Looking to build a somewhat simple variable voltage DC PS

[flinx]

Here are the major parts I have on hand.
1 - transformer pulled from an old APC Back-UPS 650, unknown ratings except that there is a 12 amp circuit breaker on the AC side and there were 2 - 40 amp fuses on the DC side (mounted to the circuit board). Transformer has 4 wires 2 purple ground and a white and a yellow wire. voltage across white to purple and yellow to purple is 8.6vac~ across white to yellow is 17.2vac~. 2 wires white and black for input AC.

1 - 40 amp Bridge Rectifier pulled from an electric wheel chair charger.
1 - Brand new 25 amp BR.
1 - 20 amp BR pulled from a different electric wheel chair charger.
1 - 8 amp BR pulled from a computer PS.
1 - 20 amp BR pulled from a server PS.
1 - new 10,000uf 50v cap.
many 3300uf 16v caps
2 - 2200uf 25v caps
Many other caps purchased from badcaps.
1 - kia78r 12api
several ssh7n90 017
1 - irf224h
1 - lm317t p+ (might be able to get another off a board somewhere)
1 - L7812cv
1 - lm324n
a couple of lm339n
Several small 50k pots
misc blade fuses up to 40 amps
many monstrous looking schottky rectifiers s60sc6m
a pile of mosfets still mounted to the PS boards possibly 40amp or greater, I cannot read the numbers without desoldering them.
whatever else I can pull from misc boards floating around.

With this transformer I could do 2 - 12 volt say 15 amp outputs, and or 1 adjustable 0-24 volt 35 amp (correct me if I am wrong). The battery in this unit was a 12v 12ah SLA.

What say you? what dangerous piece of equipment can I build with this?

Yes I will need a variable 24 volt PS to run a wheel chair motor I am looking to acquire.

skill level is novice.

[flinx]

I found this site but not sure if the circuit listed can handle the amps (aprox 35). What mods to the circuit would be needed for the amperage I am looking for?

I plan on using a 10,000uf 50 volt cap for the unregulated output, or should I put that in place of the 1uf cap? what watt rating for the resistors should I use

I have a hefty heat sink for the LM317 which I can fan cool.

If you are wondering I plan on using it for something similar to this.

It needs to be variable from between 12vdc and 24vdc.

[mariushm]

You don't put a capacitor INSTEAD of the 0.1 uF capacitor, that one should remain there to remove a bit of the noise that still goes through.

For 35A+, I'm not quite sure a single 10.000 uF will be enough.

I also can't figure out where from do you get the idea that the transformer can actually do so many amps. Is that rating listed on it?
Usually 20A+ transformers are very large and heavy.




The first capacitor right after the bridge rectifier is the one that's a large value. The diodes rectify but each of them has a 0.4-0.7v drop and also there are drops due to the nature of the rectifying so the large capacitor stores energy to keep the DC voltage output as steady as possible.

You can see the effect of the large capacitor here: http://www.falstad.com/circuit/e-fullrectf.html

note 1 : page has a Java applet, accept it when it asks
note 2 : you can actually download the java applet in a zip, and there's loads of circuits shown you can experiment with by selecting them from a menu.. see the info in the page

Right click on the capacitor and select Edit and increase the capacity to see how the output DC is much smoother.
Edit the resistance (the load) and decrease the value to see how the DC becomes less smooth because the energy stored in the capacitor is consumed much faster

The large value of the capacitor is so that you will get a DC output between a small range, say 34-36v, so that when it goes through the LM317 and its 1.7v+ voltage drop and you can still get about 32v clear regulated voltage.

The 0.1uF (probably 0.47uF would also work) needs to be there, and should be tantalum or ceramic x7r (not y5v or other crappier standards), not electrolytic.

....

I really can't understand why people keep using these LM317 regulators... they're crap, 1.5A max, large voltage drop... why?

There's better regulators out there like LT108x family which can do up to 7.5A, 1v dropout, much better regulation, at 0.1% compared to 1% on LM317 : http://cds.linear.com/docs/Datasheet/108345fg.pdf

You also have these LT chips available in Spice's database, so you can simulate the circuits before doing it.

http://www.linear.com/designtools/software/#LTspice

... and you can request samples but don't get greedy... I asked for 2x LT1084 and 2 LT1085 and got them in about 2 weeks.

...

What resistor are you talking about? If it's a resistor in parallel with the large capacitor, then the resistor there is just to bleed out the large capacitor in time. It's not really needed but a good safety measure.

The power dissipated in a resistor is I * I * R or V*V/R

Let's say you use a 4.7k resistor, then the resistor will dissipate 50v * 50v / 4700 = 0.531w so a 1w resistor would work.

I'm too lazy to do the math now on how fast the capacitor would bleed out, you can simulate that in Spice or the applet I showed you.

[budm]

If transformer can supply two 12V @15A, then total power output is 180Watts x 2 = 360 Watts, then at 24V output, the max current will be only 15A, you cannot get more power than what you have with continuous output (you use large cap to provide surge current handling for the load).
24V @ 35A = 860 Watts, I do not think your APC transformer will be that big.
For bridge rectifier, for example, if you want 1A max load, the transformer has to be rated at 1.6A.

[flinx]

Originally Posted by mariushm

You don't put a capacitor INSTEAD of the 0.1

For 35A+, I'm not quite sure a single 10.000 uF will be enough.

I also can't figure out where from do you get the idea that the transformer can actually do so many amps. Is that rating listed on it?
Usually 20A+ transformers are very large and heavy.

I really can't understand why people keep using these LM317 regulators... they're crap, 1.5A max, large voltage drop... why?

There's better regulators out there like LT108x family which can do up to 7.5A, 1v dropout, much better regulation, at 0.1% compared to 1% on LM317 : http://cds.linear.com/docs/Datasheet/108345fg.pdf
..

Well as to the transformer I don't know the amps. but the 650 UPS had 2 40 amp fuses so I am guessing that it was somewhere below that. It is rather heavy. I would LOVE to know the actual specs on it but google cannot find anything based on the numbers on it. I got circuit diagrams from APC but they say nothing about the amp ratings of the transformer. Please forgive me.

I did not chose the lm317 I happen to have some. I can look for the other ones. I am building this from scrounged parts so again, Please forgive me.

[flinx]

Originally Posted by budm

If transformer can supply two 12V @15A, then total power output is 180Watts x 2 = 360 Watts, then at 24V output, the max current will be only 15A, you cannot get more power than what you have with continuous output (you use large cap to provide surge current handling for the load).
24V @ 35A = 860 Watts, I do not think your APC transformer will be that big.
For bridge rectifier, for example, if you want 1A max load, the transformer has to be rated at 1.6A.

I am just guessing, Much googleing returned nothing for this or any other transformer I have. the output wires are 10awg if that helps. googleing 10awg says max 30 amps.

none of these older APC UPSs have a bridge rectifier like I have seen in other devices. It looks like they did them the old fashion way with 4 diodes as I found 2 sets of 4 diodes in what looks like a BR configuration. so i can't look to that to see what it can take. The diodes are 1n4006.

This is why I am asking questions instead of just slapping stuff together.

[mariushm]

Fuses are almost always rated for much higher values than what the device actually uses.

Remember, fuses are not to protect the device but rather to protect the mains wiring and your house so they're not chosen to blow when your device goes bad..

There's also the issue of inrush current which is particularly important when working with high capacity... for a short moment the capacitors act like a short circuit, until they charge up, when they're empty they simply "pull" as much energy as they can... so at that point smaller fuses could pop because for a very short time the current going through them can reach even 50A

You can learn a lot simply by googling "inrush current capacitor" but here's a good pdf to start with:

http://www.bearpwr.com/pdfs/InrushCurrent.pdf
http://www.epcos.com/web/generator/W...ationnotes.pdf

So yeah, the fuses are 40A but the transformer itself maxes out at something like 15-20A... I don't know.

What you could probably do is to go to a car parts store and buy a few light bulbs for motorbikes or whatever.

I have some around that are rated for 12v 25 watts. (basically 25w resistors... well technically not really but close enough). So if you can find these, but a dozen or so, they should be cheap - mine were around 1$ each.

The 8.6 VAC rectified is about 12.1 V DC ... the actual voltage is probably going to drop to about 10.5-11 when you actually put a load, but it would be enough to test.

Get a multimeter or two if possible, put it on the Amps scale, put it in series with the lightbulbs to measure the current... and with a second multimeter you could measure the voltage.

Keep adding light bulbs in the series and check both the voltage and the current and monitor the temperature of the transformer... if it reaches about 50-60 C on the outsides it means the insides are probably close to 80-90c and you shouldn't really go over.. Therefore, you're probably close to its maximum output or its safety limits.

So the amp meter should tell you how many amps go through it... and the voltmeter would tell you the voltage it drops to under such load.

[flinx]

Originally Posted by mariushm

Fuses are almost always rated for much higher values than what the device actually uses.

Remember, fuses are not to protect the device but rather to protect the mains wiring and your house so they're not chosen to blow when your device goes bad..

There's also the issue of inrush current which is particularly important when working with high capacity... for a short moment the capacitors act like a short circuit, until they charge up, when they're empty they simply "pull" as much energy as they can... so at that point smaller fuses could pop because for a very short time the current going through them can reach even 50A

You can learn a lot simply by googling "inrush current capacitor" but here's a good pdf to start with:

http://www.bearpwr.com/pdfs/InrushCurrent.pdf
http://www.epcos.com/web/generator/W...ationnotes.pdf

So yeah, the fuses are 40A but the transformer itself maxes out at something like 15-20A... I don't know.

What you could probably do is to go to a car parts store and buy a few light bulbs for motorbikes or whatever.

I have some around that are rated for 12v 25 watts. (basically 25w resistors... well technically not really but close enough). So if you can find these, but a dozen or so, they should be cheap - mine were around 1$ each.

The 8.6 VAC rectified is about 12.1 V DC ... the actual voltage is probably going to drop to about 10.5-11 when you actually put a load, but it would be enough to test.

Get a multimeter or two if possible, put it on the Amps scale, put it in series with the lightbulbs to measure the current... and with a second multimeter you could measure the voltage.

Keep adding light bulbs in the series and check both the voltage and the current and monitor the temperature of the transformer... if it reaches about 50-60 C on the outsides it means the insides are probably close to 80-90c and you shouldn't really go over.. Therefore, you're probably close to its maximum output or its safety limits.

So the amp meter should tell you how many amps go through it... and the voltmeter would tell you the voltage it drops to under such load.

Thank you!

I have googled for information on how to determine the amp rating of a transformer and your suggestion makes the most sense.

[budm]

The 40A fuses are for batteries (Batteries are discharged through those fuses when it runs on batteries mode, your main AC input circuit breaker is about 10~12A, see the back of the UPS), we use about the same size fuse for our UPS also, the reason it is high AMP fuses is because we are converting either 12V or 24V from batteries to 120VAC. For example, the AC output is 120V 1A, POWER = 120W (120V x 1A = 120Watts), so to get 12V batteries to produce 120Watts of power, you will have to draw 10A from batteries, 12V x 10A = 120Watts if we have zero loss in conversion process, in real world, it will be more than 10A. Batteries can high inrush current, for example, your car batteries is rated with cold crank amp in 500~800A range.

[tom66]

Also, remember if you have a 0-24V adjustable range, and want just 3A at 1V out, you must dissipate (at minimum) 72W in your LM317. (Assuming it's the 5A rated variant. It will be more because of the higher input voltage required due to ripple voltage and dropout, so figure close to 80-90W.) Which is a lot of power, in fact I think the SOA protection may operate before that.

[mariushm]

Indeed, the excess voltage will be dissipated as heat in the regulators.

Ideally, you'd switch between using a single secondary or both secondaries in parallel for lower voltages and both secondaries in series for higher voltage.

Maybe using relays to switch at certain threshold.

[budm]

For your big DC motor driver, you should use PWM type, for example:
http://www.electronickits.com/kit/co...or/CKMX033.htm

[flinx]

I have a 5 ohm and a .5 ohm resistor similar to this. Using the 5 ohm draws about 2 amps at 24 volts. I need to find something else to draw more of a load.

[budm]

If you do not have anything else to use, you can use 200W 120V lamp which has cold resistance of about 5 Ohms, just parallel them up.

[flinx]

note to self...

large .5 ohm resistors draw 25 amps at 24 volts...
This causes some of the smoke to be freed from the resistor...
do not do that again...

However the transformer mains circuit breaker did not trip nor did the transformer heat up in the 30 to 45 seconds it was on.

The resistor was not destroyed, it did get VERY hot, around 350 degrees by my infrared thermometer.

[budm]

If Voltage is 24V and the load is 0.5 Ohm (not sure how big the wire are using) then the current draw is 48 Amp (that is 1152 Watts of power), so your power supply cannot supply that current for sure. I doubt that the output was at 24V with 0.5 Ohm load. So your transformer is more like 500Watts. Is the UPS rating at about 500Watts (850VA)?

[flinx]

Originally Posted by budm

If Voltage is 24V and the load is 0.5 Ohm (not sure how big the wire are using) then the current draw is 48 Amp (that is 1152 Watts of power), so your power supply cannot supply that current for sure. I doubt that the output was at 24V with 0.5 Ohm load. So your transformer is more like 500Watts. Is the UPS rating at about 500Watts (850VA)?

OK, well the volt meter read 11.something or other at the time so...

UPS was rated at 650.

wire sizes are 10 to 12awg

I'm gonna cap the PS at 20 amp when I use it.

[budm]

Is it 650VA or 650Watts, if the model is xx650, then it is 650VA or about 400Watts which is about right, I have almost the same form factor for my old APC UPS 650VA..

[flinx]

Originally Posted by budm

Is it 650VA or 650Watts, if the model is xx650, then it is 650VA or about 400Watts which is about right, I have almost the same form factor for my old APC UPS 650VA..

I never actually looked at the back of it. It is a 650va, and next to that it says 410 watts. So I should cap it at 15 amps?

[budm]

That is about to limit of it.