PC Power Supply Mod.

[What's_nitro?]

Originally Posted by BrianG

Isolation like that could be done, but the diodes would:

- Be VERY beefy. A diode capable of handling 15+ amps each would be large indeed.

- Have a low forward voltage drop. Most diodes have a drop of 0.7v. Schottkey diodes can get as low as ~0.3v.

- Must be mounted on a heatsink. Even a 0.3v drop diode will dissipate 4.5w @ 15A.

- Expensive.

Ok, what about these?

$3.99 for 20 diodes w/ free shipping...

5 diodes needed for full PS output per pad. 0.5V drop at full load = 1.5W per diode max. Use 10 diodes per pad, so 0.75W dissipated per diode max (probably less since each diode would see 1.5A max load). A diode that size should be able to dissipate that amount of power by itself, or just mount them all under the PS fan.

It may be a bit clunky looking, but it would fit into the PS case no problem.

[BrianG]

Paralleling semiconductors is generally not advised unless you compensate for their negative temperature coefficient. What will happen is the one with the slightly lower Vf value will do most of the work until it overheats and fails. Then the next one, then the next, until all are gone. Something like this would be better; single diode with 0.360v drop @ 19A. Gonna be a little expensive.

[itbvolks]

I did manage to hit up the local "shack" and picked up both 1ohm/10ohm - 10watt resistors to see if I can load the 3.3/5v and see if that helps out at all.

Didn't get to wire them up last night though. Maybe tonight.

[itbvolks]

UPDATE! It works!!!!!


Added the 10ohm/10watters in parallel on the 5v line and picked up an immediate .3v unloaded for a total of roughly ~12.45.

So hooked up the charger and checked to see what was what... Upon moving to 2c charge rates on my 4200's (8.4A), I'm holding a rock solid 12.34v loaded with no sign of any sag at all at this current loading!!! Brian's a genius!

Now I need to find a place to mount the resistors to bleed some heat. Also need to clean up the wiring and get it all stuff in there..

Holy shiot, those resistors are HOT man! WOW! Just re-charged/storage moded the second pack. Clearly needs to get mounted to the case with some TIM or something but how hot should they be?


Interestingly though, my BRAND NEW Hyperion 4200 35c batt seems a little odd.... Set my 720i for storage (which I thought was 60%) and balanced/storage loaded both packs. I only ran one of the two packs on the first day so the other has been sitting. I hooked it up and saw the battery was at roughly ~46%!!! Doesn't sound right to me. Hoping there isn't a dead cell in that pack. Thing is completely new and hasn't even been in the truck yet.

To complicate this even more, the pack I have run (2x disharge cycles) was also run through the "storage" mode and I just checked that pack and it's at 70%....????


My missing something?

[BrianG]

Told you they would need to be on a heatsink. People don't seem to realize how hot 5w of heat really is. I assume the PS case is steel, in which case, does not work well as a heatsink. You can use a simple aluminum U-channel as long as it is in the fan's airlflow somehow, or use some type of finned sink.

As to the % issue, I am not familiar with that charger, so I can't say. If it uses cell voltage to estimate charge level, it will be a rough estimate at best. If it uses capacity and mAh put back in, it has to assume you started with a fully depleted battery. If you stick a 5Ah pack with 1Ah still left in it, it may put 3Ah (60% of 5Ah), but the result would actually be 4Ah.

[itbvolks]

Ya, pretty warm for sure.... If I stay with these I'll need to come up with a slick heatsink setup.

Any reason not to try say like a 1/4w - 1k ohm resistor for this app? Does the 5v require load or just the PS seeing it's enabled? Just curious?

Easy enough to test.

In terms of the batts, I'll just have to keep an eye on them. Doesn't sound like there is an issue right now.

[BrianG]

You'd have to experiment with the various loading to get the proper value. 1A usually is a pretty good value to use for proper regulation, but you may be able to go lower.

[itbvolks]

Just curious if it's load based regulation or more the PS is looking for the 5v to be "loaded" with something to enable proper regulation.

[itbvolks]

Not that I think the 1A is all that serious a loading. Just some heat I need to dissipate.

I've got some 8020 here (extruded aluminum channel) that while not really a HS per say, should work and has decent area. 1" x 1" x 3" directly under the fan should do the trick....

[Arct1k]

Brian - I never loaded the 5v rail and do get some drop... would a couple of these work in parrallel screwed or thermal epoxied to heatsink?

10 ohm 30w

http://search.digikey.com/scripts/Dk...MP930-10.0F-ND

http://dkc3.digikey.com/PDF/US2010/P1994.pdf

[itbvolks]

I would think a couple of this in parallel would work very well with a HS. I can't describe the difference adding the load to the 5v made. I went from running 4.2A loaded with a main line voltage of 11.3v to running 8.4A loaded with a main line voltage of 12.34v rock solid and I'm sure more is available if needed. At the 8.4A loading prior, I was in the ~10.4v range and the charger would shut down.

I was measuring some V-drop between the PS connectors and what the charger was reporting as well (up to .5v). Now I measure 12.4v at the PS connectors and the charger reads 12.34v.

Matter of fact, now that I know how to make this work, I might grab a few of those myself...

[BrianG]

Originally Posted by Arct1k

Brian - I never loaded the 5v rail and do get some drop... would a couple of these work in parrallel screwed or thermal epoxied to heatsink?

10 ohm 30w

http://search.digikey.com/scripts/Dk...MP930-10.0F-ND

http://dkc3.digikey.com/PDF/US2010/P1994.pdf

Yeah, those can work, but they seem to be a bit pricey. I would just get a couple/few 10 ohm 10w power resistors from radioshack and parallel them onto a heatsink. Even those "transistor-style" resistors need to be on a heatsink, so there's really no benefit to using them IMO.

The idea is to load them down only to the point where it is needed. Too little and the 12v line isn't rock-solid enough. Too much and it doesn't help and you are just wasting electricity (and money) as heat. Unfortunately, this "sweet spot" takes a little experimentation. Not a big deal for me since I have tons of resistor combinations and other components to test with. But, from experience, anything between 1-2A is enough.

BTW: The only power supplies I've seen that don't benefit from loading the 5v or 3.3v lines are the kind where there is an extra sense wire for each of the rails. It is easy to tell because there will be the normal larger wires going to the mobo connector, but paralleled into the same connection will be a smaller gauge wire which measures the voltage at that point to compensate for any voltage drop along the wire. In this style of PS, regulation in each rail is independent, so loading the other rails is unnecessary and useless (except as a space heater, lol).