Steve Ward's "Micro" SSTC-6 troubleshooting. / Tesla Coils / Forums

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

Steve Ward's "Micro" SSTC-6 troubleshooting.

1 2 next

Move Thread

LAN_403

Sigurthr

Wed May 09 2012, 01:08AM

Registered Member #4463 Joined: Wed Apr 18 2012, 08:08AM
Location: MI's Upper Peninsula
Posts: 597

Hello everyone,
I built Mr. Ward's Micro SSTC as my first SSTC last night and immediately I noticed some major problems. I followed his schematic to a "T" with the exact parts specified. I am waiting on the 25VAC 2A transformer so I am running the coil from my 24VDC 5A SMPS.

Steve's Page on this coil:

1st problem: Overheating of the Fet.
This coil uses a single IRFP460 Mosfet. I cooked the first of three fets today even though I had it thermal epoxied to a decent size heatsink. It overheated in about seven seconds and let the magic smoke out. I replaced the fet and used a much larger heatsink. I also limited input current to 2.5A with some wirewound resistors. Still, the heatsink gets dangerously hot from even very short run times even with the limited current. I decided to see how much current it wants to draw and clamped the heatsink down to a 1/8" thick square foot of copper plate. It runs much cooler and can be operated continuously at 2.5A now. Without the limiting resistors the coil draws just over 8 amps and the Fet gets so hot I can't touch it after about 5 seconds. This is with it being clamped to the huge piece of copper!

I am thinking of increasing the primary's turns so the impedance is higher and thus less current should be drawn. I think this may reduce output voltage though, but I'm not sure. Do you think this would help?

Steve's coil uses a measly 2A stepdown transformer, if I try to pull >8A out of that it will saturate and suffer serious voltage sag as well as possibly overheat. How did he get away with it?

Second Issue: No corona/streamer from breakout point
While I had the first fet on and full (>8A) current draw I saw a small 1/8" corona/streamer on the breakout point for all of the six seconds it lasted before the fet died. With the new fet in place I don't get any visible indication of output either with or without current limitation. The electric field from the coil is decently strong though as it can light a CFL to full brightness from about an inch away from the center of the secondary. So, I know I am getting oscillation just fine and there is HV output, just not enough for any breakout. I'm only about 10Vdc less than specified, I doubt this is enough to cause the issue. With the current limitation on there is no voltage sag from the SMPS, but without the current limitation the voltage does sag as it is only designed for 5A draw.

Third (much more minor) issue: too much reduction of 555's signal to start initial oscillation.
I have to short/bypass the 22pF DC blocking cap in the output string of the 555 subcircuit to get initial oscillation to occur. After this it continues to oscillate just fine though, even if you short the HV terminal to ground, the coil continues to oscillate just fine.

----------

Could problems 1) and 2) be caused by the Fet not turning on fully? I've never really worked with Fets before, only BJTs, but if the fet is still in a linear region it would make sense that it would run MUCH hotter than normal AND output would be less than ideal. I don't have a scope unfortunately, but it is on the list of things to get as soon as possible. I seem to remember reading something a while back somewhere that mosfets require the Gate to be at a higher voltage than the Drain to be fully on, is this correct? If so, this may certainly be the case. The voltage at pin 6/7 of the TC4420 is only 5V. Raising the input power from 24V to 36V would not affect this either as the TC4420 is powered by the 12V rail from the LM7812. The output voltage from the 555 is 11.8V but R5 and R6 form a voltage divider between that and the TC4420, not to mention that C5 blocks the DC output of the 555 anyway. If I were to bypass the troublesome C5 and then remove R1 the input of TC4420 would see close to the full output of the 555, but even then I don't know if this would increase the output voltage from it. I've never worked with gate drive chips and I don't know what determines their output voltage level.

Any and all help is much appreciated. I will settle for correcting Problem 1 (thermal), but if I could get some breakout as well, it would be fantastic.

Dr. Brownout

Wed May 09 2012, 02:14AM

Registered Member #2405 Joined: Fri Oct 02 2009, 12:59AM
Location: Wisconsin
Posts: 140

Sigurthr,
The Fet is going to get warm I built one similar to this and the FET does get really hot I am using different FETS in mine too I am using IRFP260 which is rated a little less that yours and I use two of them. You may want to play around with the coupling. I noticed with mine that I had to raise the the primary higher up the secondary before I could get any good break out. I also added some primary turns and that helped a lot. My secondary is 14" so it is twice as long as the one you have. Same diameter though.

Mine will now break out with and without a topload. However without a top load the wire gets flying around pretty good and with a top load I placed a small breakout point on the topload to aid in the breakout. My top load is 1.5" by 5"aluminum toroid.

I actually added audio to it and I let it run full steam without a fan to see how long it would go before the FETS blew out and it takes about 45 seconds then there are FET guts everywhere. However on low power it runs really good and the heating problem is actually tolerable and the breakout is still about 4" or so.

I attached a few pictures of it you can see the amount of turns on my primary and how far up the secondary it is. I still think if I could find a former to get it closer to the secondary it would have better performance. But this uses the same kind of feed back your using I am using a different interrupter but the concept is the same.

So play around with the coupling and see if that helps.

--Brian

1336529658 2405 FT138115 Audio Loop Wire

Sigurthr

Wed May 09 2012, 02:38AM

Registered Member #4463 Joined: Wed Apr 18 2012, 08:08AM
Location: MI's Upper Peninsula
Posts: 597

Wow yours looks nice!

I'll try and adjust the primary and coupling and see if that helps. Your IRFP260s have a much lower on resistance than the IRFP460 I specified though so they should actually run cooler. Their RdsON is only 0.055 ohms, where as the IRFP460s are 250mOhms or 0.25ohms.

I rigged up a big blower fan to blow on the heatsink and copper plate to see what the temp would climb to with enough cooling and I stopped the test after about 3 minutes when it reached 80C (176F), which is about half of the maximum rating for the fet. Without the limiting resistors current draw soars to 9amps and the voltage sags under load to only 14V. R = V/I so does that mean the fet's on resistance is 1.55ohms?! That would suggest it is in the linear region and not fully on, which would incurr huge losses and lots of heat.

Dr. Brownout

Wed May 09 2012, 02:54AM

Registered Member #2405 Joined: Fri Oct 02 2009, 12:59AM
Location: Wisconsin
Posts: 140

Yes I would play around with the coupling and see if that helps to get you some breakout. I was using a simlar ciruit topolgy for a flyback driver as what your using now, and I had some big heat issues with the Transistor I was using, I changed the transistor over to a IRFP260 fet and that seemes to help. I am also using a GDT to drive the FETs in both the flyback driver and the small tesla coil in the picture but like I say the FETs do get hot uness I run them on reduced power. Actually one get hotter than the other.

Run it at reduced voltage when you play around with the coupling until you get a decent spark and once you feel you found the sweet spot then you can run the voltage back up. But I am sure you will need the fan like you say. And close coupling is what these smaller coils seem to like.

--Brian

Dr. Brownout

Wed May 09 2012, 03:11AM

Registered Member #2405 Joined: Fri Oct 02 2009, 12:59AM
Location: Wisconsin
Posts: 140

Here is a picture of the driver circuit I am using for the small coil in the pitcure. This is the way it was before i added the Audio to it. There was no external interrupter. Its been changed over now so it uses an external interrupter and fiber optic cable. I run my small DVD player into it and it sounds pretty good. I dont have the external interrupter enclosed yet.

I kind of went over board on the interrupter I built it so it will handle one or two coils so it has 2 fiber optic outputs and I added LEDS for each the right and left audio channels using a dot bar generator...a total of 20 LEDs but it looks pretty good. I only bread boarded the dot bar generator to make sure it was going to ok.

I did do a PC board layout for it but I have yet to order the PCB's.

--Brian

1336533111 2405 FT138115 Small Coil Circuit

Sigurthr

Wed May 09 2012, 04:09AM

Registered Member #4463 Joined: Wed Apr 18 2012, 08:08AM
Location: MI's Upper Peninsula
Posts: 597

I increased the primary to 10 turns and raised it up about to half way up the secondary and I was able to achieve 1/2" breakout and the coil was only pulling a little under 3Amps.

Unfortunately the FET hit 90C with the massive heatsinking and blower cooling it and it died. This is the second one now! I'm thinking that I'll just find a new fet to use instead of these IRFP460s, something with way lower RdsON like the IRFP260s.

Just sucks knowing I just blew $10 in fets, lol. I still want to know how Steve got it to work at full power with only a small heatsink....

Dr. Brownout

Wed May 09 2012, 04:19AM

Registered Member #2405 Joined: Fri Oct 02 2009, 12:59AM
Location: Wisconsin
Posts: 140

What is the distance between your primary and secondary? I believe I am running about 7 turns on mine. I did try 10 and it was too much. It takes a lot of experimenting to get everything going good. At least you got the current down some. Mine draws about 3 Amps as well.

Were you able to run it with reduced power? What bus voltage are you running?

Also you should have a protection diode across that FET you may be blowing those FETS because they are getting spiked. 90C should not have taken that FET out.

Dr. Brownout

Wed May 09 2012, 04:42AM

Registered Member #2405 Joined: Fri Oct 02 2009, 12:59AM
Location: Wisconsin
Posts: 140

Another thing to check is your interrupter when the spark does break out it should look crisp, if it looks fuzzy like it does in one of the pictures I posted then its running very close to CW mode and that makes those FETS hot REALLY quick. You want to be sure the interrupter is set correct.

You can hear and see the difference when you hit the sweet spot with the interrupter because the spart looks and sounds different. Mine was picking so much noise it was running what seemed like CW mode all the time, so I made the external interrupter and added the fiber optic link and that made a pretty noticeable diference.

EDIT: I just took at the schematic and i see that one does not use and interrupter. Try adding the protection diode across the FET. Its still going to run hot but at least you shouldnt spike it dead. You could also build up an interrupter and then play with it and see if that helps ot out. Without being able to scope the signal its kind of a hit and miss.

Sigurthr

Wed May 09 2012, 05:40AM

Registered Member #4463 Joined: Wed Apr 18 2012, 08:08AM
Location: MI's Upper Peninsula
Posts: 597

The secondary is on a 3.25" coilform and the primary is on a 4.5" coilform, so there is 5/8" between the coils on each side.

I've been powering it from a 24V SMPS, so the fet is fed directly on 24V. The 555 and gate drive chip is fed from 12V. I ordered a 2A 25Vac transformer for it like it says in the schematic but it won't be shipped for another week as Mouser did not have one in stock.

I've been wondering if it being in CW mode had something to do with it. I'm new to SSTCs and haven't really looked in to an interrupter yet.

Will add a protection diode and give it a try again.

I've heard from yet another person now who subbed the IRFP460s for IRFP260s and had success, so I'll be ordering some of those in as well.

Update: I'm all out of protection diodes so I put the only thing I had left on it; a 1N4003!

To my huge surprise the fet runs cool continuously and power consumption has dropped 600%! The coil only draws 600mA now without any current limiting resistors. I now get two 1/8" streamers.

Can someone please explain to me why adding a protection diode (in this case a very slow one) would reduce power consumption?

Also, I placed an order for 4x IRFP260s, 4x TO-247 beefy heatsinks, and 6x MUR860s so I'll be well stocked if/when this fet dies.

Video: (before I removed the unecessary current limiting resistors, so only one streamer, and 500mA current draw)

UPDATE: 5/10/12
I rewound the primary directly on the secondary. The lower inductance of the smaller diameter primary increased current draw dramatically, but still the fet runs happy, but hotter. To compensate I increased the primary to 10.5 turns, which seems a happy middle ground between output and current draw (fet heat production). I haven't experimented too much with the coupling height, but I have the primary starting at the base of the secondary's winding now. It may get better output higher up, or it may not, I'm not sure. Output is stable now and the fet is happy, so I'm afraid to mess with it much. The driver now draws between 2 and 4 amps, averaging 3A, depending on capacitive coupling to ground (loading). I was able to use a smaller heatsink (than the 1sqft copper plate!) and a small 12V fan powered by the 12V rail and lm7812 to help keep the fet's heatsink cool. The coil now happily runs continuously for hours without overheating. The 7812 gets hot, but not dangerously so, and they can take quite a beating after all. There is only maybe 300mA draw on the 12V rail, but it is being fed by 24V so there is a lot of heat from that. The antenna is also less sensitive to placement now, which is great. I added a pushbutton switch that shorts the bypass capacitor / DC blocking cap between the 555 and the TC4420, which allows me to start oscillation whenever it stops (shorting output to ground or shielding antenna from secondary's E field).

Output is typically two 5mm streamers when no ground is placed near the breakout point. If you bring a ground or capacitive load (such as an ungrounded person's hand) within a few inches of the breakout point the streamers merger and grow to 15cm. From about 3cm away from the streamer you can feel the heat from the RF on your skin, it doesn't burn, just feels warm. Having a ground near the breakout causes the driver to draw around 4amps, so it should not be done for long as it increases heat generated by the fet.

There are certainly some design/schematic changes I would make to this SSTC driver, but I don't feel qualified to judge Steve Ward's work. I may draw up a schematic detailing my changes in the future though. Mainly I'd add a high speed protection diode across the fet and change the fet to an IRFP260. I'd also like to do something about the need to pushbutton start the oscillation, so perhaps the attenuation string between the 555 and the gate chip needs to be reworked.

I'll update (or reply if someone else comments between now and then) the thread with pictures and video tomorrow. I'm satisfied with how it turned out, even if I am only getting 1/3 of Steve's output. Up next will be making his "Mini" SSTC... I'm itching to get one of those going!

UPDATE:
As promised, here is the video!

Sigurthr

Fri May 11 2012, 10:20AM

Registered Member #4463 Joined: Wed Apr 18 2012, 08:08AM
Location: MI's Upper Peninsula
Posts: 597

Forgive the double post; too much text in one post for me to edit on my tablet.

UPDATE: I moved the coil to a more stable wood table and proceeded to try more turns on the primary. I fired her up at 11 turns and she promptly shut off. I check the RdsOFF resistance and it reads 0.1ohms, Rgd is 5ohms. I know what this means =/ another dead fet. I removed it from the heatsink and circuit and noticed something odd this time...

Is it just me or does that NOT look like electrical damage? Did the pad on the back actually start to melt!?!?! I used Ceramique2 thermal compound and this was on a fan cooled heatsink!

IRFP460s are forever dead to me now.

I should have my IRFP260s, MUR860s, and TO-247 finned heatsinks in the mail tomorrow or early next week. I was kind of planning on making a full H bridge to run off rectified mains with them though. Maybe I will, maybe it is time to scrap this driver. 3 fets dead and it's a strike, right? There's only about $15 worth of parts on the driver board, and everything except the 555 and TC4420 is reusable (and I have spares of those).

I'm thinking about building a full H bridge version of Steve Ward's Mini SSTC instead, and using my tesla resonator for that. I'll open up a new thread for that though.

Thanks for all the help and suggestions to all who have contributed. I can't say I recommend this circuit to anyone, maybe with 260s instead of 460s and the freewheeling diodes on there it works fine, but this as-is version has been a nightmare.

1 2 next

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