Where can i buy a VRM??

[Nathan G]

I am looking for a vender that will sell me a VRM only, without a MB or CPU. Anyone know where i can buy one? As an option, I would be interested in a used one if anyone has one they wish to get rid of!
Thanks

[Nathan]

Hi Nathan!

Will you refresh my memory on what VRM is?

[Nathan G]

Its a Voltage Regulator Module... It is used to regulate the high currrent >10A /low voltage 1.5V to 3.5V Out put for Core voltage. There have been a couple revision of the Intel spec. here is Rev. 8.1 pdf file from intel http://developer.intel.com/design/celeron/applnots/243408.htm
I think we are up to Rev8.3 now ...
Why i want VRM:
I have an old socket7 MB (P54c only)which i want to put a K6-2 in (2.2V core)
I have managed to split to I/O and core voltage planes by unsoldering some jumpers on MB. I plan on using onboard regulation for I/O voltage (3.3). But I need a external source for Core voltage. I have tried making my own voltage regulator but, it just didnt meet specs. (too much dropout, ripple blah.blah) So I will just use a VRM... umm if i can find one at a reasonable cost http://www.sysopt.com/forum/wink.gif
This is just a rainy day type project, i know it would be easier, cheaper to buy a P55 MB. but, whats the fun in that. http://www.sysopt.com/forum/wink.gif

[psyklone]

what kind of caps did you use in your filter circuit for the voltage regulator? the voltage should be relatively clean (low ripple) coming off of the PS.

[Nathan G]

Hehe at first i thought making a stable 2.2V out of 5V would be no prob. (boy was i wrong)
I checked...Turns Out That most single component Voltage Regulators stop at around 5Amps max.
(start Benny Hill Theme Now)
So i decided to use a pass transistor (200W T0-3 w heatsink) and a biasing resistor (low ohm) on the input of Voltage regulator(low dropout type). I Placed tantalum/electrolytic caps in circuit per regulators application notes. got several big ole 25W .5 ohm ceramic resistors for testing and a bench Supply. It worked fine from no Load (i had bias R) to ~2A but would suffer severe V drops >2A. Well,,umm i gave up trying to make it stable at anything over 2A. I realized that my setup was a-REAL POWER HOG/Heat generator!!! (inefficient)
The transistor would have 5V-2.2V=2.8V@ 10A (max) =28W !! EEEEEE Hot EEE Hot marshmellows anyone. http://www.sysopt.com/forum/redface.gif Thats totally wasted power too!!(turned into heat)

I decided i needed to study more. ya think hehe ?!?#@.
Went and looked at some VRM schematics. and realized most use 12V for control and 5V for drive. On top of that most VRMs are switchmode with a FET versus what i was doing with BJT transistor (more linear). Switchmode is more efficient in its power consumption (less heat generated). The down side of switchmode regulation is they are more complicated and many more components. I added my cost of all the parts and heck, i might as well get a VRM "off the shelf"...It doesn't rain here THAT much and this is after all,a "rainy day" project.
soohere i am...looking 4 a VRM http://www.sysopt.com/forum/wink.gif

BTW- one of the members here has been nice enough to offer a VRM to me... Thanks All.



[This message has been edited by Nathan G (edited 02-21-2000).]

[Paul V]

Just a thought, howsabout a regulator that regulates 4.4 V at 5 A, then a 2:1 transformer to halve your voltage and double your current?

I haven't really done much thinking on this one, just a stray idea that popped into my head so I haven't gone over any of the potential drawbacks yet...

[Nathan G]

sorry I never mentioned that all the voltages i have been talking about are DC
Paul- i believe Transformers are AC devices. http://www.sysopt.com/forum/smile.gif
Nathan "the editmeister" G


[This message has been edited by Nathan G (edited 02-21-2000).]

[dkozloski]

In order to use a VRM you must have a motherboard that has a VRM socket. I'm sitting here looking at a VRM and all I see are six 330 mfd. caps, a toroidal coil about as big around as a dime, three diodes, and three little transistors. There is some surface mount stuff on the other side. The motherboard it goes on, Supermicro P55T2S, has two voltage regulators already. The VRM just changes the voltage outputs of the existing regulators, it does not add an aditional regulator.
I have heard of regulator modules that go between your CPU and the existing socket that do what you want. That sounds like what you need.
Also the VRM I have costs as much as a whole motherboard

[Paul V]

D'oh, I knew there was something wrong with that when I said it... I was thinking of the ideal transformer equations which do not involve omega (and thus are not frequency dependant, in the ideal world at least) but in actuality having zero frequency (DC) would stop a transformer from working.

[This message has been edited by Paul V (edited 02-21-2000).]

[Nathan G]

"In order to use a VRM you must have a motherboard that has a VRM socket." usually yes- but in my case No. I stated earlier that I have split I/O and Core power planes in preparation for installing the regulator. I am planning on driving the VRM edge connector with my own little board w switches (or jumpers) for different voltage settings/enable and power in/out. Please see the Intel data sheet (.pdf) that i refered to earlier in this thread. I have seen a VRM similar to what you are describing and i dont think they conform to these intel VRM specs 8.1 or later. It is non-compliant Due to reason you mentioned- it controls onboard regulators and is not actually a VRM (its more of a filter/controller w/o the drive devices). This idea is supported by the fact that your MB is a P55 which means it allready had split core/io voltage planes(MMX)with seperate regulators on board.
I believe your VRM may be proprietary (it is made to control the Onboard regulator on THAT Companies MB only) therefore they can charge whatever they wish.
A socket adapter is avail that achieves all this and seperates the needed BF pins in place too. but they cost approx half the price (or more) of a new MB!! I could just use another PS alltogether but that seems like a waste. I just need 1 efficient regulator!
I wish to try this approach. Its much more fun this way!! dont ya think?? http://www.sysopt.com/forum/redface.gif
thanks for all your replies http://www.sysopt.com/forum/smile.gif

BTW- if anyone has a schematic of a circuit that basically meets The Intel VRM 8.1 specs or later.(0-10Amps 2.0 to 3.0V adj. Output would be ok.he) that i can build inexpensively,with at least 70% efficiency or better... Please let me know!



[This message has been edited by Nathan G (edited 02-22-2000).]

[RobRich]

You might try the follow companies and resources:
http://www.intersil.com/families/power_control.asp
http://www.vxie.com/02-models1.html

Good Luck,
Robert Richmond

[Paul V]

Have you tried www.digikey.com? (http://www.digikey.com?)

I am pretty sure they sell voltage regulators, although 10 amps is a large amount, I dunno how many regulators will handle that.