Looks to be an NMB 12v 300mA fan, which should be super easy to find a replacement for if OP is comfortable doing a fan swap.

I'm not 100% against it, wouldn't be the first time I cracked open a PSU definitely wouldn't be the last. My main concern is just how the damage occurred but I'll tell the two of you what, tomorrow I'll crack it open and take pictures of the PCB. If you guys don't see anything that looks like it was dropped from a hight (cracked solder joints, cracked PCB, bent components, surface mount chips snapped off) then I'll hammer out the housing and replace the fan. I most likely wouldn't replace it with a identical one though. I have a small pile of random 80mm fans. If it's 12V I'll solder a old one to the original fans cable.

 

I will say though if it dies and takes some of my equipment with it I'm blaming both of you.

[Guest]

2 hours ago, Windows7ge said:

 

2) When I say 2U I'm only referring to the enclosure. I'm aware the PSU itself is standard ATX.

 

I was talking about the enclosure (chassis) as well.  It doesn't look like a 2U and I haven't seen a 2U that accepts standard ATX PSUs.  This looks like a 3U.

 

2U's are only 3.5" tall.  

[iamdarkyoshi]

49 minutes ago, jonnyGURU said:

I was talking about the enclosure (chassis) as well.  It doesn't look like a 2U and I haven't seen a 2U that accepts standard ATX PSUs.  This looks like a 3U.

 

2U's are only 3.5" tall.  

2U on the spec page. http://www.istarusa.com/en/istarusa/products.php?model=MAGE212U40-PCI-E

 

It makes sense, 1U servers use 40mm fans, 2U uses 80mm fans, which is what is inside a standard ATX PSU.

 

As for inspecting the PCB for damage, it sounds like a good idea. I have had a gateway PSU that took a fall and the transformer broke off the board.

[Windows7ge]

1 hour ago, jonnyGURU said:

I was talking about the enclosure (chassis) as well.  It doesn't look like a 2U and I haven't seen a 2U that accepts standard ATX PSUs.  This looks like a 3U.

 

2U's are only 3.5" tall.  

Close enough.

I knew I bought a 2U.

 

I'll crack open this zippy PSU and inspect the extent of the damage.

[Guest]

Haha!  Ok. My bad.  

 

Back in my day, you only saw PS2 PSUs in 3U and up.  Pretty cool it's a 2U that uses a PS2 PSU.

 

[Windows7ge]

15 minutes ago, jonnyGURU said:

Haha!  Ok. My bad.  

 

Back in my day, you only saw PS2 PSUs in 3U and up.  Pretty cool it's a 2U that uses a PS2 PSU.

 

Cool yes, limiting my options, also yes. Look here:

It is perfectly the hight of the interior of the enclosure. The PSU cannot be any thicker at all or else the lid won't go on. Which reiterates my point of suffocating a replacement PSU if it relies on a gap (one a 3U enclosure would EASILY provide - if I was working with a 3U I would not have posted this topic).

 

Give me a minute while I take this PSU to bits.

[Windows7ge]

@jonnyGURU @iamdarkyoshi

Alright, got it torn apart and wow you wernt lying about zippy PSUs this thing is built like a tank. Filtering capacitors on the AC input (both box film & ceramic), bleed resistors, extra solder on almost every trace for current carrying and thermal dissipation, thermal probes, proper spade connectors, proper double throw switch (for live & neutral), extra PVC insulation on the wires, everything's glued to everything else, PCB held to the enclosure with lock washers, she's built to last.

 

I was going to take pictures but there's simply far too much circuitry to go over. I inspected as much as I could especially where the fan got pushed into the PCB and everything appears solid. PCB isn't cracked. Nothing's loose that I could find. Nothing looks bent or pushed. That grill turned out to be soft enough for me to push back into place with my hands. I replaced the fan with a random old but never used 12V 0.25A 80mm fan. I know it doesn't move as much air as the original but for what I paid if a damaged fan is the only thing wrong with it it was a heck of a deal.

 

 

What I've done is put the box back together and thrown a few HDDs in it and I'm going to let it run overnight. See if it overheats, or let's out the magic smoke, if it's fine then I'll take a gamble that it will still do it's job because the inside of that thing was impressive.

[iamdarkyoshi]

1 hour ago, Windows7ge said:

-snip-

Pretty much exactly what I would have done. The 50mA difference on the fan really shouldn't be an issue. Worst I could see happening is it runs a tad warmer and the fan controller increases the speed to compensate, assuming it has speed control.

 

My guess is someone slid it into a rack and there was a cable behind it like a power cable and it dented the fan in. Dunno.

 

Either way it sounds like it's back up and running, overnight test should give us an answer as to if it survived.

[Guest]

14 hours ago, Windows7ge said:

Cool yes, limiting my options, also yes. Look here:

It is perfectly the hight of the interior of the enclosure. The PSU cannot be any thicker at all or else the lid won't go on. Which reiterates my point of suffocating a replacement PSU if it relies on a gap (one a 3U enclosure would EASILY provide - if I was working with a 3U I would not have posted this topic).

 

Well.. it shouldn't be impossible to find a modern PSU with push/pull if this Zippy weren't to work out.

 

I know FSP and Cougar (HEC) still have some units on the market.  And they still use them in China for iCafe gaming PCs that use "desktop" (not tower) cases similar to your 2U or 3U.  I believe Great Wall and Huntkey are making those units.

 

But if the Zippy works, go with it.

 

[Windows7ge]

13 hours ago, iamdarkyoshi said:

assuming it has speed control.

That's actually the one thing I don't like about this PSU. It only uses a 2 wire fan. Now it could just be directly driving the motor by limiting the input voltage and reading off the thermal probe but I think that's a but crude so I think it's just set to run 12V 100% of the time.

 

2 hours ago, jonnyGURU said:

Well.. it shouldn't be impossible to find a modern PSU with push/pull if this Zippy weren't to work out.

 

I know FSP and Cougar (HEC) still have some units on the market.  And they still use them in China for iCafe gaming PCs that use "desktop" (not tower) cases similar to your 2U or 3U.  I believe Great Wall and Huntkey are making those units.

 

But if the Zippy works, go with it.

 

Hard to find, no. Hard to find from a reputable brand without being exorbitantly overpriced yes. If the zippy should die pre-maturely (and hopefully not take anything else with it) I'll look into those brands.

[Guest]

7 minutes ago, Windows7ge said:

That's actually the one thing I don't like about this PSU. It only uses a 2 wire fan. Now it could just be directly driving the motor by limiting the input voltage and reading off the thermal probe but I think that's a but crude so I think it's just set to run 12V 100% of the time.

Crude?

 

FYI:  99% of power supplies on the market today use two wire fans that are controlled via varying DC voltage controlled by a thermistor based controller.

 

The only reason to use PWM is if you need the granular speed control that allows a fan to spin well below what would be a normal startup voltage for a DC fan.  And for a "server" PSU, "low noise" is not a sought out feature, so there's no reason to add the cost of PWM. 

 

And even when PWM fans are used, I've never seen the fan controller use the tach wire to report back to the controller.  Even when an MCU is used to control the fan.

 

 

 

[Windows7ge]

2 hours ago, jonnyGURU said:

Crude?

 

FYI:  99% of power supplies on the market today use two wire fans that are controlled via varying DC voltage controlled by a thermistor based controller.

 

The only reason to use PWM is if you need the granular speed control that allows a fan to spin well below what would be a normal startup voltage for a DC fan.  And for a "server" PSU, "low noise" is not a sought out feature, so there's no reason to add the cost of PWM. 

 

And even when PWM fans are used, I've never seen the fan controller use the tach wire to report back to the controller.  Even when an MCU is used to control the fan.

Calm down there it's just an option. I understand the concept of if it can be done simpler than it should be but you would think with RPM reporting & PWM that the extra few dollars it costs to include would be offset by the $200+ prices I see similar wattage units going for. I would say though with only two wires it can't report a fan failure. Not unless it went open-circuit. Of course your argument to that will be if that happens in a business environment you just replace the PSU but the issue is how do you know it failed when it can't tell you other than shutting down from overheating? Kind of makes RPM reporting a worth-while feature so it can notify you even if it's only driven directly by limiting DC voltage so I do find that a bit crude.

[Guest]

1 hour ago, Windows7ge said:

Calm down there it's just an option. I understand the concept of if it can be done simpler than it should be but you would think with RPM reporting & PWM that the extra few dollars it costs to include would be offset by the $200+ prices I see similar wattage units going for. I would say though with only two wires it can't report a fan failure. Not unless it went open-circuit. Of course your argument to that will be if that happens in a business environment you just replace the PSU but the issue is how do you know it failed when it can't tell you other than shutting down from overheating? Kind of makes RPM reporting a worth-while feature so it can notify you even if it's only driven directly by limiting DC voltage so I do find that a bit crude.

I'm curious as to what you use for a desktop PC PSU. 

 

Fan fail alarms is something simply not implemented in ATX power supplies as they're not meant for mission critical application.  Literally none of the ATX PSUs on the market have that feature.  Like I said:  Even when a PWM fan is implemented, there's no reporting back to the fan controller of fan status.  Those sort of features are typically reserved for server specific PSUs.

 

Simply put, the way the circuit works is if the fan doesn't spin, the PSU gets hotter, and eventually OTP kicks in and shuts the PC down.  Again:  NOT good for mission critical systems.  

 

Ironically, even the ATX PSUs that have RPM fan speed reporting can't tell the difference between a failed fan and a fan that's not spinning because it doesn't need to.  So unless you're actually looking at the reporting software to see if the fan is spinning or not, you have no idea if the fan has failed.

[Windows7ge]

2 minutes ago, jonnyGURU said:

I'm curious as to what you use for a desktop PC PSU. 

 

Fan fail alarms is something simply not implemented in ATX power supplies as they're not meant for mission critical application.  Literally none of the ATX PSUs on the market have that feature.  Like I said:  Even when a PWM fan is implemented, there's no reporting back to the fan controller of fan status.  Those sort of features are typically reserved for server specific PSUs.

 

Simply put, the way the circuit works is if the fan doesn't spin, the PSU gets hotter, and eventually OTP kicks in and shuts the PC down.  Again:  NOT good for mission critical systems.  

 

Ironically, even the ATX PSUs that have RPM fan speed reporting can't tell the difference between a failed fan and a fan that's not spinning because it doesn't need to.  So unless you're actually looking at the reporting software to see if the fan is spinning or not, you have no idea if the fan has failed.

A Corsair AX1200i but that's besides the point. I'm only referring to server PSU's. I don't expect a desktop one to sound a buzzer and light up an LED next to some text that says PSU fan failure (although that would be nice). I would think even if it is using the ATX form factor that a server PSU brand would include server PSU features. A lot of 1U/2U servers from brands like Dell or HPE include indicators for when case fans fail so it's something I would have liked to see considering the server application of the device.

[Guest]

2 hours ago, Windows7ge said:

A Corsair AX1200i but that's besides the point. I'm only referring to server PSU's. I don't expect a desktop one to sound a buzzer and light up an LED next to some text that says PSU fan failure (although that would be nice). I would think even if it is using the ATX form factor that a server PSU brand would include server PSU features. A lot of 1U/2U servers from brands like Dell or HPE include indicators for when case fans fail so it's something I would have liked to see considering the server application of the device.

Most of Zippy's server PSUs have that feature.  This Zippy ATX PSU is a corner case.  It was built for a US company to market a cheap server chassis.