If you need assistance, please send an email to forum at 4hv dot org. To ensure your email is not marked as spam, please include the phrase "4hv help" in the subject line. You can also find assistance via IRC, at irc.shadowworld.net, room #hvcomm.
Support 4hv.org!
Donate:
4hv.org is hosted on a dedicated server. Unfortunately, this server costs and we rely on the help of site members to keep 4hv.org running. Please consider donating. We will place your name on the thanks list and you'll be helping to keep 4hv.org alive and free for everyone. Members whose names appear in red bold have donated recently. Green bold denotes those who have recently donated to keep the server carbon neutral.
Special Thanks To:
Aaron Holmes
Aaron Wheeler
Adam Horden
Alan Scrimgeour
Andre
Andrew Haynes
Anonymous000
asabase
Austin Weil
barney
Barry
Bert Hickman
Bill Kukowski
Blitzorn
Brandon Paradelas
Bruce Bowling
BubeeMike
Byong Park
Cesiumsponge
Chris F.
Chris Hooper
Corey Worthington
Derek Woodroffe
Dalus
Dan Strother
Daniel Davis
Daniel Uhrenholt
datasheetarchive
Dave Billington
Dave Marshall
David F.
Dennis Rogers
drelectrix
Dr. John Gudenas
Dr. Spark
E.TexasTesla
eastvoltresearch
Eirik Taylor
Erik Dyakov
Erlend^SE
Finn Hammer
Firebug24k
GalliumMan
Gary Peterson
George Slade
GhostNull
Gordon Mcknight
Graham Armitage
Grant
GreySoul
Henry H
IamSmooth
In memory of Leo Powning
Jacob Cash
James Howells
James Pawson
Jeff Greenfield
Jeff Thomas
Jesse Frost
Jim Mitchell
jlr134
Joe Mastroianni
John Forcina
John Oberg
John Willcutt
Jon Newcomb
klugesmith
Leslie Wright
Lutz Hoffman
Mads Barnkob
Martin King
Mats Karlsson
Matt Gibson
Matthew Guidry
mbd
Michael D'Angelo
Mikkel
mileswaldron
mister_rf
Neil Foster
Nick de Smith
Nick Soroka
nicklenorp
Nik
Norman Stanley
Patrick Coleman
Paul Brodie
Paul Jordan
Paul Montgomery
Ped
Peter Krogen
Peter Terren
PhilGood
Richard Feldman
Robert Bush
Royce Bailey
Scott Fusare
Scott Newman
smiffy
Stella
Steven Busic
Steve Conner
Steve Jones
Steve Ward
Sulaiman
Thomas Coyle
Thomas A. Wallace
Thomas W
Timo
Torch
Ulf Jonsson
vasil
Vaxian
vladi mazzilli
wastehl
Weston
William Kim
William N.
William Stehl
Wesley Venis
The aforementioned have contributed financially to the continuing triumph of 4hv.org. They are deserving of my most heartfelt thanks.
Registered Member #902
Joined: Sun Jul 15 2007, 08:17PM
Location: North Texas
Posts: 1040
wylie wrote ...
At one point i had an assortment of old Dell boxes (mostly pentium1&2's, 1 pIII lol) all running seti@home. They couldn't compare with the northwood p4 i had at the time in throughput though. Even my Octane could complete more work units in a day than they did in a week. Heck, my lil sparcIPX wasn't much slower than the pentium2's for seti@home. I think the IPX beat them all if you consider watts vs. performance.
I've been thinking about my own rendering farm since i got into using blender and synfig studio, but i've been considering using playstations. Once the ps3 is old enough to be had to for around $200, a 10-node cluster would be pretty attractive. Seems like a poor-mans alternative to server blades. Do they have enough RAM though? The 32mb in a ps2 seemed too small for use in a render farm. (And putting linux on them seemed like a nontrivial task last time i looked.) Is the 256mb in the ps3 going to be enough for each node? I have to go google the typical throughput of a render farm now dangit. But seems like you'd get alot more bang-for-your-kWh going that route. Or maybe build a custom cabinet and use a bunch of pentiumM boards. With electricity prices going up, the power requirements become a bigger concern.
Really, running them at all is a cost/benefit analysis. Home File Server? Sure, probably worth the electricity. Home Automation? Yeah, would be cheap, but you could get something new thats much more efficient and not that expensive. I mean, running a pent2 all the time, even with minimal drive activity (ie, letting em sleep) is using 20-30W just for the CPU. I don't want to try to quantify the overhead of the rest of the system, but you could use a pic micro or some commercially available embedded board that used under 10W total. Just throwing out a number there, might be able to control a home automation system with way less power. I guess it depends on the features.
For whatever use though, you want to consider the power consumption of those old beasts. Even if i hadn't messed up the network settings on my Octane so that i couldn't telnet in anymore, i still would be hard pressed to power it up. 730Watt powersupply. I doubt it actually draws that much, but based on the space-heater like characteristics of the air from its fan....it's using a good amount of power. Now maybe if i needed a space heater, i could use its computational abilities on the side. But it just doesn't make sense to run it only for the computing power.
@Dinges
Very quality work on those mods! The induction heater one is really neat, got more pics/info on that project? I tried winding a gdt on those yellow toroids with poor results. Anyone got a guess as to what material# those are? I only tried ones that had xformer windings on them to begin with, so pretty sure i wasn't using a choke-grade material.
currently the most powerful supercomputer in the world is actually made from the PS3 design (pretty much a PS3 without the blu-ray drive etc) - 256MB sounds low for what I recall, but may be right - but if you used a small Solid State Drive (and for a linux based cluster it doesn't have to be very big) and SSDs should be fast enough to act as a Swap Disk, I know they are fast enough to boot Vista at lightning speed, and load (without the programs already being in memory) 45 or so memory heavy programs in a few seconds; even make a REAL TIME stop motion from Windows Explorer Slide Show
it may be worth a try, but I don't think the slims still have a swappable hard drive unless you void the warranty, so go with the original version
-Jimmy
EDIT: also, there is a program similar to what rp181 suggested regarding rendering individual frames on computers called Render Queue - I have used it on Windows and it was extremely simple, I imagine the Linux version would be just as good
Registered Member #1232
Joined: Wed Jan 16 2008, 10:53PM
Location: Doon tha Toon!
Posts: 881
> Anyone got a guess as to what material# those are?
Yellow and white? Looks like Micrometals Type #26 to me. Alright for noise-suppression and buck smoothing chokes in SMPSUs but very lossy at high-frequencies unless used with low AC flux swing.
Registered Member #2511
Joined: Mon Dec 07 2009, 02:46AM
Location:
Posts: 36
Thanks for the heads-up on the powdered-iron toroid core; knew it wasn't ideal but was the first thing I found in the box of cores. The plan was to simply try and replace it with something else if it got too warm. It's supposed to be just a quick experiment before I start construction of a PLL controlled unit with voltage/current control.
As for the PC power supply modifications, I have searched a lot on the internet. And found a lot, but not the current-limit modification that is so great. Variable voltage modification is a piece of cake, hardly worth mentioning. But the variable current makes the difference between 'just a supply' and a short-proof bench supply.
Variable voltage is very nice.... but variable current-limiting is the icing on the cake.
Oh, btw, the easiest way to disable the overvoltage shutdown is to disconnect the one or two small diodes (1N4148) that go pin 4 of the TL494 (deadtime control). It cuts out the entire LM339/393 circuit, but you'd better have current-control in place to prevent loud bangs and escaping magic smoke when you short the outputs. BTW, don't remove the resistor and capacitor at pin 4, as they soft-start the PSU on powering up.
Some other small things that need to be changed is suitably rated elcos (voltage rating). I also routinely add a 100n capacitor over the voltage potmeter wiper (to ground); one PSU kept blowing up without it when adjusting voltage, the cap cured that - have been adding them ever since, just out of precaution.
I have *really* come to like these PC PSU modifications and think they deserve more attention - free junk that can easily be turned into valuable equipment. It's by no means new or rocket science, and I don't claim any originality - am just copying what others have done.
Registered Member #30
Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706
Dinges wrote ...
Variable voltage is very nice.... but variable current-limiting is the icing on the cake.
Indeed it is :) I didn't really understand your current limiting mod, though. Looks like you're using a length of wire as a resistor, and one of the TL494's error amps?
When I did it, I used an 0.01 ohm sense resistor, amplified by an op-amp, to give enough voltage for a digital panel meter. The amplified signal then went through another op-amp where it was compared with the voltage from the current limit knob, and the resulting error signal got integrated. The integrator output went to the TL494's compensation pin through a diode, to combine with the existing error amp signals.
This was quite slow-acting, but it gave a true 0 to 7A current limit that could be used for battery charging.
My modded SMPS can be seen in this heap of test gear in our lab. It wasn't an ATX, it originally powered a broadcast TV camera, but it still followed the classic TL494 blueprint.
Registered Member #2511
Joined: Mon Dec 07 2009, 02:46AM
Location:
Posts: 36
Took me a while to find the PSU in your picture.... but looks very nice!
Yes, correct, the trick is to use the 2nd error opamp in the TL494 which is normally unused (I've only seen one PSU out of about 20 which used it - I simply requisitioned the opamp for my own purposes anyway). A small bit of wire in the common lead (flying lead) of the transformer acts as sense resistor; it feeds a small negative voltage (more output current -> more negative) to the 2nd opamp. Output of it becomes positive and limits pulsewidth, thus limiting current. Whichever control is more stringent (output or voltage) controls the output of the TL494. I've actually come away with using existing PCB traces as the sense resistor, by cunningly adding the ground reference (pin16) of the TL494 to another part of the ground trace PCB.
The idea is not by me but by someone called Oztules on an another forum I frequent; if you google his name you'll find many more references. He uses it as battery charger for 48V battery banks too, for example.
Yesterday, I came along this same idea in the datasheet for the TL594:
Check out the 1R resistor, 240R and how it goes to pin 15 & 16 (2nd opamp of the TL494).
Another idea hit me this morning - rewind the transformer for 300Vdc for audio valve experiments? And perhaps add a little separate small 6.3V winding on the transformer with its own linear regulator (LM317) for the filaments?
Up to 300Vdc adjustable, and with current-limit; along with 6.3Vdc for the filaments... does it get any better than this?
Registered Member #30
Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706
Sounds good!
However, as you increase the output voltage, you start to run into trouble with self-capacitance of the transformer secondary winding and the output filter choke winding. You might want to study Ralph Hartwell's design:
This is possibly the most adventurous DIY SMPS I've ever seen. It puts out 1kW at 1500V DC, so 300V should be easy in comparison. He used the same basic topology as a PC power supply, and solved the self-capacitance issue by just throwing power and snubbers at it, but more elegant methods are available.
I've also seen those electronic 12v halogen transformers rewound for HV output:
Registered Member #1232
Joined: Wed Jan 16 2008, 10:53PM
Location: Doon tha Toon!
Posts: 881
I agree with Steve here. In theory what you have done is fine for making a constant load current mode. However in practice there may be a couple of potential problems depending on how you use the supply:
Firstly the constant current mode can only command the drive side to turn down the duty-ratio, this means that whatever current is flowing in the output choke will continue to flow until it ramps down. Also a shorted output will still result in a bright flash as the output reservoir electrolytics discharge into the fault.
Secondly, floating the centre-tap of the secondary and then grounding it through a current sense resistor potentially diverts a lot of HF hash that is capacitively coupled from the primary side into this current sense resistor! This won't necessarily be a problem if your current limit threshold is high enough and the control loop bandwith is low enough. I'm just pointing out that the centre-tap of the secondary is usually grounded for a good reason, and then bypassed back to the mains side with a Y-cap or two in order to provide a path for the power side HF hash to get back to the power side.
Thanks for that link Steve. That is the one I was looking for when Stella was asking about possibly using a Solid State Microwave Oven supply to power an amateur radio amplifier.
SMPSU modding seems to be becoming quite a popular hobby for some, although i wonder how many people actually go to the trouble to test their new "bench power supply" in terms of its output ripple, regulation, load-step response and stability!
There might be a lot of PC power supplies with the outputs modded from 12V down to 4 volts that are on the brink of instability and just waiting to blow up someone's delicate design during testing!
If I was designing a SMPSU with output of 300VDC or more at a few hundred watts i'd personally go for the split secondary approach. Use two seperate secondaries, with their own rectifiers, each generating half of the required output voltage. This also allows you to split the output capacitance into two series halves and use lower voltage rated electrolytics. Use a coupled buck choke to achieve good cross-regulation and ensure voltage sharing across the two outputs. That is how I would do it.
This site is powered by e107, which is released under the GNU GPL License. All work on this site, except where otherwise noted, is licensed under a Creative Commons Attribution-ShareAlike 2.5 License. By submitting any information to this site, you agree that anything submitted will be so licensed. Please read our Disclaimer and Policies page for information on your rights and responsibilities regarding this site.
What to do with old computer parts?
