Choosing QCW switches / Tesla Coils / Forums

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

Choosing QCW switches

1 2 next

Move Thread

LAN_403

zzz_julian_zzz

Sat Feb 15 2014, 05:03AM

Registered Member #3964 Joined: Thu Jun 23 2011, 03:23AM
Location: Valenzuela City
Posts: 332

Hi everyone,

I'm currently building my QCW drsstc bridge at the moment, and would like to seek some thoughts about choosing the switch in the full-bridge inverter of my 320khz coil. I plan to put not more than 150 Amps on my bridge.

Main requirement - is that it is fast IGBT or MOSFET
like these:

I'm bit curious on what characteristic to look for from the data specs of theses switches in the field of current capacity.

1. Should we look for high (at least 50 Amps) Continuous Collector Current@25*c & 100*c??
2. How about Pulsed Collector Current ? (sometimes there is 1ms period indicated)
3. Is paralleling them makes the capacity doubled?

There are switches that has high Icc but low Pcc. And also, low Icc but high Pcc. So which are more appropriate and requirement of QCW DRSSTC system?

The IGBTs that I'm inclined to procure for are:
1. STGW20NC60VD -

2. IRG4PC50UD -

3. IRG4PF50W -

Thank you very much, any thoughts related to this topic is highly appreciated.

Steve Ward

Tue Feb 25 2014, 06:57AM

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

The thermal performance is probably one of the most critical specs for the IGBT, it determines how much power it can dissipate.

Ive use the FGH60N60SMD from fairchild for QCW coils. I choose to run them at only 75A and less for "high reliability" projects, but my test bridge showed them hard switching 150A just fine at 350khz. Cooling the IGBTs will be critical for high duty cycle operation.

The IGBTs you are looking at have about half the power capacity of the 60N60SMD (based on the Rth junction-case spec). They are similar in switching performance, so they should work well for QCW, but you may need ot parallel them for 150A.

Paralleling requires care to make the inductance/resistance for each IGBT to be equal, otherwise they will not share the current and you will not get the most from them.

zzz_julian_zzz

Wed Feb 26 2014, 11:03AM

Registered Member #3964 Joined: Thu Jun 23 2011, 03:23AM
Location: Valenzuela City
Posts: 332

Hi Steve,

First of all, thanks for your astute comment. Surely, will take note of them.

For the 150 Amp bridge current I'm aiming, I will be constructing a parallel bridge to achieve it. (I'm looking for other designs and i found the photo below)
Plus, the cooling/ventilation like what you said, must be designed properly.

For the current rating of about half of what you have used, not to mention, your modest IGBT usage;
I guess the current rating to be considered more would be the Ic (Collector Current) capacity of the switch to be used on this type of application.

Just a curious question, On your test bridge, is there any dummy load used? What is its characteristic? capacitive or inductive? Thank you.

Goodchild

Wed Feb 26 2014, 03:43PM

Registered Member #2292 Joined: Fri Aug 14 2009, 05:33PM
Location: The Wild West AKA Arizona
Posts: 795

I agree with Steve thermal performance is huge. The TO-247 and it's variants like the Plus247 are great thermally, in addition the topology can also influence the switch you choose.

For example I take a great deal of care in choosing a switch with the proper parasitics for ZVS. Low channel capacitance ect help a great deal when hard-switching.

IXYS has some really good parts as well that are supper fast and have a 1200V rating like the IXYR50N120C3D1 in addition a lot of the IXYS parts have a positive tempco making them ideal for reducing the chance of thermal runaway when paralleling.

zzz_julian_zzz

Thu Feb 27 2014, 10:18AM

Registered Member #3964 Joined: Thu Jun 23 2011, 03:23AM
Location: Valenzuela City
Posts: 332

Hi and Thanks to you too Eric,

Actually, i don't want to put an over stress (in the field of temperature) on the switches i will be using on my bridge. I'll just run it for a couple of minute(2 -3 max) then shut it down.. I just want to see long sword sparks (x5) coming from my 7 inch coil..

I'm worried on the current capacity of these IGBTs, but ideally, paralleling them will allow me to expect more (twice) from them & well considering the thermal state of these switches.

Thank you.

Goodchild

Fri Feb 28 2014, 03:52PM

Registered Member #2292 Joined: Fri Aug 14 2009, 05:33PM
Location: The Wild West AKA Arizona
Posts: 795

zzz_julian_zzz wrote ...

Hi and Thanks to you too Eric,

Actually, i don't want to put an over stress (in the field of temperature) on the switches i will be using on my bridge. I'll just run it for a couple of minute(2 -3 max) then shut it down.. I just want to see long sword sparks (x5) coming from my 7 inch coil..

I'm worried on the current capacity of these IGBTs, but ideally, paralleling them will allow me to expect more (twice) from them & well considering the thermal state of these switches.

Thank you.

You may be surprised at the performance you get thermally out of low end switches, even without water cooling. If you keep your break rate low <1 or 5pps you can probably run indefinably with 4 or 5 feet even with air cooling. Water cooling make it so you can run higher break rates and for longer periods of time.

Usually, the peak rated current in a QCW will be dictated by your thermal impedance and capacitance from junction to heatsink. This is why it's so important to choose good thermal properties when selecting an IGBT because almost no amount of fancy cooling external of the switch will help you with peak currents if the heat can't get out of the switch fast enough.

So when you parallel switches, this will help with your peak current, but will do nothing for you when it comes to your long term heat removal. This is because your still dissipating the same amount of power into the same cooling solution.

On the other hand your steady state power dissipation over time is dictated more by your cooling solution (thermal impedance from case to ambient), heatsink fans etc. A better cooling solution external of the switch like water cooling will allow you to run at higher break rates and for longer periods of time without overheating.

I hope this helps!

zzz_julian_zzz

Sun Mar 02 2014, 07:47AM

Registered Member #3964 Joined: Thu Jun 23 2011, 03:23AM
Location: Valenzuela City
Posts: 332

Salute to you Eric :)

Great explanation, Thank you very much..

I've tried paralleling switches(MOSFET) for testing purposes first, so I dont get blow more expensive IGBTs.

I paralleled 8 pcs of 25N60 mosfets, but bangs when I inputted 150v ramp from my buck. I'm currently troubleshooting it now. I guess insufficient decoupling caps is the error. I'm starting to think that paralleling mosfet is different from paralleling IGBTs...

Do you have any tips or notes I should be checking when paralleling MOSFET/ IGBTs for DRSSTC use?

Thanks again Eric!

zzz_julian_zzz

Mon Mar 10 2014, 11:00AM

Registered Member #3964 Joined: Thu Jun 23 2011, 03:23AM
Location: Valenzuela City
Posts: 332

UPDATE:

Hi Eric and Steve,

As of this moment, I am waiting for the delivery of my ordered FGH40N60SMD IGBTs from RS.

This have positive temp co, fast, high ampacity, Rthj-c is 0.43 (low), and max Junction case is 175*c, Power dissipation is about 350 watts for 25*c.

Question is, if I'm going to use this IGBT, do i still have to parallel it for longer spark? or there will be a certain ampacity of switches in the bridge that will not increase the QCW coil spark output even if I put/parallel higher current ratings of IGBTs?
Thank you.

Sulaiman

Mon Mar 10 2014, 01:33PM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3140

When paralleling mosfets DO NOT just connect all of the gates together
... a classic way to get >>Mhz parasitic oscillations.

use one gate resistor per gate,
sometimes also need a NiZn ferrite bead in series with each gate resistor.

P.S. not tried QCW myself, but from general principles;
it is good to have excellent thermal conductivity to good heatsink
but reducing switching losses reduces the heatsinking requirements.

Steve Ward

Thu Mar 13 2014, 02:22AM

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

Question is, if I'm going to use this IGBT, do i still have to parallel it for longer spark? or there will be a certain ampacity of switches in the bridge that will not increase the QCW coil spark output even if I put/parallel higher current ratings of IGBTs?
Thank you.

The IGBTs will not "limit" the spark because they represent a relatively small resistance to the whole system. Adding more in parallel (and changing nothing else) will likely not produce any change in spark performance. The question is "can this IGBT produce X spark without failure?".

I would think 2 IGBTs in parallel would handle the 150A provided low duty cycle (low rep-rate) and relatively cool heatsink temperatures (say less than 40C). If you want higher rep rate, you either need much better cooling system, or more devices in parallel to spread out the heat.

I suggest looking at manufacturer application notes about paralleling IGBTs to get some idea of the possible problems you can encounter. The FGH...SMD series says its easy to parallel... hopefully its true

.

I built a bridge from 4-parallel FGH60N60SMD. I used small ferrite EMI suppressors in series with each gate as well as a seperate gate resistor and reverse diode (for fast turn off). I cant really say how well the IGBTs shared the load because i only loaded the bridge to 200A and had no way to measure each IGBT current. I did note, however, that when i did fail the bridge (due to my own negligence, hard switching 500A or something...) that only certain IGBTs failed... the ones that had the least inductance and resistance in the circuit. Making a completely symmetric parallel circuit is pretty difficult, so i changed to 4 seperate H-bridges, and tie their outputs together through some means of balancing.

1 2 next

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