Shop OBEX P1 Docs P2 Docs Learn Events
IGBT / Circut suggestions, advice needed — Parallax Forums

IGBT / Circut suggestions, advice needed

Hey everyone, I've got a somewhat interesting project on my desk. Need a little direction to get started.

The goal is to power magnetic coils with + and - 3-12ish volts @ 1.5 to 3 Amps. Pulses are likely to range from a simple DC pulse in the hundred to few thousand millisecond range. We may find a duty cycle works or is needed during the pulse too, and that would appear to be on the order of tens to a few hundred hertz range.

I will have two source supplies for this, both 20A or so as the voltage may be different on the plus and minus rails, and sufficient for a few active coils at any given time.

This is an R&D project, which is why I've given wide ranges. The engineers aren't sure where things, or the sweet spot power / heat / performance will be. They aren't sure the thing they have planned will even work. I'm supplying a test bed for them to explore options in.

There are a fair number of coils and most will be dormant at any given time. What I need to do is apply power to a given coil at a given time. The test bed being manufactured contains a handful of coils. My initial plan for the project test bed is to trigger power using a Propeller pin per power, per coil. This will handle a few, and is fine for now. Later, other I/O expander type devices and serial comms are likely to be used to expand the range of addressable power. I'll get there when I get there. For now, I'm seeking simple digital I/O type signals easily generated with a Prop chip. Worst case, two prop chips to keep it all really simple.

There will be a third appropriate power rail for the Prop chip to run on and supply the signal used to trigger the IGBT chips as needed, obviously. That one I have covered and understand.

I've been looking at IGBT chips and the selection range is daunting to me to say the least!! Any suggestions or personal favorites capable of handling power in this range? I'm looking for small devices, simple operation. IGBT sees signal, supplies one or the other power, that's it. So far, at this stage, it's all fairly light duty cycle. Bursts of power where and when needed. No continuous power operation. I'm also, at this stage, looking for dead simple application with as few extra components as necessary.

I'm still looking at devices, but if there are ones capable of handling either power, plus or minus on a signal, that would reduce component count. Likely important for the next stage of the project, but not too important right now.

How can I reduce my search scope on these things?

Comments

  • ErNaErNa Posts: 1,752
    I would go to power fets. IGBT's are to prefer at higher voltages when saving a cent is important.
  • tonyp12tonyp12 Posts: 1,951
    edited 2016-04-28 21:16
    >is to power magnetic coils with + and - 3-12ish volts

    Why would you need a negative voltage rail, why would magnetic coils care what the ground reference is?
    Or are you referring to flipping the polarity using a hbridge?
  • evanhevanh Posts: 15,916
    ErNa is right, use power FETs when below 100V. You can cheaply have much higher current ranges and smaller heat-sink because the switching is full rail, there is no diodic drop.
  • Good stuff so far.

    The dual polarity is used to overcome another magnet in some use cases.

    I'm looking through Power FET devices now. :D
  • jmgjmg Posts: 15,173
    potatohead wrote: »
    The goal is to power magnetic coils with + and - 3-12ish volts @ 1.5 to 3 Amps. Pulses are likely to range from a simple DC pulse in the hundred to few thousand millisecond range. We may find a duty cycle works or is needed during the pulse too, and that would appear to be on the order of tens to a few hundred hertz range.

    I will have two source supplies for this, both 20A or so as the voltage may be different on the plus and minus rails, and sufficient for a few active coils at any given time.
    ...
    I've been looking at IGBT chips and the selection range is daunting to me to say the least!! Any suggestions or personal favorites capable of handling power in this range? ...
    How can I reduce my search scope on these things?
    3-12V is low, and only if you have high flyback voltages would you need to even consider IGBT.

    Can you drive this full bridge, like a Motor-Driver ?
    That allows MOSFETS of lower voltage ratings, and also allows you to tap into the existing Motor-driver ICs out there, which
    would find those numbers a walk in the park.
    The better ones allow 0-100% PWM drive (charge pumps on upper N-FET gates)


  • potatoheadpotatohead Posts: 10,261
    edited 2016-05-07 15:36
    I didn't see this jmg.

    The answer is likely, "yes" Some early tests just using switches and a couple bench supplies were favorable. They also revealed the definite need for dual polarity.

    Got a device recommendation or two I could gawk at?

    The final production circuit will end up replicated a large number of times. Simple, low component count is primary. The test bed is likely 32x.

  • evanhevanh Posts: 15,916
    I've got a friend I trust a lot for his designs who has just recently chosen a http://www.allegromicro.com/en/Products/Motor-Driver-And-Interface-ICs/Brushless-DC-Motor-Drivers/A3930-1.aspx

  • evanhevanh Posts: 15,916
    There's also solutions that include the power transistor in the same package but they are quite limited in current rating. Best example I see there is http://www.allegromicro.com/en/Products/Motor-Driver-And-Interface-ICs/Brush-DC-Motor-Drivers/A4950.aspx
  • evanhevanh Posts: 15,916
    I would also probably advise using optocouplers or at least some in-line resistors on the control lines to allow the drivers to voltage bounce without stressing the microcontroller pins.
  • jmgjmg Posts: 15,173
    edited 2016-05-07 23:19
    potatohead wrote: »
    Got a device recommendation or two I could gawk at?

    The final production circuit will end up replicated a large number of times. Simple, low component count is primary. The test bed is likely 32x.

    As mentioned above, Allegro are a pretty good place to start. (and also OnSemi and Infineon)

    If you have many channels, and low voltage, I would also consider just using PFET & NFET parts & gate driver/level shifters.

    The '+ and - 3-12ish volts @ 1.5 to 3 Amps' could need some care at the low end, as if they decide to go below 3V, gate drive level shifters are not as simple.
    If you wanted to offer 0~20v, then charge pump and all N_FET may be better, with many channels, charge pump can be shared and does not add much cost.

    Addit: Since they seem unsure on what they want, I'd maybe go more general and the simplicity of a SOT23 high side driver like IRS25752 appeals.
    Small and cheap, and you can use that on all N-FETS, (both hi and lo), and use one as a charge pump.
    Needs ~12V to be above UVLO and give enough Gate drive for lowest Rds, but can level shift from Prop fine.


  • Good call on 3V. Good news is simple tests indicate 4 and above should do nicely.
  • jmgjmg Posts: 15,173
    potatohead wrote: »
    Good call on 3V. Good news is simple tests indicate 4 and above should do nicely.
    If you go with a separate gate drive supply, then you can vary the main power rail, right down to 0V, with no side effects on Rds.
    The IRS25752 SOT23 driver suggestion I added, would support that method.
  • evanhevanh Posts: 15,916
    You could use PWM, or similar, to effect the lower voltages. Coils are good with that practice.
Sign In or Register to comment.