tomfiorentino

Eliminating the switch...for now..has helped a little bit. I am now getting 5 volts on the single white (?), purple (fan-off), yellow (-5V) and blue (-12V) when I test voltage using grey (auxgnd) as the other lead.

No fan. No 12 volts.

Does anyone have any additional thoughts for me?

Tom

tomfiorentino

OK guys...not feeling too much love but that kept me working at it and Ifigured this out pretty good. I went to a different power supply unit and had much more luck. Maybe the other one was bad!

Anyway, I have a supply that is operational now. I have a switch, LED, fan and all voltages (3, 5, 12) working properly. I also learned that the wire labeled "SB" on the circuit boardis astand-by line...always5 volts running throughthat when my unit is plugged in regardless of switch position.

Interestingly, I have a blue wire that is -12 volts when the unit is on. When I turn it off, the voltage slowly decreases all the way to 0 volts and then increases up tp .34 volts and stays there.

Here is my question:

I would like to havefour total outputs.

1.One 5.05 volt line (purple wire) that I can use for driving servos etc.
2. Three 12.12 volt lines (yellow wires) that I can use for charging batteries; one yellow wire to each output.

I have a quick question on the yellow wires. They are 18 guage and I have three of them; each rated for 90 degrees C and 300 volts. I was reading a wire guageamperage table which indicated 18 guage wire can handle 16 amps (I believe I am interpreting that correctly). The biggest batteries Ihave are 11.1 volt bind and fly stuff and usually I am just charging some 6 volt A123's and some 4.8 packs for my son's trainer. I'm sure the power supply can handle any of those items individually.

I know its been talked about here before, but I am still unclear on how I can anticipate using these outputs at the same time. For example:

Output 1: Quick field charger for the 4.8 volt NiCad
Output 2: Multi 4 charger for the 6.0 volt A123
Output 3:The cigarette lighter type charger that comes with the bind and fly stuff to charge the 11.1 volt Lipo

How do I calculate the total demand and Iguess Iwould compare that to the 8A that the power supply is rated for on the 12 volt line (250W total for the supply)? Sorry guys. I know this stuff has been kicked around before but I am just not smart with this.

What can you tell me please about calculating the total demand?

Thanks very much!

Tom

bpenner697

My Feedback: (5)

Tom, it's good to hear that you are having more success...I unfortunately can't answer your question. The only thing I would know to do is compare the consumption in terms of watts of the chargers (should vary with the batteries being connected I would think) to the output of the power supply. I understand that most PSUs are overrated for marketing purposes so a 250w PSU may only actually output 200w. A question on your setup: Did you combine your 3 12v outputs into a singe terminal or do you have them as separate terminals? I would be interested to see what kind of performance you are seeing...12v voltage with load vs voltage without load.

On my PSU there are 2 separate 12v lines so I made 2 separate 12v terminals (3 yellow wires combined to 1 post, 1 yellow wire to the other post). I was using 1 output for one of my chargers and then connected a second charger to the second output. The result was my PSU shutting off immediately because I think I created a short. What I believe is going on is that the two 12v outputs are not the separate lines they are on the same line so I need to figure out the 12V1 and 12V2, which I just haven't had a chance to do yet.

BGOSSELIN - thanks for the info...I am going to try the diodes this weekend. I notice there are several kinds of diodes (zener, barrel, rectifer, etc). I found a rectifer diode...will this work?

link: http://www.radioshack.com/product/in...ductId=2036270

the product information

Your project may not be simple, but finding the right parts for it is. 30 amp surge. Type: 1N4004. Peak Inverse Voltage (PV): 400.

Peak Reverse Voltage: 400V (Max)
DC Blocking Voltage: 400VDC (Max)
Forward Rectified Current: 1A
Silicon Rectifier
Diffused junction
Maximum Average Forward Voltage Drop: VF= 0.8 Volts
Non-repetitive Peak Surge Current: IFSM= 30 amps for 1 cycle
Maximum Reverse Current @ 25 degrees C: IR= 10uA[/i]
If I read the tech information, I should bump up 0.8 x 2 = 1.6v + 12v = 13.6v projected output...does this seem right?

One last question, you show a squiggly (resistor I think) on the diagram after the diode to 12v sense connection before the connection to ground. What's the purpose and do I need one? If so, what size?

Thank you in advance!!

bpenner697

My Feedback: (5)

How do I calculate the total demand and I guess I would compare that to the 8A that the power supply is rated for on the 12 volt line (250W total for the supply)? Sorry guys. I know this stuff has been kicked around before but I am just not smart with this.

TOM - It's a bit of a slow day at work so I am going to make a run at your calculation for my own edification:

DEMAND:
quick charger = 4.8v x 1500ma battery = 4.8 x 1.5A = 7.2W
multicharger = 6v x 2000ma battery = 6.0v x 2.0A = 12W
lipo = 11.1v x 2200ma battery = 22.2W
consumption to operate chargers = no clue but they use some amount of power...assume 12v x 1A each = 12W x 3 chargers = 36W

So adding up the total DEMAND = 7.2 + 12 + 22.2 + 36 = 77.2W

SUPPLY:
8A max on your 12v line x 12v = 96W
20% overstated capacity = 96W x .8 = 76.8W

I would say the chances of operating all 3 chargers at the same time is very questionable...but that's my math. I'd like a guy who knows this stuff better to chime in...

tomfiorentino

Thank you for the response!

I am using each 12v line going to a different terminal. So I have three outputs total. One is on a female cigarette lighter so I can use my car charger for the bind and fly stuff. The other two just go to a red and black binding post.Each 18 guage wire is rated for 16 amps so I left them separate.

Last night,using the bind and fly car charger on a 11.1 volt LiPo alone gave me 12.18 steady output volts from the supply. When I added my Multi 4 on an 1100mah A123 and my Hobbico Field charger on a 700 mah NiCad...so all three outputs were going at the same time...my voltage at the supply dipped to 11.78. But at the end of the day, all three is too much for this particular supply. My A123 wasn't that low so the charger never got to pulling 3 amps and I'm not sure the state of charge of the NiCad. I am going to do the test again with batteries that need more of a charge so I can see what happens when all the chargers start competing for the amperage. Mostly though, I'm having a lot of fun with this and learning a whole bunch.In fact, I talked with our IT guy today about being on the lookout for old computers that we can pull the power supplys out of...

I have copied and pasted in bold belowa portion of "The Fix" document that is part of the permanentresource on this; but I think this explains your unexpected shutdown. The same thing happened to me!

Unexpected Shutdown I've made the recommended changes, but when I connect my charger, the PS still shuts down. Now what?? The usage of a PC powersupply as a substitute for a field charger power source falls far outside the intent of the original design specifications. Once the PS is running and stable, the overload circuitry is tuned to detect high current sinks and shut the PS down under normal usage, these sinks would be indicative of an internal short in the PC. When a microcomputer is running, powersupply demands change, but these are minimal and are usually associated with optical, hard or floppy drive usage or USB devices being attached. Some field chargers produce a high current sink when first attached and generate a latch into the overload state as designed. The specifications state that the PS will remain latched until the load is removed from the rail the PS will either automatically reset or may require at least one PS_ON cycle. Even though your PS may have sufficient wattage to effectively drive the charger, the load change is the problem. One method that has been effective is to attach the charger before powering up to the PS, the charger now appears as a high draw motherboard and not as a potential short occurring after the PS is stable.


Thanks for the calculations too...I'll look at them later.

Tom

bgosselin

bpenner697, The diagram I posted is for a supply where there is a feedback line specifically for the 12v rail. This is where you would use the diodes and resistor. Only a few supplies have a system like that.

bgosselin

Tom: Yes, you definitely hook up the charger to the supply, THEN turn on the supply.

Andrew

Have you metered a supply to see if this is actually the case? In theory, you would expect all voltages to move in the same direction, however, this is not my experience. Rather than deal with discreet voltage drops, I simply dropped a pot inline with the 3.3v sense wire to see how changes in downstream voltage levels would impact the three primary rails. Initially, I expected all rails to increase/decrease together. By increasing the resistance on the sense wire, I pushed the 3.3v rail over 3.7v before the supply latched off. However, both the 5v and 12v rails dropped in voltage - admittedly, not what I initially expected. Is this always the case or just a design fluke in this particular model PSU. We do know that combined power on the 3.3v and 5v rail is limited and less than the individual current ratings added together. Therefore, the regulation circuitry quite likely is designed to rob Peter to pay Paul and increased voltage demands on the 3.3v rail must be compensated for by reducing available voltage on the 5v rail.

Soooooo..... If the supply ramps up the 3.3v rail (and decreases the 5v and 12v rail) when it perceives a drop in voltage at the connector, I tested the supply to determine how it would react to over-voltage at the sense wire. I tied the sense wire with pot to the 5v rail - it turned out to be pretty sensitive and took a little time to get the supply to latch. In the end, I found that the supply would latch with the 3.3v sense wire attached directly to the 5v rail. Here are my findings: SPI ATX12V2 350 watt; 10 ohm preload resistor on the 5v rail; no load on the supply.

Standard setup with 3.3v sense tied to 3.3v rail: 3.3v, 5v and 12v rails metered 3.39v, 4.84v and 13.06v. This is a little lower on the 5v rail and higher on the 3.3v and 12v rail than usual.

Modfied setup with 3.3v sense tied to 5v rail: Rails metered as 2.78v, 4.95v and 13.25v.

Under load, we expect the 12v to drop, but if the 5v rail holds its voltage, the regulation may be improved. However, if the 5v rail falls in conjuction with a drop in the 12v rail, the feedback on the sense wire would cause a further drop in the 12 rail. Without loading it, I don't know if it would spiral down or not.

Would I recommend tying the sense wire to the 5v rail? Without further testing, I can't say, so you would be on your own. My supply did not seem harmed in any way and if this does in fact boost the output voltage, it may be a freebee in getting higher voltages from these conversions.

andrew
http://web2.murraystate.edu/andy.bat...OWERSUPPLY.HTM

bgosselin

Andrew, I've only had the opportunity to do one PS the way I described, and yes, it did work as I said. For the other two, I did different things (one had an adjustment pot, the other had a separate feedback line for each rail). It just goes to show that every supply is different and these manufacturers all go about things differently with regards to regulation. I guess every one is a matter of trial and error.

Andrew

I'm in complete agreement, the only consistent factor with these supplies is inconsistency in design. I've run across several supplies with internal pots, but have seen only one with multiple sense wires. I've got some other new supplies on the shelf and am a little hesitant to clip the sense wire just to find if they behave differently, but my curiosity is up and it's a simple repair. I'll see how they work out.

andrew

tomfiorentino

http://www.playtool.com/pages/psucompat/compat.html

This link was good reading for me...maybe not for some more knowledgeable though.

Check it out and sorry if it is a repeat!

Tom

marcin-rc

Perhaps you dont know.
This
http://shop.ebay.com/i.html?_from=R4...All-Categories

Is much better

tomfiorentino

Probably but a lot less fun!

Actually, I have been toying around with a couple gaming machine power supplies because some of them can be pretty heavy duty. But nothing to write home about.

Tom

bgosselin

What's nice about those gaming power supplies is that they are meant for 12v only (maybe an extra power line too, but it's much smaller) so there's fewer wires to deal with and smaller form factor. What's not so nice is that they can get pretty toasty when running close to their max output.

nitrohawker

I am having some issues with my conversion. I have a 300 watt ATX V1.3 Seasonic Super Tornado that was working fine. I had a sense wire on pin 2, 3 and 7. Pin 2 was a 22ga orange with a 18ga orange, pin 3 was a 22ga black with a 18ga black, pin 7 was a black 20ga along with another 20ga black. Most all the other wires in the connector are 18ga. The supply powers up but the 12v voltage is up and down and not very high. You can see the corresponding pulsing in the auto bulbs I've hooked up to the 5 volt rail. I can't get a voltage reading on the 5v because it is pulsing so fast my meter won't register fast enough. I have the small orange sense wire connected to a bigger orange wire, the small black wire connected to a big black. The green is connected to a black and the unit turns on fine. I have tried running the supply with a load on the 12 volt but it doesn't make a difference. I'm confused because the sense wires are different than everything I've read online, usually on pin 11. I also have a blue and a black wire coming off the fan wires but not sure what to do with those or what they are yet. I think they might hook up to a fan controller in the computer. I started out with one auto bulb but had same issues. I added another in hopes that it would help but it didn't. I made a short video of what's happening in hopes that somebody might have an idea or suggestion.

[youtube]http://www.youtube.com/watch?v=nOf3Si9TZd4[/youtube]

jermpool

Just converted mine over and plugged it up for test and it worked, 12.6 volts, yay! Then I noticed that the spot I strapped the resistors (2x 1 ohm 10 watt) was extremely hot, too hot to touch. What should I do? Thought about getting a heat sync and putting it on the outside of the case.

Andrew

Are they connected in parallel or series. If you're using 1 ohm resistors, they should be connected in series. They will get very hot, but with 12.6 volts on the 12v rail and given the amount of heat you're describing, it sounds almost as if they are paralleled. Usually, strapping to the vent louvers with a little heat sink compound will provide sufficient heat dissipation.

andrew

karolh

My converted 250 watt ATX unit with a 10 ohm 10 watt resistor strapped to the vent louvres with heat sink compound works quite well and even though it only puts out 11.6 volts on the 12 volt rail it powers my Turnigy Accucell 6 50W charger okay.

Karol

jermpool

it is parallel, so I need to change it to series? And that should lower the heat.

Andrew

That's correct. There's seldom much advantage in dropping the pre-load below 2 ohms - it simply results in an excessive load on the PSU and high heat (which you found out). By connecting your two resistors in parallel, the effective load is .5 ohms, which is simply too much. When connected in series, the pre-load is two ohms.

You will likely get somewhat lower voltage on the 12v rail, but heat will be significantly lower and the PSU will not be stressed nearly as much.

andrew

bgosselin

Wow, I'm surprised the resistors lasted. You must not have run it very long.

5v @ 0.5ohm (two 1 ohm resistors in parallel) is generating 50w, which is 25w per resistor. Putting the resistors in series is 2 ohms. At 5v that's 12.5w, which is 6.25w per resistor, and is much better to extend their life (you usually select a resistor that has double the power dissipation than required).

jermpool

Switched it to series and only lost about .5 volts,
Next question, I bought an illuminated switch and I'm not sure how to get it to light, i thought the power going through and a ground would work but it didn't. I am currently looking on the internet but at the rate I am going, may get a response on here quicker.

jermpool

Had it on long enough to test the output, so less than a minute. I used to know all kinds of electronic stuff, I took a tv repair course, but I never used it and it's been about 10 years. I can't even remember how to use a multimeter, I'm such a loser.

bgosselin

To your illuminated switch question:

First, where is this switch located? Is it "interrupting" the main 120v AC line, or is it attached to the green wire? If attached to the mains wires, the bulb has to be rated for 120v AC or it will blow immediately (if not sooner). If it's attached to the green wire, it doesn't light because you are essentially shorting the green wire to ground when on; there will be no voltage on the switch when it's on. If you want an "on" indicator (and if the fan is not a good enough clue), simply add an LED and the appropriate resistor to any of the 3.3v, 5v, or 12v rails.

And referring to your last post:

Not to worry, it's been around 20 years since tech school for me, and I know if you don't use it, you lose it. Which is why I'm always tinkering with projects to keep me from forgetting stuff. The biggest things to remember about multi-meters are: 1) Do NOT hook the meter in parallel with the device to measure when configured to measure current, and 2) Do NOT use the ohms setting when the circuit has power. There are a few more "don'ts", but only for more advanced meters (like ones that have transistor, diode, capacitor, etc functions).

jermpool

It is on the green wire. I am going to have the ps located on a shelf under my work bench and run a lead to the top and mount a pair of jacks on the bench for when I am home and have a seperate switch on the bench to turn it on and off. Thought that if I had the light I would have the reminder to flip the switch off when I was done.