brtaman's spare parts DRSSTC - build thread. / Projects / Forums

Forums

4hv.org :: Forums :: Projects

« Previous topic | Next topic »

brtaman's spare parts DRSSTC - build thread.

Move Thread

LAN_403

brtaman

Wed Jul 10 2013, 05:08PM

Registered Member #2161 Joined: Fri Jun 05 2009, 03:36PM
Location:
Posts: 247

Hello everyone,

It has been a while. :) Hope all has been good on your end.

I've finally managed to attain some free time lately and was cleaning my "lab", when I came to the realization that I am a hoarder with far too many components gathering dust. So I've decided to slam them together into a semi-decent DRSSTC and since I've always loved reading build threads but haven't actually done one myself, I figured it is time to slap one together.

Specs are still hazy at this point, but I will update them and add pictures as the project progresses. I am however attempting to recycle as many things as possible, due to an all around lack of time.

Specs so far:

Secondary: 15x50cm, AWG30, with recycled torroid(big boy 60cm,17cm): Frequency ~80kHz

Primary: Pancake

MMC: 2-strings of 4 capacitors Arcotronics Snubber Capacitors (Specs: 1400V, 1000A pulsed, 70ARMS, 0.6ESR :))
5600V 2000A pulsed and 140ARMS

Controller: Ward DRSSTC3 (recycled)

Power section: Half -bridge BSM150GB120 bricks - Perhaps with doubler.

Basically the coil will be nothing new, just another variation on a proven design.

Here is a photo of the form and toroid, with the 30AWG wire ready for winding. While it might not look it...that toroid took a very long time to complete. It started out as two square pieces of hard foam glued together and was sanded down to the form and covered with alu tape. The "project" took far too long to complete, in addition my lab looked like it was caught in a perpetual snow-storm. Unfortunately it is quite frail, I really should have covered it with some lacquer or something similar upon completion...

FormTorroid

Here is the current testing rig on my desk. Had a lot of work to do fixing the driver board after the spectacular racing spark fail caught in the video at the end of this post. Also seen is the brick to be used in the power section of the coil.

Testing

Finally here we have the caps to be used on the project, I will likely be running the 2200uF caps in parallel for a voltage multiplier. Also seen are the three options for the MMC. The RIFA FKP(4000v 0.022), RIFA FKP Snubber (4000v 0.33uF) or the Arcotronics GTO Snubber (1400V 70A RMS, 1000A pulsed).

Capactiors

For a preview of the toroid and controller in action here is an old video I dug up, of the experimental coil I was working on years ago.

PROJECT STATUS:

18.07.2013
-Wound new GDT, using a more appropriate ferrite. Leads are still quite long though and will be trimmed during assembly, pushing the UCC gate drivers though :/ (some serious gates), might have to rethink the use of the UCC and switch to FETS.
-Secondary almost wound
-Did some driver testing at 72kHz, will be adding and extra torroid or two the secondary to lower Fres for operation.

Photos:

Waveform3

Anyone know what exactly is going on with the notch in the last picture?

Testing the controller has made me realize how much I have subconsciously missed the lab and my o-scope. :)

Regards,
brtaman

dude_500

Wed Sept 04 2013, 11:36PM

Registered Member #2288 Joined: Wed Aug 12 2009, 10:42PM
Location: Cambridge, MA
Posts: 179

Is the waveform the gate drive? The only obvious thing would be the miller plateau, although it doesn't really look like a miller plateau if that is indeed the gate waveform.

Overall, the gate switching looks pretty slow and will likely result in substantial losses when running the coil. As you mentioned, I think FET driving brick gates is indeed a good idea and would probably dramatically improve your switching speed.

brtaman

Thu Sept 05 2013, 08:18AM

Registered Member #2161 Joined: Fri Jun 05 2009, 03:36PM
Location:
Posts: 247

Hi,

Thanks for the interest. Indeed, as you said a more substantial Gate Drive is required (very slow switching, would probably melt the heatsink or make a nice bang in short time). Project is on hold atm (more or less) as I am finding no time between training and work.

The main issue I have with using these bricks is that they are a complete mystery. According the the internet (chinese semiconductor suppliers), these bricks (BSM150GB120DN11 ) are supposedly the equivalent of the BSM150GB120DN2E3166. (

However, my gate drive trials have proven that they have a much higher gate capacitance. I have tested the gate drive with an old MG100Q2YS42 brick (Cies=12nF) vs. DN2E3166 (Cies=11nF) and have found the gate drive waveform to be perfect, with switching times becoming more than adequate. It really feels as though I am attempting to drive a 500amp brick in terms of capacitance, when I install the BSM onto the gate drive there simply isn't enough "juice".

This worries me. As much as I would like to use the BSM bricks, I am not a fan of operating in the dark without a datasheet and based on experiments, clearly the DN11 part differs significantly from the DN2E3166/DN2.

In other news, the secondary is almost complete, as is the base etc. I will be uploading pictures soon.

Have a nice one,
brtman

Inducktion

Thu Sept 05 2013, 06:31PM

Registered Member #3637 Joined: Fri Jan 21 2011, 11:07PM
Location: Buffalo, NY
Posts: 1068

I think you can actually parallel UCC gate drivers without any problems.... Though don't take my word for it!

And it sounds like they may have paralleled some mosfets inside of the brick to compensate for their actual ratings. Would explain the increased gate capacitance for sure.

brtaman

Mon Sept 09 2013, 06:19PM

Registered Member #2161 Joined: Fri Jun 05 2009, 03:36PM
Location:
Posts: 247

Inducktion, could be. I have no explanation of my own.

As for paralleling, the gate drive waveform was generated using 3x drivers paralleled. Paralleling only 2x resulted in a sharktooth/sine waveform. I would normally dismiss the bricks as defective, however all four samples exhibit exact same gate drive properties...insane.

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