|
|
Hardware and Case Mod's You Paint it, Cut it, Solder it, bend it, light it up, make it glow or anything like that, here is your forum. |
Thread Tools |
02-19-2003, 04:41 PM | #76 |
Cooling Savant
Join Date: Mar 2002
Location: Oregon, USA
Posts: 230
|
Update on the joining circuit: I have the article largely completed, I now need to finish a few little things, before I submit it.
__________________
Signing out... Yo-DUH_87 If it works, fix it until it's broke! Then, after it's broke, add duct tape! Affordable webhosting! |
02-20-2003, 09:22 PM | #77 |
Cooling Savant
Join Date: Mar 2002
Location: Oregon, USA
Posts: 230
|
Article has been submitted, now it is up to Joe and the rest of the staff to evaluate it...
Hope I get picked
__________________
Signing out... Yo-DUH_87 If it works, fix it until it's broke! Then, after it's broke, add duct tape! Affordable webhosting! |
03-17-2003, 08:57 AM | #78 |
Responsible for 2%
of all the posts here. Join Date: May 2002
Location: Texas, U.S.A.
Posts: 8,302
|
A minor bit of progress.
I forced the PSU on, to find 5.2 and 11.71 volts on output, with no load. The 12V line drops to 11.66 with a CCFL tube on. I guess this one's gonna need tweaking, along with all the other ones! Seeing that the innards of this PSU is quite voluminous, the watercooling part will have to be put on hold, along with the slimming down. At this point, I would much prefer to find an external solution to mounting these in an array. Yu_duh's solution is actually very feasible, and relatively easy, so I might go with that, until I find or come up with something better. [edit] The PSU fan is very quiet, and effective. Last edited by bigben2k; 03-17-2003 at 10:40 AM. |
03-24-2003, 11:41 AM | #79 |
Responsible for 2%
of all the posts here. Join Date: May 2002
Location: Texas, U.S.A.
Posts: 8,302
|
Update:
I loaded the 12V line with a 10 Ohm, 25 Watt resistor, and measured 11.58 volts. It's still within 5% of 12v, but I think I'll try to tweak it back up to 12. In the mean time, here are some pictures! #1: the back of the PSU. |
03-24-2003, 11:45 AM | #80 |
Responsible for 2%
of all the posts here. Join Date: May 2002
Location: Texas, U.S.A.
Posts: 8,302
|
#2: The innards.
|
03-24-2003, 11:47 AM | #81 |
Responsible for 2%
of all the posts here. Join Date: May 2002
Location: Texas, U.S.A.
Posts: 8,302
|
#3: the micro linear chip
|
03-24-2003, 11:52 AM | #82 |
Responsible for 2%
of all the posts here. Join Date: May 2002
Location: Texas, U.S.A.
Posts: 8,302
|
#4: two other chips (the dark one is relevant, I don't believe that the other one is significant)
|
03-24-2003, 11:54 AM | #83 |
Responsible for 2%
of all the posts here. Join Date: May 2002
Location: Texas, U.S.A.
Posts: 8,302
|
And finally, two of the seven MOSFETs attached to the heatsinks.
|
03-24-2003, 04:54 PM | #84 |
Cooling Savant
Join Date: Sep 2002
Location: Omaha, NE USA
Posts: 216
|
Try loading the 5v line with the resistor and the 12v with the ccfl and then see what the 12v line is with the meter. Give it a try and see what happens.
Just a thought. |
03-24-2003, 05:09 PM | #85 |
Responsible for 2%
of all the posts here. Join Date: May 2002
Location: Texas, U.S.A.
Posts: 8,302
|
I did, and the 12v line read 11.68, with the inverter turned on. It was 11.71 with it off.
I can't guarantee the accuracy of my meter, so I'll re-measure with another one. Last edited by bigben2k; 03-24-2003 at 05:18 PM. |
03-28-2003, 12:42 PM | #86 |
Responsible for 2%
of all the posts here. Join Date: May 2002
Location: Texas, U.S.A.
Posts: 8,302
|
(edited out, at Yo-Duh's request)
Now you'll have to PM or e-mail me your thoughts! Last edited by bigben2k; 03-28-2003 at 02:50 PM. |
03-28-2003, 01:00 PM | #87 |
Cooling Savant
Join Date: Mar 2002
Location: Oregon, USA
Posts: 230
|
Yes
If you would edit your post, I would apreciate it
__________________
Signing out... Yo-DUH_87 If it works, fix it until it's broke! Then, after it's broke, add duct tape! Affordable webhosting! |
03-31-2003, 06:13 PM | #88 |
Cooling Savant
Join Date: May 2002
Location: Canada
Posts: 247
|
Not sure if it's any help to you, but if you want them I can take pictures of the innards of various Compaq/HP servers here at work. We just got a ML350 G3 about a month ago, and though we only have a single PSU in that one (it's a development server) it supports dual-redundant and will have the PCU-load-balancing PCB in it. I also have very good access to a pair of DL380 G3's (both dual-PSU-capable) right now as I'm right in the middle of testing some stuff on them and I'm the only person using them (so I can power them off whenever I want to open them up).
Also, I haven't seen it mentioned in this thread yet, but one thing to consider is that many ATX PSU's require a load on most (all?) of their voltage lines before they will output a proper voltage. (I think this is what murray13 was getting at in his last post). Due to this reason, I usually recommend against running one PSU for the MB and a second for HD's/CD's - I don't trust the outputs of the second PSU to be accurate with an unusual load like that on it. If you can get this all to work, I'm also interested in a copy of the schematic. I'm running a 300W + a 250W right now, but I've balanced them with everything that requires an accurate voltage (MB + HD's + CD/DVD) on the 300, and everything that could care less (fans, lights, etc.) on the 250. |
03-31-2003, 07:16 PM | #89 |
Responsible for 2%
of all the posts here. Join Date: May 2002
Location: Texas, U.S.A.
Posts: 8,302
|
Cova: yes, that could prove to be interesting!
Specifically, a pic (and number) of the ICs. Also, I'm still trying to spec out a good connector for it, so it'd be nice to see what "The Pros" use! Thanks! |
04-01-2003, 01:20 PM | #90 |
Cooling Savant
Join Date: May 2002
Location: Canada
Posts: 247
|
I'll try and remember to bring my camera to work with me tomorrow. By the end of tomorrow I should be able to have pics of the DL380 G3 stuff. I'll keep the camera at work (after downloading those pics off it) then and when I get a chance to get inside the ML350 I'll do that too - but the ML is rack-mounted already with a bunch of developers using it, and the DL is all mine to play with for the next 3 weeks. I don't think it will make too much difference since I doubt you actually have the exact same PSU as what goes into that ML. (If you do, and you can get them for $20 each, I want one to upgrade this server to dual - it costs us about $400 each for these from Compaq)
|
04-01-2003, 01:24 PM | #91 |
Responsible for 2%
of all the posts here. Join Date: May 2002
Location: Texas, U.S.A.
Posts: 8,302
|
LOL! no my PSU is not the dual model, but it does come from an ML350 server.
Kewl, I'm looking forward to the pictures, thank you! |
04-01-2003, 01:27 PM | #92 |
Cooling Savant
Join Date: May 2002
Location: Canada
Posts: 247
|
My point was - there only is one model of PSU for ML350's. There is a small PCB inside the case that the PSU(s) plug into, and that plugs into the MB. Doesn't matter if you only have a single PSU (as we have in ours, with a metal placeholder in the other bay for airflow-control) or a dual PSU system, you get the same ones (and they cost $400 each from Compaq).
Anyways, I'll get the pics as soon as I can. |
04-01-2003, 01:32 PM | #93 |
Responsible for 2%
of all the posts here. Join Date: May 2002
Location: Texas, U.S.A.
Posts: 8,302
|
Hum... I've seen the same PSU on the net for ~$100.
I could have sworn that I saw more than one PSU model for the ML350. Mine looks like a normal PSU, complete with cables. Oh well, I'll just wait for the pic. |
04-01-2003, 01:39 PM | #94 |
Cooling Savant
Join Date: May 2002
Location: Canada
Posts: 247
|
I was just looking on Compaq's site - it could be possible you have the PSU for the original ML350 (the model that existed a year or two ago) - which didn't have the Dual PSU option (I'm looking at the manual for such server here: ftp://ftp.compaq.com/pub/supportinfo...201345-001.pdf)
So - it's quite possible you do have the exact same PSU as used in that box, but that server doesn't even support dual PSU's. Compaq/HP server model numbers are broken up into 3 bits - the letters (ML, DL, BL) for the line they come from (ML - big standup servers, DL - dense rack servers, BL - blade servers), the number (350, 380, etc. - bigger numbers are more powerful CPU-wize), and the generation (nothing = G1, G2, G3, etc.) - the ML350 we have is a G3, 3rd-generation - and uses different (dual-capable) PSU's than the G1, which doesn't support dual PSU's. Anyways - cleared up any confusion I think, pictures still coming when I can get the camera here. (and if ya have other questions about Compaq servers I'm happy to answer 'em if I can, I work with the stuff every day here - or other pictures if you want, we have DL360's, DL380's, DL580's, ML350's, all across a mix of G1 to G3) |
04-02-2003, 04:40 PM | #95 |
Responsible for 2%
of all the posts here. Join Date: May 2002
Location: Texas, U.S.A.
Posts: 8,302
|
An invisible update...
I've been corresponding with Yo_duh about his circuit, and I'll be trying it out. $50+ worth of parts: I hope this works well! In the mean time, maybe his article will be published? Joe? |
04-03-2003, 08:44 PM | #96 |
Cooling Savant
Join Date: May 2002
Location: Canada
Posts: 247
|
BigBen - you have a PM. I'm too lazy right now to recompress all the pics from my camera down to forum-attachable size, and I don't want to use as much bandwidth as it would likely take to host them from my house, so the URL is in the PM, filesize about 3MB, 11 various pics none of which show the internals overly well (it's very cramped as you will see). They are all of the DL380 G3.
As far as I can see, there are no IC's inside of the part thats covered by the acrylic top, but the PCB extends through under the fans (towards the front of the case) and there are some IC's 1/2-sticking out from under there, as well as some IC's on the other side of the fans completly. However even with the thing right in front of me, good light, and being able to choose my angle of view I can't read anything off any of them without completly dissasembling the entire box, which I don't want to do. If anyone else wants a copy of the pics, PM me or post here and I'll send you the URL as well. Sorry for being a day late with 'em - the lock broke on the door to the room with all the test stuff in it (including that box) at 4:00 and the facilities people had all already left, so I had to wait to the next day to take them. |
04-04-2003, 09:38 AM | #97 |
Responsible for 2%
of all the posts here. Join Date: May 2002
Location: Texas, U.S.A.
Posts: 8,302
|
Nice pics!
I can see a regulation circuit, passed the PSU, and from the looks of the connector on the PSU, it's definitely a hot-swap design: notice how the copper lines on the edge connector are at different height. The short "fat" one is the power line. What I don't get is how come there's only one of those fat connectors (maybe there's another one on the backside?). It makes me think that those PSUs provide a base DC source, and that the rest of the required voltages are created in the clear topped box.:shrug: Nice pics, Cova, thank you very much!!! |
04-04-2003, 01:31 PM | #98 |
Cooling Savant
Join Date: May 2002
Location: Canada
Posts: 247
|
They are hot-swap (notice the purple tab on the back where the power plugs into it? Everything on a compaq/HP server that is purple is hot-swappable, eg all the fans in that case) assuming you have the second one plugged into it. I didn't even think yesterday to get a pic of the bottom of the connector, but I can get that without pulling the server out of that mini-rack again if you want to see it. I can just pull the PSU straight out of the back of the server. If the connectors on the PSU itself interest you, I can also pull a PSU out of one of the production servers and look at it's connector (the production servers all actually have their redundant PSU's in them). Hmm..., now I'm curious if they even all have the same number of pins.
|
04-04-2003, 02:11 PM | #99 |
Responsible for 2%
of all the posts here. Join Date: May 2002
Location: Texas, U.S.A.
Posts: 8,302
|
It would be interesting to know, yes! A pic may not be necessary.
|
04-04-2003, 04:03 PM | #100 |
Cooling Savant
Join Date: May 2002
Location: Canada
Posts: 247
|
Don't have the camera with me right now - but this is what the pin-layout on the bottom of that PSU connector looks like. A lot more complicated than the top is.
I'm guessing that the longer pins are grounds, the one on the top is +12V, the big one on the bottom here is +5V, and most of those small pins are for feedback from the regulation circuit back into the PSU, perhaps also one for the PSU fan RPM's to be monitored as well. Reading the label on the PSU (there is a closeup pic of it in that zip) says that the PSU only puts out +12, +5, and -12V, so that reguilatoin/load-balancing circuit must be building the other voltages from those. |
Currently Active Users Viewing This Thread: 4 (0 members and 4 guests) | |
|
|