Announcement

Collapse
No announcement yet.

Mosfet or Transistor

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Mosfet or Transistor

    It seems people are always using mosfets to drive the coil in there metal detectors,
    Is there really any advantage over a transistor?
    more efficient?
    less capacitance?


    Thanks!
    Jon

  • #2
    Easier to find high current and high breakdown voltage, plus C is probably lower.

    Comment


    • #3
      Something from my experience,and may be a good idea.

      Typical low voltage bipolars (100V or less rated, like TIP31, MJE 3055 etc) common in some MD designs are worst possible solution. MOSFETs are tradeoff between high breakdown voltage, low loss and easy to drive (also cheap and widely available). High voltage (1200V rated and up) IGBT devices are typically too large, unnecessarily large chips, large output capacitance will slow down whole thing. Interesting to consider here are HV bipolar devices, like S2000, BU2508, BUL310 or similar, breakdown voltages 1.5KV or so, designed for CRT TV/monitor flyback stages. It is impossible to replace them directly in MD designed for bipolars, they have tricky drive requirements (actually, it is hard to turn them off, special base drive circuit i needed).

      However, cascode connected MOSFET + HV bipolar seems to be win-win combination so far. This can withstand 1.5-2 KV flyback pulse, no problem, and output capacitance is lower compared to anything else, and it is fast and cheap enough...and works just fine.

      Comment


      • #4
        Originally posted by Tepco View Post
        However, cascode connected MOSFET + HV bipolar seems to be win-win combination so far.
        Do you have a schematic for such a configuration? Sounds interesting.

        Comment


        • #5
          I've tried a cascoded coil switch and don't recall the details, but I think it was heck to get a stable bias for the cascode device. All I remember is I abandoned the idea.

          Comment


          • #6
            I abandoned the idea of stacked MOSFETS driving a coil because I could not figure out how to not exceed the maximum ratings for the gate voltage on the second MOSFET. I looked at various designs at the time, and none seemed suitable. The higher breakdown voltage and lower capacitance would be an advantage, if it can be done. A fast switch-off speed would be a requirement as well, which may be another issue to overcome with such a configuration. Just my 2ยข.

            Comment


            • #7
              Stacking together multiple MOSFET or IGBT devices is nothing new, actually, not tens but hundreds of kilovolts at kiloamp levels are achieved already, all solid state.

              But this cascode connection is nothing to deal whit such things, and not a rocket science really, just 3 additional parts, like this:

              http://www.4shared.com/photo/k0t3fUkz/cascode.html

              Most noticeable feature of HV bipolar transistor is the ridiculously low current gain (10 or even less), so massive base current is needed to turn it on (15 Ohm resistor in this case). Naturally, removing excessive base charge later to turn it off is a problem, some fast diode (HER307 or something similar) will do the job. Lower NMOS can be low voltage type, it will never see more than 12V across its D-S terminals.

              How good is this? Well, depending on coil parameters and pulse width, can be 40% or even more faster compared to, say, IRF740 alone. And this is why:

              Normally, MOSFETs tend to "zener" something close to 10-15% above rated voltage. And then, most of the energy stored in coil will end up as heat dissipated at switching device, lost in the form of highly nonlinear dissipation process. Far cry from ideal LCR network anyway, and in order to get rid of energy as fast as you can, something as close as possible to ideal LCR is needed. This configuration will tolerate much higher flyback voltage across the coil without "clipping"(1.5-2KV), resembling LCR as close as possible, enabling shorter sampling time.

              Cautionary words: if you want to modify any existing PI detector this way, be aware that you may blew up the original coil due to dielectric breakdown. Normally, insulation, cable, shielding, epoxy etc will routinely withstand 2KV DC, or even RMS at, say, power line frequency, but submicrosecond flyback pulse of similar amplitude (high rate of change)means death for insulation (i managed to blew up seahunter coil doing this, for example)

              Comment

              Working...
              X