First DRSSTC

[pawlox98]

Hi
I'm building a medium DRSSTC based on Mads Barnkob DRSSTC I
Here's how I see that:
 MMC: 6 strings in parallel of 2 in series 942C20P15K-F capacitors for 0.45uF at 4000VDC
 Bridge supply: 230V AC through a variac, full bridge rectifier, 2x ALS30A332NF450 KEMET capacitors 3300uF 450V ESR: 40mΩ I_ripple: 21.2A for 10kHz
 Primary: Spiral coil made of copper tube ∅10, 9 turns
 Secondary: 2400 windings ∅160x768mm ∅0.3 double coated enamelled copper wire on PVC pipe
 Topload: 130x630mm aluminum flex tube
 Resonant frequency: 67-68kHz
 Driver: Finn's Hammer Predikter + GDT (if someone wants driver schematics, i can send them)

and here some numbers from JAVA TC :



For now I finished driver and almost topload, I just have to tape it with aluminium tape. Some pictures soon.

Questions:
 1- Transistors I want to use IXGN400N60B3 because they are easy to get for me (datasheet: http://www.mouser.com/ds/2/205/DS100156A(IXGN400N60B3)-311982.pdf) but
  they are rated for 40 kHz switching. According to Mads's DRSSTC design guide IGBT selection for Tesla coils part they can go faster even 4 times because of soft switching. So can I use them?
 2- Varnish  It is better to just use Polyurethane varnish for secondary or it will be better to use electro-insulating varnish for example mg chemicals red insulating varnish ?
 3- Primary Is acrylic a good material to hold primary coil ?

[Garyf]

It appears that the 400n60 doesnt have an anti-parallel diode.

[Mads Barnkob]

1: It depends entirely of the switching current and the temperature rise you can accept, you have to calculate all this, with help from the guide. 4 times faster is just a very rough estimate that does not take switching higher currents than specifications into account, switching losses just higher and only a temperature rise calculation can give you a good estimate.

2: I have only used polyurethane and 4-5 layers have always been enough, so far, the harder you plan to push things, the more layers are needed.

3: acrylic is not good, it easily deforms and melts from mildly high temperatures. So be sure to water cool it, only do short run times or switch a pretty low current to prevent that from happening. See a example of melted acrylic primary holder at the bottom here: http://kaizerpowerelectronics.dk/tesla-coils/drsstc-design-guide/busbar-and-primary-circuit/
That being said, I also used acrylic in my DRSSTC1, but that has never been used for more than 2 minutes at a time and only at 500Aprimary

Regarding the missing diodes, you will just have to buy some large and very fast diodes to mount externally and they will perform better than a built-in parasitic diode would.

[pawlox98]

When it comes to diodes, I was thinking the same but with
calculations, I got stuck witch P_diss I don't know how to get T_J and T_C

[Mads Barnkob]

It is explained in the part "Fmax1 and Fmax2 switching speeds" of http://kaizerpowerelectronics.dk/tesla-coils/drsstc-design-guide/igbts/

We want Tjunction to stay below 80ºC and have Tcase cooled to stay below 50ºC. A advisable rule is to keep the temperature ripple below 30ºC, we often use second hand IGBT bricks and we do not know their usage history. To avoid real fast degeneration of die and bonding wires inside the IGBT, we will aim to reduce large temperature swings.

Further down where I do the calculations with real numbers is where it all happens.

[pawlox98]

I'm sorry I do not know how I could have missed it
I've done my calculations but Im not 100% sure about them

[Mads Barnkob]

That looks much better, the 15V on the upper part is the voltage for which the data is given in the datasheet, which typically is 15V, your supply voltage should be on the lower part. So the result should be 0.3mJ if the data is given at 15V and you drive them at 15V.

These maximum values does not take gate drive current into account! Just because the thermal calculation shows you can dissipate this heat at up to 1,1MHz, does not mean that you can actually drive the IGBT gate that hard But it is a valuable tool to examine if you will blow up your silicon or not.

[pawlox98]

So with this voltage correction Fmax2 is 726 366.67Hz which is way above, so it is possible but not 100% sure they will not blowup ?

[Mads Barnkob]

So with this voltage correction Fmax2 is 726 366.67Hz which is way above, so it is possible but not 100% sure they will not blowup ?

It means that there is no thermal limitations to run these with 500 A at 68 kHz, as 726 is many times higher. I would not worry about gate drive either at 68 kHz, as that is a very small amount of gate charge on these IGBTs.

Just get this build and let us see some sparks 

[pawlox98]

Hi Mads
In "Snubber capacitor selection" You said  ΔV/Δt should be at at least 500-5000 V/μs or above but 40L snubber You used have only 50 and all snubbers I have seen had only 30-400. I will probably use 2x WIMA MKP 5uF 1000V they had 300 V/us

Zener diodes on transistors gates They should be 1W diodes for 15V, correct me if I am wrong

[Mads Barnkob]

The 40L snubber capacitors are likely a left-over from the SCR days where the switching frequency of a inverter was lower, so by todays standards for a snubber capacitor the 40L isn't doing that well, not something I knew when I made the DRSSTC1, I used Steve Wards plans.

The switching spike in a IGBT inverter is somewhere around 200ns long, so that gives quite a steep dV/dt, but also remember that there is also high dI/dt and thus a good RMS rating on the snubber capacitor is just as vital to handle/dissipate the energy.

Gate voltage is directly proportional with the current that can flow through the C-E junction, this is why we normally overdrive the gate with 24VDC instead, to gain that little extra fast switching to limit the switching times when we are exceeding the pulse current ratings of the IGBT. I use 33V zener diodes back to back.

[pawlox98]

Update
I have made secondary 2 layers at this moment and here few problems starts. There should be 2400 windings but there are 2358 if I remeber . I dont know why pipe has 833mm, wire has 0.32mm times 2400 equals 768mm So i should have 65mm free space. I left 20mm gap on one side but that was not enough. There are 2 options, 1. Arduino has missed some turns (program  or slotted transducer) it wasnt very precise device, but it worked when I've tested it. 2. option (the most possible) there is some space between windings . 42 windings its not much this shouldn't be a problem right ?
Bigger problem I think is dust and hairs from brush which created holes in varnish. What should I do about it ?

[Mads Barnkob]

It is a nice looking secondary coil, don't worry about 2350 or 2400 turns, that is the same in terms of Tesla coils, we have so many assumptions that a little off will not make a difference. I think you forgot to add the thickness of the varnish on the wire. A 0.32 mm wire will typically have 0.032 mm varnish if you count both "sides", for almost any enamelled wire you can assume 10% extra on the given diameter.

To remove air bubbles, gently sand the area down with fine grit sand paper, then varnish it again.

[pawlox98]

Wire diameter is 0.3mm but I was lazy and I measured it with caliper. Measure was 0.32mm but it probably have about 0.329mm

Can this be a problem?  It happened 1 or 2 days ago. I was winding it about a month ago . Luckily wire is not broken

[futurist]

Secondary looks nice! Don't worry about that gap if the turns are not overlapping

[Mads Barnkob]

I agree, as long as there is no point "taller" than the rest of the coil, it will be fine. But as it does present a tiny amount of edge compared to the long flat coil, you give it some extra varnish right at that point on the coil.