Totem-pole gate driver VS Active pull down VS Fast Fet turn off

[dbanner]

I am of the view that flyback voltages can be a function of the soil medium as well as intrinsic circuit characteristics. However small, but it's there.
There were some Japanese fet transistors way back, l would be curious to see how they would perform in a pl circuit.

[Ferric Toes]

Scope ground is at +12V so - supply on gate is off. Gate goes from zero V(on)to -supply(off)in less than 40ns with the 3906. How much faster do you want it? https://www.geotech1.com/forums/atta...8&d=1615911700

In a paper by F. B. Johnson in 1956, it states "so long as the time constant of switch-off is less than about one tenth of the decay time of eddy currents in the object, the behavior of the eddy currents is practically the same as if the field had been removed instantaneously". It then goes into a lot of maths. I would say that provided the switch-off is at least a fifth of the object Tau then there is little benefit of going for more. Faster switch-off than necessary will cause higher flyback voltages, and depending on the device, more avalanche situations and more ringing.

Eric.

[green]

Hi Green is it possible to change the 3906 to a Darlington to speed up turn off

Scope ground is at +12V so - supply on gate is off. Gate goes from zero V(on)to -supply(off)in less than 40ns with the 3906. How much faster do you want it? https://www.geotech1.com/forums/atta...8&d=1615911700

[Carl-NC]

Hi Green is it possible to change the 3906 to a Darlington to speed up turn off

That made a fair difference in one of my designs.

[6666]

Agree. I had purchased some different Mosfets to try. Tested them to see if I could see a difference. Added 10 ohms in series with coil(coil to B, R=10)to prevent avalanche. 3904, 3906 driver, no added gate resistor.
IRF740 did avalanche, 400V rating. Others 500V rating. Coil 328uH about 1MHz resonance. Rd=1000 ohms, slightly overdamped.

Hi Green is it possible to change the 3906 to a Darlington to speed up turn off

[green]

My take from this discussion, is that it is very difficult to correctly measure the gate oscillation with the tools that we have.

Agree. I had purchased some different Mosfets to try. Tested them to see if I could see a difference. Added 10 ohms in series with coil(coil to B, R=10)to prevent avalanche. 3904, 3906 driver, no added gate resistor.
IRF740 did avalanche, 400V rating. Others 500V rating. Coil 328uH about 1MHz resonance. Rd=1000 ohms, slightly overdamped.

[Monolith]

Some tests with different gate resistor values turning Tx on and off. Guessing some of my problem is the circuit is hard wired. Different gate resistors effected oscillation across .1 ohm resistor at Tx on. Not much effect at Tx off.
Coil connected to A, avalanches. Should different gate resistors effect oscillation at Tx off. Would a good circuit board layout look a lot different?

I use a coil with different Tx and Rx coils. Noise level at integrator out seems to be the same whether Tx is on or disabled. What negative effect does the oscillation at Tx on or off have? Should I be seeing integrator out noise increase if Tx is enabled?

If your noise level is the same with TX on and TX off, I would say it is as good as it gets.
My take from this discussion, is that it is very difficult to correctly measure the gate oscillation with the tools that we have.

[green]

Some tests with different gate resistor values turning Tx on and off. Guessing some of my problem is the circuit is hard wired. Different gate resistors effected oscillation across .1 ohm resistor at Tx on. Not much effect at Tx off.
Coil connected to A, avalanches. Should different gate resistors effect oscillation at Tx off. Would a good circuit board layout look a lot different?

I use a coil with different Tx and Rx coils. Noise level at integrator out seems to be the same whether Tx is on or disabled. What negative effect does the oscillation at Tx on or off have? Should I be seeing integrator out noise increase if Tx is enabled?

[Tinkerer]

Gate resistor

Something about Mosfet gate resistors

[green]

Some recordings, IRF840. Coax with 50 ohm resistor picks up less noise than x10 probe. Middle left picture, coil connects to +12V. Shouldn't be any current thru .1 ohm resistor(some signal CH2, CH1 not good). Like Eric, I get less oscillation if Mosfet doesn't avalanche. Rd to +12V not coil. Not posting how to measure oscillation, posting the problems I'm having.
UCC27518 driver

[green]

Thanks for the replies. Don't have an oscillator. Thinking a steep response might work. Used UCC27518 output(connected to CH1 and CH2 only). 50 ohm resistor soldered to coax center wire.

[Mr.Jaick]

yes indeed
I have to make a T-shirt --> "know your instrument"
they can sometimes easily fool you, either it's the humble DMM, an oscilloscope or even a simple power supply!

here is a very good representation from Doug Ford about scope probes, it's a joy to read.
THE SECRET WORLD OF PROBES OCt09.pdf
warning: read this with Australian accent

[Qiaozhi]

using bare caox i.e transmission line is a tricky business
those ringings are reflections caused by a low impedance source which is the power rail in your case and high impedance load = 1M scope input

It is quite likely that the unexplained ringing is indeed due to transmission line issues. This is a big problem in high-speed digital design, or anywhere where signal edges change very quickly, because this emphasizes the behaviour of passive circuit elements. A very good book to read is: High-Speed Digital Design (A Handbook of Black Magic) by Howard Johnson and Martin Graham.

[Mr.Jaick]

reply#122

Still trying to measure oscillation. More recordings. CH1(normal scope probe) CH2(cut BNC cable soldered across .1 ohm resistor).
Four left pictures compare CH1 connected to scope ground or point A(oscillation is noise pickup?). Any suggestions on how to determine if oscillation on CH2 is actual current oscillation? Think it isn't but don't know to prove it.
Third row, labeled CH1 wrong. Should be point E(555 out), scope trigger for all recordings. Forth row, coil connected to 1 ohm resistor in series with .1 ohm resistor. Ch1 connected to scope common or point B(noise pickup on both?)

If oscillation CH2 is caused by BNC cable resonance, any suggestions on how to solve it or what is causing it?

using bare caox i.e transmission line is a tricky business
those ringings are reflections caused by a low impedance source which is the power rail in your case and high impedance load = 1M scope input
caox has quite high capacitance, for low frequencies it has high impedance and it's ok just remember you are putting a capacitor in the circuit by using coax(27pF / feet)
but when you get higher in frequency like your rise time it causes all sort of reflections, oscillations and crap due to your scope 1M input impedance and low source impedance
terminate the coax with a 50ohm termination(if coax is 50ohm) on the scope input to get rid of reflections i.e widening the bandwidth
you have to use series resistor with your coax to lower the loading effect if the source impedance is higher
for example:
if you're using 50ohm coax
use 450 ohm resistor in series with the core so you get 500 ohm overall impedance and also your signal will get divided by 10:1 at the scope so watch out for that
you can also use a 950 ohm resistor to get 1k input impedance so lower loading of your signal but 20:1 ratio is bad for your low level measurement
also try to get shortest tip length to the board to minimize tip inductance
my solution for the last one is to solder a coax connector to the board and then connect your coax to that
but it's not that important for this application
hope it helps a bit

[Ferric Toes]

reply#122

Still trying to measure oscillation. More recordings. CH1(normal scope probe) CH2(cut BNC cable soldered across .1 ohm resistor).
Four left pictures compare CH1 connected to scope ground or point A(oscillation is noise pickup?). Any suggestions on how to determine if oscillation on CH2 is actual current oscillation? Think it isn't but don't know to prove it.
Third row, labeled CH1 wrong. Should be point E(555 out), scope trigger for all recordings. Forth row, coil connected to 1 ohm resistor in series with .1 ohm resistor. Ch1 connected to scope common or point B(noise pickup on both?)

If oscillation CH2 is caused by BNC cable resonance, any suggestions on how to solve it or what is causing it?

It is difficult to find out what is going on at switchoff. I found out that there is a lot of variation between different Mosfets. A lot of ringing disappears if you avoid avalanche. If you are in avalanche, put some series resistance in the coil circuit to reduce peak current. It is worth substituting a thick film resistor for the coil and seeing if it still rings. I tried a 5W resistor that is not metal film and it has some inductance because I was getting about 20V of flyback voltage. Thick film gives none. I still get one overshoot spike on the scope across the 0.1ohm resistor. Can't explain that as yet. By the way, I only ground one of my four scope probes as all the BNC sockets are connected together in the scope. This avoids a ground circuit ring. I'll post some plots by Monday.

Eric.