Looking for input.

[GonzoEngineer]

So, I am going to design a Buck Regulator Power Supply, and I was looking for input and advice.

Here are the spec's I need:

Input voltage: 400-480 VAC, 3 phase.
Output voltage: adjustable to 400VDC
Output current: up to 100A. (40kW)

Am thinking about using a PWM controller (TL594?), and a single IGBT switch rated at 1200V / 450A. (Infineon art # FF450R12KT4, switching at 20kHz)

The supply is essentially a capacitor charger, (26000uF) that feeds a Plasma Pulse Generator. I don't need it to be isolated, as the plasma generator is an IGBT H-Bridge driving a step up transformer, so I will be isolated there.

My question is, I have never used a PWM Controller, are they designed to work to give you an adjustable output? Or are they more suited to a fixed output supply application?

I have the data sheet, and will try to decipher it later tonight.

I just wanted to know if anyone has any input or advice on this.

(BTW. Don't worry about my skills or the high voltage dangers. My plasma generator currently puts out pulses of over 1MW, and I know what I am doing.)

I am currently wasting a lot of money using off the shelf AC/DC supplies, and they are overkill and cost way too much!

I estimate I could end up building this for a lot less that what I pay now.

 

[BobK]

Why not use a switching regulator controller chip? A converter needs a feedback loop, this is all built into that type of controller, you would be doing it yourself if you used a PWM controller. There are two kinds of chips out there, ones that have an internal switch and those that control and external switch. You would obviously want the second kind.

Bob

 

[GonzoEngineer]

Why not use a switching regulator controller chip? A converter needs a feedback loop, this is all built into that type of controller, you would be doing it yourself if you used a PWM controller. There are two kinds of chips out there, ones that have an internal switch and those that control and external switch. You would obviously want the second kind.

Bob

Thanks Bob.....I will google that!

Never too late to learn something new......especially if it can save me money!

 

[(*steve*)]

For this kind of power I'd also be looking at one of the "power factor correction" (aka "green"!) devices. However you may already know how you're going to get the input DC rail and you may be well prepared for the issues there.

Essentially they're a pretty standard boost regulator that goes between your rectified mains and the input filter capacitor that tries to maintain current draw throughout the full half cycle rather than a massive current spike neat the top of each cycle.

I kinda feel embarrassed suggesting this because I'm sure your other work would incorporate something like this.

I only suggest these because I see them in switchmode power supplies and have gone to look at the specs of the devices -- I've not actually designed used one. A quick google for a representative device suggests benefits in higher powered supplies.

 

[GreenGiant]

if you want to buy one I would look at Cosel's line up, they have some power supplies that will provide those powers and then some and youre looking like 2K per? (give or take I dont have quotes)

Im not sure what price range building would be in, but good luck!

 

[GonzoEngineer]

if you want to buy one I would look at Cosel's line up, they have some power supplies that will provide those powers and then some and youre looking like 2K per? (give or take I dont have quotes)

Im not sure what price range building would be in, but good luck!

I looked at their site......nothing there with the spec's I have....

 

[GonzoEngineer]

For this kind of power I'd also be looking at one of the "power factor correction" (aka "green"!) devices. However you may already know how you're going to get the input DC rail and you may be well prepared for the issues there.

Essentially they're a pretty standard boost regulator that goes between your rectified mains and the input filter capacitor that tries to maintain current draw throughout the full half cycle rather than a massive current spike neat the top of each cycle.

I kinda feel embarrassed suggesting this because I'm sure your other work would incorporate something like this.

I only suggest these because I see them in switchmode power supplies and have gone to look at the specs of the devices -- I've not actually designed used one. A quick google for a representative device suggests benefits in higher powered supplies.

I don't need to be "Green"....I just need to get the cost down.

 

[(*steve*)]

I don't need to be "Green"....I just need to get the cost down.

It's not about being green

 

[GreenGiant]

I looked at their site......nothing there with the spec's I have....

My apologies, I thought they had something in that range, their power supplies can (most of the time) be connected together and run as one. I'd say give a shout and see if they have anything, they will come out to your work and sit down with you and go over everything to see if they have anything, free of charge mind you

 

[GonzoEngineer]

Just as an update....they don't make a controller chip for this application.

Most PWM controllers are designed for a fixed voltage output. They are not suited to making a supply that you can adjust from 100V out p to 600V!

So we are generating the PWM...with a microcontroller that costs 2 bucks.

OK, so I have to pay the programmer to write the software....but he has only worked for a total of maybe 12 hours....and he is about 90% there!

It works like a charm!

The real cost to do this, is in the IGBT's, and the Inductor!

I went with a Synchronous Buck Regulater design, so I need 2 IGBT's.

We are using an Infineon half-bridge.....50 bucks!

I have to go with with a custom inductor, with a Metglass core. It might end up costing me 100 bucks, but it does the job beautifully!

The other cost is the Laminated buss....maybe another 60 bucks.

I now have a 20kW supply that we can build for less than 800 bucks, and it works better than the ones we buy for $4,000!

I love my job!