4
\$\begingroup\$

I've searched over the net on this issue but found no answers. In your own experience or thought, is it possible to use igbt as a switch?

If yes, please provide schematic as an example.

EDIT:

This is what I actually wanted:

enter image description here

I wanted to turn on the 220V bulb when the logic state is '1', and off when '0'. The problem is, I couldn't make it work.

Do you have a workaround for this?

\$\endgroup\$
  • \$\begingroup\$ What do you mean by "couldn't make it work" exactly? Did you try it for real, or simulate? Was it on all the time or off? What voltage was applied at the gate? \$\endgroup\$ – Oli Glaser Sep 9 '11 at 19:19
  • \$\begingroup\$ I couldn't make the bulb turned on. Yes, it's a simulation. I tried making it on and off alternately. The gate is 5 volts when '1' and 0 volts when '0'. \$\endgroup\$ – Neigyl Noval Sep 9 '11 at 19:23
  • \$\begingroup\$ Is that circular component connected across the bulb a voltage source? If so the IGBT will do nothing as it can't break the current loop through the bulb. The diagram is not very clear. \$\endgroup\$ – Oli Glaser Sep 9 '11 at 19:26
  • \$\begingroup\$ It's actually a voltmeter. I'm sorry, I failed to put labels. I actually used Proteus ISIS for that. \$\endgroup\$ – Neigyl Noval Sep 9 '11 at 19:31
  • \$\begingroup\$ See updated answer regarding gate voltage. I used a spice model I found on the web for the part number shown in your diagram. \$\endgroup\$ – Oli Glaser Sep 9 '11 at 19:44
5
\$\begingroup\$

Yes, in fact this is what they are most commonly used for. AN IGBT is just a BJT driven by a FET.
Basically you just raise the voltage at the gate above a certain level (defined in the datasheet, around a few volts) to turn them on just as you would with a MOSFET, then current will flow between the collector and the emitter.

IGBT

I would avoid any use with high voltage/current (if intended) if not sure of things. If you tell us some more about what you want to switch then a better example could be given.

Edit - here is a Gate voltage (blue) vs Drain-Source current (green) for the IRG4BC10UD, notice at 5V it is still firmly off:

IGBTslope

\$\endgroup\$
  • \$\begingroup\$ Thank you for the answer. I actually like to create a PWM driven 220V bulb dimmer. I plan to take the output of the PWM to the gate of IGBT and the bulb on the collector. Please tell me if this is possible. Thanks. \$\endgroup\$ – Neigyl Noval Sep 9 '11 at 18:23
1
\$\begingroup\$

Where did you get your IGBT model from?
Is it a model for the actual part of a generic IGBT model?
The part that you have chosen should typically work with 5V drive but the curves in fig 3 below show that you typically need at least 6.5V drive and that say 8V+ drive would be better. Notice that the maximum gate drive voltage shown on the graph is 13.5V even though the device is 20V gate voltage rated. Operation well below rated gate voltage is a good safety aim as long as the device works well at the chosen gate voltage

The IRG4BC10U IGBT datasheet can be found here

This is a nice part that is well suited to your need IF PWM frequency is not too high and as long as you provide enough gate drive.

Data sheet page 2 says VGE(th) = gate threshold = minimum turn of voltage is 6V max (worst case) so you MUST use at least 6V drive in a prope design.

Data sheet fig 3 page 3 shows you can get 10A at 9V drive - which should be ample for you. 10A x 110 ~= 1 kiloWatt!. For other reasons you would have trouble reaching that current but you don't need to. A 100 Watt lamp requires about 1 Amp at 100 VAC once warmed up.

enter image description here

HOWEVER

  • Fig 1 shows max current with frequency. It's happy enough at 1 kHz, getting unhappy at 10 kHz and getting bad much above that. 1 kHz is quite slow PWM, but ore than fast enough to avoid lamp flicker.

  • Fig 5 shows that you will get under 2V drop at 1Amp - maybe 1V drop.

    At 1A x 2V you get ~2 Watt dissipation plus there are switching losses.

  • fig 10 shows that switching losses are small at 1 kHz compared to energy loss due to the 1V to 2V turned on voltage at 1A.

So, overall - with about 10V gate drive and 1 kHz PWM and a modest heatsink this IGBT will work OK at 110 VAC, 1A.

BUT a MOSFET may be a better choice - higher PWM without the losses of the IGBT.

The Infineon IPA60R520E6 MOSFET is in stock at Digikey at about $2/1 Datasheet here.

  • 600V, 8A.

  • Up to +/- 30V AC drive on gate so more rugged than most.

  • The "FullPAK" TO220 type case is fully insulated, making it safer to use.

  • It will run OK at 5V drive with <= 2A load - much happier at say 6V or more drive.

  • Rdson = on resistance is slightly over 1 ohm at up to several amps load at 5V drive. So say 1.5W losses at 1A and 3W at 2A. (Page 10, various figs).

    A similar heatsink to that for the IGBT is required.
    For say 30C temperature rise at 3 at a 30/3 = 10 C/W heatsink is needed.
    The FullPAk pkg adds another 4.3C/W thermal resistance.

  • BUT switching times of well under 100 nanoseconds mean it is suited to far higher PWM frequencies than the IGBT.

Overall, the FET would work well and allow wider choice of PWM frequencies.

\$\endgroup\$
  • \$\begingroup\$ Not sure where the OP got model from, I assumed it was in the Proteus Library. I got mine from here \$\endgroup\$ – Oli Glaser Sep 9 '11 at 23:02
0
\$\begingroup\$

IGBTs need a gate voltage that is specified in their datasheet. Most IGBTs behavior is specifed at Vge = 15V, with absolute maximum of Vge = 20V.

If you're below 15V, the IGBT may not perform as specified (Vce voltage drop may be higher than at 15V Vge); if you're above 20V the IGBT may be damaged.

In practice this means design them with slightly higher than 15V drive so that it's definitely more than 15V but you have plenty of margin below 20V.

You also need to know that IGBTs aren't as fast as MOSFETs: most IGBTs work fine to 10kHz PWM, but if you want to use 20-30kHz PWM then you need to buy special IGBTs that are optimized for fast switching.

\$\endgroup\$
  • \$\begingroup\$ Thank you for the answer. What MOSFET do you recommend that could supply 220V dc and drive +- 5V, and suitable for at least 30KHz PWM? \$\endgroup\$ – Neigyl Noval Sep 10 '11 at 11:19
  • \$\begingroup\$ you need to specify current and package and heatsinking, it's not something I can do for you. International Rectifier and Fairchild Semiconductor both have 300V MOSFETs available (don't use 250V MOSFETs -- not enough voltage margin for inductive spikes) that perform very well and are reasonably cost-effective. \$\endgroup\$ – Jason S Sep 10 '11 at 12:33

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.