High current DC power Supplies

[fr0st]

I'm looking for a more efficiant alternative to the ESP ones (if there is one). Volatge needs to be regulated adjustable from about 10-16 volts. How much do variable Voltage regs (the lm317 types) disapate with added transisters?
I need about 50 amps in total (continious) for car audio testing puposes.
Thanks

[Sch3mat1c]

Hmm. I'd guess voltage needs to be variable for testing the voltage tolerance.

I say, use a BHT (Big Honking Transformer) and some hefty diodes and capacitors and vary the primary with a variac. "Only" 100W wasted in the diodes...lol. But not 200+ in pass transistors. Ack!

Tim

[JOE DIRT®]

frost if you have the transformer for it and the transistors....I have the basic design...you can keep adding from there.....Just dont blow up the planet!!!!..lol...basicaly you can just use a 317 reg and bypass it with transistors...each with there own 0.1 ohm current limiting...then do the math on how many you need to acquire....if your using this for testing car audio....after all these years I`ve learned to go get a good battery and just charge it....especially if your into competion


Cheers!!The DIRT®

[fr0st]

The lm317 datasheet specifies how to add to the current capabilities of the package but the heat production of it worries me. Has anyone built something similar and can specify real world efficiency?
The variac would be a good idea but the only ones I have access to are only 500w which isn't enough.
i'll keep looking for a solution.

[jackinnj]

if you are really getting into that power arena you are, at the same time, getting into a new and higher level of protecting yourself.

some of the things which I would suggest:
safety interlock on the cabinet
bleeder resistors on the filter caps
soft start

Further, if you are going to spend the time and money on a big honkin' power supply like this you owe yourself better than to use a LM317 as the error amplifier/reference. After all, the heat sinks are going to cost more than the PASS transistors. If you search the DIYAUDIO archives you will see a proven ARRL design (that's American Radio Relay League) that you can employ with a lower noise error amplifier and voltage reference.

[JOE DIRT®]

personally Jack...I use the LM 723...its an old regulator but has a built in band gap reference...I have a 50 amp supply I use for testing car audio and It current sensing so I dont blow up the world..LOL

[jackinnj]

the ARRL circuit uses the '723 (in the 2001 handbook) they also have a 40 amp switching supply using the LM3524 driver chip -- the important thing about this article is they give you step by step instructions for building the transformers and snubbers (and as an ARRL design it's doubltess lower noise than that featured on the ESP site.) I'd still prefer current mode control, however, particularly with BJT's.


For $30, the ARRL Handbook is a must own. In OZ, they probably have RSGB Handbooks in the local library with similar designs. In fact, a push-pull switcher is going to be one heck of a lot more efficient than a linear design --

[fr0st]

Originally posted by JOE DIRT®
personally Jack...I use the LM 723...its an old regulator but has a built in band gap reference...I have a 50 amp supply I use for testing car audio and It current sensing so I dont blow up the world..LOL

Is that one variable?
I've found plans for a 40 amp 13.8v psu using the lm723 which think is variable. found here
It has a voltage adjust.... to what extent I don't know. I'll try and calculate how far it can adjust using the formulas off the datasheet *fingers crossed*.

the ARRL circuit uses the '723 (in the 2001 handbook) they also have a 40 amp switching supply using the LM3524 driver chip -- the important thing about this article is they give you step by step instructions for building the transformers and snubbers (and as an ARRL design it's doubltess lower noise than that featured on the ESP site.) I'd still prefer current mode control, however, particularly with BJT's.

I'm not sure if i'm comfortable rectifying mains... I's rather a linear design so I can play with a lower "less lethal" voltage. If there a switching design thats more efficient than a linear one and accepts <20vdc I'd be all for it.