300 amp power supply

On Sat, 19 Sep 2015 20:22:35 -0500, Jon Elson Gave us:

Just use an old slot car controller! :-) Hehehehe..

I'm really enjoying this discussion, watching everyone offer solutions that favor their areas of expertise and experience. That also includes me. This is not a criticism, just an observation.

I think all the DC to DC converter electronic solutions have already been mentioned. Well, nobody mentioned a 4500 watt shunt regulator, but that can be forgiven.

I threw in a series dropping resistor as the "hillbilly" solution. If regulation is required, just make it a PTC power resistor. I would be hard pressed to find one that will do 300 amps and 0.01 ohms, but maybe something can be contrived by connecting a huge number in parallel.

I haven't seen any suggestions for a motor-generator, or a gasoline engine driving a big alternator, but those would also work. Regulation would be by controlling the motor speed. I'll spare you my motor driven mechanical pulse with modulator regulator idea.

Charging a small battery with a fully charged much bigger battery would also sorta work.

Of course, such a powerful machine demands a proper electrical panel. I think something in the mad scientist tradition would be cool:

Other than a nuclear powered solution, did I miss anything?

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
Santa Cruz CA 95060 http://802.11junk.com 
Skype: JeffLiebermann     AE6KS    831-336-2558

It costs north of $10k for engineering samples. We're hoping it'll come down to less than half that after about a year. The final product should sell for six figures, as there's no competition in our market (at least initially). We don't need to sell many for the project to break even.

those.

Yeah, me too. The first proto should be back soon. It will have a dummy load in place of an FPGA.

[anecdote] This reminds me of place I worked at in about 1990, when a technician was tuning up some K(?)-band amplifiers that used space- qualified GaAsFETs. He managed to blow up about half his annual salary worth of FETs one afternoon. The terrestrial amplifiers had overcurrent limits, so that the FETs would be protected as the bias was adjusted, but the space-qualified units had the protection circuit removed to make them more reliable.

Allan

One of my guys blew up eight GaN fets one day, at about $100 each. I went into his office and he told me about it, and asked if I wanted to see him blow up one more.

Sorry, I didn't mention that of course I'd disconnect everything from the car first. But don't worry, I'm (almost) sure this Isuzu Elf truck from 1984 wouldn't give a crimp even if Obama (to name someone evil enough to come up with the idea) would light up an EMP device right above us. I guess I'd only have to replace the diodes and regulator IC in the alternator if that would happen.

And put the bundle right beside where the exhaust pipe connects to the manifold, a spot where *my* pipe broke, causing the exhaust gasses to blow right on the wires, melting the insulation and causing a short circuit.

Happened just last week, same truck, and I still have to clean up the mess (always full tank and boiling engine, according to the indicators).

joe

Nah, maybe the ripple voltage after rectifying the alternator's voltage.

joe

I'd use a synchronous rectifier for the transistor/diode combination. Buy one inverter leg MosFet, use the upper Mosfet as switching transistor, put the lower one in conduction when the ds voltage threatens to become negative. Of course you'd put lots in parallel to keep the voltage loss low, otherwise one could as well use some IGBT's. But putting those in parallel with their negative temperature coefficient could become quite dangerous for one's budget.

And pay very very very much attention to the layout to keep all current paths that involve the mosfets short and low inductance.

joe

On Sat, 19 Sep 2015 19:41:37 -0700, Jeff Liebermann Gave us:

Resistors used on electric trains.

I had something like that when I worked at a technical university.

My supervisor asked me whether I'd like to reverse engineer the control circuitry of a 10 kW brushless motor drive.

I said yes, went to work, produced a satisfactorily schematic of the whole system--luckily everything still discrete components at that time-- and then he asked me to 'take over' the control. His aim was to bypass the whole control circuitry to be able to test all kinds of con trol strategies, including operating the drive as a generator.

So I went to work, but unluckily (for him) bypassed also the protection circuits.

When we switched on my contraption there was a short hiss and all power modules were destroyed. Took almost a year to find the budget for the repair so I wasn't able to ask him whether he'd like me to blow up some more. :)

I'm smiling now but felt terrible at the time...

joe

On Sat, 19 Sep 2015 19:41:37 -0700, Jeff Liebermann Gave us:

Apparently you did not examine all of them either.

You don't need litz wire for this, you use copper tape wire.

You need to look around because they make copper tape with one side sticky that can act as a insulator. Roll out the length you need and layer it. Bond the ends. Or you can use paper layers between.

The other issues require low on R mosfets, back feed diodes and controlled switching.

I made a inverter welder once :)

jamie

ll do

me fires up the gas burner, goes & gets filament lamp, some glass, thermome ter & an old dead valve or 2. Oh... no, a mercury arc won't go down to 3v d rop. Sorry.

Definitely one of the less reassuring electronic devices I've used. If one of those lets go when hot in a confined space... not sure which is worse, t hat or the toxic fire extinguisher. Carbon tet isn't great cold, but in a f ire it gets rather more toxic. I didn't plan to use the fire extinguisher u nless there was no other way out. But TBH there were more immediate issues to keep an eye on, CO in particular.

NT

Umm, obviously wrong! But, a SEPIC seems to have a number of components that are not NEEDED for a plain step-down only regulator. At this current level, you REALLY want to reduce the parts count to the bare minimum, both for cost and power dissipation reasons.

Now that we THINK we know more about what is being asked for, you really need to be careful about providing 300 A to a dead battery. This can be a good way to get a face full of sulfuric acid! Unless these are railroad starting batteries (or for other LARGE diesels) putting 300 A into a dead battery could cause damage to the battery, and if the battery is actually completely failed, it could explode if Hydrogen is allowed to build up.

(Don't ask how I know this!)

Jon

Oh, that can be remedied easily! 1940s-era welders were big motor/DC generator combos, and I can recall using a Navy-surplus

have been scrapped; an eight-foot-tall wheeled welder isn't a familiar sight, this century.

On Sun, 20 Sep 2015 14:10:28 -0700 (PDT), whit3rd Gave us:

What the hell do you call that shit? A "tiny" morph? Angstrom sized?

As suggested elsewhere, you could use a linear regulator with perhaps three

200 watt 0.02 ohm resistors to drop 2 volts at 100 amps each, and bipolar or MOSFET transistors to regulate the remaining 1 volt, dissipating 300 watts. Probably 8 MOSFETs in parallel carrying 40 amps each would be easy to deal with, and probably only a couple dollars each. The heatsinks and resistors will likely cost a lot more than the MOSFETs.

A buck regulator would be more efficient, but I would use MOSFETs rather than IGBTs because you don't need high voltage but you do need low voltage drop at high current. As for the inductor(s), you can get 150 amp 25 uH inductors for about $35 each:

I am working on a 10kW EV battery charger that supplies up to 50 amps at up to 300 VDC, using a single inductor for the input PFC and another for the output. It is a custom wound toroid with about 40 turns of multi-strand twisted magnet wire for 50 amps. You could use the same design with six windings of six turns each to get 300 amps in parallel. Here are some pictures of the contraption (which I did not design):

(The toroid buck output inductor) (a 50A 220uH choke used for PFC) (Overall schematic)

You can buy a complete 12kW charger for $2300 or 25kW for $2600:

The chargers I am working on were kits assembled by customers and there were many design problems, which is why I'm trying to fix them. So I would not especially recommend EMW but they might be a reference point as to what can be done.

I have designed test sets that can produce as much as 100,000 amps AC and I've worked on DC supplies capable of 30,000 amps, so 300 amps does not seem like so much. But it is a challenge to make using solid state switching technology.

Paul

Paul, thanks for sharing this design.

Since the time I first heard about synchronous rectification I have always been of the impression that controlling the gates of the 'freewheeling' fets in order to put their channels in low resistance mode when the diode would be freewheeling, the resistance and hence voltage drop and hence dissipation could be lowered compared to just connecting the gate to source.

Do you think I've always been wrong with that?

If I was right, this would overall heat up the leg's module to a lesser extend.

joe

Yes, cutting forward losses is the motivation for synchronous rectification. Vf=1v * 300A is a serious motivation.

A multiphase synchronous buck might be the simplest way to move big amps from

18V to 15V (without warming the planet).

Cheers, James Arthur

Like this?

Hmm. I don't see the ripple reduction yet, but I'm still thinkin' on it. Oh, yes, I think I see the potential. Thanks.

Yup. They're magic.

Cheers, James Arthur