Flyback driver on mains? / High Voltage / Forums

Forums

4hv.org :: Forums :: High Voltage

« Previous topic | Next topic »

Flyback driver on mains?

1 2 next

Move Thread

LAN_403

Dr. Dark Current

Thu Nov 30 2006, 06:20PM

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

What I want is a self-tuning half-bridge flyback driver, that will run directly from rectified mains. I know this is possible and probably an easy circuit. I remember older PC PSUs that had a kind of startup circuit that worked in that way. When you removed the ferrite transformer (thus disconnecting all low-voltage stuff) and connected a flyback primary where the original primary was, the flyback made pretty decent and very hot arcs.I remember seeing a self-tuning half-bridge flyback driver schematic here on 4hv, I believe it was made by cbfull, maybe it could be used with some modifications to run directly from mains? I know asking for schematics is against rules, so any ideas are appreciated.

thanks, J.M.

EDIT> Found the circuit

click Any ideas how to modify it would be great!

cbfull

Fri Dec 01 2006, 09:53PM

Registered Member #187 Joined: Thu Feb 16 2006, 02:54PM
Location: Central Ohio
Posts: 140

Yes that was me. I've been digging through my archives to find the file. I have a paper schematic in front of me but not the file. It works quite nice, nice enough that I will not be using push-pull anymore. I have tried it on both half and full bridge with great results. I had to do a LOT of research and reconfiguration to get this circuit. I plan to scale it up to air core when I get some work space.

I'll get it to you as soon as I can find it.

Thanks for the mention!

Edit- Oops, looks like you found it! They key to getting the bridge to self-tune is gate blocking capacitors, and the pull-up resistors are necessary to make it self-starting. It's a really great circuit. I was even a bit reluctant to share it, but then I "got over myself". I hope you get some great results.

I am sending you the app note that got me on the right track. I'll link to it if there is any interest.

Keep me updated!

Dr. Dark Current

Sat Dec 02 2006, 08:09AM

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

cbfull wrote ...

Thanks for the mention!

Edit- Oops, looks like you found it! They key to getting the bridge to self-tune is gate blocking capacitors, and the pull-up resistors are necessary to make it self-starting. It's a really great circuit. I was even a bit reluctant to share it, but then I "got over myself". I hope you get some great results.

I am sending you the app note that got me on the right track. I'll link to it if there is any interest.

Keep me updated!

Do you think that just increasing the value of the base pull-up resistors will work to run it from mains? Or maybe some isolated low power low voltage power supplies will work to supply current to the transistor bases? I'm also quite unsure about how many feedback turns to use.
Thanks for your help!

Marko

Sat Dec 02 2006, 09:44AM

Registered Member #89 Joined: Thu Feb 09 2006, 02:40PM
Location: Zadar, Croatia
Posts: 3145

If I remember, very old OSRAM energy saver bulbs used a similar circuit with BUT11A's.

They also got some kind of small 3:3:5 base drive transformer wich now believe to be a current transformer.

Advantag of halfbridge is that devices never see more than supply voltage and you can run such a circuit under 230V mains with 500V devices. (with royer you would need 1200 or so V MOSFET's or IGBT's wich spoils the fun).

Problem I see here is that primary part itself isn't resonant, and base drive is hard to get well since coupling is low, so I expect to see quite big losses with that circuit; low operating current may help but even those saver bulbs running at some 10 watts did get very hot.

I was thinking if such a circuit could be maxed up by adding a series resonant capacitor and big iron powder choke on + supply?

Cfbull, do yu have any pics or data of working circuit over there?

RickR

Mon Dec 04 2006, 05:42PM

Registered Member #93 Joined: Thu Feb 09 2006, 04:11PM
Location: Houston, Texas
Posts: 64

CBFULL:

I'd be interested in the app note link, if you wouldn't mind posting it.

Thanks,

Rick

cbfull

Mon Dec 04 2006, 06:32PM

Registered Member #187 Joined: Thu Feb 16 2006, 02:54PM
Location: Central Ohio
Posts: 140

Sorry for the delay, I realized that all my files were here at work, so I had to wait.

Firkragg, why do you say that the coupling is low? It should be the same as it is with any of the other flyback circuits (push-pull, single transistor, etc.), it uses the same feedback windings on the same ferrite core, just different circuitry.

The only picture I have is of a 12" plasma globe operating at reduced pressure with argon, didn't take any of the rat's nest I made. The picture of the globe is nice, but that's at home! I'll see if I can get it posted.

Here is the pdf of the App Note that got me started. Keep me updated on your findings.

] 1165257075_187_FT18570_an1329_self_oscillating_res onant_cfls.pdf[/file]

Marko

Mon Dec 04 2006, 07:20PM

Registered Member #89 Joined: Thu Feb 09 2006, 02:40PM
Location: Zadar, Croatia
Posts: 3145

Yup, that's exactly the same circuit I talked about

And I fear you might be in trouble running such circuit at considerable power levels without any resonant cpacitor, firstly because transistors will suffer heavy losses since they are hard switching, with poor base drive signal since feedback windings still have relatively low coupling.

Also, it seems that such a circuit is much more likely to follow core-saturation oscillation (just like '3055 driver) than fundamental frequency of flyback because it's Q factor is low (unless it's tuned for it).

With compact lamps at some 10 watts and tiny currents both of these aren't problematic, and for driving small stuff like plasma gloes circuit is OK.

As described in app, circuit works that way that it gets triggered by RC and diac, and goes into high-Q, high frequency oscillation with small capacitor placed over lamp therminals (since lamp yet repreents open circuit).

Secondary voltage rings up to kilovolts and filaments get hot from tank curent.

Lamp then strikes, lowers the frequency and Q greatly so circuit falls into lossy 'core saturation' oscillation wich is essential here in order to keepsome kind of reactance between circuit and lamp.

If circuit continued to resonate it would blow itself up from overcurrent.

I had an idea to use similar circuit, with a series resonant cap with flyback priamry.

This would require big choke on supply rails to act as a current source, although I don't know how well would it work.

In that case circuit would operate in ZCS and hopefully have minimal losses, although it's a just theoretial circuit for now.

vasil

Mon Dec 04 2006, 08:22PM

Registered Member #229 Joined: Tue Feb 21 2006, 07:33PM
Location: Romania
Posts: 506

This is exactly the schema I wanted to use in a SSTC (didn't work because wrong RC values in the diac circuit). I was inspired by Mazzilli Vladimiro that used it for a small but working SSTC. The schematic is here:

http://www.pupman.com/current/vladi2/index.html

Look in pdf files at mosfet.pdf (paralleled fets for higher current).

It works at least for air core SSTCs.

cbfull

Mon Dec 04 2006, 08:53PM

Registered Member #187 Joined: Thu Feb 16 2006, 02:54PM
Location: Central Ohio
Posts: 140

It's starting to make sense. I can see that we will start running into problems as we push more power into this circuit. Notice that the drawing below does not have the diac start-up circuit as in some of the App Notes. That's the purpose of the base blocking capacitors and the pull-up resistors.

I wanted the circuit to function as close to the classic push-pull as possible, so that I could keep all variables (components) consistent between the two types and evaluate performance. It's been a while since I ran it but it worked beautifully with bipolars, both half and full versions.

The one thing I remember is that I had to put twice the supply voltage to the H-bridge (full) to get the same output. This I did not expect but I assumed that it was because with two switches on during one swing, the switches may be adding resistances as though they are in a series circuit. The next logical step would then be to double up each switch to cut the resistance in half, but since I was using bipolars I didn't want to waste my time. I'll wait for the MOSFET version.

Here are my original sketches. I have only used it with bipolars powering my bigger plasma globe (it was grossly overdriven and the central terminal wire started to burn a hole in the glass. I jumped out of my seat and shut off the power before it burned all the way through.

Bipolar half and full drawing without component values:

1165264517 187 FT18570 Half And Fullbridge Bipolar Work In Progress

I also redrew the circuit with MOSFETs in place of the bipolars, but I was not sure what to make of the gate capacitors since the FETs already have tiny capacitors built in. Any thoughts on this?

Thanks for all your comments guys.

EDIT- I should mention that the circuit in question is figure 7 in the app notes.

Marko

Mon Dec 04 2006, 08:53PM

Registered Member #89 Joined: Thu Feb 09 2006, 02:40PM
Location: Zadar, Croatia
Posts: 3145

How you intended to limit the current, and also be in tune with secondary?

My guess would be, that with so low coupling to secondary circuit would start to oscillate with it's primary circuit, with virtually no impedance and destroy itself. (Or was the saturation of ferrite core dictating the oscillation, if so, coil would be impossible to tune).

I was looking for some kind of such small halfbridge converter, wich would be ideal for small unregulated SMPS's; up to some 25 watts of power, simple saver-bulb configuration may even work well; but I would really like something powerful wich I wouldn't need 1200V devices in order to run from 240V mains!

cfbull: generally, bipolars are better bang/buck for such low power systems; equivalent mosfets used in CFL's have ON resistances in order of few ohms and probably aren't much more efficient than bipolars at that power level.

Only thing I think that needs to be done is to somehow employ soft switching topology; it would automatically allow much more power trough the circuit.

I found some more saver bulbs. They all use practically same topology with little differences in circuits:

1 2 next

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