QCWDRSSTC

[曹靖]

Hello everyone~This post is about the QCWDRSSTC I created two years ago~Due to time constraints, I am no longer able to provide a complete production process and photos. Currently, only arc snapshots and some fragmented circuit physical structure photos are available. Overall, it is quite chaotic~I only registered for this forum a few days ago~If I registered here two years ago or earlier, I can display the complete production process~Unfortunately~

[davekni]

Nice performance!  Most interesting data would be basic coil parameters:
Frequency at start of ramp.  Frequency at end of ramp.
Primary inductance, secondary inductance, coupling factor.  Or, if you know diameter and height and turns count, JavaTC will estimate those parameters plenty well.
MMC capacitance.
This would be interesting data, but no pressure to answer.  Time constrains all of us.

[曹靖]

Nice performance!  Most interesting data would be basic coil parameters:
Frequency at start of ramp.  Frequency at end of ramp.
Primary inductance, secondary inductance, coupling factor.  Or, if you know diameter and height and turns count, JavaTC will estimate those parameters plenty well.
MMC capacitance.
This would be interesting data, but no pressure to answer.  Time constrains all of us.

This QCW DRSSRTC is located several hundred kilometers away from my current residence in my other home, so I can only rely on the photos and memories left to describe some coil parameters. The longest arc recorded is approximately 94 inches.
The secondary coil uses a white PVC pipe with a diameter of about 6 inches and a height of about 10 inches. I used AWG27 red enameled wire to wrap the PVC pipe with a 5mm gap at both ends
I can't recall the specific size of using AWG14 ordinary single core copper cable for winding at the beginning - the diameter of the grading ring is 12 inches, and the height is about 3-4 inches.
After installing the grading ring on the secondary coil, measure the resonance frequency between 430KHz and 450KHz
The primary resonant capacitor is connected in series with two 22nF4000V capacitors to obtain 11nF. Combined with the primary coil, the range of 395KHz to 420KHz is obtained
The input voltage of 420VDC is rectified and filtered by the mechanical voltage regulator, resulting in a pulse width of 23mS
I can only remember these vague information. There are also some inverter bridge switch waveform pictures and arc extinguishing device modulation waveform pictures. Additionally, there are some physical pictures of the arc extinguishing device and host circuit structure. These are all taken as souvenirs during the production process

[davekni]

The input voltage of 420VDC is rectified and filtered by the mechanical voltage regulator, resulting in a pulse width of 23mS
I can only remember these vague information.

Great information.  Thank you for sharing what you have.  I'm still experimenting with my first QCW (not counting earlier low-frequency experiment).  Produces many curved arcs and relatively few straight ones.  I appreciate data on QCW coils that make straighter arcs.
One question:  Scope images show roughly 55ms an 45ms ramp times (including initial flat portion).  Do you recall if you later reduced that to 23ms?  Or does the 23ms value refer to something other than the ramp time?

[Mads Barnkob]

That is a very lovely and simple QCW coil! Do you have any pictures of the IGBTs / bridge?

Regarding room temperature and straight/bent arcs, this sounds much more like a humidity issue. With warmer room, you are lowering the relative humidity of the air, making it more dry and less damp/conductive. So "wet" will properly help sparks with a better coupling towards ground.

[曹靖]

That is a very lovely and simple QCW coil! Do you have any pictures of the IGBTs / bridge?

Regarding room temperature and straight/bent arcs, this sounds much more like a humidity issue. With warmer room, you are lowering the relative humidity of the air, making it more dry and less damp/conductive. So "wet" will properly help sparks with a better coupling towards ground.

The inverter full bridge power tube uses eight 160N60UFD to step-down converters, which are also composed of eight identical IGBTs divided into two groups, one group for switching and the other group for continuous current

[曹靖]

The input voltage of 420VDC is rectified and filtered by the mechanical voltage regulator, resulting in a pulse width of 23mS
I can only remember these vague information.

Great information.  Thank you for sharing what you have.  I'm still experimenting with my first QCW (not counting earlier low-frequency experiment).  Produces many curved arcs and relatively few straight ones.  I appreciate data on QCW coils that make straighter arcs.
One question:  Scope images show roughly 55ms an 45ms ramp times (including initial flat portion).  Do you recall if you later reduced that to 23ms?  Or does the 23ms value refer to something other than the ramp time?

The slope width in the above photo was randomly adjusted during controller debugging. The slope width during machine operation was 23mS. In addition, I found during the experiment that the QCW linear arc was very heavily dependent on the surrounding environment, including air temperature and humidity, atmospheric pressure, and the arrangement of surrounding conductive materials such as wardrobes, tables, beds, and walls. When I moved the machine to another position, the arc was no longer straight but curved towards the ground. When I moved the machine When approaching my wardrobe, the arc will return to a straight state. My wardrobe is made of wood and does not conduct electricity. Additionally, I have a large stove in my room. It was particularly cold in winter, and when the temperature in the room drops, the arc will no longer be straight but will bend towards the ground. When I light the stove for heating, the arc will return to a straight state when the temperature in the room rises to a comfortable level

[davekni]

The slope width in the above photo was randomly adjusted during controller debugging. The slope width during machine operation was 23mS. In addition, I found during the experiment that the QCW linear arc was very heavily dependent on the surrounding environment, including air temperature and humidity, atmospheric pressure, and the arrangement of surrounding conductive materials such as wardrobes, tables, beds, and walls. When I moved the machine to another position, the arc was no longer straight but curved towards the ground. When I moved the machine When approaching my wardrobe, the arc will return to a straight state. My wardrobe is made of wood and does not conduct electricity. Additionally, I have a large stove in my room. It was particularly cold in winter, and when the temperature in the room drops, the arc will no longer be straight but will bend towards the ground. When I light the stove for heating, the arc will return to a straight state when the temperature in the room rises to a comfortable level

Thank you for the information!  I've been wondering how much of my arc curving might be due to environment.  Ramp adjustment mostly affects branches in my testing, not much change in curvature.  Mads suggested my thin wide top load may be causing some of the curving.  That's something I'm planning to test eventually, though I'm skeptical top load shape is a critical factor.
BTW, though wood is not very conductive at low frequency, it can be significantly conductive at 12MHz.  And dielectric materials should have some effect even if not conductive at all.

[曹靖]

Thank you for the information!  I've been wondering how much of my arc curving might be due to environment.  Ramp adjustment mostly affects branches in my testing, not much change in curvature.  Mads suggested my thin wide top load may be causing some of the curving.  That's something I'm planning to test eventually, though I'm skeptical top load shape is a critical factor.
BTW, though wood is not very conductive at low frequency, it can be significantly conductive at 12MHz.  And dielectric materials should have some effect even if not conductive at all.

I personally believe that the thickness of the grading ring does not have a significant impact on whether the arc is bent. Although I have not conducted relevant experiments on QCWDRSSTC, I have used very thin grading rings on some ultra small linear arc coils and later replaced them with thicker grading rings. The difference is not significant. Additionally, I also used very thin grading rings on my other ultra small linear arc coil, but its arc is very straight and not bent

[曹靖]

I found another photo of me measuring the arc interrupter. In the photo, I can see two channels of probes, one of which is attached with a low-pass filter to restore PWM and see the slope. Another example is that when I tested the arc on a small QCWDRSSTC, I found that as long as the coil is close to the iron frame, the arc tends to be more straight. When it is far away from the iron frame, the arc will bend. Look at how straight the arc is

[davekni]

Another example is that when I tested the arc on a small QCWDRSSTC, I found that as long as the coil is close to the iron frame, the arc tends to be more straight. When it is far away from the iron frame, the arc will bend. Look at how straight the arc is

Great information!  Not quite what I would have guessed.  I'm surprised that being near the metal didn't make the arc curve towards the metal.  This gives me ideas to test for my QCW coil.

[Rafft]

Nice info!

Regarding ramp controller/interrupter, any early pictures with still not using fiber optics? Like just plain sheilded wire. From controller to tesla coil.

On my qcw setup, interrupter sugnal is galvanically isolated thru an optocoupler(on the coil side) ,so GND of controller still runs going to tesla coil base. The ramp control signal though, is electrically connected(GND) to the buck. I wonder if this is a 'problem' for generating straighter arcs.

Planning on using cheaper fiber optics both for interrupt and ramp. Toslink I can get fairly cheap.

[曹靖]

Nice info!

Regarding ramp controller/interrupter, any early pictures with still not using fiber optics? Like just plain sheilded wire. From controller to tesla coil.

On my qcw setup, interrupter sugnal is galvanically isolated thru an optocoupler(on the coil side) ,so GND of controller still runs going to tesla coil base. The ramp control signal though, is electrically connected(GND) to the buck. I wonder if this is a 'problem' for generating straighter arcs.

Planning on using cheaper fiber optics both for interrupt and ramp. Toslink I can get fairly cheap.

My arc suppressor uses two HFBR-1414TZ fiber optic heads for output. One set is used for PWM output to the step-down converter. The other set is used for normal interruption of the DRSSTC driver board. Since I directly welded with an engineering board and did not draw a circuit diagram, I can only draw a rough diagram using trivial memory, which is similar to the following. In addition, I believe that for straight arc, in addition to providing suitable bus voltage with good slope width and QCW, a special environment is also needed To make the arc more straight, as I mentioned earlier

[Rafft]

Yes. Correct. Thanks.

As you mentioned in your earlier posts, you dont like micros and prefer analog for the ramp generation.

Im using arduino (ramp/int) but mine is not isolated (buck  gnd connected to controller gnd). Yes microcontrollers are more sensitive to plasma. My next move is to use optical isolation. With toslink fiber cable.

[曹靖]

Hello Rafft! This is the picture I provided you with. These are all the circuits I actually use. I hope it will be helpful to you

[Rafft]

Thank you!

Nice! So your buck is simply just a switch! Only pwm inputs and not analog! The size of those main IGBTs though 

This is indeed simpler. 3rd time seeing this actually used and done! (Just plain switcher, not self-osc)

What was your pwm switching freq for this?  Output LC 80uH Im guessing

[曹靖]

Thank you!

Nice! So your buck is simply just a switch! Only pwm inputs and not analog! The size of those main IGBTs though 

This is indeed simpler. 3rd time seeing this actually used and done! (Just plain switcher, not self-osc)

What was your pwm switching freq for this?  Output LC 80uH Im guessing

My PWM modulation frequency is 25KHz, and the main filter LC is 40uF and 100uH

[Steve Ward]

Thanks for sharing many of your great projects!  Really incredible spark length without branches forming!

Nice! So your buck is simply just a switch! Only pwm inputs and not analog! The size of those main IGBTs though 

This is indeed simpler. 3rd time seeing this actually used and done! (Just plain switcher, not self-osc)

I'd like to point out that the buck control can be quite simply provided from a microcontroller using the ideal transfer function for a buck converter and also by having continuous ADC measurement of the buck input voltage. Just compute the duty cycle at some switching frequency, once per cycle, and apply it to a PWM timer:

Duty_Cycle = Vout/Vin

In reality, there is ESR in the buck inductor that will cause the Vout to be somewhat smaller, and this also ignores any transient effects of the LC filter, but so long as your ramp rates are not too aggressive, it should not be a problem.  It also assumes the inductor current reaches continuous conduction mode.  This method is simple because it requires no feedback loop compensation, which would be the case if you used Vout as a feedback signal for the PWM generation.  I use this method on the "Fat" QCW coil.

The self-oscillating type buck modulator was neat in that it provided feedback control and switching all in one shot, but its not necessarily the best way to go.  I originally adapted it from some audio amplifiers i was designing, where finer control of the output voltage did matter a lot to the sound.  Those amps did sound quite good, but switched at 300khz or so.  Its more trivial to control switching for tesla coils, fidelity and bandwidth are not as crucial.

[曹靖]

Wow, the founder of the great Steve Ward QCW! The fact you have stated is worth us repeatedly pondering and practicing (● ●)

[Rafft]

@Steve
Valuable info. Thank you for confirming this.


@曹靖
I notice that your build 'lacks' metal sheilding. All is 'open' and even the platform is made up of non-conducting materials (wood/acrylic/nylon). Could you explain why?