Beefy ZVS / High Voltage / Forums

Forums

4hv.org :: Forums :: High Voltage

« Previous topic | Next topic »

Beefy ZVS

previous 1 2 3 next

Move Thread

LAN_403

Adam Munich

Mon Oct 04 2010, 08:04PM

Registered Member #2893 Joined: Tue Jun 01 2010, 09:25PM
Location: Cali-forn. i. a.
Posts: 2242

Yeah I know, these dip switches are rated at 100mA each, but the load will be spread across them. Also, switches tend to be dramatically overspec; I've pushed dip switches up to .5A per microswitch before without any problems.

Also, my Jameco order ships tomorrow, while the estimated charge date for my Allied order is October 31st. Wow, at least there'll be nothing left for them to charge.

genious 7

Mon Oct 04 2010, 09:23PM

Registered Member #2887 Joined: Sat May 29 2010, 11:10PM
Location: Panama City, Panama
Posts: 107

Why dont you do this?

Another thing you can do is use a separate 12v supply for the gate pull-up, thus meaning less voltage drop across the pull-up resistors. I have done it with no problems, and it only needs to be something like a 500ma supply.

I've done it before and works quite well (Although, I did it because I had forgotten to buy zeners

). Instead of connecting the two 470 ohm resistor to the main power supply, connect it to a secondary 12 V one, and connect the rest as normal. That way, the resistors will only dissipate those 12v, while the flyback + MOSFETs obtain the bulk of the voltage.

Just look at my paint-modified schematic.

Adam Munich

Mon Oct 04 2010, 09:53PM

Registered Member #2893 Joined: Tue Jun 01 2010, 09:25PM
Location: Cali-forn. i. a.
Posts: 2242

Well I already ordered the parts, but if the DIPsistor doesn't work out as planned, I'll try that method. A 7812 could be used I suppose.

Adam Munich

Tue Oct 12 2010, 04:59AM

Registered Member #2893 Joined: Tue Jun 01 2010, 09:25PM
Location: Cali-forn. i. a.
Posts: 2242

Success!

I got my ZVS working and I like it. Here's a pictorial build log.

Designing the PCB on the fly.

Board's done. Front, Back. It's a clever design IMO and doesn't require any wire jumpers. I'll make a bitmap of it in case anyone wants to copy it.

DIPsistor front.

DIPsistor back.

The board is populated. Here's the back. I thickened traces that will handle a lot of current with a boatload of solder.

A closer look at the ZVS. The heatsinks may be a little small, but a fan could fix any heating problems. They were the only identical pair I had. I also made the inductor removable.

And the test. It's an 18" computer monitor flyback. The primary was 5+5 turns of poorly wound 20AWG magnet wire, the inductor was 29 nice turns of 20AWG wire. Power supply was a 12V AGM tractor battery, current flow unknown but likely under 3A. Both the DIPsistors were set to 460 ohms. I'm too tired to play with it any more tonight, but I'll push it 'till the flyback fails tomorrow.

tritium

Tue Oct 12 2010, 04:53PM

Registered Member #2591 Joined: Fri Jan 01 2010, 11:58AM
Location: netherlands
Posts: 76

I run this circuit with 100 ohm resistors, in LTspice i see 10 volt gate charge, and work perfect, I use the 7812 for the gate driving.

why it is 470 ohm? i think it is without voltageregulator and running at 40+ volts.

Arcstarter

Tue Oct 12 2010, 08:21PM

Registered Member #1225 Joined: Sat Jan 12 2008, 01:24AM
Location: Beaumont, Texas, USA
Posts: 2253

Glad to hear it works

One problem i always had with ZVS drivers is the voltage on the gate getting too high, and shorting the protection zener (normally used 18v zeners). That would cause that mosfet to turn off, but the other would be on, and since it was not oscillating, it would kill the mosfet by shorting it across the supply through the primary.

My way of solving this, as well as letting me get by with one 1/2 resistor per gate, was using the aux 12v supply. I never had those awful parasitic oscillations, which often requires higher resistance to fix. I am using 500 ohms, but i need something lower, as it heats a bit at 40v. The gates are being turned on too slow, and too little voltage present on the gate.

You could easily wind an aux supply on the LOPT. Maybe 10-15 turns on the core, a limiting resistor, and a zener. Also a small reservoir capacitor may be necessary, as well as a resistor to charge the capacitor through the main supply to start the oscillations. This is how most SMPS's from CPUs and TV main boards seem to be made. Uzzors' site has a good example:

Antonio

Tue Oct 12 2010, 11:33PM

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

Why high-power resistors? Their only function is to charge the gates.
This circuit has several possible modes of failure, and no protection against them. It will self-destruct if the series inductor is too small, if it gets saturated, if there is too low load at the output, and so on.

Adam Munich

Wed Oct 13 2010, 02:00AM

Registered Member #2893 Joined: Tue Jun 01 2010, 09:25PM
Location: Cali-forn. i. a.
Posts: 2242

I know I should've used a 7812 to supply the gates, but I just followed Mazilli's bare bones design. I just wanted to play with my new parts. This circuit will likely not be permanent, and I will rebuild it sometime using a 7812 to supply the gates when I find a bigger piece of copper clad. I'll also try and get the new board CNC'd.

The relatively high value resistors are to prevent parasitic oscillations, and with the DIPsistor they can be adjusted anywhere from 350ohms to 1.1kohms.

1286935249 2893 FT97679 New Bitmap Image

Edit: I hooked it up to 2 100Ah ups batteries, something I like to call my backbreaking (150lbs) 24v supply. Using a regular computer monitor flyback I got 1cm thick, 3.5" long arcs. They started at 1.1cm, so the flyback is likely putting out 10kV. After about 5 minutes of constant arcing, my capacitor was about 100*F and the MOSFET heatsinks were about 90*F. A fan may be necessary, but the MOSFETs don't generate very much heat at all. Playing with the DIPsistor didn't affect performance at all. Arcs were the same at 350 ohms as they were at 1.1kohms. I like the way the arc squeaks before it breaks.

Edit2: @36v 6" arcs, 1cm thick, 13kV. Cap and FETs get really hot really quick.

Tomorrow I'll put on a fan and try my superflybacks.

Edit3: Awesome, I just found a 48V 8A power supply with a medium sized 7krpm fan. Figure I can take it out if I leave the cover off the PSU and have it passively cooled. I can squeeze a 7812 to power it in between the two heatsinks. No heatsink needed as the fan draws only 200mA. Not sure what to do about the hot cap though...

Arcstarter

Wed Oct 13 2010, 03:35AM

Registered Member #1225 Joined: Sat Jan 12 2008, 01:24AM
Location: Beaumont, Texas, USA
Posts: 2253

48v would kill a typical 12v regulator. Not to mention, at 200ma, the regulator would dissipate about 11 watts. Remember, rectified 48v will be about 67 volts. That's too much voltage for 200v mosfets i believe, and say goodbye to your LOPTs

.

Killing an LOPT may not be so terrible, but superflyback or not, they cannot handle that much power, unless you add more primary turns which in the end would not have much better performance (other than the fact that the current for a given power is lower when higher voltage is used).

Just don't want you to kill anything, i know how bad that sucks. I killed my driver, and 3 flybacks yesterday

Two of which where from rear projection TVs.

Adam Munich

Wed Oct 13 2010, 03:51AM

Registered Member #2893 Joined: Tue Jun 01 2010, 09:25PM
Location: Cali-forn. i. a.
Posts: 2242

Yeah I was just thinking about that... I might just make the fan a battery powered snap on thing. I'll only need it at >24v anyway.

"rectified 48v will be about 67 volts"
I'm powering this from lead acid cells. I'm pretty sure they aren't measured in RMS voltage. :p

Actually I'm aiming to kill at least one computer FB. Maybe 48V will do it. I'm not going over 48V though.

previous 1 2 3 next

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