Tesla Coil Build Thread

[Down with Umbrella]

I will post some schematics and other details soon.
I NEED to familiarize my self with 555timer circuitry and function. If I'm not mistaken PED started a 555 timer thread I can review.

 

[petters]

Hey, could someone give me some advice on my SSTC build?

I have a 7 turn primary (pancake coil) made of quarter inch copper tubing and a secondary (2ft, 34awg, 3415 turns).

I am currently running it at 12v, which at resonance brings the voltage across the primary much much higher and it can create up to 1-2 inch sparks only if I bring a metal object close enough. I have previously thrown 170V into it, but it blew my FETs (likely due to me starting at resonance and current draw being high since the copper tubing has such a low resistance).

My question is basically what are my options to get larger sparks? Here is what I see my options as (assuming I upgrade my FETs, which I plan on doing):

-If I wanted to throw 170V in to get larger sparks, operate below resonance to keep current draw low
-Increase coupling by changing the pancake coil to a helix

Those are my only options as of now. I do need to check the current draw from the battery during operation to get a better understanding of where I stand

Thanks

 

[DashApple]

Running and SSTC with a helix run primary is best , its the inductance of the primary that limits current drawn by the unit in CW operation

Running below or above resonance will casue your arcs to be smaller and harder switching in the mosfets , You need to rate the bridge according and run at resonance with a high input voltage to get long sparks .

This was an old SSTC I did , https://www.youtube.com/watch?v=_T6RZCv1he8

Input is 330V DC smoothed , Full bridge @ 500Khz

 

[petters]

Running and SSTC with a helix run primary is best , its the inductance of the primary that limits current drawn by the unit in CW operation

Running below or above resonance will casue your arcs to be smaller and harder switching in the mosfets , You need to rate the bridge according and run at resonance with a high input voltage to get long sparks .

This was an old SSTC I did , https://www.youtube.com/watch?v=_T6RZCv1he8

Input is 330V DC smoothed , Full bridge @ 500Khz

But changing to the helix would also generate larger sparks in comparison just because of better coupling, right? Something I want to be sure of before I dismantle the pancake coil which took me days of frustration to get right :cryyy: oh well live and learn

Yeah, will definitely want to run it close to resonance then, can already see the wave getting tarnished when I move too far out of resonance. Looks like I'll have to go for higher rated components, probably IGBTs with sufficient fall times to handle the current.

Thanks!

 

[Sigurthr]

Realistically there are two options:
1) Run at resonance and adjust primary inductance to the desired current draw, adjust input voltage to the desired arc length.
2) Run above resonance (inductive region) to find an acceptable combination of output and current input. This stresses the bridge the further you run from resonance via inductive kick voltage.

You should never run a bridge in the capacitive (below resonance) region because it causes massive current commutation through the anti-parallel/body diodes. I've built bridges to withstand the capacitive region successfully (my USSTCC recommended bridge is one such one) but it is a lot more expensive than what is needed if you never enter capacitive region. Also, capacitive region causes funky harmonics in the waveform.

 

[Down with Umbrella]

I don't mean to switch topics here but I have progressed in my quest for repairing my SSTC.
Driver testing.
I ordered some 555 timers for the bread board but I found a packet on Amazon.
It's basically a large 555 timer circuit with a few pots and options. I soldered up the board and selected an Astable setting as close as I could to match to produce a square wave.
I connected this to my o-scope and after some trial and error I determined the 555 timer produces a nice square wave at 4.50 kHz.
Hmm. I'm going to make some assumptions here the ground probe(s) on my o-scope will always be connected to "ground" I have run a jumper from the timers ground to the drivers ground (j1)
The timer output has a jumper into the drivers antenna terminal. The timer is powered by a 9v battery and I'll need to use the step down driver to power the driver.
The drivers GDT output I have two probes. Do I only need one or do I use a math function here?
Is this the best way to test the driver?

My goal here is to determine the TC driver is oscillating. As described in sig and loneocreans above posts

 

[Sigurthr]

Looks like you have it hooked up fine in the photos. Power the driver with either the transformer or any 12V dc supply. The 555 board is electrically floating thanks to the battery, and you've made the ground continuous with the white jumper, so you're good.

as far as scope math functions: you want to add the two channels and invert one of them. This gets you a differential measurement across the scope probe pins that isn't referenced to ground.

show the waveform for differential measurement, and for regular of each channel (with both channels displayed but no math functions) and I should be able to see what's up with the driver.

 

[loneoceans]

I don't mean to switch topics here but I have progressed in my quest for repairing my SSTC.
Driver testing.
I ordered some 555 timers for the bread board but I found a packet on Amazon.
It's basically a large 555 timer circuit with a few pots and options. I soldered up the board and selected an Astable setting as close as I could to match to produce a square wave.
I connected this to my o-scope and after some trial and error I determined the 555 timer produces a nice square wave at 4.50 kHz.
Hmm. I'm going to make some assumptions here the ground probe(s) on my o-scope will always be connected to "ground" I have run a jumper from the timers ground to the drivers ground (j1)
The timer output has a jumper into the drivers antenna terminal. The timer is powered by a 9v battery and I'll need to use the step down driver to power the driver.
The drivers GDT output I have two probes. Do I only need one or do I use a math function here?
Is this the best way to test the driver?

My goal here is to determine the TC driver is oscillating. As described in sig and loneocreans above posts

DownWithUmbrella,

The ground pin on your scope is always connected to earth ground through your mains cable. When scoping this driver, there isn't much to be worried about since everything is isolated (your driver is powered via a small transformer which is already isolated and most low voltage supplies are all isolated supplies).

The driver is extremely basic so there shouldn't be much troubleshooting needed. Once you have everything hooked up and the UCCs enable pin tied high (pin 3 of both UCCs) and your 555 hooked up to the antenna feedback, you should see the following:

1) 4.5kHz square wave from your 555 at antenna input
2) Same square wave at pin 1 of the schmidt trigger IC
3) 12V square wave at pin 4 of the schmidt trigger and pin 2 of the UCCs

If you scope Pin 7/6 of each of the UCCs with a different probe, you should see the same square wave but out of phase of each other, because of the UCCs should be inverting and the other is not. Your ground clip should be clipped to the negative rail of your circuit.

Why did you choose 4.5kHz? By the way this driver has no 'osciallation' to speak of since it's simply an inverter hooked up two two gate driver ICs. If not for the DC blocking cap on the antenna input, the outputs of your UCCs will simply be HI/LO or LO/HI depending if you pull pin1 of the 14-pin IC high or low. What follows from this is - probably not a good idea to get your sparks touching ground. This circuit depends on the self oscillation of your secondary coil for feedback and touching the sparks will severely 'mess up' the radiated frequency and this can cause your driver to latch high or low for some period causing the GDT the saturate and your FETs to blow up. If you can, always a good idea to test the circuit at the desired operating frequency which I imagine would be in the 00s of kHz. If this works I'd recommend connecting your GDT and scoping the outputs of the GDT after to ensure your GDT is suitable.

On a side note, why do you have a heat-sink on the inverter (the 14pin DIP)? It shouldn't be getting remotely warm at all otherwise something really bad is going on.

Finally, I noticed that the inverter used is a 12V inverter, which will require a relatively high voltage (maybe >7V?) to transition. This requires a fair bit of induced voltage; i.e. antenna needs to be relatively near the coil / long to pick up a strong signal. Typically in most simple drivers, a 5V part is typically used which would help antenna sensitivity.

 

[Down with Umbrella]

DownWithUmbrella,

Why did you choose 4.5kHz? By the way this driver has no 'osciallation' to speak of since it's simply an inverter hooked up two two gate driver ICs. If not for the DC blocking cap on the antenna input, the outputs of your UCCs will simply be HI/LO or LO/HI depending if you pull pin1 of the 14-pin IC high or low. What follows from this is - probably not a good idea to get your sparks touching ground. This circuit depends on the self oscillation of your secondary coil for feedback and touching the sparks will severely 'mess up' the radiated frequency and this can cause your driver to latch high or low for some period causing the GDT the saturate and your FETs to blow up. If you can, always a good idea to test the circuit at the desired operating frequency which I imagine would be in the 00s of kHz. If this works I'd recommend connecting your GDT and scoping the outputs of the GDT after to ensure your GDT is suitable.

4.5kHz was the highest frequency I could achieve out of the resistors I had available.
Other than that I have no specific reason.

"Sparks to ground", can you elaborate here

The coil ACTUALLY quit working while I was holding my xenon bulb close to the secondary. For longer than usual. Could this extended draw off feedback freq. have caused a chain reaction that caused a problem?
The coil when working properly resonanted at 222-225kHz

Ps how long can I leave the driver powered up for testing with out causing issues sig?

I can answer more of these questions in depth after I do the testing on the driver. Should be soon.
I really appreciate everyone's help and I'm researching the best I can in my free time.

 

[loneoceans]

4.5kHz was the highest frequency I could achieve out of the resistors I had available.
Other than that I have no specific reason.

"Sparks to ground", can you elaborate here

The coil ACTUALLY quit working while I was holding my xenon bulb close to the secondary. For longer than usual. Could this extended draw off feedback freq. have caused a chain reaction that caused a problem?
The coil when working properly resonanted at 222-225kHz

Ps how long can I leave the driver powered up for testing with out causing issues sig?

I can answer more of these questions in depth after I do the testing on the driver. Should be soon.
I really appreciate everyone's help and I'm researching the best I can in my free time.

It depends a lot on what happened since there can be a lot of reasons for a failure. Though typically an antenna feed-back type system like this, especially one which requires a very strong antenna feedback, can be quite problematic if not set up and operated properly. The system works fine if the radiated field from the secondary system induces sufficient voltage in the antenna for feedback. But when you have ground arcs or similar, this can change the frequency of the system significantly and also change this radiated field.

For example, lets imagine that for some reason the antenna starts to get some noise picked up due to a ground arc, and jumps around the threshold voltage of the inverter IC. If this is greater than the hysteresis gap, it can cause a whole bunch of unusual fast switching which can cause issues in terms of inverter switching since typically the frequency is higher than baseline. This usually leads to the FET running for a high duty cycle in the 'transition' linear region and will lead to a blow fet.

Another example is that your feedback might suddenly drop leading the driver to miss a few cycles. It doesn't sound like a lot but this will typically cause your GDT to saturate, often leading to a blown fet. To avoid these problems, many drivers such as PLL drivers drive the coil at a certain pre-set frequency and tracks the freq change as the spark forms. Alternatively some coils employ discrete drivers which avoids the GDT saturation issue.

You can leave the driver powered up for however long you want it to be and it will be fine. It shouldn't be getting remotely warm at all. You'll only start to see the UCCs get a bit warmer if you're driving it at a high frequency (few 00s kHzs) and into a load.

 

[Sigurthr]

Ps how long can I leave the driver powered up for testing with out causing issues sig?

You can leave the driver powered up for however long you want it to be and it will be fine. It shouldn't be getting remotely warm at all. You'll only start to see the UCCs get a bit warmer if you're driving it at a high frequency (few 00s kHzs) and into a load.

If the GDT isn't connected it can be powered indefinitely. The only issue occurs when there is a current path on the GDT output and the antenna input is left floating and susceptible to noise. The UCC chips self-resonate on the board at around 8MHz, which is well above their Safe Operating Area, causing quick overheating.

Btw, thanks to the dc blocking cap on the GDT primary there's little worry of the GDT core saturating. I've tested these cores from 60Hz up to 30MHz. For the 1-30MHz section it was at 100W of power through the GDT itself too. No signs of saturation.

The only observed mechanism of failure I've seen on this design is thermal (once construction errors are removed from the picture). What happens is a component or connection overheats and severs while a high current is flowing, this causes a huge high voltage spike from inductive kick which then takes out other parts in the area.

 

[loneoceans]

Btw, thanks to the dc blocking cap on the GDT primary there's little worry of the GDT core saturating. I've tested these cores from 60Hz up to 30MHz. For the 1-30MHz section it was at 100W of power through the GDT itself too. No signs of saturation.

Unless your GDT is huge, it will typically saturate around a few tens of kHz, especially one designed for ~200kHz operation. What core is being used and how many turns?

The DC blocking cap prevents 'flux walking' so the core will not saturate in the long run so it's an essential part, but the core flux density will increase throughout each half of the driving cycle, so once your frequency drops (not increases) below a certain point, the core will saturate and no longer act like an inductor. This is rather common failure mode for GDT drives leading to sad transistors and gate drivers.

Looking at DwU photos, it looks like a reasonably sized core with 12-15? turns. A very rough calculation with say a 1x1cm core area and a saturation flux density of 0.2T typically used for ferrite (some of them go up higher though), it seems like the GDT will saturate around 40kHz or lower. Depending on the material, can be a fast or slow roll-off though most ferrites are typically very steep so it's important to stay well below the sat. limit. It looks just fine for DwU's coil though

 

[Down with Umbrella]

Driver 90% scoped.
I scoped The drive as photo'd with my lead inverted and sure enough the output to the GDT measures 4.5 kHz. With a square wave. @ a hair over 12v


Next I hooked up the probes to the two ICs pins 7 and the neg rail WITHOUT inverting one channel and got a similar waveform. (I'll spare you the duplicate photo but they were the same)
Last 10% are you referring to the 14dip Ic as the Schmidt trigger? Are these steps in reference to ground or neg rail? vv
Quote:
"2) Same square wave at pin 1 of the schmidt trigger IC
3) 12V square wave at pin 4 of the schmidt trigger and pin 2 of the UCCs"

Looks like the driver is working fine.
Next maybe its back to the drawing board on the GDT(15T) and bridge. All the parts are new and should work.

 

[loneoceans]

Driver 90% scoped.
I scoped The drive as photo'd with my lead inverted and sure enough the output to the GDT measures 4.5 kHz. With a square wave. @ a hair over 12v


Next I hooked up the probes to the two ICs pins 7 and the neg rail WITHOUT inverting one channel and got a similar waveform. (I'll spare you the duplicate photo but they were the same)
Last 10% are you referring to the 14dip Ic as the Schmidt trigger? Are these steps in reference to ground or neg rail? vv
Quote:
"2) Same square wave at pin 1 of the schmidt trigger IC
3) 12V square wave at pin 4 of the schmidt trigger and pin 2 of the UCCs"

Looks like the driver is working fine.
Next maybe its back to the drawing board on the GDT(15T) and bridge. All the parts are new and should work.

Looks like it's doing fine (i.e. one of the scope measurements is inverted right?). I'd imagine the outputs of the board are basically connected to Pin 6/7 of each UCC with one of them having a cap in series. Is there a schematic for this you can post? I think it would help me a lot instead of just guessing what's going on.

What do you mean about ground / negative rail? They should be connected together when you scope it.

 

[Sigurthr]

As long as one of the channels is inverted in that photo it looks fine. If one channel isn't inverted there we have a problem, as the two points should be out of phase 180deg. I'd still like to see a differential measurement where the channels are added but it isn't necessary.

Here's the schematic, Loneoceans: https://app.box.com/s/7kvqvvs0sx8504g2de4j
The core number is on there too.

 

[Down with Umbrella]

Looks like it's doing fine (i.e. one of the scope measurements is inverted right?). I'd imagine the outputs of the board are basically connected to Pin 6/7 of each UCC with one of them having a cap in series. Is there a schematic for this you can post? I think it would help me a lot instead of just guessing what's going on.

What do you mean about ground / negative rail? They should be connected together when you scope it.

Yes in the photo one one channel was inverted to the other. When I tested pin 6/7 I connected the ground clips to the (-) pin on the board rectifier.





im not exactly sure how to attach full res images to a thumb nail. Hope you can see what you need to here.

I got this wave form too

Maybe a mistaken pin