Bassblaster
Super Member
Nope. single primary. there model TD300-124-28. there all over eBay. the 2 i have here i got for free so cant complain.
Yet another DIY Power Supply, Almost done
- Thread starter Bassblaster
- Start date
petehall347
the brandy coffee man
lowering volts on primary ought to be easier as less current to deal with .
audiodummy
Well-Known Member
Actually... you have two identical transformers right, connect the two primaries in series and the two secondaries in parallel... make sure they're parallel and not inverse parallel else you'll get a big molten mess. It might be less risky to hook them up via the halfbridge fullwave rectifier if the wire in the transformer is fairly beefy. Then you'll get 12VAC RMS and will need to drop less for 14.4VDC.
Bassblaster
Super Member
That sounds perfect. with the two primaries in series i assume i would still get the 12.5A pre secondary on the output? for the full 25A? if maths done me right thats only 64W lost as heat with 17V after the rectifier. on to rectifier, i could use the half bridge but i already have a 25A full wave bridge in by parts bin i get from a PC power supply i thought about using. Is there any downside to using the full over the half? i wouldn't think there would be. other than maybe cost.
petehall347
the brandy coffee man
not something i ever thought about . but thinking of transformer balance may be of concern .. amp meters should suffice here .
I LIKE MUSIC
Super Member
This is some general information to help you with your calculations.
Remember that when you are in the design stage the 34 volts is what the filter capacitors charge to, it is the peak voltage. What you are concerned with is the RMS voltage in terms of power dissipation. As soon as current is drawn from the power supply the voltage on the filter capacitors will start to drop. There will also be a voltage drop due to the losses in the transformer.
It is not likely that at full current (25 amps) that the voltage drop across the pass transistors will be 19.6 volts.
Below is a ball park example of what you might expect.
Note in the picture below The current (green line) is about 24 amps and the usable voltage is about 25 volts. The would cause about 275 watts of dissipation for the pass transistors. Somewhat less than your 490 watts. This is based on ballpark numbers for you transformer and filter capacitors.
It shows the voltage on the filter capacitors under your full load.
If the current drawn is reduced by about half, the available voltage will increase to maybe 28 volts and the power dissipation of the pass transistors would be about 168 watts.
Again these are just ballpark numbers, but they will give you an ideas of what is happening.
Again you can not count on the peak voltage being available under full load Below is an example with a 12.6 VAC transformer.
Note the actual voltage available at full load, about 24 amps.
Again, the peak voltage that you are using is only going to show up in the filter capacitors when there is no current being drawn from the power supply. This is a ballpark example with a 12.6 volt transformer.
Again, note the available voltage (the red trace). And again these are ballpark examples using normal numbers for the transformer, diodes and filter capacitors, but they give a good indication of what to expect.
Half bridge will not work well with two transformers - each one is supplying half-wave output. With two windings on the same core, the DC flux cancels, but not when on separate cores. I think you'll be best off limiting output current for continuous use - consider it a 10A PEAK rating.
Some big linear supplies used a variac ganged with the output voltage control to limit the pass device dissipation.
Some big linear supplies used a variac ganged with the output voltage control to limit the pass device dissipation.
Bassblaster
Super Member
before i mount everything in an enclosure, i thought,how would i wire somithing like a quick connect terminal strip so i could select between both transformers in parallel (12V 25A) or just one for 24V at 12.5A. so i could switch between low voltage high current and high voltage low current? im not having much luck trying to figure that out
Not that difficult. Primary leads in parallel. Red connection for 24 volts. Green connections for 12 volts. Notice reversal of A and B leads. Test the series (red) first to make sure phases are correct. (if wrong, no AC voltage difference between out1 and out2
Last edited:
Bassblaster
Super Member
i got the board assembled, but my dumbass has one of the outputs behind a filter cap so mounting that one will be fun.
but i think im going to stick with just 1 transformer for a few reasons.
1: at 25A each pass transistor would be passing over 8A, they should have no issue with that, but the emitter resistors would be melting.
2: Im obviously not an expert in circuit design, so im not too sure how the 150mil trace feeding the collectors of the pass transistors would hold up under 25A.
But i can always make a bigger supply, with even more pass transistors. There's currently a 93A 16V transformer on eBay right now, anyone down for a linear 100A power supply? i figure about 12 pass transistors should do it
but i think im going to stick with just 1 transformer for a few reasons.
1: at 25A each pass transistor would be passing over 8A, they should have no issue with that, but the emitter resistors would be melting.
2: Im obviously not an expert in circuit design, so im not too sure how the 150mil trace feeding the collectors of the pass transistors would hold up under 25A.
But i can always make a bigger supply, with even more pass transistors. There's currently a 93A 16V transformer on eBay right now, anyone down for a linear 100A power supply? i figure about 12 pass transistors should do it
Bassblaster
Super Member
Got to the end of this thread to find out you have pcbs done already, I was going to suggest to add some means of current limiting, since this is, I assume, to be a lab supply so you do not want a shorted load to smoke your supply..
The old LM/uA723 is still in production by TI, it was the std for so long, probably still running on some space probe somewhere.
If you want I have some PS stuff done in ltspice to use as examples.
Look at IPC or net for trace current capabilities. you can use to sides of copper since they are basically free these days. can also open up solder mask to add solder to increase current capacity.
The old LM/uA723 is still in production by TI, it was the std for so long, probably still running on some space probe somewhere.
If you want I have some PS stuff done in ltspice to use as examples.
Look at IPC or net for trace current capabilities. you can use to sides of copper since they are basically free these days. can also open up solder mask to add solder to increase current capacity.
audiodummy
Well-Known Member
Apparently those chinese analog bench PSUs commonly use those µA723's, and since they're still making these crappy PSUs, incentive for TI and such to keep making more as these PSUs fry and go to the garbage heap...
Bassblaster
Super Member
i actually have a supply that HAS the uA723 in it. its a fixed 13.8V 10A jobby that i modded to be adjustable voltage. was going to do constant current as well, but i dropped that idea. but i dont think chinese supplies is the reason the uA723 is still being made. its simply one of the best choices for a better linear supply.