"Post your fail here!" - thread / Tesla Coils / Forums

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

"Post your fail here!" - thread

first previous 2 3 4 5 next

Move Thread

LAN_403

ubuntupokemoninc

Sun Oct 10 2010, 02:28AM

Banned on 1/22/2011 for repeated rule violations after multiple warnings.
Registered Member #3299 Joined: Sat Oct 09 2010, 08:11PM
Location: Bantown, USA
Posts: 220

my 811a vttc!, don't even get me started!!grr

Steve Conner

Mon Oct 11 2010, 09:03AM

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

Crunchy Frog wrote ...

Incidentally, what IS the right way to ground in the MHz range?

Put the whole setup on top of a big plate of non-ferrous metal. In fact, preferably put it inside a complete Faraday cage. There is an old cartoon in the archives showing a guy with a Class-E Tesla coil being run out of town by radio hams with pitchforks.

This is an old one, but a good one. Bad heatsinking and massive overcurrent were probably to blame. I found pieces of it more than 10 feet away. I still have the asploded IGBT somewhere.

iceowl

Sun Oct 17 2010, 03:12AM

Registered Member #2834 Joined: Thu Apr 29 2010, 02:04PM
Location:
Posts: 22

How many failures can you have in one life?
I suppose it's not true engineering if things don't break. Unfortunately, coils are like mechanical clocks. One thing out of balance, and nothing works.

I just put together a pig power coil. That represents about 2 months worth of weekends, right there. Eight hour days. Building a housing for the pig so it will be safe. Building a control center with the appropriate staging for power up and power down. Kill switches. Not to mention ballast for the pig.

First light was last week. But like most pig projects, you get a couple KVA working for you and it doesn't look much better than your NST coil. So tweak tweak tweak. Folks on TCML advised me to move away from my SRSG and go for an ARSG. I spent most of my weekday evenings machining aluminum and drilling G10. Now I have an ARSG.

Was running 0.09uf tank cap. 6.5" secondary 24" of 26ga. Two toroids. 3x12, and 4.5x18. All powered by a modest 5kVA pig.
Primary is 1.25" wide spring brass wound around 0.25" weather stripping. Inner diameter 12". k=0.12, measured empirically.

Now I can't say I wasn't advised against the primary I used, inevitably. Experienced coilers told me they would never use a ribbon primary again. But I figured, I have a new design. And you know, it worked for me with all the NSTs. As I amped up the power supply month after month, I had no problems. My primary was sandwiched between two polycarbonate disks. No racing sparks were going to get me.

Then today I decided - because with my pole pig the sparks were so unremarkable, that I'd take off the top polycarbonate disk and increase the k by a couple hundredths to see what that would do. And then I kept it off while I adjusted the tuning.

Results. Things got better and better.

I learned that ballast is everything. I burned up a surplus 100A arc welder. Lucky for me, when I bought the welder - I actually bought 2. One was rated at 100A and one at 160A. Tossed the 100A transformer in the trash and went for the 160A.

Things were better.

I remember reading somewhere that adding a small series ballast resistor gave better results than the ballast inductor alone. At the same surplus store they were having a blow out sale on power resistors. No kidding. I got a bunch of them for $20.

There was a 0.15 ohm 300W resistor in the bunch. I added it in series with the 160A arc welder and suddenly it was a whole different coil. Arcs went from 4' to about 8'.

I probably should have stopped there. I'd still have a working coil if I had. But we never do, do we? Enough is never enough. There's gotta be more.

I'd won a maxwell cap on an eBay auction. The thing had one stripped terminal with a broken off screw still in it. I couldn't get the screw out, so I soldered on a piece of copper and used that as the terminal.

Adding that cap brought my total tank cap to 0.12uF.

I adjusted the tuning to account for the additional cap. Fired up the ARSG, running about 2000 rpm with 8 electrodes and a propeller gap. Turned the knob on the arc welder. Wide open. Max current.

Twisted the variac. The sparks were incredible. Bigger than I'd ever seen.

But there was one that didn't come from the toroid.

The last picture, is the last picture.

Fabio

Mon Oct 18 2010, 10:38PM

Registered Member #122 Joined: Fri Feb 10 2006, 12:55PM
Location: Milano Italy
Posts: 148

Hi All!

This is the history of an EPIC FAILURE!

I was in a friend's home to debug on his first Tesla Coil; the problem wasn't hard to find, he use a poor quality capacitors in his MMC so he had a great loss and many capacitors failures.

I connected his oscilloscope and his frequencymeter in parallel of the primary coil, i also used his signal generator to excite the LC circuit for evaluate the quality factor of the resonant circuit; as expected the oscillation was too damped.

i temporarly connected one of my capacitors instead of his poor quality capacitor and i repeat the measures, the results was great as expected.

Some more time to adjust the tuning and everything was ready for the first light!

My friend - "OK i turn it ON!"
Me - "NOOooo....!"
BZZZZZZZZZZZzzzzzzzzzzzzzzzzzzzzzzzzztttttttttttt ttttttttttttt.............

moral of the story:

- is a really bad idea to leave the instruments connected on the primary coil when the coil is powered up!

- Is also a bad idea to remove the jumper on the spark gap before removing all instruments!

- modern instruments are too delicate for us, they cannot withstand even 11kV peak to peak with few hundreds of amperes peak!

- Waveforms are best if viewed on oscilloscope's screen instead of on oscilloscope's smoke (no mention of course if smoke is provided from the signal generator or the frequencymeter!)

- Now my friend have a new oscilloscope, a new signal generator and a new frequencymeter, imagine why!

Ciao!
Fabio.

iceowl

Tue Mar 01 2011, 06:17AM

Registered Member #2834 Joined: Thu Apr 29 2010, 02:04PM
Location:
Posts: 22

Well, several months and several secondaries later...we have this.
The specs are: f0=63khz.
12.25" secondary, 56" tall (60" total height) - wound with 18ga copper
12"x48" hoop topload
9 turns ribbon primary
ARSG 300bps
running ~6kva from a pig
And now the key - 0.09uf 50kv rated Maxwell cap model #31173
I didn't think these things were supposed to detonate, but they do. I think these red ones aren't supposed to experience more than 1 pps.

The coil was running swimmingly. Then - blamo. A pop and brown smoke. It took me about 3 seconds to hit the panic button during which time the coil continued to run beautifully, even though the capacitor had spilled its guts all over.

Joe

dude_500

Fri Mar 11 2011, 03:20PM

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

While I was running my CM400 coil, the high frequency isolated gate drive power supply dropped out causing all the gates to run right below saturation. Interestingly, the coil actually continued to work (although the arc length went from 5 feet down to barely visible, but could still be heard). I later took the bad brick apart and found this interesting destruction. Was running about 200VDC when this occured, with around 15,000uF capacitance on the filtering.

I now use 60hz gate drive supplies.

Marko

Sat Mar 12 2011, 02:39AM

Registered Member #89 Joined: Thu Feb 09 2006, 02:40PM
Location: Zadar, Croatia
Posts: 3145

Hi guys, thanks for interesting posts.

Owl: I feel most sorry for that maxwell cap venting it's guts on you. General consensus today, though, seems to be that MMC's beat even the toughest pulse caps in price vs goodness.

Dude: That is a worrying kind of failure, even if GDT's are used. In my new controller design I have a relay that goes off as soon as the supply voltage drops low, but I'm not sure that it will make an effective protection in case controller supply is lost. I think that including a faster UVLO circuit would be a really good idea here.
For isolated drivers, I'd definitely include an UVLO circuit on every driver board.

Marko

Steve Conner

Sat Mar 12 2011, 09:13AM

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

Yes, I agree. Richie B. persuaded me to add an independent UVLO to the Odin gate drive boards I designed. Each board ended up with zeners and optocouplers to report the status of the rails back to the main board, which would lock out the drive unless all the rails were present.

I've since found a source of tiny and cheap 50/60Hz transformers, 0.9VA!

Really even ordinary drivers should have a UVLO.

PSCG

Fri Jun 17 2011, 11:06AM

Registered Member #3792 Joined: Sun Mar 27 2011, 06:07PM
Location:
Posts: 136

I was trying a homebrew oven-controlled crystal oscillator for my HP5328A (purchased from Ebay for 100$). I installed a 7815 voltage regulator to power the oven, but i didn't notice that the instrument was still powered from the mains. With a little "help" from Murphy's law, a cable that supplied +25v from the PSU touched circuits that where not meant to be powered from that kind of voltage. The result: an exellent frequency counter was ruined and a worth of 100$ was lost. I'm still trying to fix it (without too much success) and i don't think that i will repair it soon!

Nah

Fri Jun 17 2011, 05:36PM

Registered Member #3567 Joined: Mon Jan 03 2011, 10:49PM
Location: USA, 1960s
Posts: 260

I connected a dvm directly to a MOT..........

Battery was still good.

I vented by throwing some compactrons at the wall.

first previous 2 3 4 5 next

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