buck converter / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

buck converter

1 2 next

Move Thread

LAN_403

IamSmooth

Wed Oct 06 2010, 03:53AM

Registered Member #190 Joined: Fri Feb 17 2006, 12:00AM
Location:
Posts: 1567

I have seen the IGBT gate for the Buck converter run off high-side drive of a ir2125. Is it possible to use TC4421 and TC4422 to drive a gate-drive transformer to switch the IGBT?

Sulaiman

Wed Oct 06 2010, 04:24AM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3140

The main thing to remember about inductors & transformer windings is

The integral of voltage x time must sum to zero over time

or there will be "flux walking" (google it) leading to core saturation.

So you can use a gdt but unless you have some particularly unusual problem(s) to solve
(e.g. operating voltage higher than any high-side driver capability)
the added complications are not worth it.

If you MUST use a gdt I would suggest a single transistor drive for the primary,
exactly like a flyback circuit, remember the neccessary flyback voltage!
e.g. 90% duty, 15V gdt supply, the flyback voltage will be at least 15 x 9 = 135V
and the transistor must be rated for at least 15V + 135V = 150V plus 'spikes' from gdt leakage inductance.
The igbt gate circuitry (gdt secondary side) must also allow for -135V 'kickback'

Steve Conner

Wed Oct 06 2010, 09:12AM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

In other words... No... GDTs are no good for signals whose duty cycle varies much away from 50%.

You can make them work by adding the kind of extra circuitry Sulaiman describes, but it's easier to use an IR21xx, or a TLP250 if you need isolation. I see TLP251s and those little isolated DC-DC converters used a lot in student projects here.

IamSmooth

Thu Oct 07 2010, 05:12PM

Registered Member #190 Joined: Fri Feb 17 2006, 12:00AM
Location:
Posts: 1567

Is a boost converter the way to go if I am currently using a variac and a voltage doubler? The output voltage for a buck converter is D = Vo/Vi, so the output voltage, at most, will equal the input voltage. The voltage for a boost converter is D = 1- Vi/Vo, so the lowest voltage, it seems the output starts out equal to the input and increases. So, if I want to control the power to an inverter from zero to twice the voltage, what is my best choice? A buck-boost?

Second follow-up question. I found the IR2117 series which has a 600v boot-strap high side. What does one do if the voltage is higher than 600v?

Sulaiman

Thu Oct 07 2010, 10:51PM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3140

As Steve noted, TLP250 etc. are opto-coupled igbt drivers. They seem reliable.
For large igbts switching at high frequency they're a bit weak,
a normal opto-coupler with e.g. TC4421 may be required.

I have very little confidence in small commercial DC-DC converters
having replaced so many of them,
in fact at work when we get a faulty item with a dc-dc converter
that's one of the first things that we check.

The most reliable isolated dc supplies are from small pulse transformers
used as isolating transformers at 10's kHz with rectification, smoothing etc.
Line frequency transformers are quite good but the inter-winding insulation and capacitance are not ideal for high frequency high voltage bridges.

IamSmooth

Fri Oct 08 2010, 01:02AM

Registered Member #190 Joined: Fri Feb 17 2006, 12:00AM
Location:
Posts: 1567

I just looked at the TLP250 and it certainly has the isolation cabability. If I put a bootstrap capacitor/diode combination on the output, would that be sufficient? It has a switching time of 0.5us, so would running it at 100-150khz be reasonable?

Steve Conner

Fri Oct 08 2010, 11:35AM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

Are you asking whether you can use a bootstrap circuit to supply the gate driver side of the TLP250?

If so, I guess maybe you can, but remember that the bootstrap circuit fizzles out at 100% duty cycle. It needs those negative periods to keep its capacitor charged, so you have to limit the duty cycle to 99%, say.

Also, the bootstrap may do weird things in a buck converter, because there's no low-side switch to force the voltage down to zero, just a flywheel diode. If the output current is very low, the flywheel diode may never conduct, so the bootstrap would never get charged. You could get in a situation where the buck converter can't start up, because it needs output current to charge its bootstrap capacitor and generate gate drive, but it can't have output current until it has gate drive, etc. Or it mysteriously blows its IGBT on light loads or whatever.

I would use a proper isolated power supply. I've had reasonable success with a "ZVS" Royer oscillator driving little pulse transformers, rectifiers etc.

IamSmooth

Fri Oct 08 2010, 01:40PM

Registered Member #190 Joined: Fri Feb 17 2006, 12:00AM
Location:
Posts: 1567

These replies have been helpful. So am I correct to presume that for a buck converter (step-down) one never wants to be able to increase the voltage to unity? If one wants a larger range a buck-boost is in order?

iJim

Sat Oct 09 2010, 12:55PM

Registered Member #2662 Joined: Fri Jan 29 2010, 10:14AM
Location:
Posts: 36

I once built a buck convertor using an ir2181, to avoid issues with the low side output I substituted the buck convertor diode for a second mosfet. The arrangement looked something like a half-bridge with inductor in place of the usual transformer; however, unlike a half-bridge the duty cycle was not fixed at 50%, I varied it from 1-99% to control the output voltage...

Not the easiest way, but those where the components I had to hand.

iJim.

Steve Conner

Sat Oct 09 2010, 01:54PM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

That's a synchronous buck converter. It's very widely used at low voltages, where the low-side MOSFET is more efficient than a diode. But above roughly 12V, it gets more cost-effective to use a diode.

There's nothing stopping you from using a synchronous buck converter at any voltage you want, it just wouldn't be done in a commercial design that had to be cost-effective.

When using the synchronous buck converter, you have to remember that it's bidirectional. Power can flow through it backwards, and then it acts as a boost converter. One useful application is when driving a DC motor off a battery: if you set a lower speed than the motor is actually running, the converter will brake the motor and feed power back into the battery. But if this happens by accident with a power supply that can't accept power forced back into it, things can get blown up big-time.

1 2 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob