2

Author Topic: Square wave pulse (0-2.5kV)&(0.2-100msec)  (Read 8266 times)

0 Members and 1 Guest are viewing this topic.

Offline ellendortTopic starter

  • Beginner
  • *
  • Posts: 6
  • Helpful? 0
Square wave pulse (0-2.5kV)&(0.2-100msec)
« on: March 23, 2006, 07:36:03 PM »
I intend to use my power supply (3000V - 130mA).

I need some guidance and reference on how to design a pulse generator to produce square wave pulses where the pulse width (0.2-100msec) and the pusle height(0-2500V) is adjustable.
Perhaps maybe by using a microcontroller or by charging a capacitor or high voltage amplifier.

My aim is to apply a manually triggered square wave pulse on my sample load resistance in between 2 electrodes with electrode distance 0.1cm or 0.2cm or 0.4cm. The voltage and my load samples are generally divided into two category as below,

1) Voltage: 0-2500V
    Load sample resistance: 2000ohm - 5000ohm

2) Voltage: 0-500V
    Load sample resistance: 20ohm -200ohm

Please advice

Offline Admin

  • Administrator
  • Supreme Robot
  • *****
  • Posts: 11,666
  • Helpful? 169
    • Society of Robots
Re: Square wave pulse (0-2.5kV)&(0.2-100msec)
« Reply #1 on: March 23, 2006, 09:36:22 PM »
Ok then what you want is a MOSFET:
http://www.societyofrobots.com/electronics_tutorial.shtml#mosfet
Ok I admit that tutorial kinda sucks, been too lazy to rewrite it though . . . here is another tutorial that uses MOSFETS:
http://www.societyofrobots.com/schematics_h-bridgedes.shtml

Basically you apply your signal to the gate of the mosfet, and put the other two leads (drain and source) in serious with your 3000V power supply. Basically the gate acts as an electronic on/off switch for your 3000V power supply.

So all you do then is apply your square wave pulse to the gate. You can do this with a simple function generator if you have it, or with a 555 timer IC, or with a microcontroller with very simple programming. Simple as in it is the same as the code to make an LED blink fast.

You probably should amplify the voltage (with an op-amp) from your signal to the mosfet gate to like ~20V just to make the mosfet more efficient. Higher gate voltages means higher efficiency and less overheating.

quick schematic:


                                        3000V Source
                                                |
signal generator ---- op-amp ----- mosfet
          |                                     |
          |                           Load resistance
          |                                     |
      Ground                            Ground

Just one side comment, you do realize that your circuit will be drawing an insane 1250 watts?!? (power=2500V*.130A=1250W)

Anyway let me know if you have questions, I kind of breezed by on this and didnt explain it that well.

Offline ellendortTopic starter

  • Beginner
  • *
  • Posts: 6
  • Helpful? 0
Re: Square wave pulse (0-2.5kV)&(0.2-100msec)
« Reply #2 on: March 26, 2006, 10:01:20 PM »
I will charge a capacitor and discharge through cascaded MOSFET to generate square wave pulse.
Any reference assume i wants to cascade the MOSFET(600V) in series for switching?

As i mention before i need to manually triger the pulse. 1 or 2 pulse or perhaps more than that. How do i control this?
Can i use this circuit for the manual triggering and to drive the MOSFET to generate square wave?
http://www.edn.com/article/CA236418.html

But i'm not sure the frequency adjust here as to my application because different square wave design has different controllable fequency and as for the pulse width will it be possible 0.2-100msec?

Please advice

Offline Admin

  • Administrator
  • Supreme Robot
  • *****
  • Posts: 11,666
  • Helpful? 169
    • Society of Robots
Re: Square wave pulse (0-2.5kV)&(0.2-100msec)
« Reply #3 on: March 26, 2006, 10:39:08 PM »
I will charge a capacitor and discharge through cascaded MOSFET to generate square wave pulse.
Any reference assume i wants to cascade the MOSFET(600V) in series for switching?

Hmmm I dont understand your first question.

If you want that much variable control, I still recommend using a microcontroller. You can have a single manual button attached to trigger a pre-programmed event, such as your square wave.

I dont know anything about cascaded MOSFET's so I cant help you there.

Offline ellendortTopic starter

  • Beginner
  • *
  • Posts: 6
  • Helpful? 0
Re: Square wave pulse (0-2.5kV)&(0.2-100msec)
« Reply #4 on: March 27, 2006, 08:49:35 PM »
Pls refer to the attached figure.
100Kohm*1uF=Approx. 100msec

How can i control the time so as the triggered pulse will stop at the preset value which is controllable from 0.2-100msec.

Pls advice.

Offline Afroman

  • Expert Roboticist
  • Full Member
  • *****
  • Posts: 50
  • Helpful? 4
Re: Square wave pulse (0-2.5kV)&(0.2-100msec)
« Reply #5 on: March 28, 2006, 11:45:22 AM »
Putting several fets in series won't do anything other than blow them up.
For these kind of voltages, what you want is a suitable IGBT
http://www.fairchildsemi.com/sitesearch/fsc.jsp?command=eq&attr1=AAAFamily&attr2=IGBT+Discrete
or an SCR
(Search digikey.com)
Or make a mechanical switch somehow.
What are you trying to do anyway? It sounds like you are using a NST?

Offline ellendortTopic starter

  • Beginner
  • *
  • Posts: 6
  • Helpful? 0
Re: Square wave pulse (0-2.5kV)&(0.2-100msec)
« Reply #6 on: March 29, 2006, 01:32:07 AM »
Why would they explode? assume i use 5 MOSFET in series each rated at 600V? My max. voltage is only 2500V.

IGBT's are very expensive.

I need to apply a constant voltage in between 2 electrodes. In the other word need to apply electric field to my load samples.

Depending from my load samples i know for sure they are categorized into 2 as below,

1)   In general my cell resistance vary from(Depending on the cell): 2000ohm-5000ohms
      Voltage(Depending on the cell): 0 - 2500V
      Time(Depending on the cell): 0.2-10msec

2)   In general my cell resistance vary from(Depending on the cell): 20ohm-200ohms
      Voltage(Depending on the cell): 0 - 500V
      Time(Depending on the cell): 0.2-100msec


Please advice

Offline Afroman

  • Expert Roboticist
  • Full Member
  • *****
  • Posts: 50
  • Helpful? 4
Re: Square wave pulse (0-2.5kV)&(0.2-100msec)
« Reply #7 on: March 29, 2006, 04:18:56 PM »
Putting mosfets in series simply doesn't work like that. It is not the same as capacitors. If one mosfet is off while all the others are on, kablam, you have the entire 2500V across 1 fet.

For your 2500V thing lasting only a few milliseconds and doesn't have a heavy load... just use an automotive ignition coil as a transformer to step up the voltage or something. Or build a high voltage DC supply with a large capacitance, then trigger it with an SCR.

Offline ellendortTopic starter

  • Beginner
  • *
  • Posts: 6
  • Helpful? 0
Re: Square wave pulse (0-2.5kV)&(0.2-100msec)
« Reply #8 on: March 29, 2006, 07:25:07 PM »
Assume 5 MOSFETS are on. The entire 2500V across them and the 6th MOSFET will be used to trigger?

On building a high voltage DC supply with a large capacitance,
i was thinking of using
High voltage - single capacitor (25mfd/3.5kV)
Low voltage - Multiple capacitor's in parallel (390MFD/500V)

Please advice

Offline Afroman

  • Expert Roboticist
  • Full Member
  • *****
  • Posts: 50
  • Helpful? 4
Re: Square wave pulse (0-2.5kV)&(0.2-100msec)
« Reply #9 on: April 02, 2006, 07:32:13 PM »
I am telling you for the last time, you cannot put fets in series like that! You will start a fire.

Offline ellendortTopic starter

  • Beginner
  • *
  • Posts: 6
  • Helpful? 0
Re: Square wave pulse (0-2.5kV)&(0.2-100msec)
« Reply #10 on: April 02, 2006, 07:59:41 PM »
But i need a "square wave" which implies turning off the pulse at some point.  Once triggered an SCR stays on until
the current drops below its holding current, usually about 10mA, so my 2 to 25A pulse would continue for many time constants,
gradually dying away to nothing. As for the IGBT's they are too expensive.

Please advice

Offline Afroman

  • Expert Roboticist
  • Full Member
  • *****
  • Posts: 50
  • Helpful? 4
Re: Square wave pulse (0-2.5kV)&(0.2-100msec)
« Reply #11 on: April 02, 2006, 08:05:58 PM »
If IGBTs are too expensive, it sounds like the only choice you have will be to make a suitable mechanical relay switch yourself. There really is no way to get the kind of precision you want without paying for it.

 


Get Your Ad Here