High Voltage Forum
Tesla coils => Solid State Tesla Coils (SSTC) => Topic started by: costas_p on August 30, 2020, 05:00:49 PM
-
Hello all,
I have built a half bridge sstc, with 555 as an interrupter, antenna as a feedback, and the driving is done with the ucc3732x chips, my gdt core is n30 material.
When I apply voltage through a variac , around 60-70 Vac nothing happens, until i close the switch, where there is a 10cm spark momentarily on the topload, and the same length of spark on the antenna.
When i change the leads of the primary, at around 50 Vac it works as a CW sstc and blows both mosfets and and the gate resistors after 2-3 seconds (the visible spark output is 1-2cm approximately)
Unfortunately i do not have an oscilloscope.
Any ideas about what is wrong?
I can post schematic and more photos if someone want to help.
Thanks in advance :)
-
It might be possible to get through debug without a scope, but more information is needed to be of any help. Most important, please post schematics of your circuits. Then reference specific elements, rather than just "the switch", which could be anything.
Look through other threads and links to tutorials here on SSTC debug. If that doesn't help, ground the antenna and walk through your circuit with a multimeter recording voltages. List voltages for nodes within the circuit and I'll look through your results. Either annotate another schematic copy with voltages, or list voltages with specific node labels such as U2-1 for pin 1 of IC U2.
-
Hello davekni, thank for your reply.
Grounding the antenna meaning on negative rail, or earth ground??
By switch I mean the mains switch cutting the power to the bridge.
Also before putting the chips into the sockets, I did a voltage test and found all 12v connection good, with the lm7805 supplying only 4,6v , althought the 74hc14 can operate on this voltage (referring to the datasheet)
So for the test, how can I measure the voltage of a signal (eg 555 output) with a multimeter ?
I will post voltages and schematic with grounded antenna later today
-
This is the pcb layout,
I do not have the schematic in digital form, if you prefer that, i can draw it by hand and upload it.
-
Yes, a schematic would be helpful. Thank you for the layout - that will help too. I could trace the layout, but if you upload a schematic, then everyone will be using the same reference for discussion. Perhaps it makes sense to just post your schematic before taking time to probe, in case there's some clear schematic issue.
Is your control circuitry grounded to earth somewhere? My thought for initial meter probing was to tie antenna input to the low-voltage negative (ground) rail. For probing with a meter, set to DC volts, negative meter lead to low-voltage negative (ground), and touch different circuit nodes with the positive meter lead. (I often use a standard T-pin held in an alligator-clip for probing, with tape around it to avoid touching the circuit with my hand. The sharp steel tip makes it less likely to slip off an IC pin while probing. Even sharpen the pin or needle further with sand paper or whatever.)
4.6V isn't an issue unless it indicates some broader problem. One such possibility is unstable regulator operation. The 4.6V might be oscillating from 3V to 5V with an average value of 4.6V. Check that node with your meter set to AC volts.
What value are the X2 caps? With those two caps, the DC blocking cap is serving no function, but not hurting either.
There have been a few cases here of cheap Chinese counterfeit UCC2732x chips. Not sure if there are similar issues with the UUC3x versions. The counterfeit chips have completely non-functional enable pins (always enabled). Non-functioning enable looked like a possibility from the initial behavior you posted.
-
Attaching fast hand drawn schematic, its a classical 555 based interrupter, 74hc14 as hex inverter and the driving is done by ucc3732x with a GDT.
Also the 2 videos of its operation mentioned in first post:
1) />
2) />
Indeed the UCC driver chips and the LM7805 regulator, are bought from a questionable retailer at ebay.
But with both 3 stuck on enable, the coil should blow its fuse due to CW mode and maxing amps on fuse, but the fuse seems to blow because both mosfets make a short circuit between the + - DC rails. (fuse in installed before the bridge rectifier).
Probing the coil will be done later this week (possibly Friday).
PS. Hand drawn schematic is a mess, i can maybe make a proper schematic using an online tool also later this week
EDIT: Rail caps are WIMA MKP x2 150nF 305V and dc block capacitor is WIMA FKP 100nF 1600V (on schematic i put reversed values by mistake)
-
Schematic
-
Thank you for the schematic, that is helpful.
Looking back at your first post, were you expecting something to happen before closing your mains switch? It almost sounds like that was working properly for a moment. Sparks to the antenna may have damaged your HC14 input and/or the 1N4148 clamp diodes. The antenna needs to be farther away. The input diodes designed to protect the HC14 work under normal conditions aren't designed to handle spark currents. Damage may induct high input leakage current without causing a complete hard failure of the chip. One quick way to check is to power your control board only, then measure voltage from HC14-1 to ground, then HC14-1 to +5V. Both should read about 0V, as the meter should be able to pull HC14-1 either high or low.
To check ucc3732x enables, power the board with 12V and set duty cycle to minimum. Connect HC14-1 to ground (negative side of 12V/5V supplies). Measure voltage from ground to enable (ucc37321-3 and ucc37322-3). Measure voltage from ground to ucc37321-6/7 and to ucc37322-6/7. Then repeat with HC14-1 to 5V. In all cases, the ucc3732x output voltage should read no more than the enable pin voltage. If the output voltage is higher, then you have some of the non-functional-enable counterfeit chips.
Transistors usually fry faster than fuses. Continuous enable could easily explain your problems, as could the initial spark to the antenna (partially) frying the HC14.
The half-bridge cap values are quite low for normal SSTC designs. Have you ran JavaTC on your coil design, or some other calculation of inductances and frequencies? You may have accidentally built a DRSSTC, which may be overloading the half-bridge given long enable pulses (long for DRSSTC use).
-
Coil secondary/primary coil geometry is copied from kaizer sstc 3 so don't think there is a problem as being a drsstc.
Secondary 165mm length / 75mm diameter, 36awg enameled copper wire.
Primary 10 turns of 1,5mm2 wire, 80mm in diameter.
I try use javatc on chrome for android with some errors
(Same errors on chrome for windows desktop - the coil picture is not centered, and also javatc show my coil as 255khz resonant, while Mads is around 180khz with same coil)
-
The key difference between SSTC and DRSSTC isn't in coil design, but rather in the capacitor(s) in series with the primary. For an SSTC, the capacitance is large enough to make primary resonance way below secondary. The capacitance is typically referred to as DC-blocking, with a value around 5uF. For DRSSTC, the series capacitance is much lower, making the primary resonant frequency close to that of the secondary.
You have roughly a DRSSTC setup. Primary capacitance is the two X2 caps in parallel (0.2uF) in series with the 0.15uF "blocking" cap, for 0.086uF. Combined with 8.23uH primary produces roughly 200kHz primary resonant frequency. That's close to the secondary resonance, so dual-resonant or DRSSTC. The driver's enable pulse width designed for SSTC is much too long for this DRSSTC. Primary current will ramp too high.
In short, the X2 caps and "blocking" cap need to be much larger. As I mentioned previously, the blocking cap isn't needed, as the two X2 caps perform that function fine by themselves.
-
So essentially , after probing the driver circuit, finding out if pin 3 on both ucc chips is working, removing dc block cap and replacing the X2 caps, the coil should work.
What is your suggestion for new X2 caps? Will two Wima MKP X2 2.2uF work??
https://www.tme.eu/en/details/mkp-x2-2.2u_305/x2-y2-polypropylene-capacitors/wima/mkx2aw42206f00kssd/
Or should a chose caps with higher voltage rating than the one on the link in order to suppress transients?
Unfortunately, i cannot buy a lot a variants in order to experiment, local market has no such stock, and the shipping from the company I buy online is a bit on the high side comparing most prives of general electronic components.
-
Those 2.2uF 305VAC/560VDC X2 caps may work fine, but X2 caps aren't generally intended for high current at high (200kHz) frequency. This part has no ripple current specification. From a very quick search of WIMA parts, MKPF3Y42206B00KSSD 2.2uF 450V is specified at 3.5A RMS 10kHz, so could likely handle a bit more at 200kHz. Should be within spec up to ~20% duty cycle for your coil. Of course, larger parts (either more capacitance or higher voltage) generally handle more ripple current. DC link capacitors are good for SSTC use, but even small ones tend to be more than you need (and physically larger).
If you don't mind waiting, another option for caps is Chinese induction cook-top caps. I use an array of such 0.33uF 1200V resonant caps for my DRSSTC MMC. They perform well in my abuse testing. The same line is available in what they call "X2" caps, at 5uF 400VDC, 275VAC. Although they are called "X2" caps, given the induction cook-top application, they must handle high-frequency AC current well. (Just ordered a few for future abuse testing.) A couple EBay links for these:
https://www.ebay.com/itm/10PCS-New-BM-Capacitor-MKP-5uF-275VAC-400VDC-for-Induction-cooker-repair-P-31/272271652027?epid=25006519205&hash=item3f64a7b4bb:g:-EAAAOSwmtJXXQmN
https://www.ebay.com/itm/2pcs-5uF-400VDC-MKP-Capacitors-Smoothing-Filter-For-Induction-Cooker/392262782819?hash=item5b54af0f63:g:lOgAAOSwc3ZUnp7E
Not thoroughly specified, but my guess is that they will perform quite well, and the price is low.
Your antenna may need to be a bit farther away to avoid arcs. I'd suggest bending a small loop at the top to make it less pointed, so less likely to arc. Of course, too far away doesn't pick up enough signal to function. It's also possible that the one arc you had damaged the HC14 or diodes. Hopefully not. If it doesn't work, that's something to check. You'll also may need to reverse primary coil polarity depending on how it's left at the moment.
Hopefully the 4.6V regulator is DC, not oscillating. Another hopefully-not issue.
Yes, with the cap changes, you should be close.
-
Your antenna may need to be a bit farther away to avoid arcs. I'd suggest bending a small loop at the top to make it less pointed, so less likely to arc. Of course, too far away doesn't pick up enough signal to function. It's also possible that the one arc you had damaged the HC14 or diodes. Hopefully not. If it doesn't work, that's something to check. You'll also may need to reverse primary coil polarity depending on how it's left at the moment.
Maybe consider using secondary coil feedback, with a 1:50 current transformer around the bottom of the primary. Alternatively maybe you should use a resistor in series with the antenna to take some load off of the diodes and power rails.
-
Gentlemen success!! :D
Replaced the 5v regulator to a stable one (seems the 5v rail was a problematic lm7805 transistor)
Replaced the 1n4148 diodes, the 74hc14, the UCC chips, the 555, all from texas instuments ( no advertising :P )
Made antenna a bit longer and a circle on its end as davekni suggested.
I link the video of the sstc working @ 50v AC max, have not tested it at full power yet.
Strange thing is that when the sstc gives a spark, the variac seems to zap me with no pain or any "electrical" feeling on my hand, its like a big speaker with bass, but the effect feels only on the fingers turning the knob. I cant explain it better because I am not a native English speaker. Anyone felt anything like that before?
-
Congratulations! :)
Strange thing is that when the sstc gives a spark, the variac seems to zap me with no pain or any "electrical" feeling on my hand, its like a big speaker with bass, but the effect feels only on the fingers turning the knob.
Tesla Coils usually create high EMI which can induce voltages on nearby conductors as your variac's casing is.
Especially if you ground the secondary via main's earth. That can even cause problems. Once i destroyed my interrupter because i grounded the secondary to the mains earth, and the interrupter was powered from mains, which caused arcs on the microprocessor of the interrupter, eventually frying it :)
-
I'd recommend grounding everything close to the coil that has a metal case. Still, if you are close to the coil, you will pick up some field yourself. You are acting as an antenna. So grounding yourself by touching something grounded results in some current that you may feel as you touch a grounded object.
-
I tried testing the coil with 230V mains voltage (since my 50v ac test was done, the variac was borrowed from a friend and no longer in my possession).
I applied power to the driver, then I applied power to the bridge, but it resulted in an explosion cutting power to the house because the plastic box holding the fuse in the pcb melted and Line touched the earth lead of the mains connector.
-
Can you take some photos of the damage?
-
Here are the photos
EDIT: on the first photo, on the IEC filter/mains connector, there is a breakdown of the metal casing, bottom right side, unfortunately i can't focus my phone to take clearer picture.
-
The fuse is the only thing that blew? Can you measure the VBUS and MOSFETs for continuity to see if they're shorted?
-
Mosfets and gate resistors are dead (the reason of fuse blowing)
Vbus, voltage splitter caps, smoothing cap, bridge rectifier, mains filter seems ok.
Nothing is shorted to ground also.
-
Looks like the fuse was higher-current that the holder was rated to handle, or the fuse wasn't rated for the fault current available in your outlet.
Without a variac available, another option for gentle bring-up is to wire a ballast load such as an incandescent light bulb in series with the line power. The light bulb acts as a fuse and current limiter. Incandescent bulbs are great because there resistance is lower when cold. But, if you need higher current than available by paralleling available bulbs, a heating appliance such as toaster or waffle-iron or ... also works for ballast.
There are several possible causes for MOSFET frying. It will be difficult to tell what's causing trouble without scope traces. If you decide to replace parts and try again, you could just run at lower power with the ballast, or look to borrow or acquire a scope.
-
I will replace the parts in the next days, and will try soft switching probably with a ballast.
When i have news, i will update the post
-
Costa's
74hc14 ic is sawtooth output, maybe a schemtie hex inverter that is square wave might work better.
-
Costa's
74hc14 ic is sawtooth output, maybe a schemtie hex inverter that is square wave might work better.
74HC14 output is a square wave (or pulse with non-50% duty cycle). It is often used as an oscillator, with the input having a sawtooth waveform, not the output.
-
I was wandering, maybe apart from incandescent lamps/ballast/variac, is it correct practice to use a NTC rush current limiter, as a means of soft starting at 230v directly from the wall, so i can keep the "compactness" of my 100x100mm pcb design ?
Also i have a question: i have 2 voltage divider caps in series connecting + and - DC rail, and another one in the middle as dc block. In my understanding in each circle of the sstc the 2 caps active are the middle one and either the one on + rail, or the other in - rail depending on which mosfet is ON at the time. So in order to make calculations about F/L/Xl to find the primary peak current, should I calculate the active caps total capacitance, or calculate the 3 caps capacitance and consider it primary cap on all calculations ?
I attach 2 pictures, the one I have on my mind how a half bridge works, and a clear template for someone to correct me.
-
The drawing is correct. But regarding the peak current calculation, the capacitance of the caps isn't really relevant. The inductance of the primary is. I have suggested before that you omit the DC blocking capacitor as the capacitive divider acts as a DC blocking capacitor and poses a risk to resonating.
Now that I mention the primary, is it in phase with the secondary? I have attached an image that shows how they should be phased. The top of the primary should be connected to the MOSFET side in the schematic, providing the primary and secondary are wound in the same direction.
-
Yes, as Zipdox said, an unintended primary resonance is one of many possible reasons for your FETs frying. Didn't you fix that already per my previous suggestions? I thought you had removed the "DC blocking" capacitor and increased the other two from 100nF to 2.2uF or 5uF or something around that.
-
Since new parts have not arrived yet, I though I should give it a shot since the low voltage test was completed succesfully.
I designed a new pcb layout for 2.2uF caps and removed completely the "dc block" cap, as davekni and Zipdox correctly suggested. (the explosion video is with the first setup 100/150nf)
Will using an NTC as inrush limiting, be able to soft start the charge to capacitor? Or should I use another NTC into the + DC rail after the smoothing cap in order to feed the mosfets "softer" current? Or by just adding the larger caps this problem will be solved if nothing else is wrong)
I am looking into a solution where i can run the coil directly from the wall, as a show piece, without having to use variac/balasts/incadencent lamps all the time (of course i understand those extra components play a critical role in tuning and troubleshooting the sstc)
Another question I have is, the 2 cap voltage divider supplies the primary with only half of the rectified voltage, in my peak current calculations, i should use 320v or 170v?
Forgive my questions, I am a HVAC techician, with limited knowledge regarding LCR circuits.
EDIT: the new pcb layout has oval slots for caps, so i can easily choose caps and not have to design and print again and
again pcbs
-
>> Or should I use another NTC into the + DC rail after the smoothing cap in order to feed the mosfets "softer" current?
No. DC rail is not a good place to add resistance on purpose.
>>I am looking into a solution where i can run the coil directly from the wall, as a show piece, without having to use variac/balasts/incadencent lamps all the time...
Wish there were a picture handy of my Ballast Box. Electric utility box with two ordinary 120V receptacles that are wired in series, clearly marked as such on the outside. For example, a toaster or lamp plugged into one outlet serves as a ballast for anything plugged into the other outlet.
>>Another question I have is, the 2 cap voltage divider supplies the primary with only half of the rectified voltage, in my peak current calculations, i should use 320v or 170v?
If the capacitance is much larger than resonant value, primary coil sees + or - 170 volts.
I found current calculations and measurements in this well written document by Loneoceans: https://loneoceans.com/labs/sstc2/
-
Yes, everything kludgesmith said. An NTC inrush-current-limiter is a good idea, in series with the bridge rectifier. It can be in one of the AC input legs or in one of the DC outputs of the rectifier. Before the bulk cap, not after. The NTC will avoid burning switch contacts or plug prongs when turning your coil on during normal use.
On the bright side, knowing now that you hadn't changed caps yet, that is likely why FETs fried. So there's more chance that your next try will be successful.
I agree that ballast isn't desirable for any final design. Ballast is very helpful during bring-up. Tektronix used to make a variac/ballast box with internal incandescent bulbs and switches to select the amount of ballast (100W to 600W total):
[ Invalid Attachment ]
Besides that old ballast/variac box (with unfortunately non-functioning meters), I also have a couple series-connected outlet setups as kludgesmith does. Very useful.
Good luck!
-
Thank you all for your replies and suggestions, currently I will be waiting around a month for the new pcb to arrive from china in order to proceed.
I will update this post once the new tests will be done.
-
For the inrush limiting you could use an NTC thermistor, but you have to make sure it's rated for the current it'll be passing without overheating. But a better solution would be using a power resistor and relay to bypass it once the cap charges.
-
For the inrush limiting you could use an NTC thermistor, but you have to make sure it's rated for the current it'll be passing without overheating. But a better solution would be using a power resistor and relay to bypass it once the cap charges.
What I find in most larger power supplies is an NTC thermistor bypassed by a relay once DC power comes up. I use that combination in several of my projects too. The low resistance of the NTC after power-up limits relay inrush current.
-
New (correct value parts) are arriving to my home, and in the next days I will replace the capacitors of the bridge.
Also I finally found a good offer for a 3kVA variac, because the 1kVA i am using is blowing its fuse when i apply more than 100V.
Meanwhile a video with 85V ac test:
-
-
Nice to see success! Thank you for sharing.