Anyone try a Forward Converter?

presslab · 07-31-2001, 11:49 AM

Last weekend I modified a line powered ATX supply to work off 12V. I wound a transformer, and built up the PWM control circuit. So basically half of the supply is kept intact, and the other half (AC side) is replaced. It works great to 8V, but below that it shuts down. I think I've got that problem solved, so it can go down to about 6.5V for a brief moment. (Starting the car, I measured my 2000 VW to drop the battery voltage to 7.5V.

)

Anyway has anyone seen plans for such a supply? If there was enough interest I could draw up some schematics and post them here.

Presslab

marsjell · 07-31-2001, 12:34 PM

oeh oeh me me I want!

if it is really that easy.... but shure, I'd like to see how you did this

pinkysqueaks · 07-31-2001, 01:57 PM

PressLab,
We are always looking for new, better, and cheaper powering solutions. I do not know much about power supplies, but your solution sounds very interesting. I, as I'm sure many others, would like to know more

mpt · 07-31-2001, 06:20 PM

Hey this is great! Someone really did it! I tried to modify my AT supply about a year ago so it worked with a 14.4V - 10V supply, but i found out i nedded a very large transformer to transform those 14.4 to 48V so it could work on the PMW circuit. But it didn´t work out and i give in. Maybe i fried something before trying it. I guess ill never know...

Please post that circuit here on the forum (if you want to of course).

cenwesi · 07-31-2001, 06:22 PM

PressLab, we are all waiting and can't wait no MORE.....

magnetik · 07-31-2001, 06:22 PM

looks like you could solve alot of problems PressLab. I would be very interested to see this too....

presslab · 07-31-2001, 07:45 PM

Okay, here is a little info.

The transformer is a large E-Core that I stripped from a 300W AC-DC supply I bought at the surplus store for $7. I need the big ferrite core so it doesn't saturate at the 30A+ peak primary current. The Forward Converter topology helps with this too. I put the transformer in the oven to loosen the glue and carefully disassembled it. I counted the turns ratio and did some quick inductance calculations and I changed the secondary turns by a factor of 2. Meaning for each seconday turn that I unwound, I re-wound it with twice the turns. I also used a slightly smaller gauge wire but I doubled it up. Less eddy currents in the wire this way. I noted which windings were on which layers so as to rewind it in the same fashion.

The transformer has three primary windings, each 3 turns, 18GA doubled up. The secondary windings are all 22GA, they are 8 turns doubled up, 4 turns tripled up, and 2 turns tripled up. It is VERY important to keep the turns ratio, and have exactly 2 turns for example. (Bend the wire at 90deg when done with the winding back to the bobbin pin.) Also the layer of each turn is important too. I will post all this info when I make it pretty.

The PWM chip I used is the SG3525A, it is only $0.75, but now that I see the 7.5V dip I think I will use another chip. I'll post the current design for now.

It operates at 125kHz which seems to be pretty efficient, but I will play with it to see if I can get the duty cycle lower. (i.e. more efficient.)

I used two IRF540N switch transistors, they are excellent, with Rds(on) of only 0.04 ohm! They are also popular and cheap. Their Vgs should be good at 6.5V. I will try the IRL540 to see if it help it to go lower in input voltage.

I used a National Semi. LM2671M-ADJ simple switcher to power the +5V standby, as I had this available and the current demands are low. (720mA) I will probably design some simple circuit instead of the LM2671, but for now it works fine.

That's all for now, I'll post more when I get it ready.

Presslab

magnetik · 07-31-2001, 08:49 PM

sounds good!!!!

so what does the transformer windings do? raise the input voltage higher??? like an inverter???? or am I missing something.???

If you have a camera, some pics would be great!! and a circuit diagram...

... hey, we dont ask for much!!!!

zootjeff · 07-31-2001, 08:50 PM

I thought about doing that too! Where did you learn about transformer design? That’s my biggest stumbling point. Can you recommend any books? As for your 7.5-volt problem, I used a Schottky Diode and a Capacitor to filter the input of the supply, and worked just fine for me with the 10.5 - 14 volt laptop supplies I was using. I'm not sure how long your battery stays below the threshold, but I used an MBR3045 and a 3300 uF Cap and it solved my problems. (I know how it feels to want to avoid that though..)

Jeff_

zootjeff · 07-31-2001, 09:07 PM

[QUOTE]Originally posted by magnetik:
[QB]sounds good!!!!

So what does the transformer windings do? raise the input voltage higher??? like an inverter???? or am I missing something.???
QUOTE]

Basically what a ATX AC-DC power supply does is takes the 120 in, converts it to 170 with a bridge rectifier and caps. It is then switched on off into the primary of a high frequency transformer. The secondarys have a turn ratio with the primary such that 170 goes to 12, 5, 3.3 etc..

If I'm not mistaken, and correct me if I'm wrong, all that Presslab did was rewind the transformer so that the secondary windings look the same as they did before but the primary has a lot less turns so that an input of 12 volts can be used instead of the 170 volts. He scrapped the old transformer because the current on the primary previously was fairly low at 170 volts but now to get the same kind of power rating, he needed a lot more current and the old transformer would saturate (cause problems under heavy load). He scrapped the pwm circuit for controlling the 170 volt transistor switcher and used his own for the new 12 volt one. Is that more or less it?

Jeff

presslab · 07-31-2001, 09:36 PM

[quote]Originally posted by Jeff Mucha:

Quote:

Originally posted by magnetik:
[QB]sounds good!!!!

So what does the transformer windings do? raise the input voltage higher??? like an inverter???? or am I missing something.???
QUOTE]

Basically what a ATX AC-DC power supply does is takes the 120 in, converts it to 170 with a bridge rectifier and caps. It is then switched on off into the primary of a high frequency transformer. The secondarys have a turn ratio with the primary such that 170 goes to 12, 5, 3.3 etc..

If I'm not mistaken, and correct me if I'm wrong, all that Presslab did was rewind the transformer so that the secondary windings look the same as they did before but the primary has a lot less turns so that an input of 12 volts can be used instead of the 170 volts. He scrapped the old transformer because the current on the primary previously was fairly low at 170 volts but now to get the same kind of power rating, he needed a lot more current and the old transformer would saturate (cause problems under heavy load). He scrapped the pwm circuit for controlling the 170 volt transistor switcher and used his own for the new 12 volt one. Is that more or less it?

Jeff

That's pretty much it. The secondary windings are actually all doubled in the number of turns. Otherwise I'd need 1.5 turns on the primary. Talk about a peak current! The output cross regulation is within 5%, so doubling the windings didn't droop the output.

I don't really want to go the schottky diode route, at the best it's lost efficiency (heat), at the worst it's little fluffy clouds of black smoke (surge current). (120W / 0.80%) / 8V = 18.75A. 18.75A * 0.5Vd = 9.4W diode loss

I planned on getting a book from Amazon for $90, but I gave it a shot first. I guess I got lucky, I never bought the book. I downloaded all the app notes on flyback and forward converters I could find. I'll post references to them when I get time.

Quote:

If you have a camera, some pics would be great!! and a circuit diagram...

I have a camera, and I'll probably just throw up a web page to stick all the stuff on it.

Presslab

magnetik · 07-31-2001, 11:38 PM

Jeff, PressLab thanks very much for your explinations....

And since we are ****ing off the original tranny, the same winding ratios would work on ATX supplies designed for 240Volt Australian usage too, would it not?

How small is the final powersupply. Can you fit the re-designed transformer back into the ATX case shell..... maybe this is how the ARISE guys build their powersupply units!! hehehehe!

I guess if some schematics were posted, and the ferrite core could be purchased of the size required (it is a core, not a rod?) im guessing you would have the problem solved for us guys like myself still using a PS/inverter setup.....

this is exciting stuff...

magnetik · 07-31-2001, 11:41 PM

Quote:

Originally posted by PressLab:
The transformer has three primary windings, each 3 turns, 18GA doubled up. The secondary windings are all 22GA, they are 8 turns doubled up, 4 turns tripled up, and 2 turns tripled up.

another thing... does this double/tripple windings explain the multiple 12volt/5volt/3.3volt outputs needed?

zootjeff · 08-01-2001, 03:10 AM

Quote:

Originally posted by magnetik:


another thing... does this double/tripple windings explain the multiple 12volt/5volt/3.3volt outputs needed?

No, the double/tripple windings just ups the ammount of current that can be induced. Plus it lowers the eddy currents, (which I don't understand, brain dumped that from my undergrad education, is it something to do with skin effect or the way current flows through a round wire, or hysterisis?? Arg, too much digital design knoledge pushing out the analog stuff.) :-)

The number of turns dictates the voltages.

Jeff_

presslab · 08-01-2001, 03:23 AM

Quote:

Originally posted by magnetik:
Jeff, PressLab thanks very much for your explinations....

And since we are ****ing off the original tranny, the same winding ratios would work on ATX supplies designed for 240Volt Australian usage too, would it not?

How small is the final powersupply. Can you fit the re-designed transformer back into the ATX case shell..... maybe this is how the ARISE guys build their powersupply units!! hehehehe!

I guess if some schematics were posted, and the ferrite core could be purchased of the size required (it is a core, not a rod?) im guessing you would have the problem solved for us guys like myself still using a PS/inverter setup.....

this is exciting stuff...

Yeah, the transformer will work for the 240VAC ones as well. If your power supply is set up like mine, then the same transformer could theoretically work without any winding differences. For example, the old transformer had a primary winding of 41 turns (120VAC). Yours would probably be the same, if you have the switch in the back for 120/240, or if it says 'autoswitch'. If it is 240VAC only, it might be 82 turns on the primary. They basically just do a full-wave or half-wave rectification based on what the switch is set to.

The modified power supply is small enough to fit into the same microATX power supply case, although the fan needs to be moved up to allow the bigger transformer to fit. I just took out the funny looking fan, and replaced it with a regular case fan, if fit fine. When it is done a good dose of hot-glue will keep the pieces from rattling and breaking in the car.

The ferrite core is called the "E" core. It is two E's back to back like this "E3". There is a bobbin, like the sewing kind but square, that you wind the wire onto. The two E's then slide over the bobbin, and they touch in the center and outer edges. Sometimes the E core center leg has been gapped, so it doesn't touch in the middle. This allows a magnetic flux bias in the core without saturation. We don't need that with the Forward Converter, it just hurts the maximum flux capacity of the core. In a continous-mode flyback converter, there is a magnetic bias, therefore the core is probably gapped. Any power supply over about 200W is not a flyback, as those don't work well with that much power.

So to scavenge parts go with a nice beefy supply. I bought two, good thing because I broke one of the E cores getting it off!

Good thing I had a spare.

I just did some testing, and at 7V I can do 60W (probably more, my 12V supply only does 11A) It is about 60-80% efficient, based on load (best efficiency is at about 50W load). I played with the frequency a little to get it better, but I think I can get more out of it.

I am thinking more and more on getting a better PWM chip that goes down to 6V or so and using the IRL540s. Any suggestions on a cheap flexible PWM chip with mosfet sink/source drivers?

Presslab

User Name		Remember Me?
Password