A few questions reguarding my first NSTC / Tesla Coils / Forums

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

A few questions reguarding my first NSTC

previous 1 2 3 next

Move Thread

LAN_403

Steve Conner

Fri Apr 21 2006, 09:36AM

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

Hey hazmatt,

If you're ever wanting to collaborate on that digital spark gap controller, I'm up for it. The reason being that I have a spark gap with a universal motor too, and it would be nice to run it sync. I've written PIC code for digital PID controllers and phase locked loops before, and the motor controller is basically a PLL. Incidentally, that suggests you could do it all with a 4046 PLL chip and some op-amps, and not have to bother with a uC.

If the RPM on your rotary gap changes at .5Hz, that suggests it's not syncing properly. This situation of "almost 120bps but not quite" is about the worst thing you could do to an NST. A sync gap needs to be just that: sync. The electrodes should present exactly 120 times per second (or 240, 360, 480, whatever) and always at exactly the same instant relative to the waveform of the line voltage. (which is why I suggest controlling it with a PLL.)

So I would look at getting that fixed. I have a 1.5hp induction motor here that was modified for sync use by grinding the rotor, and if I run it at low voltage, I can hear it slipping at about .5Hz like you said. If I increase the voltage on the variac it locks in properly. The motor is for sale if anyone in the UK wants it :P

Jrz126

Fri Apr 21 2006, 12:54PM

Registered Member #242 Joined: Thu Feb 23 2006, 11:37PM
Location: Erie PA
Posts: 210

So you pretty much need to be dead on accurate. What if there is just a slight phase shift? (Electrodes fire at 120Hz but they are slightly off from the peak of the Line voltage)

My idea:
We used embedded C for programming pics in my microcontrollers class, so I wouldnt think it'd be too hard to program. My inital thought was to seperate it into 2 smaller problems. First would be to just have the motor run at 3600 RPM. Second would be to get the electrodes sync'd to the line voltage. As long as the motor is spinning at exactly 3600 RPM, it would just be a matter of syncing to the AC.
A simple hall effect senor and a magnet would be used to detect the RPM, and maybe a Zero crossing detector for the AC voltage. The hall effect sensor would be placed 90degress away from the stationary electrodes, when the zero crossing is detected, by the time the flying electrode moves the 90 degrees to the stationary electrodes the AC voltage should be at the peak.
The uC would control the motor through a simple PWM. Once the motor was up to speed, it would then sync up the electrodes by skipping a pwm cycle or 2 (turning the pwm off or full on depending on the electrode position).

It sounds like it should work, in my mind anyways...

Some problems with this idea:
I dont know if there is any extra force involved when the electrodes fire. Im pretty sure there is, but I dont know of the significance.
I'm also assuming 2 flying electrodes running at 120 BPS (although you could easily use 1800RPM and 4 electrodes). How are the electrodes positioned when running at a higher BPS? Still equally spaced? or do you want them so that you get multiple firings per Vpeak?

Opp time for class, I hope it all makes sense, dont really have time to proofread.

Oh, and about the Terry Filter, What is a good cutoff frequency to use? I calculated about 3mhz or something from the schematic. (although I dont know if the primary coil inductance plays a part in it).

Steve Conner

Fri Apr 21 2006, 01:46PM

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

Hi jrz

The slight phase shift you're talking about should be adjustable by the user, and the system should accurately hold the phase shift the user sets. You probaby want the adjustment to have a range of 90 degrees, so the electrodes present at a line voltage peak at one extreme, and at a line voltage zero at the other extreme.

The phase shift on a traditional SRSG is usually adjusted (by trial and error for best performance of the coil) by turning the motor in its mountings, moving the electrodes, or using a fancy L-C phase shifter circuit invented by John Freau.

If you want to do this, you'll basically be writing a digital phase-locked loop algorithm, and it will include a PID controller like hazmatt mentioned to compensate for the characteristics of the motor and inertia load. One peculiarity of PLLs is that you may not be able to use any "I" in your PID at all for stability reasons, but the "I" is what keeps the phase shift constant. You may have to use a fake integrator whose gain doesn't tend to infinity at DC, for this reason.

Hazmatt_(The Underdog)

Fri Apr 21 2006, 09:23PM

Registered Member #135 Joined: Sat Feb 11 2006, 12:06AM
Location: Anywhere is fine
Posts: 1735

yea....lots of work. I will hopefully use the basic stamp's data logging capability to get my motor inertia and time constant parameters. After that, transfer function, woopiee...*sigh* Then to Matlab and simulink for simulation and compensation. Probably have to write some custom code for feedback, I dunno yet.

Jrz126

Fri Apr 21 2006, 10:14PM

Registered Member #242 Joined: Thu Feb 23 2006, 11:37PM
Location: Erie PA
Posts: 210

look what I found:

The code (asm and C)

It'll need to be modded alittle, but atleast there is a good example.

Heres what Im planning:

The zero crossing will detect both 0 crossings per Hz of the AC voltage. The difference between the zero crossing and the hall effect will be used for the error. Then I could probably add a phase shift by just adding a constant to the error.
error = ZC time - Hall time +offset.

There will be a timer running at 8.3333mS (1/120hz). The difference between the zero crossing and the timer will be the error for the PID. Then the hall effect will just be a phase shift issue.

Terry Fritz

Sat Apr 22 2006, 02:47AM

Registered Member #393 Joined: Tue Apr 18 2006, 12:30AM
Location:
Posts: 297

Hi,

A microprocessor based sync motor sounds cool, but there are some wildly nasty fields around the gap and all that is going to play havoc on the electronics.

Why not just convert an AC motor? It is not hard and it is well proven. I am new here so fogive if I messed the point or something.

Cheers,

Terry

Jrz126

Mon Apr 24 2006, 01:28AM

Registered Member #242 Joined: Thu Feb 23 2006, 11:37PM
Location: Erie PA
Posts: 210

I could use a sync. motor, but I already assembled a DC motor and I'm sure there are plenty of other people that would find it useful.
I should be able to fix the field problem by shielding everything in a metal box.
Hopefully I'll have a prototype by the end of the week.

For the terry filter, do I really need to use a 100W resistor? I have several 12 ohm 200W resistors, would I be able to use these and just adjust the cap value?

Terry Fritz

Mon Apr 24 2006, 07:45PM

Registered Member #393 Joined: Tue Apr 18 2006, 12:30AM
Location:
Posts: 297

Hi,

Ok cool! Try to make all the wire leads of any significant lenth "twisted pair" to help regect stray fields. That is even "indside" the enclosure too. I usually use die cast metal baxed with the lid sealed with copper tape.

The filter needs say 500 to 2000 ohm resistors. The exact value is not important, but it can't be "too" far off. The 100W resistors are pretty over rated so you can use say 20W but plan on them getting really hot. Those ceramic resistors run "at full rating" are 300C!!! They glow dull red in a dark room!! The filter's resistors are far over rated since folks just don't accept how hot they get. The 100W versions run at say 100C which is far more normal.

Cheers,

Terry

Dr. Drone

Mon Apr 24 2006, 08:11PM

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

Jrz126

Wed Apr 26 2006, 03:41AM

Registered Member #242 Joined: Thu Feb 23 2006, 11:37PM
Location: Erie PA
Posts: 210

I found some 750 ohm 50W resistors locally. they're kinda small, so Im planning on running 2 in series on each leg.

I made some decent progress on the motor controller...unfortunitly, I burned up my pic...guess I should have turned the powersupply off...
Hopefully I have another pic somewhere. I'll try to get some pics posted sometime soon

previous 1 2 3 next

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