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

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

"Post your fail here!" - thread

1 2 3 4 next last

Move Thread

LAN_403

Marko

Fri Sept 17 2010, 05:24PM

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

Hello to all ladies and gentlemen here.

I'm sure that one of things that bug us most in our Tesla coiling lives, are, of course, failures. We are all frustrated when our creations arc over, catch fire, blow mosfets at random, suffer ''reliability problems'' or just die in some unimaginable way.
But one other thing that bugs me much more is how easy it is to make your projects, posted on the internet, appear reliable simply by not telling how many times they blew up.
We like to document our successes, but I think we're way too ashamed of failures - even if they're not our completely own design.
But I also know how frustrating it is when we find out that some our unique creation, into which we have put lots of money and effort will simply never work because we overlooked something that can't be fixed anymore. Or, we just keep blowing mosfets in Steve Ward's Mini SSTC and don't know why!

Hence I attempt to start a hopefully revolutionary thread - where instead of posting all our glamorous project success, we can post about things that didn't work - it might sound like a really, really depressing thread at first, but it's purpose is to put an end to those failures. Adding some humorous comment along the posts is certain to make the thread less depressing, though!
I decided to limit the thread just to Tesla coils for now, in hope to prevent it from getting too chaotic - perhaps if it proves a success it could be expanded to other topics later.

Also, in order to avoid this thread being just another "post pics of your blown components here", I made up a few general rules of what posts should look like.
Basically every post it a small story, preferably accompanied by photographs, with following points:

- Brief description of the project (or a link to project thread if it exists)
- Description of the failure
- Possible explanation of cause of failure, if known.
- Fix of the failure, if known, and do's/don't's following from it

I'll start by a few examples of mine. I won't be posting on this forum for a while after, and I hope to find a full load of posts in this thread when I come back! Any TC projects, from tiny single mosfet ones to largest DRSSTC's are welcome.

1. *** Homemade heatsinks fail

Description:
Back in 2007, when working on this project,

, I didn't have too much understanding on thermal management of power devices, too much time and little money on my hands. I couldn't find a source of heatsinks for my double-H-bridge, and ventured into building my own. I found some 1.5mm aluminum, and cut it up into wanted shapes. The end product took way too much time to construct, looked pretty, and was absolutely terrible at dissipating any heat. I had IRFP450 mosfets coupled to the plates by sil pads, two on each plate.

Explanation:

-The plates were so thin they hardly conducted heat away.
-They were horizontal and close together, preventing natural convection almost completely.
-Sil pads added lots more of thermal resistance.
-Sil pads were easy to displace and damage, causing mosfets to short to the heatsink.
-I tightened everything using 3mm screws through the mosfet holes, causing their 'behinds' actually being lifted from the heatsink.
-The setup was an excruciating pain to disassemble and reassemble to replace mosfets when they blow.

Fix: The heatsink was thrown away.

Really, if you want to keep your H bridges, buy four heatsinks as big as possible, and mount the devices straight to them. Or, use a single heatsink, with isolated devices, or devices mounted on a large heat spreader with insulation underneath.

I would classify this failure as a Epic Heatsink Fail, RTFDS Fail and Listening to Richie Burnett Fail.

2. *** Wooden primary supports and parts

In Big Bad the SSTC,

I used a ring of thin plywood to secure the top of my primary form (which is solid polypropylene studs) and make it look pretty.

While running at not all that high power levels, I noticed the zip ties holding the topmost turn of my primary mysteriously burning up. Another day I experimented more to find out what causes it, only to withess a massive flashover emanating from the wooden ring and hitting my secondary. The secondary was licked by plasma but it survived - mostly thanking to PLL (Pure Lady Luck).

Explanation: Wood conducts high voltage. I thought I might get away with a wooden ring isolated from ground, but I didn't.

Fix: Never, ever use wood to support your primary coil. I've seen many people doing it, and I still don't understand it. You may get away with it for 10 minutes, a hour, until you don't... wood always contains some moisture, and it a conductor for the hv. At first the current may be small, but still enough to cause significant heating in the wood. The zip ties on my primary were cut by partial discharge between the sharp edge of the plywood ring and my primary coil!
At one point a hotspot will appear where wood will start to carbonize, and seconds later you will get a fire and possibly a ruined coil.

This could be classified as a HV and Water Mixing Fail and Secondary Fail Near Miss.

3. *** Primary connection cables in a high power CW SSTC

Another one from the same project... after running the coil for a several minutes, a burning smell and smoke appeared. I realized wat is happening and shut down in time - the audio cable with 3mm+ thick core simply melted it's insulation under extreme RF current. It turns out that audio cables have rather thick insulation with low melting point, which impedes dissipation of heat.

Cause: Unexpectedly high primary current, miscalculated cable thickness.

Fix:

It would probably be more advisable to use litz wire or flexible silicone insulated cable for the tapping point cable, since this one is rather rigid.
If you do not have a tapping point, and you run same power level like I do, maybe it would be advisable to run the copper pipe all the way to the bridge.

I classify this failure as a Cable Overcurrent Fail, but also as Epic High Current Destruction Win.

Now, let's see what do you have for me! I'd be especially glad to see all MOSFET and IGBT failures you can gather for me!

Cheers,

Marko

803

Fri Sept 17 2010, 07:45PM

Registered Member #2807 Joined: Fri Apr 16 2010, 08:10PM
Location:
Posts: 191

I have 2 fails that I had in the last month or so

1. when cleaning my NEW 803 tube I broke it.
2. when I was working on my tesla coil, I used 2 pvcs connected with tape on my primary. long story short, it had sparks burn up the secondaryt coil

lesson, use a whole piece of pvc
3. I was making a e cap bank. I connected the ps wrong and BANG in my bedroom

Patric

Fri Sept 17 2010, 09:17PM

Registered Member #2899 Joined: Wed Jun 02 2010, 06:31PM
Location: Deinze, Belgium
Posts: 255

I thought this capacitor with plexiglas was indestructible! How wrong was I!

iceowl

Fri Sept 17 2010, 09:34PM

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

I had been getting all sorts of great ideas from some of the tesla coil websites (and I still do).
One idea I had not yet tried was the high speed asynchronous rotary gap using this motor: AMETEK AC/DC 19,500 RPM, and a propeller gap arrangement with a UHMW mandrel and a 7" tungsten welding rod as the rotating conductor.

I would post pictures of this disaster, but the destruction was rather complete. In addition, my state of mind at the time of the failure was one of survival and safety, keeping my home from burning to the foundation, and not so much that of photography.

Here's what happened.

The motor comes with a mounting bracket on 1 side only. It's some sort of vacuum cleaner motor, meant to be used in a cleaning function, so it has a place it wants to be. An ARSG is not where it naturally feels like going, so I coerced it by opening the body of the motor and drilling holes in the "reverse" side, and mounting 2 8-32 bolts. Thus there was now a mounting bracket on one side and bolts coming from the obverse side of the motor housing. I then used these with some "L" brackets to attach the motor to a G10 plate. I had put 8 tungsten electrodes in the plate, each on its own insulator, and then wired them together, 4x4, on the back side with some conductive strips. My setup was very similar to the one on the website where I got the idea.

I should have read the website details a little more thoroughly, though, because the owner of that site told of a spectacular failure with that motor and had consequently taken to surrounding the spark gap with bullet proof polycarbonate shielding whenever it was in use.

Of course, the little coil designer in my mind assured me my design would be in some ways superior to his, and so I would need no bullet proof shielding. This little coil designer voice needs to be stomped.

On my first run, I noticed with some great degree of fear that the motor spun with a great deal of vibration and fanfare. Vacuum cleaners are not noted for their silent, stealthy sucking behavior. Motors like these are part of that characteristic. So when my propeller gap spun it did so with all the aplomb of a spastic hoover, encircling itself in a shower of white hot sparks.

Well, at some level this was quite rewarding to see. Controlled electric violence is, in a way, what coiling is all about. Though the spark output from my coil with this ARSG left a lot to be desired, the ARSG itself was putting on a show for the ages.

And it was right before my second run, after I made my tweaks, that the perpetually silent voice of reason spoke up within me. I saw images in my mind of the other coiler's ARSG encased in steel and lexan, and the onlookers staring in mute horror while they awaited some awful clash of electrons and metal to precipitate.

I happened to have beside me a 4'x6' sheet of 3/4" plywood, and it is only because at the last minute I decided to prop that up against my coil power supply between me and the ARSG that I am here today to type these words, unmamed.

For when I alternately twisted the knobs on the variacs - one to control the ARSG speed and the other for primary voltage. And when the sparkage failed to grow to lengths I desired, I maxed the voltage on the primary - and it was when I was spinning up the ARSG to speeds that sounded, behind the plywood wall, like the life was being rung out of several cats by random predatory killer voles - that there was a great flash of white hot light, a sound of metal shattering, and then sudden silence.

One or two seconds later, the plywood sheet I had propped up against the power supply case fell over backward, impaled by a couple small shards of tungsten rod.

As my vision was thankfully blocked by the plywood I did not see what happened. But I surmise my mounting system for the ARSG motor was inferior, and as the motor spun up it went through several vibrational modes, one of which cause the motor to shake free of its mooring. With the gaps adjusted very tightly, it didn't take much movements to cause the rotating tungsten electrode to smash into one of the stationary electrodes and splinter.

The force of the crash sent the tungsten flying. The motor broke free of the bolts I had inserted, and now lacking any sort of constructive support, they slid into the rotating armature and caused it to stop suddenly, bending the bolts and forever destroying the motor. Thank ye gods.

The flashes of white hot light, I surmise, were my safety gaps firing, protecting my NSTs and caps. For none of them were harmed in the crash.

There was, however, some smoke, and so I dashed for the fire extinguisher mounted on the wall of the garage, but by the time I figured out how to yank out the plastic certification-of-charge tag, the smoke had dissipated.

The electrodes, their mounts, and the G10 plate were all salvageable and went into my next ARSG attempt, which is still working swimmingly.

I have also blown up 8 IGBTs building a MINI SSTC (not Steve's), but those were rather unclimactic, though significantly more expensive than the ARSG disaster.

Cheers,
Joe

Blackcurrant

Sat Sept 18 2010, 01:26AM

Registered Member #2989 Joined: Sun Jul 11 2010, 12:01AM
Location: UK
Posts: 94

This was part of my DRSSTC H-bridge until it melted
IGBTs 4 of them cross conducted and shorted out a 6.8mF 400vdc cap
I think this was due to the tuning being too far out
i tried fixing it with copper braid but it looked like a dogs dinner

Marko

Sat Sept 18 2010, 01:59AM

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

Hey guys, thanks for the quick replies!

803 - could you tell a bit more than a few sentences, though? The idea of the thread is to give others a complete picture of what happened. (what are e caps after all?) I'm sorry for your tube - not sure if it counts as an electrical failure, but I'm sure many people would feel awful about it.

iceowl - your post was an interesting story, thanks for it. May I ask though, why did you use a high speed motor along with 8 electrodes? That looks like way too high BPS. Were you using DC resonant charging? Running AC coils with asynchronous gaps is generally unadvised.
Can you tell more about the blown igbt's though?

Blackcurrant: well, yeah, mistuning too far can cause the primary current to go off the roof in a DRSSTC, if you don't have an OCD circuit to guard against it.
Still, I know there are countless people here who have blown countless silicon and never found out why - don't be ashamed to post even if you have no clue why it happened. But make sure to post at least few main details of the project and circumstances which preceded the failure - more the better! In short, not a post like "I've blown 10igbt's last week and meh!"

Cheers,

Marko

iceowl

Sat Sept 18 2010, 04:21AM

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

DC resonant charging...I could only hope, but no. It was foolishness. It was stubbornness.
ASRGs not advised for NSTs...yes. Badness all around.

But basically, I was trying to see what would happen. I'm one of these experimental engineering types.

Don't get me wrong. I'm a firm believer in simulation. I'm an apostle of JavaTC - the best thing that could happen to tyro coilers. It's saved my bacon more than once. But every now and then, which is almost always, I want to see what would happen if I make this tweak or that tweak. I had a perfectly serviceable SRSG, but once I figured out how to make that work, I wanted to hear the sparks whine. So after I watched the videos on the other websites, I did the ARSG.

And the sparks did roar. Whine they did, till Blammo.

Note - just recently I build another ARSG based on a Dayton universal motor - a slower motor, only 5000RPM tops. And I've been running my 15/240 coil with that, just to hear the sparks whine. And whine they do. Though it seems I've blown at least one of my NSTs... Safety gaps set too wide, and terry filter accidentally disconnected.

We get what we pay for in life, and coiling.

On the blown IGBTs - I built Dan's microBrute. Worked perfectly the moment I switched it on. Of course, that wasn't good enough for me so I had to start playing with the current limit and the tuning. The 1st set of IGBTs lasted about 10 minutes under my torture.

Unfortunately, after that I tweaked the circuit & the resonator into a state where it was difficult to get it functional again. My ground connection was faulty - and apparently a bad ground on your secondary is tantamount to mortal sin in the bible of DRSSTCs, as I have come to learn first hand. So the second set of igbts went about 5 minutes after the 1st.

Now, anticipating that I would destroy my igbts I had bought 3 sets. And it wasn't long before through overvoltage, over and undercoupling experiments, tuning way past resonance, trying different breakout point combinations, that all 3 sets, plus the original set that came with the kit were gone.

Having learned my lesson I bought a final set, popped them in, and now the system works just fine. Of course, the in demoing the coil to some of my daughters friends I was holding a fluorescent bulb to do the classic - bulb-lighting-at-a-distance experiment. The little coil will keep a 4' bulb lit at about 2' distance. I turned to the kids to see how they were reacting - when suddenly I felt a spot on my leg catch fire.

It seemed that even though I was holding the tube by the glass part, I was able to conduct an arc through the light (or over it) then through or over my body, to my calf muscle, which happened to be located within 1/4" of a metal garbage container, to which my calf then arced.

Needless to say I was "shocked" (pun intended) and I immediately pointed out to the youngsters how much it hurt (it did) and that it was through lack of attention that this sort of accident happened - and that this is why we always say, "clear" when we start the coil - to make sure we're away from living creatures we don't care to zap, especially me.

Regards,
Joe

803

Sun Sept 19 2010, 03:32AM

Registered Member #2807 Joined: Fri Apr 16 2010, 08:10PM
Location:
Posts: 191

Marko wrote ...

Hey guys, thanks for the quick replies!

803 - could you tell a bit more than a few sentences, though? The idea of the thread is to give others a complete picture of what happened. (what are e caps after all?) I'm sorry for your tube - not sure if it counts as an electrical failure, but I'm sure many people would feel awful about it.

Cheers,

Marko

OK Marko,

Well I was too cheap to add a tube cap so I soldered the wire to the cap. When I got the cap, I tried to unsolder the cap. Long story short, glass doesn't like the heat from a 100watt soldering gun

ByBy 1941 Nos tube. Second, I was making the final adjustments to my vttc and I ran out of PVC. I found that I had 2 smaller pvcs that had the hight I needed. The sparks arces to the primary winding through the crack and burned up both coils. I'm too cheap, eh?

Steve Ward

Fri Sept 24 2010, 05:03PM

Registered Member #146 Joined: Sun Feb 12 2006, 04:21AM
Location: Austin Tx
Posts: 1055

Here's one related to DRSSTCs:

On a clone of my DRSSTC-2 system a ringy gate drive circuit caused a CM300 failure, but not in the normal sense of "my waveforms look ringy!". The issue was that at the end of the drive cycle, the GDT was to shut off and be at 0 output, but instead the GDT would ring at around a few volts. This was enough to turn the IGBTs back on momentarily when the currents were still quite high. I suspect hard switching of 1200A or so to be the reason for failure. The IGBT failed just seconds after ramping up the power to maximum.

The fix: adding a parallel 50 ohm resistor with the DC blocking cap on the GDT to damp out this ringing (which is caused by GDT leakage inductance and blocking capacitor). With a clean turn off gate signal at the end, that system went on to run for years without ever failing another IGBT, eventually that whole drive was scrapped/recycled for an upgraded set of electronics, though the same CM300 modules were used and continue to work.

TheBoozer

Fri Sept 24 2010, 09:45PM

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

A picture is worth a thousand words :

Please bear in mind that this was back when I had no idea what I was doing.

Rich

1 2 3 4 next last

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