Building your own power supply: An actual example

[Th3_uN1Qu3]

Re: Building your own power supply: An actual example

You're welcome, and good luck

Uh, i forgot you're, you know, alive on this forum. Coz i sure know you're alive on YM... Or actually, it could just be too much for tonight. I'm sure Chris knows how much i had... Not enough to be drunk, but certainly enough to not be fully operational.

@ all others: In either case, we'll be around, and always willing to help you with whatever we can. Switching power supplies aren't THAT hard to understand, i suggest anyone willing to learn to start with an unregulated PSU and then try to close the loop.

Before closing the loop, it's not that different from a regular mains transformer, just working at a higher frequency, hence enabling a smaller size. After closing the loop, it's not that different from a linear regulator, just with an extra inductance in the mix... sometimes. When working with current-mode controllers, the inductance goes out of the frame and things are quite a bit easier.

[Khron]

Re: Building your own power supply: An actual example

You're welcome, and good luck

[rogfanther]

Re: Building your own power supply: An actual example

Ok, understood. Thank you very much. Now back to reading about UC3842 psu design ...

[Khron]

Re: Building your own power supply: An actual example

Eeeeeeeexactly Actually, measuring the inductance of a few turns of wire, and dividing that by the number of turns will get you exactly that Al you're looking for

And then you can compare that with whatever datasheets you can find, and see which "mix" is the closest match to the result you got.

[rogfanther]

Re: Building your own power supply: An actual example

Yes, i can measure them. Do you mean using some turns of wire in the unknown core, measuring the inductance of this coil, then calculating the other parameters from it ?

[Khron]

Re: Building your own power supply: An actual example

I wouldn't Don't you have any possibility of measuring inductances?

[rogfanther]

Re: Building your own power supply: An actual example

sorry if I cannot explain it clearly, but just to be sure I don´t misunderstand you, what I mean is if I could take a Al value say for a EI-33 core from Ferroxcube, and use it for some salvaged EL-33 core from undetermined origin...

[Khron]

Re: Building your own power supply: An actual example

Oh, yeah, core sizes are quite standard. Ferroxcube have a pretty comprehensive and nicely-classified set of datasheets on their site.

[rogfanther]

Re: Building your own power supply: An actual example

Thanks, but I mean, if I have a no-name core, thus no official datasheet from the factory, can I use some "typical" values for that core size ?

[Khron]

Re: Building your own power supply: An actual example

The ferrite mixes might also be slightly different, for a given core size, so just to be on the safe side, you might wanna measure the Al (measured in nH per turn of wire) of the core (big A, small L); it's a parameter given in the datasheets.

[rogfanther]

Re: Building your own power supply: An actual example

I mean, the properties are just all geometric, different makers ( like, Siemens or Ferroxcube versus Chinaboom Inc. ) won´t change them that much ?

[Th3_uN1Qu3]

Re: Building your own power supply: An actual example

A theoric question :

Suppose I buy a core from a store, or take one from some China atx psu. Probably they will not have references for the magnetic parameters for transformer design.

Is there ( and can it be used ) a "generic" table for core properties , so that one would use it, some conservative values for flux density and the like

All cores are standard sizes. You can simply take a ruler to it, note down the dimensions and you're likely to find it on Ferroxcube's site. If not, try Epcos, TDK and Micrometals.

[rogfanther]

Re: Building your own power supply: An actual example

A theoric question :

Suppose I buy a core from a store, or take one from some China atx psu. Probably they will not have references for the magnetic parameters for transformer design.

Is there ( and can it be used ) a "generic" table for core properties , so that one would use it, some conservative values for flux density and the like, and the little differences would be taken care of by the control/compensation part ?

Or the bigger cores have so much bigger specifications that one just need to calculate number of turns/wire diameter and be done with it ?

Thanks

[ben7]

Re: Building your own power supply: An actual example

That Class-D amp module i was talking about, i still have it around, it works, and i still haven't gotten around to building the higher power version even though the PCB design is 100% complete since.... last year.

Um, lol!

[Th3_uN1Qu3]

Re: Building your own power supply: An actual example

It's still sitting around somewhere. Made a couple other power supplies in the meantime. That Class-D amp module i was talking about, i still have it around, it works, and i still haven't gotten around to building the higher power version even though the PCB design is 100% complete since.... last year.

[ben7]

Re: Building your own power supply: An actual example

Did you ever finish the design of this? and did you ever take pics? :/

[Th3_uN1Qu3]

Re: Building your own power supply: An actual example

An IR2110 will be happy to drive those MOSFETs for you.

For one, IR2110 costs the same as one IRFP360. Which means "not quite cheap". Secondly, it's easy to blow up. I've followed lots of class D experiments on diyaudio and all of them have resulted in several failed 2110s. This is why in my amp i went for discrete drivers. Third, my controller will be on the secondary side. The gate drive transformer gives me isolation from the primary. An IR2110 would not.

Different FABs and different lines will have different laser marking equipment so the finished parts look different. Often the case plastic, leads, tabs, and shear marks are different too.

Thanks for clearing my doubts on this.

[severach]

Re: Building your own power supply: An actual example

An IR2110 will be happy to drive those MOSFETs for you.

Different FABs and different lines will have different laser marking equipment so the finished parts look different. Often the case plastic, leads, tabs, and shear marks are different too. It comes from buying companies and from buying from different equipment suppliers through the years.

[Th3_uN1Qu3]

Re: Building your own power supply: An actual example

I now have my first protection circuit. I put a LM339 and a TL431 on my protoboard, made an undervoltage protection circuit and set it to cut in at 70 volts. I then shorted the power supply. Click. Zero volts. No popped fuse.

Cut the power to the LM339... and it came right back up! Wooohooo! It'll also need a delayed powerup, because due to the soft start the power supply will not start at all if the LM339 is connected from the beginning , but hey, it works.

I'll also decrease the value of the soft start capacitor, 4 seconds is way too long. It'll be more like 1 second or so. A slow soft start won't help if the power supply is trying to start on a short circuit. In the final version the delayed powerup of the protection circuit won't be a problem because current limiting will step in before the LM339 is enabled.

[Th3_uN1Qu3]

Re: Building your own power supply: An actual example

200W and they aren't even getting warm. Even at 400W level they stay comfortable with only a slow running fan. Now all i need is a nice PCB design and i'm all done.