Most efficient transformer drive design / High Voltage / Forums

Forums

4hv.org :: Forums :: High Voltage

« Previous topic | Next topic »

Most efficient transformer drive design

1 2 next

Move Thread

LAN_403

Turkey9

Fri Jul 17 2009, 04:30AM

Registered Member #1451 Joined: Wed Apr 23 2008, 03:48AM
Location: Boulder, Co
Posts: 661

Ok so I've dug around the forum and the rest of the internet and I can't seem to find the most efficient driver. I know the ZVS is great... is it the best one? I would assume that you could build a Half-Bridge design that could get to the same power and efficiency. I've messed around with both circuits and no matter what I get the best performance with the Mazilli driver, unless the half-bridge is on a sstc. Is the reason that a bridge works so well with a Tesla coil is the feedback?

I have a function generator that I used to tune the half-bridge so I'm sure it's a very close to resonance. How can the ZVS use the same primary turns and not heat nearly as much as the half-bridge?

Ok so the base line is that I'm looking to build the most efficient driver I can get that is also at least fairly high power. I can build what I need as long as I can get some background on how it works.

Sorry for all the questions and I hope I wasn't too confusing.

Antonio

Fri Jul 17 2009, 01:24PM

Registered Member #834 Joined: Tue Jun 12 2007, 10:57PM
Location: Brazil
Posts: 644

Considering that flyback transformers are designed to work in CRT TVs and monitors, operating for long periods without failure and with good efficiency, a look at the circuits of these devices gives a hint about the correct way to use them. Invariably is the flyback configuration, with several kinds of protection.

Proud Mary

Fri Jul 17 2009, 02:29PM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

"Efficiency" is by no means a simple, or singular quality.

If your idea of 'efficient' includes a long Mean Time Between Failures, then you would do best to follow the LOPT manufacturer's data sheet, to benefit from their own guarantees and warranties.

If you mean to push the LOPT to the very edge of destruction (and often beyond!

) as is enjoyed by 4HV-niks interested in arcs n' sparks, you will find many circuits such as the popular ZVS in the 4HV archive.

Dr. Dark Current

Fri Jul 17 2009, 02:57PM

Registered Member #152 Joined: Sun Feb 12 2006, 03:36PM
Location: Czech Rep.
Posts: 3384

You mean what is the most efficient way to drive a diode-split high voltage transformer? Flyback topology is easiest on the transformer, bridge probably dissipates less power in the driver (but the transformer itself will probably get hotter).

The most efficient way to drive an unrectified or fullwave rectified HV transformer is a bridge circuit.

Turkey9

Fri Jul 17 2009, 06:20PM

Registered Member #1451 Joined: Wed Apr 23 2008, 03:48AM
Location: Boulder, Co
Posts: 661

Sorry I should have been a little bit more clear in my first post. I'm driving a transformer that I built to use a flyback core. I don't use the spacers so I guess that it's just a step up transformer. I've driven it with both the ZVS and a bridge but the bridge seems to lose a lot of energy as heat. That's where I want the efficiency, transferring the energy to the transformer and not loosing it in the drive stage.

Dr. Dark Current

Fri Jul 17 2009, 06:40PM

Registered Member #152 Joined: Sun Feb 12 2006, 03:36PM
Location: Czech Rep.
Posts: 3384

Turkey9 wrote ...

Sorry I should have been a little bit more clear in my first post. I'm driving a transformer that I built to use a flyback core. I don't use the spacers so I guess that it's just a step up transformer. I've driven it with both the ZVS and a bridge but the bridge seems to lose a lot of energy as heat. That's where I want the efficiency, transferring the energy to the transformer and not loosing it in the drive stage.

Well, the "zvs" circuit will waste a lot of heat in the primary and core, I think you'll find out:-)

If you want a good performing half-bridge, you need to assure that the gate signals will have fast rise-fall times e.g. high current gate driver and low-leakage GDT. And of course, MOSFETs rated for higher current will dissipate less heat in the driver. Go with transistor voltage rating of 1.2-1.4x your DC supply voltage, higher rating is unnecessary and just increases the ON-resistance.
Then you need to fiddle with the frequency (around 40-50kHz seems to work well) and the primary turns to get the desired performance.

jonny5

Sat Jul 18 2009, 02:11AM

Registered Member #1807 Joined: Tue Nov 11 2008, 07:36AM
Location: San Luis Obispo, California
Posts: 19

Just to clarify, does the resonant half-bridge you speak of accomplish zero voltage switching? This can be a big energy saver...one of the reasons why the Mazzilli is incredibly awesome!

Other points to ponder: the Mazzilli topology possesses high circulating resonant tank currents. Fortunately, the tank current stays locked in the resonant capacitor/transformer magnetizing inductance. If you are interested in efficiency, I would consider looking into acquiring/making Litz wire for the primary. Magnet wire + wire gauge tables to find the skin depth frequency can yield very good results! Of course, it's a royal pain to make decafilar wire (or worse) by hand.

If you are rectifying the transformer output, the resonant secondary voltages/currents allow the output diodes to softly recover. This may or may not improve efficiency, but it will ultimately be easier on the output diodes.

Lastly, the Mazzilli topology is somewhat protected from an output short circuit! The reflected impedance of an output short shunts the transformer's magnetizing inductance. Since the Mazzilli requires inductance (and capacitance

) to accomplish resonance, a (relatively) small amount of leakage inductance can be beneficial. It sounds like you are constructing the transformer on a U shaped core (with primary on one side and secondary on the other)? The low(er) coupling may be sufficient to guarantee short-circuit protection. For LOPT transformers with integrated voltage multipliers, the capacitive ballasting of the CW voltage multiplier may also serve to protect against shorts. I keep harping on the short-circuit protection, because 1) accidents happen 2)drawing arcs is fun 3)accidents will happen!

Turkey9

Sat Jul 18 2009, 02:37AM

Registered Member #1451 Joined: Wed Apr 23 2008, 03:48AM
Location: Boulder, Co
Posts: 661

There isn't going to be much danger of a short circuit. The transformer is going to be used to charge capacitors. It's not for a coil gun, so I thought it would be better in the high voltage thread. The capacitor is going to be charged to around 1150V and will hold about 290J.

I'm going to try with both an E core and a U core to see what I get best results out of. Are there any sites out there that describe how to do zero volt switching with a bridge? I think my transformer will resonate at a couple thousand hertz at least. I like the idea of a Litz wire primary... I have plenty of small awg magnet wire.

So I did some experimenting today and ran into a problem... the mosfet driver that was feeding the GDT does fine with a .68uf cap in between it and the GDT but when I increase that cap to 2uf the driver gets extremely hot after a couple seconds. Same with a 1uf cap. Is the driver getting too much current drawn from it? It's rated for 9 amp pulse and I doubt the GDT is drawing that much.

Sulaiman

Sat Jul 18 2009, 06:07AM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3141

You've just changed the ball game !

The posts above assume a constant VOLTAGE output,
charging capacitors requires constant CURRENT.

Look through the forums for 'CCPS'

Bjørn

Sat Jul 18 2009, 06:36AM

Registered Member #27 Joined: Fri Feb 03 2006, 02:20AM
Location: Hyperborea
Posts: 2058

There isn't going to be much danger of a short circuit. The transformer is going to be used to charge capacitors.

If the capacitor is any good it will be a short circuit when you start charging.

1 2 next

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