schmitt trigger / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

schmitt trigger

1 2 next

Move Thread

LAN_403

IamSmooth

Tue Aug 25 2009, 08:36PM

Registered Member #190 Joined: Fri Feb 17 2006, 12:00AM
Location:
Posts: 1567

Here is a link to a schmitt trigger using a comparator

This trigger goes from HIGH to LOW when the signal goes above V2; it goes from LOW to HIGh when it goes below V2prime

Is there a variation that goes from LOW to HIGH when the signal goes above V2, and HIGH to LOW when it goes below V2prime?

I guess one way would be to feed the output into the base of an NPN common emitter and take the output at the collector. Any other suggestions?

CT2

Tue Aug 25 2009, 09:01PM

Registered Member #180 Joined: Thu Feb 16 2006, 02:12AM
Location: Ontario, Canada
Posts: 187

I think if you switch the inverting and non-inverting inputs you could achieve this result, however I have not looked at the schematics in any detail, so wait for someone else to comment. If anything you could always just put an inverter on the output.

Dr. Slack

Wed Aug 26 2009, 02:54AM

Registered Member #72 Joined: Thu Feb 09 2006, 08:29AM
Location: UK St. Albans
Posts: 1659

Two comments

The scmidt comparator bascially has differntial inputs, so using ref to -ve and signal to +ve (through a resistor large compared to the feedback resistors) results in non-inverting behaviour. However, as -ve input is high impedance so easier to use, and the +ve input already has an implicit rail-related reference connected to it by the feedback, most people only use it inverting.

It's bad practice to follow a wimpy stage which uses feedback (like this comparator) with a load which needs considerable current to drive (like an IGBT gate) (cookies for why), so a current gain stage is really needed after a scmidted comparator. This can conveniently be inverting.

Steve Conner

Wed Aug 26 2009, 08:51AM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

I'll just take this opportunity to point out that gate driver chips such as the UCC37321/37322 are available in both inverting and non-inverting flavours. :P

Also, your bang-bang control is not complete until you put a low-pass filter somewhere in the loop to limit the maximum switching rate. Without this, it can still take off at an arbitrarily high switching speed and burn itself out.

If you put a RC lowpass just before the Schmitt trigger, then the maximum switching rate will conveniently be set by the RC time constant and the Schmitt's hysteresis band.

IamSmooth

Wed Aug 26 2009, 12:01PM

Registered Member #190 Joined: Fri Feb 17 2006, 12:00AM
Location:
Posts: 1567

Thanks Steven. I was actually trying to find a method of slowing the rate. Right now I was only considering widening the hysteresis band. The internal resistance of my generator is about 5.5ohms. If I trigger at 240vdc with a 30ohm load the voltage will immediately drop to close to 205v due only to the voltage divider effect.

So you are saying that even if the voltage falls below my trigger voltage the RC filter will not let the state change until the time constant has elapsed? I guess this is because of the time it takes to charge the capacitor.

Steve Conner

Wed Aug 26 2009, 12:26PM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

That's right, yes.

Note that you don't want to slow the gate drive, you want to delay it. The IGBT must still switch snappily to prevent lingering in the linear region, but it musn't switch too often.

Dr. Slack

Thu Aug 27 2009, 12:46AM

Registered Member #72 Joined: Thu Feb 09 2006, 08:29AM
Location: UK St. Albans
Posts: 1659

The internal resistance of my generator is about 5.5ohms. If I trigger at 240vdc with a 30ohm load the voltage will immediately drop to close to 205v due only to the voltage divider effect.

Will it? What about the effect of your DC bus capacitor? I'm sure your inverter will have one on its input, even if you haven't added extra externally. This will source the extra current and your terminal voltage will drop smoothly.

You may need to consider the ripple current and consequent heating that this capacitor will experience in switch mode shunt limiting. If your shunt is pulling 240/30 = 8A, and your generator is supplying 4, that 4A rms sloshing in and out of the DC bus capacitor. That's not a big deal with suitable caps, but it could be with unsuitable ones. What is on the input of your inverter?

Plastic film ones to cope with HF filtering? - OK, but you may want to add extra capacitance for smoothing
Largish PSU types? - probably OK
Small 'lytics - maybe worth checking their ripple current, and adding extra for smoothing

IamSmooth

Thu Aug 27 2009, 02:24AM

Registered Member #190 Joined: Fri Feb 17 2006, 12:00AM
Location:
Posts: 1567

Yes, I do have 1000uf/900v of electrolytic capacitors to let the voltage fall smoothly. I did not mean to imply that it would suddenly change values.

Dr. Slack

Thu Aug 27 2009, 02:31AM

Registered Member #72 Joined: Thu Feb 09 2006, 08:29AM
Location: UK St. Albans
Posts: 1659

OK, it must be a new use of the word immediately that I am not familiar with

4amps from 1000uF is about 4kV per second. If your hysteresis band is equivalent to (say) 20v at the DC bus, then it's going to self-oscillate at in the 10s of Hz, without any extra LPF.

The voltage *will* drop immediately due to capcitor ESR, but for 1000uF I assume that's going to be in the order of mohm or 10s of mohm, so by not very much at all.

Steve Conner

Thu Aug 27 2009, 09:00AM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

There's also the stray inductance of the wiring though, which can be pretty high.

When the switch comes on suddenly, with fast gate drive, it can easily drop a sizeable proportion of the DC bus voltage in L*di/dt across the wiring. This would cause any comparator to turn itself immediately back off, whereupon the resulting kickback of stored inductive energy would send the voltage sky-high, turning the comparator straight back on again, and so on, resulting in a vicious high frequency oscillation and extreme voltage spikes.

This is why I recommend a low-pass filter in the loop, to stop that kind of thing. A nice big capacitor strategically placed in the dump load unit wouldn't hurt either.

1 2 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob