Re: It works! Solid state driver
From: Alan Sharp[SMTP:100624.504-at-compuserve-dot-com]
Sent: Tuesday, December 09, 1997 5:51 PM
Subject: Re: It works! Solid state driver
>I have to tell you and the people on Chips list that the circuit has had
>a catasthophic failure due to what Mr. Buylund called "Flux Walking".
>This is a push-pull config. and is prone from what I've read to this
>sort of failure. The circuit is very stable at up to, and including
>24vdc. But beyond that the power transistors begin to share the current
>unequally. This causes one of the transistors to short out due to
>excessive current draw. The only way to correct this is to use one of
>the bridge configurations, or to use a flyback configuration at lower
>voltage, or duty cycle. I am in the process of rebuliding from the ashes
>at this time! The energy in the shorted transistor was so high it
>actually punched a hole in the metal casing of the transistor the size
>of a pencil lead! It also took out the pwm, and driver.
Yes been there done this.
Push pull can be made to work - its got its problems but so has
My pushpull runs at 150V, 5A to give 14 inch arcs -
using an interupter so peak currents are many times
this. Circuit is a 3825 PWM chip, a pair of the 9A output
buffers driving 3 IRF740's on either side.
Coil is a 5 litre bucket wound with 0.4mm wire
Toriod is about 18" diameter 3" high.
1) the two halves must be as equal as
possible - the PWM chips give a symetrical output.
2) Current limiting is essential - most PWM chips have
a SD - short detect input. You can directly sense the voltage
across a low value non inductive resitor in the common ground
path - followed by an RC filter to take out the intial spike.
Start sensitive and gradually desensitise.
I'm now experimenting with a current transformer for the half bridge.
3) More primary turns in the output transformer to reduce the
risk of saturation. For the primary strands of insulated thin wire
have less AC resitance than a single thick wire.
4) A pair of 1A schottky diodes reverse biased
PWM +Vc ---|<---- output-----|<------ ground
The output refered to is the driver output for the FET
gates. A 10 ohm resistor also helps.
The gate drive can ring - going below ground -
chips don't like this.
Resistor and diodes also help in the event of a
Alan Sharp (UK)
Rev Alan Sharp BSc BD
St Ninian's Church of Scotland
Doing hivoltage services in Corby, Northamptonshire.