Calling the Forward Converters Gurus - Please Help !!

**Zznakie** · 10-03-2004, 03:15 PM

Hi everybody,

I'm looking for help on building my forward converter. For people not knowing what a forward converter is, basically it's the modification of an AC-DC power suply into a DC-DC unit. There's a very extensive and old thread discussing the subject here:

Well, I've already read all the info on the above thread and also on marsjell's website , read some theory on transformers and am willing to begin the job. So I unmounted my ATX PSU and did tried to identify the parts of th PSU. For my surprise (and if I'm not wrong) there is a clear separation of the AC part of the PSU from the DC part. I handled to identify the transformer wich I will have to modify, but some things called my attention, and since I don't know much about electronics (did take basic electronics at University) I decided to ask for help of all the gurus that have already done this job and from other electronic gurus that always hang around here. I've taken some pictures to help me explain my doubts: (Please scroll down in order to check my questions written after the pics)

So, my doubts are the following

:

1-. I've got 6 components (3 IC's, 2Transformers and 1 ??) joining the 2 parts (ac and dc) of my PSU. After removing them and re-winding the transformer will I just leave them out??? Is it going to work fine???

2-. I searched the internet for info on any of the 6 components above and found nothing. Does anybody have any idea of what those 3 Ic's do???

3-. My PSU has 2 transformers (as can be seen in the pic). So far I have only heard speaking of 1 transformer. What's the purpose of the smaller transformer??? I believe that I'll be working (re-winding) the bigger one).

4-. After building the circuit (I belive it's Presslab's circuit design that I have) and re-winding the transformer shall I connect the output of the transformer back to the PSU?? The circuit says that the output is 12v, 5v and 3.3v?? Do I connect it back to the PSU in order to regulate these voltages and obtain the negs??

Sorry if my questions seem to lame

but I really need some help.

Thanks in advance,

Zznakie.

**Rob Withey** · 10-03-2004, 03:57 PM

I've not built one of these myself, but I'll help where I can:

Originally Posted by Zznakie

1-. I've got 6 components (3 IC's, 2Transformers and 1 ??) joining the 2 parts (ac and dc) of my PSU. After removing them and re-winding the transformer will I just leave them out??? Is it going to work fine???

2-. I searched the internet for info on any of the 6 components above and found nothing. Does anybody have any idea of what those 3 Ic's do???

3-. My PSU has 2 transformers (as can be seen in the pic). So far I have only heard speaking of 1 transformer. What's the purpose of the smaller transformer??? I believe that I'll be working (re-winding) the bigger one).

Those small ICs are optocouplers. They close the feedback loop from the primary switching side to the secondary regulation side. They are needed. The second small transformer may be one of two things, either a small transformer for the 5v standby (the most likely explanation), or another part of the feedback loop. You may need to rework the 5v standby section accordingly.

Originally Posted by Zznakie

4-. After building the circuit (I belive it's Presslab's circuit design that I have) and re-winding the transformer shall I connect the output of the transformer back to the PSU?? The circuit says that the output is 12v, 5v and 3.3v?? Do I connect it back to the PSU in order to regulate these voltages and obtain the negs??

The PWM circuit will provide the regulation. It hooks up to the optoisolators on the AC side. Did you forget this circuit?

This is not a trivial mod, you really do need to know what you are doing, most power supplies are slightly different.

**Zznakie** · 10-04-2004, 12:42 PM

Hi Rob!

First of all thank you for replying and for your valuable knowledge and help.

Where can I start? Let's see...

I did some research on optocouplers and understood that this are safety components used for circuit separation (I believe in this case separation of the AC circuit from the DC circuit) please correct me if I'm wrong.

I also checked back on what you call the PWM Circuit ( I was refering to it as Presslab's circuit) and realized that there are some connections to optocouplers but there are only two of them ( on/off opto and voltage opto) and my PSU has three of them.

How can I identify wich is the on/off and wich is the voltage one?? What to do with the third one??

I also analized marsjell's picture of the modified PSU and there you can see the three optos, but when you refer to his PCB and component placement this 3 optos dont seem to connect to anything.

Is it possible that a "wire bridge" is conecting R3 and R10 to the opto as in presslabs circuit says??

This is not a trivial mod, you really do need to know what you are doing, most power supplies are slightly different.

I don't really know much of electronics and that's why I'm asking for your help and comprehension, but I'm really studying a lot so I can keep up on the discussions. I've done a lot of things on the past that I couldn't believe I could do. You just have to try hard enough, study a lot and have some help

Best regards,

Zznakie.

**MrPerfectionest** · 10-04-2004, 02:55 PM

I don't really know much of electronics and that's why I'm asking for your help and comprehension, but I'm really studying a lot so I can keep up on the discussions

I know many diploma'd EE students that it would take them quite a bit of study time to understand how a forward converter works, then take an existing one meant for AC supplies, characterize the circuit, and change it to work with DC only, would be extremely difficult.

I am not trying to discourage you but what you are attempting isn't easy by any means. Not even close.

**Rob Withey** · 10-04-2004, 04:33 PM

Originally Posted by Zznakie

I also checked back on what you call the PWM Circuit ( I was refering to it as Presslab's circuit) and realized that there are some connections to optocouplers but there are only two of them ( on/off opto and voltage opto) and my PSU has three of them.

How can I identify wich is the on/off and wich is the voltage one?? What to do with the third one??

Trace the tracks on the board and understand how the power supply works.

Originally Posted by Zznakie

Is it possible that a "wire bridge" is conecting R3 and R10 to the opto as in presslabs circuit says??

Yes.

**Zznakie** · 10-08-2004, 01:04 PM

Hi!

Since there may be some other people interested in acomplishing this project, I decided to get together all the posts that I have found about presslab's project of building a dc-dc converter from a regular ATX PSU.

Anyone try a Forward Converter?
Presslab I have a problem... again...
presslab's powersuply..
presslabs forward converter
Presslab, I have a problem
Presslab a problem... again
Whoo! Got Presslab's DC-DC convertor working!
Finished Tutorial for PressLab's DC-DC
Presslabs Psu Modification

Anybody interested on this subject can mail me. I'll be glad to share this project with anybody.

Zznakie.

**Zznakie** · 10-13-2004, 09:41 PM

Hi all!

Well, yesterday I had prepared a very big post and when I pressed the "submit" button my computer crashed, so I'll try to resume here all the points that I had written yesterday. I'll be posting a lot of questions, some of them have already been asked but will be posted again since I haven't completely understood the answers given and expect Presslab to drop by, as well as the contribution of all the members of this forum. BTW I'm working with an ATX 150W PSU that has a switch on the rear for selection of the input voltage (115-240).

1-. I haven't been able yet to identify the optocouplers, so I don't know wich is the on/off and wich is the voltage. Worst of all, I have a third opto and don't know wich is it's function. Probably for the 5V stdby? (see the next point in order to understand this 5V stdby issu).

2-. I've got a second smaller transformer. Rob said that it might be for the 5V stdby and that makes a lot of sense after reading all the previous posts on this subject. My doubt here is: After I reap it off, I'll need a simple circuit for the 5V stdby won't I? Marsjell included this in his project but I don't really understand how this simple circuit connects to the secondary side of the PSU.

3-. I did some measurements before reaping of my transformer and found only 12V and 3.3V no 5V. Is this possible or did I measure something wrong?

4-. I reaped off my transformer and this is what I found:

Primary Side (3 Pins)

Pin 2 => 26 turns CCW (watching from Top to Bottom) doubled => Pin 1

Secondary Side (6 Pins)

Pin 2 => 5 turns CCW (watching from Top to Bottom) tripled => Pin 1
Pin 5 => 4 turns CCW (watching from Top to Bottom) tripled => Pin 3
Pin 6 => 4 turns CCW (watching from Top to Bottom) tripled => Pin 4

Back to Primary Side

Pin 3 => 26 turns CCW (watching from Top to Bottom) doubled => Pin 2

This is what I did:

Primary Side: Followed Arby's tutorial, so I made 3 coils, each having 3 turns of 18 gauge wire doubled up.

Secondary Side: Got a little bit confused about the pins and the polarity, so I just got the 22 gauge wire and wound up everything just like it was but with the double of turns so now it is:

Pin 2 => 10 turns CCW (watching from Top to Bottom) tripled => Pin 1
Pin 5 => 8 turns CCW (watching from Top to Bottom) tripled => Pin 3
Pin 6 => 8 turns CCW (watching from Top to Bottom) tripled => Pin 4

Here I have a lot of questions: What kind of glue should I use to join back the two parts of the transformer's core? I thought of using loctite's crazy glue but since I may need to apart it again for rewinding it (if something is not OK) maybe this is not a good choice. Since I'm having some trouble finding some of the PWM circuit, I would like to know how can I test the transformer alone? Where should I connect the voltage and what result should I expect? Where shall I do the measurings?

Well guys I believe this is a pretty long post for today, I'll probably be posting some more questions and maybe some photos of the transformer mod (if needed) some other day.

Best regards,

Zznakie.

**Zznakie** · 10-18-2004, 05:05 PM

Arby made a post on the original thread asking:

Resistors - Normal 1/4 watt metal film resistors good?
Capacitors - 100V, electrolytic fine?

No body answered, neither did he ask again. May I assume that the answer to both questions is yes?

I'll keep on posting.

Best regards,

Zznakie

**Rob Withey** · 10-18-2004, 05:24 PM

Yes to both questions, but for the high frequency sections, you are better off using low ESR electrolytics rather than conventional.

**presslab** · 10-20-2004, 02:11 PM

Originally Posted by Zznakie

Hi all!

1-. I haven't been able yet to identify the optocouplers, so I don't know wich is the on/off and wich is the voltage. Worst of all, I have a third opto and don't know wich is it's function. Probably for the 5V stdby? (see the next point in order to understand this 5V stdby issu).

2-. I've got a second smaller transformer. Rob said that it might be for the 5V stdby and that makes a lot of sense after reading all the previous posts on this subject. My doubt here is: After I reap it off, I'll need a simple circuit for the 5V stdby won't I? Marsjell included this in his project but I don't really understand how this simple circuit connects to the secondary side of the PSU.

You can not use the voltage feedback optocoupler if you like. Because the supply is not isolated (in GND = out GND) you can just tie the feedback to the 3.3V output of the supply. Make sure the ground connection between the input and output is very stout though.

For the on/off opto, follow the traces on the circuit board and try to figure out which one goes where. The on/off opto will probably connect to some voltage brownout chip, followed by the on/off line on the ATX connector.

That small transformer and opto are probably for the 5VSB power. If you want standby power (a full working ATX supply), you will have to replace that supply with your own. A simple buck converter or even a linear regulator (5V @ 750mA) will suffice. The linear regulator will get really hot though. IIRC, I used a National Semiconductor Simple Switcher LM2675-5.0.

Originally Posted by Zznakie

3-. I did some measurements before reaping of my transformer and found only 12V and 3.3V no 5V. Is this possible or did I measure something wrong?

I'm not sure where you were measuring, but sometimes the output windings of the transformer are put in series to supply the various voltage outputs.

Originally Posted by Zznakie

4-. I reaped off my transformer and this is what I found:

Primary Side (3 Pins)

Pin 2 => 26 turns CCW (watching from Top to Bottom) doubled => Pin 1

Secondary Side (6 Pins)

Pin 2 => 5 turns CCW (watching from Top to Bottom) tripled => Pin 1
Pin 5 => 4 turns CCW (watching from Top to Bottom) tripled => Pin 3
Pin 6 => 4 turns CCW (watching from Top to Bottom) tripled => Pin 4

Back to Primary Side

Pin 3 => 26 turns CCW (watching from Top to Bottom) doubled => Pin 2

This is what I did:

Primary Side: Followed Arby's tutorial, so I made 3 coils, each having 3 turns of 18 gauge wire doubled up.

Secondary Side: Got a little bit confused about the pins and the polarity, so I just got the 22 gauge wire and wound up everything just like it was but with the double of turns so now it is:

Pin 2 => 10 turns CCW (watching from Top to Bottom) tripled => Pin 1
Pin 5 => 8 turns CCW (watching from Top to Bottom) tripled => Pin 3
Pin 6 => 8 turns CCW (watching from Top to Bottom) tripled => Pin 4

Here I have a lot of questions: What kind of glue should I use to join back the two parts of the transformer's core? I thought of using loctite's crazy glue but since I may need to apart it again for rewinding it (if something is not OK) maybe this is not a good choice. Since I'm having some trouble finding some of the PWM circuit, I would like to know how can I test the transformer alone? Where should I connect the voltage and what result should I expect? Where shall I do the measurings?

Sounds like you did everything right. To test the supply, I just put a couple turns of electrical tape around the cores to hold them tightly together. Epoxy adhesive should work just fine once everything is working right.

You can't really test the transformer alone, at the least you will need the PWM driver.

One thing is for certain, test the completed ATX supply before hooking it up to something you care not to burn up!

Originally Posted by Zznakie

Well guys I believe this is a pretty long post for today, I'll probably be posting some more questions and maybe some photos of the transformer mod (if needed) some other day.

Best regards,

Zznakie.

Good luck! Let us know if you need more assistance.

Presslab