Ward's controller @ low frequency and short on times / Tesla Coils / Forums

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

Ward's controller @ low frequency and short on times

1 2 next

Move Thread

LAN_403

Finn Hammer

Sun Apr 19 2009, 03:12PM

Registered Member #205 Joined: Sat Feb 18 2006, 11:59AM
Location: SkÃ¸rping, Denmark
Posts: 741

Steve Ward's DRSSTC-1 controller soon became the industry standard for DRSSTC's, including the current most powerfull version, the mighty FATBOY, of Arizonean fame.
I visited Arizona recently, and Chris being the generous man he is, handed me a board for a Ward controller, which I have just populated.
Having experienced the Fatboy in person, I admit to being flabbergasted by the performance of the beast. It's ability to keep ionised channels simultaneously open for later attack by streamers is nothing short of increadible, and I want one too.
It is hard to argue against something that works, and Chris says: "c'mon, just clone and make sparks all day".

Well, good point there, Chris!.

However, my audio modulating scheme handles the volume aspect of the sound by changing the on-time according to the envelope curve of the instrument being played, so I need to be able to handle short on-times like 10ÂµS of so.

The J-K FlipFlop in the controller needs the PRE input to be held high till past the sync shutdown to the clock has been performed, and this delay is created by an RC time constant formed by R9 and C14 in the controller schematic.

On the scope pic here you see the signal from the interuptor on channel 3(as it enters CLR in inverted form).
Channel 1 shows the resulting waveform in between R9/C14 and the schmidt trigger.
The resulting waveform, -after- the schmitt, is seen in trace 4, and it is this waveform that enters pin 5 of the flipflop: PRE. The cursers show the breakpoints where the yellow waveform pass the schmitt trigger levels.

In the next scope pic, you can read out the delay time. It is this delay time the flipflop needs to be able to sync the shutdown of the enable pin on the gate drivers, to the next rising edge on the clock, _after_ channel 3 went high. This delay forms the window, where there should be room for at least 0ne full cycle at Fres.

You will see that I am aiming at full sync ability down to 20kHz.....

But what happens if I shorten the on-time as seen on trace 3?
Next scope shot, the on time is down to 100ÂµS and the sync-window has shortened to 28ÂµS which means that a sync is only possible with a 40kHz signal. The reason is, that the trace on signal 1, yellow, does not have time to settle down on 0 volts.

At 72 ÂµS on time, the game is lost, no signal change at all is passed on to PRE.

Since the problem arises because the yellow trace never has time to reach zero, the solution is to enable it to do so, rapidly, and independently of the RC time constant, by bypassing the resistor, R9, with a diode to allow the schmitt trigger output to pull the cap to ground fast, in this case it takes about 200nS.
After this litle fix, the waveforms look like this, even with a short on-time of 10ÂµS:The sync-window has been restored.

I therefore recommend this litle mod for everybody who runs this board with lowish frequences, if you like to start the coil up with short on-times and if you like to keep your snubber resistors cool.

You may want to solder the 1n4148 to the board, perhaps on the rear side.

Cheers, Finn hammer

Dr. Drone

Sun Apr 19 2009, 03:36PM

Registered Member #290 Joined: Mon Mar 06 2006, 08:24PM
Location:
Posts: 1673

Finn Hammer

Sun Apr 19 2009, 04:04PM

Registered Member #205 Joined: Sat Feb 18 2006, 11:59AM
Location: SkÃ¸rping, Denmark
Posts: 741

Cool,

May I suggest another mod, then.
The LM311 sits at the end of a 5V supply line, which is supplying the gate drivers up stream.
It seems to me that this analogue comparator should have it's own separate trace from V-reg, since the gate drivers, no matter how well decoubled, are going to induce voltage spikes being fast and power hungry.
I have no data to support this opinion, just a hunch, but I butchered my own board and added a separate strip in the form of a piece of wire.

Also, if you feel like it, the Industrial Fiberoptics IFD95T

is a perfect drop in receiver, hooks right up with V+ gnd and input, which you might consider to fit to the board instead of the RCA jack. I did, and it works flawlessly

It is the black receiver with the yellow dot. It contains a TTL output, which is active high.

Cheers, Finn Hammer

Dr. Drone

Sun Apr 19 2009, 04:20PM

Registered Member #290 Joined: Mon Mar 06 2006, 08:24PM
Location:
Posts: 1673

TheBoozer

Sun Apr 19 2009, 10:49PM

Registered Member #1535 Joined: Wed Jun 11 2008, 11:37PM
Location: Northeastern Pennsylvania - USA
Posts: 117

In regards to Steve Ward's DRSSTC-1 driver board:

If I may ask, Dr. Spark, when you mentioned "In the midst of changing .brd file to fix the ground on pin 7"....

I bought 20 boards from BatchPCB a while back. I never checked for errors, but during populating the board and paying close attention to detail I noticed pin 7 of the 74HC14 wasn't connected. This is the ground pin and I figured it was bug. If this what you were referring to? Are there any other bugs I'm unaware of?

I've searched high and low and couldn't find any mention anywhere in any forums. I made my own trace and never brought it up to anyone. I haven't found any other apparent bugs.

I've populated the board. For testing, it's being used to drive an induction heater using Steve's latest interrupter. It appears to work very well. 350a peak through CM100DY12H. The board drives a daughter gate drive board. Eventually It'll drive my first DRSSTC.

Finn Hammer: Thanks for your board tweak. I'll look into that later. Any tweaks would be quite useful to me since I have quite a few of these boards now...

Finn Hammer

Mon Apr 20 2009, 08:26AM

Registered Member #205 Joined: Sat Feb 18 2006, 11:59AM
Location: SkÃ¸rping, Denmark
Posts: 741

TheBoozer wrote ...

Finn Hammer: Thanks for your board tweak. I'll look into that later. Any tweaks would be quite useful to me since I have quite a few of these boards now...

Actually, pin 3 and 4 on 74hc14 are not connected either. This means that an input, pin 3, is left floating, and it should not, so ground to that one too.

Btw, you are all aware that the pins on the board and the pins on the schematic are different.
They go like this:
schematic 1,2 = board 1,2
schematic 3,4 = board 13,12
schematic 5,6 = board 11,10
schematic 9,8 = board 5,6
schematic 11,10 = board 9,8
schematic x,x = board 3,4

Hope this helps.

Cheers, Finn Hammer

TheBoozer

Mon Apr 20 2009, 09:18AM

Registered Member #1535 Joined: Wed Jun 11 2008, 11:37PM
Location: Northeastern Pennsylvania - USA
Posts: 117

The fact that the pin numbers didn't match the schematic made it slightly confusing. It was a little difficult to ensure inputs and outputs and all the other traces were proper. I became worried that the board might not work. I'm glad it does. :) I was to lazy to correct my schematic. I intend to do so. Your notes will make it easier. Thanks. I've read a million times that leaving inputs float is a bad idea. I forget why, but thanks for spotting that one. I'll ground it. :)

Dalus

Mon Apr 20 2009, 09:34AM

Registered Member #639 Joined: Wed Apr 11 2007, 09:09PM
Location: The Netherlands, Herkenbosch
Posts: 512

Dr. Spark wrote ...

Hi Finn,
Good input will also add optics and add a fat trace to 311.

Will order boards the week after next, claim you boards now as only 10 for sale!

rgs,
ch

Edit: 8 sold already so will bump up order to 30. (More boards = less cost to all).

May is suggest to bulk buy all the needed parts to populate the board. That would make it even more cheaper for us all to build a nice DRSSTC

Finn Hammer

Mon Apr 20 2009, 02:17PM

Registered Member #205 Joined: Sat Feb 18 2006, 11:59AM
Location: SkÃ¸rping, Denmark
Posts: 741

Dalus wrote ...

May is suggest to bulk buy all the needed parts to populate the board. That would make it even more cheaper for us all to build a nice DRSSTC

And you could ask someone to solder all the parts in for you, then it would be easier too.

ROFL Finn Hammer

Dalus

Mon Apr 20 2009, 02:22PM

Registered Member #639 Joined: Wed Apr 11 2007, 09:09PM
Location: The Netherlands, Herkenbosch
Posts: 512

Uh . . . Yeah that might be easier but not cheaper. I'm not lazy . . .

1 2 next

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