High Power Home Brew Royer Based DC Power Supply -HPHBRBDCPS / Projects / Forums

Forums

4hv.org :: Forums :: Projects

« Previous topic | Next topic »

High Power Home Brew Royer Based DC Power Supply -HPHBRBDCPS

Move Thread

LAN_403

...

Sat May 06 2006, 03:26AM

Registered Member #56 Joined: Thu Feb 09 2006, 05:02AM
Location: Southern Califorina, USA
Posts: 2445

The name isn't the only thing that is long... With a (estimated) 750w output the sparks pull out to about 6" long

Driver
The circuit is a based on the Mazilli flyback driver using a pair of 40n60's (to-247 package on a small heatsink that does not get very hot) and a tank cap made of 3 3ufd/450v afps (high frequency/high current) aerovox caps in series. The caps stay warm, but the soldered joints get hot enough to melt the solder if they are not tightly twisted before soldering. After fixing that the hottest part of the circuit is the primary made of 4+4 turns of 10awg stranded wire, which gets not enough to soften the insulation (but not seaming hot) during extended runs. The thing runs at about 100khz.

Transformer
It is the one I described here, in post 2. Just your ordinary home wound flyback transformer, about 500 turns of 32awg wire heavily insulated with OH transparencies and rtv. The core saturates at about 20v p-p, so my 500 turn primary only gives 10kv out. To make this more impressive (heck, you hardly get corona with 10kv) I added a way over sized string of sf1600's and a 2n2 6.4kv cap as a level shifter to get ~20kv (I guess, at 50v ac-about 70v in the driver-the ark initiates at ~2cm).

Plasma
As you can see in the video, at 80v ac in (~110v into the bridge) the ark initiates at ~3.5cm, so I guess about 35kv

. It draws 10a at 80v when the ark is about to extinguish, so generously assuming 50w losses in the primary/fets/tank cap (those are the only parts I felt getting warm, and they were not very warm at that) that is about 20ma in the ark. But looking at the scope it looks like the voltage drops about 30% when I draw arks, that is in the 25-30ma range... I still think that I am missing something, these are FAT arks. Here is a screen capture of the afterplamsa...

If for whatever reason you can't play the other file (or are too lazy to click on it)

For those of you that can download 4mb files and play .wmv here you go :)
]1146885560_56_FT0_80voltin.zip[/file]

sorry about it being rotated 90degrees... Just turn your computer so that the plasma rises upwards

peter

Steve Ward

Sat May 06 2006, 04:15PM

Registered Member #146 Joined: Sun Feb 12 2006, 04:21AM
Location: Austin Tx
Posts: 1055

Very nice results given the setup you are using.

But looking at the scope it looks like the voltage drops about 30% when I draw arks, that is in the 25-30ma range... I still think that I am missing something, these are FAT arks.

The current is probably in fact much higher. You can't assume that when drawing an arc that the power in the arc is equal to the open circuit voltage X the arc current. Rather, when you draw an arc, the voltage across the arc drops considerably (maybe 1-2kV), so the current is that much higher accordingly.

...

Sat May 06 2006, 04:38PM

Registered Member #56 Joined: Thu Feb 09 2006, 05:02AM
Location: Southern Califorina, USA
Posts: 2445

That is what I thought, but I had a single turn around the core to monitor the voltage (I also had a current transformer on one half of the tank circuit to see when the core saturates) which only dropped ~30%

I can't imagine I am getting very much voltage drop in the windings themselves, even if the coil was 100ohms and I was drawing an amp that is only 100v, not enough to notice with 10kv out... Unless there is something else I am missing...

I was thinking of pulling an ark through some led's to see it it is 20ma or 200ma, but I killed my last 40n60 yesterday (the solder on the caps melted, the mmc fell apart and the fets died

). I might give it a go with irfp460's... I worry about their on-state resistance, the 40n60's were getting warm but not hot on the heat sink, which seems about right if they are dropping 1v @ 80v x2 =~15w. If there is 10a going through a 460 that would be ~2v drop, so ~30w--way to much for my little heat sink. I suppose I could put them on an old cpu one XD

I can't mess with it now though because I have the ap chem final in 0 days, 4hrs, 21 min, 26 sec...

Steve Conner

Sat May 06 2006, 05:46PM

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

IFACMAAFE: I Fear Americans Can Make An Acronym For Everything

Now onto what Ward was saying: An arc isn't stable without a ballast, so if you can draw an arc from a circuit, it must have ballast. In circuits like these, the ballast is the leakage inductance of the step-up transformer.

This leakage inductance causes the power factor to drop when you pull an arc, so calculating the output voltage and current is more difficult than you think. The output voltage will droop more than what the sense winding suggests, because some of the leakage inductance is "after" the sense winding in the magnetic circuit.

I'd estimate that the output voltage is about 1/3 the open circuit voltage. The voltage rises as you pull the arc out, and the maximum power transfer theorem says that the arc will become unstable and go out when the voltage gets to around half the open circuit voltage.

So I'm going to guess the voltage and current are about 10kV and 150mA when the arc is as long as it's going to get.

...

Sat May 06 2006, 06:06PM

Registered Member #56 Joined: Thu Feb 09 2006, 05:02AM
Location: Southern Califorina, USA
Posts: 2445

Actually knowing the amount of current there is... I believe the ballast for this is the 2n2 cap in series with the output... In any case, I see you point. I really need to get a hv voltmeter/ammeter going...

UPDATE:
It looks like the 40n60's were killing performance...
I wanted to get some plasma out... So I took a pair of irfp250's and fired it up. Since they are only 200v devices I couldn't turn the variac past 50v, but I decided to mess with them anyway... I think I hit some sort of resonance, it was acting pretty normally then I noticed more than normal corona (more even than with 80v in) so I turned off the lights to investigate... Then the think arked over, about 5cm! I am unsure if it was tracking or just air, but that is the farthest it has arked so far... There is about a 2v range where I get a 3/8" plume of corona off my output wire; and that is DC

At that point I could initiate the ark around an inch (estimate, no way to measure that)

Steve Ward

Sun May 07 2006, 04:44PM

Registered Member #146 Joined: Sun Feb 12 2006, 04:21AM
Location: Austin Tx
Posts: 1055

So I'm going to guess the voltage and current are about 10kV and 150mA when the arc is as long as it's going to get.

Exactly how it works

, and thats why the arc eventually breaks (you cant supply the juice to keep it going). But dont go thinking you can supply 10kV and 150ma for long periods of time from that small transformer! Drawing arcs is actually a lot easier on it, than say a high power tesla coil supply, where you are continuously drawing maximum current, and the voltage is ramping up linearly to its max at a high rate.

If you had a method to watch the instantaneous power output when arcing, you could see that the power increases greatly when you drag the spark out. If one end of your HV transformer is grounded, you could put a current transformer on that wire and watch secondary current. You could also use a HV probe to watch the voltage, but the HF might mess with a resistive divider.

And to reinstate what Conner said, your transformer has leakage inductance on all of the windings (this means some of the magnetic flux from the primary is not intercepted by the other coils). NSTs have rather high leakage, which is why you can short them and they will current limit, and the output voltage will drop to what the load impedance presents. This is also why your 1 turn measuring coil (while a neat idea indeed) isnt terribly accurate. It does give a pretty good idea of the voltage waveform present (just not the output voltage of the secondary necessarily).

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