Switched-mode PSU turned into simple adjustable lab PSU

[Thomeeque]

 Hi,

as I currently need to measure dependency of one circuit (EM clock namely) on it's supply voltage, I had to solve lack of PSU with adjustable output voltage in my humble "lab" somehow. First I planned to do something simple with LM317, but during digging for suitable transformer I did found nice light (and safely boxed!) switched-mode 24V/1A PSU module, extracted from friend's expired inkjet printer few months ago. I knew almost nothing about switched-mode PSU technology, but I knew there is some kind of feedback based voltage regulation inside. So I did some quick internet research (Díky, Jardo! ), then I did open it and long story short:





My questions:

• I have replaced original R_X (R25+VR21) by adjustable R_X (POT1+R1+R2) with preset range of regulation from 5 to 18V - does PSU work cozily in this whole range or is it tuned for 24V/1A sweetspot and I'm giving it hard-times?
• Do I get 1A in whole range?
• So by altering R_X I cannot get less then cca 5V (see V_OUT/R_X graph, it goes up with smaller R_X again), is there a way I could switch whole range somehow to get below this limit (e.g. to let's say 0.5 ~ 5V, or at least to 2.5 ~ 5V)?
• I plan to add additional filtering at the output to get rid of switching rubbish (not the one you see on photos, it is just quick temporary hack) - may I take V_REF from the filter output for the best regulation/stabilization or does it have some drawbacks?

Thank you very much, T.

PS: I know there are dangerous voltages inside, I am careful.

[R.G.]

As long as you realize that you're working with AC power line voltages stepped up to about 180-340Vdc by rectification, and adding a control on wires inside that may short to those high voltages. And that the added control invalidates the safety set up under the "double insulation" rules that this unit used, based on the AC plug and conformation of the plastic barriers in the photos.

Sure it works. But it's like owning a pet poisonous snake. 

It is a Bad Idea to open up and tinker inside small packaged switching power supplies unless you already have the skills and knowledge to do it in ways that will be safe for the long term. I urge you to not make long term use of this after you get this experiment done.

[Thomeeque]

 Thanks, but this is not the final state. Once the basic concept was proved, I'd box it to another outer (plastic) box with panel-meter, all controls, output filter and output jack.

[R.G.]

Just recognize the hazard and be careful. I hate to lose people who make music and have some technical smarts.

[Thomeeque]

Just recognize the hazard and be careful.

Will do.

people who make music and have some technical smarts.

I may assure you this is not the case.

Sure it works.

So, is this YES to my first two questions? Is PSU comfy in whole 5 to 18V range? Can I load 1A in whole 5 to 18V range?

And another question popped in my head, how would PSU react/survive in case of accidental output shortage?

Thanks, T.

[R.G.]

Is PSU comfy in whole 5 to 18V range?
Can I load 1A in whole 5 to 18V range?
... how would PSU react/survive in case of accidental output shortage?

Unfortunately, it is not possible to tell.The schematic you show has all of the answers tied up in the controller IC on the primary side and the parts that program it to work. There are quite a number of these that are all very similar, but slightly different inside. Each generation of similar devices gets better at what it does and more "polite" in its operation, but there is no way of knowing exactly what it will do with a salvaged unit.

This works by turning on the power transistor on the primary side and letting a carefully-measured time of current flow into the isolation transformer, loading the primary up like an inductor; then it turns off. When it turns off, the magnetic energy is transferred to the secondary and charges the output filters.  The optoisolator and other parts feed back a signal to the chip that says in effect "still need more..." or "OK, that's enough for now" and the chip tries to balance between getting those two answers.

The only control over the output the controller chips has is how wide or narrow the turned-on time slice it gives to the power transistor is, and there are limits on that. If it can't gets skinny enough, the output rises out of regulation. If it can't get wide enough, the output voltage goes below regulation. And exactly how wide or narrow depends on how much current is going out the secondary as well as how much DC power from the AC power line is available.

It is impossible to say exactly what the thing will do without detailed testing, however I can **guess**. These are **guesses** and must be backed up by testing.

Is PSU comfy in whole 5 to 18V range?

This is a solid "maybe". It depends on the control chip, how clever it is, and how much margin it has to do skinny pulses, or to change frequency and operating mode at very low output currents/voltages. You may have to put an extra dummy load resistor on the output to make it happy, and it still may give problems at certain loads and input voltages. Or it may be fine. Can't really tell.

Can I load 1A in whole 5 to 18V range?

Probably. Higher load is probably better for this thing than zero load.

... how would PSU react/survive in case of accidental output shortage?

There's only one way to tell. Go short it and see. But short it in a way that YOU can control, and have it set up in a safe environment so that if it dies or goes up in flames or melts, you don't start a fire or electrocute yourself. With that as a warning, it's probably OK. Most controller chips have short-circuit protection built in. But you can't be sure. So test it yourself. I would put it in the middle of about a square meter of concrete or tile surface in a place that can be well ventilated for this test. In the worst case, it could melt, explode, or got into flames. I have personally seen all of these happen on switchers. In the probable case, it will simply shut down with no ill effect, or it might blow that fuse on the PCB.

Again, remember that this is *dangerous*, not least because many of the things you can do are not dangerous. This creates a false sense of security until you touch the wrong thing.

[Thomeeque]

 Thank you very much for this explanation, I really appreciate it!

(Sorry, I was on business trip, I could not respond earlier)

Primary side looks like this (sorry for the quality, I had to use camera as "scanner" this time):


Used control chip is ON 1200AP100.

I did study the datasheet briefly, even I'm still switching mode rookie it seems to me like it is already pretty smart chip with nice features including various "fuses".

I will study your answer and specs more yet and than maybe venture to ask more questions (but I will probably not try to modify this primary side, I'm actually quite happy with actual state), I did put this reply here mainly to say "thank you" and for the case you or somebody else was interested in details missing in the first post.

Thank you! T.