• Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Rebuilding My Power Supply

Status
Not open for further replies.

kinarfi

Well-Known Member
images?q=tbn:ANd9GcQfgkZy_gktN01xPwhRlkSS6NaTipYgo_wA_2ROagaMr3RWoEDOBgthis is like the pot I AM using, unfortunately,(For you guys cause it don't bother me :), it is a 10K pot and to get the 20 volts out, I have to use 500 ohms from out to adj, and I don't need the accuracy and usually pulling much more than 10ma, that was the design I plan to use. I struggle to think of a place or situation where having the 10ma drawn by the bias resistor is important, the only thing I can think of is if you watch the meter that tells you what your voltage before you apply your load.
 

audioguru

Well-Known Member
Most Helpful Member
With a regulated power supply, I EXPECT its output voltage to be what I set it to be regardless of its load.
A 10k pot's resistance is much too high for 20V. You should use a 120 ohm resistor and a 2k pot.
 

kinarfi

Well-Known Member
I made a mistake today and fried another TIP 35 or 36 transistor, they don't seem very tough to me, maybe I should have parallelled 2 or 3 of them, oh well, I think this design should work all right, it's similar, but improved from what I had before.
One strange thing I ran into with spice was in the + half of the power supply design, I used a dual opamp and only needed one, so I used the other one as a buffer and if I include it in the simulation, spice finds single matrixes in other areas of the design. Another strange thing is that if I try to sim with a LM317 and a LM33, it won't sim the LM337 half of the design. It will if I delete the LM317 though.
I have some LM317s that are rated for 1.5 amps and some for .1 amps, if you look at my drawing, that's why the LM317s are labelled. This new design appears to be very stabile, as you put more load on it the voltage stays constant.
I did find where Audioguru was correct, when there was no load and the dial was set for 0, it would put 4 to 8 volts out and drop back to .7 when loaded. To keep it from floating up, a 250 resistor from out to ground seems to work, but it needs to be a 2 watt for when the dial is set to 20v. Another nice thing about setting the compare voltage, 10V, with a lm317, .1 amp, is you can tweek the output so that when I set the dial at 5, I get 10.0 volts and at 7.5 setting, I get 15 volts.
I hope to finish it up tomorrow, again, and I'll take a few photos of it.
 

Attachments

audioguru

Well-Known Member
Most Helpful Member
The datasheet for the TIP35 shows that it can dissipate 125W if its heatsink is perfect and does not even get warm. Maybe if it is huge, made of silver and you use a very powerful fan.
But using a HUGE normal heatsink it will be VERY hot dissipating 60W.
With a 24V supply and a 5V output then it dissipates 60W when the current is only 60W/(24V - 5V)= 3.15A.

EDIT: Your schematic has a TIP35 NPN connected to the LM337 instead of a TIP36 PNP.
 
Last edited:

kinarfi

Well-Known Member
EDIT: Your schematic has a TIP35 NPN connected to the LM337 instead of a TIP36 PNP.
Yes, that is correct, but instead of a TIP35C, I have a 163H72 NPN transistor which seems to be able to handle all the current that the transformer can deliver. I have split the power supply into pos and neg so they can be simmed because spice won't sim if both sides are present, I also had to remove half of the LM324 op amp which was use as a buffer because spice wouldn't sim with it in.
Yes, things get warm, I'm usually using about 14 volts out so my heat isn't quite so bad and I do have lots of fan cooled surface area for each 163H72 transistor plus another fan pushing fresh air into the enclosure.
 

Attachments

audioguru

Well-Known Member
Most Helpful Member
Sorry, I was wrong. The transistor in the high current negative supply should be NPN.
Your base resistor value is 100 ohms but datasheets show only 3 ohms.
 

kinarfi

Well-Known Member

audioguru

Well-Known Member
Most Helpful Member
Many schematics on that website are WRONG. Somebody posted one yesterday that is horrible and it does not work.
 

MrAl

Well-Known Member
Most Helpful Member
Hi,

I would recommend careful testing of any power supplies found on the web no matter where they come from.

Also, i saw a power supply design using LM324 IC's for part of the control circuit for the feedback. Although these parts can work, they are a little slow for power supply designs. The better choice is the TL082 or similar op amps which have faster response. In fault modes we want the circuit to be able to respond as fast as possible.
 

kinarfi

Well-Known Member
I've been working on this power supply and I have 2 requisites, one is NPN transistors for the outputs, and the other is 10K pots. The positive half was no problem, the negative half has been a pain, but I did figure out one thing about the configuration I'm using, something I probably should have known, is the resistor between the LM337 input and the -24v supply creates the bias voltage (~.7V) for the NPN and as the load current increases, the LM337 passes more current, the bias voltage increases and the pass current increases. The model link in post 47 uses a 3.3 ohm resistor to bias a TIP142 darlington, but I using a single 163H72 and I have also found that the higher the bias resistor, the lower the current through the LM337, I've settled on 10 ohm 1 watt, so far. My LM337 shares a heat sink with the pass NPN which makes direct contact (threaded) and when the insulation between the LM337 tab (input) fails, it sure messes up my readings, took me a while to find that problem, I was beginning think that this design wouldn't work, but seems ok now. Hope to finish up tomorrow.
 

Attachments

Last edited:

kinarfi

Well-Known Member
Hi,

I would recommend careful testing of any power supplies found on the web no matter where they come from.

Also, i saw a power supply design using LM324 IC's for part of the control circuit for the feedback. Although these parts can work, they are a little slow for power supply designs. The better choice is the TL082 or similar op amps which have faster response. In fault modes we want the circuit to be able to respond as fast as possible.
I agree with you about testing, that's where Spice is great. As for using LM324, that's what I'm using for the positive side, on the negative side, I'm using a MC34071 ( best I can remember) because they must be able take 24 volts and my application don't need to be that responsive and should be plenty fast for my uses.
Thanks for the suggestions.
Kinarfi
 
Last edited:

MrAl

Well-Known Member
Most Helpful Member
Hi,

Sure, you are welcome. I've used LM358's too but they are not ideal for power supplies either. Nice thing is they dont require large power supply voltages like the TL- type op amps do. That's one thing i have always hated about op amps like the TL082 and similar. But there are many more op amp choices out there these days too with very fast response speeds and they are not super expensive for one-off projects either, so it pays to look around a little.

Did you mention what you are going to test with this? I ask because the schematic seems to have gotten a lot more complex then at first.
 

kinarfi

Well-Known Member
No, never said what I was using this on, mostly just puttering around, like I'm doing with this power supply and yes, I tend to go from simple to complex and add some bells and whistle as I go, like the negative half that I'm adding now. When I acquired this thing, it was just a 20 amp 13.6v battery replacement for high powered 2 way radios.
This is what I dreamed up when I decided to make it a variable power supply, then I added a second 163H72, but I let it get so hot, it melted the solder and one of the balance resistors fell down and made contact with something that fried the base, I tore the transistor apart and it appears to be 3 transistors in one, that's probably why it can handle a lot of current. I drew this up about 8 years ago, before I found lots of help from Electro-tech-online. After I flew a plane into the ground, I bought a Joyner Trooper T2 and built a power steering assist system for it and then started improving it over the next few YEARS. Started with a set of contacts that pulled in one or the other relay to drive the gear box of Power Wheel tied to my steering shaft, Now I have a much more powerful motor that's PWM. I also have a Lambda power supply, but it can only deliver 6 amps, and some of my stuff needs more like some of my LED driving lights that I built.
So, mostly it's puttering around, something to keep me busy since my accident retired me.
Speaking of getting more complex, I may go with digital read out amp meters.
 

Attachments

Last edited:

kinarfi

Well-Known Member
Looking at ways to get the voltage I want from a 24V source, here's one that works, but I was hoping for a SMPS, but it seems it is just another linear design. It won't sim for me, but the actual actual out has a saw tooth pattern.
I think the FET is dropping into the linear region which creates heat, there for no better than the some variant of a LM317 other than it can handle far more current.
I would like to here what others have to say about this design, maybe how to cause it to actual switch via hysteresis maybe, I haven't tried it yet, but will later.
Thanks,
Kinarfi
 

Attachments

MrAl

Well-Known Member
Most Helpful Member
Hi,

Well the cause of heating is that circuit is just a pulser, not a true switcher. That means that somewhere it is dissipating the overhead voltage as heat. Even thought the transistor may turn on and off, it will still dissipate heat as if it were just a resistor. So if the input voltage is 24v and the output is 14v and the output current is 1 amp, the transistor and R6 together have to drop that 10v and at 1 amp that comes out to 10 watts. The fact that the transistor turns on and off only means that it can regulate the output, not do it efficiently. So there will be lots of heat especially at low output and high current, just like a linear.

To get the dissipation down (and efficiency up) there has to be an energy storage device somewhere in the circuit and this works out to an inductor in the buck circuit. So if you need a step down operation then you would be better off just using a buck circuit to do it. That would include the inductor which should have a rating to match the max output current and value depending on the frequency.

Buck circuits are one of the easiest switching circuits to build and understand. They work almost the same as a linear except the pass device turns on and off, and they dont waste all that power due to the overhead voltage because the inductor can store all the excess energy.
 

kinarfi

Well-Known Member
I have looked for some simple switchers to like the LM2576, but they are all limited as to how much current they can handle. My previous (heater) design appears that it can handle some current, but I haven't tested it to destruction, yet. I played around on spice and I don't know if it can be done or if it's something I can do, but, would it be possible to kind of copy the design of the LM2576 far enough to allow for the use of a IRF4905 P FET and handle something like 10+ amps.
 

Attachments

MrAl

Well-Known Member
Most Helpful Member
Hi again,

The LM2576 chip is a nice chip to work with. It's very easy to use and it works great. One thing that always bugged me too though was that it wont go beyond 3 amps, so i thought of ways around this apparently just like you are now.

My workaround idea was to still use the LM2576, but instead of having it as the control and driver, just use it for the control. In other words, use it to drive a hefty MOSFET. I think a P type might be just right too. So we'd still be using the LM2576 but now with a few extra high current stage parts. If you are interested that would be good because i always wanted to try this but never got around to it. If you are in fact interested in trying this then i'll draw up a schematic as soon as possible.

It's always possible to build a buck circuit similar to the LM2576 by using comparators and op amps, but if we have all the control on one chip that would be nice.
 

kinarfi

Well-Known Member
I gave it a quick try already by tying an inverter to the output and the inverter to the gate of a PFET, that didn't get me any where, but the out of the lm2576 seemed to have a small spike, probably some kind of test to see if it had an inductor tied to it. Any way, what I tried failed, so if you would be so kind as to share your idea with me, I'd be very glad to try it and let you know what happens.
Kinarfi
 

ronv

Well-Known Member
Most Helpful Member
Interesting discussion.
The problem I've always had with big buck regulators is that all the current has to go thru the inductor so they get really big (read expensive) really fast. The chips get around this with high frequencies which makes the inductor smaller but presents it's own set of problems. I have a bunch of 4905's. PM me your address and I can send you some if you like. Mr. Al, what do you think of winding a big old air core inductor?
PS. With a 24 volt transformer you are getting dangerously close to the 36 volt rating on a 339 comparator.
 

kinarfi

Well-Known Member
I do appreciate the offer of 4905s, but I also have plenty of them. I have some 33uH inductors that look like they were made with #10 fine strand wire that I salvaged out of some server poweer supplies. About my 24 volt transformer, 24 is the rectified, filtered voltage.
Another idea I tryed was to use the 555 in a method similar to using them for boost, but I couldn't come up with a with a method to invert and apply a signal to Cv as the derived voltage changed.
 
Status
Not open for further replies.

Latest threads

EE World Online Articles

Loading
Top