Simple FET Driver

e:

the

ge

by

vel

or.

or

If you'd like to get involved in designing the motor control loop you woul d be VERY welcome. Right now they seem to have little skill in that area a nd are initially banging the motor with full current for some period at the beginning of the cycle. I'd love to hear you discuss this with the two gu ys designing the PID, your head would explode!

ned

is.

What part of what I've explained is not clear? Didn't I indicate the reaso n for wanting to use a push-pull arrangement of a CMOS logic device rather than the slow pull up on an open drain FET driver? There is 40 us of shoo t through. With a push-pull drive on the gates of the power FETs I expect this can be reduced to nanoseconds.

Are there any specific suggestions to the two schematics I uploaded?

Lab tests on what exactly? You mean you want to prototype the design? We ship things around. That's not a problem. Oh, I'm not in the same countr y as much of the team. Some in the UK, some elsewhere in the EU, one in Tu rkey... or India... I can't recall... then some in the US.

Well if you don't have people with the required skills then recruit them. I think I'd prefer not to get personally involved.

I'd use a purpose designed gate driver if there was a good reason to have very fast switching speed. I think you can do without that for a power on/off switch.

Consider a redesign of the power switches using this page as a reference.

I've seen more than a few projects cancelled because the hardware was done in one country and the software in another. Having the team distributed around the world isn't necessarily an issue as long as someone who knows what they're doing is in overall charge. I don't think there is much else I can do for you.

te:

ts

the

bly

eds

is

too

en

od

Ok, armchair engineering.

ich

ard

Yes, that's why I am looking at CMOS logic gates, not fast, not high curren t drive, but maybe good enough. I don't know if I'll find an adequate simu lation model though.

e).

ce

I don't know where you are getting this 8 volt number. You need to check t he data sheets again.

I don't see how that web page is useful. The intent is not to allow curren t flow in both directions but to prevent it. To prevent the body diode fro m conducting current a second FET is used. Then I would point out the circ uits on that page use a 10K pull up resistor which is what I am using. Thi s is the cause of the slow turn off time. They don't really discuss the ci rcuit other than at a fairly superficial level.

As opposed to what? You ask questions like this rather than making suggest ions other than tossing the whole thing out and starting over.

I had actually started with rather large resistors because of current drain on the battery, until I realized the only time they matter is when BOTH ci rcuits are off when the driver FETs are off and there is no current flow. So 10K got something that worked ok. As is true with many LTspice simulat ions change one component value and the circuit can't find it's starting po int. Right now the thing is stuck looking for a way forward and won't resp ond to the esc key. That's from changing the 10K values to 4.7K. lol The current doesn't get larger, but the time does narrow, but not so fast. Th at's why I'm thinking a pullup transistor would be a good thing.

Yes, but "quicker" may be better, but is it good enough? At some point the resistor passes as much current as the FET.

et

Independent as in how? They are independent from the state of the comparat or. It's even worse to have a time of no conduction to either power source . In reality, the DC in will only drop out when power is lost or something similar, so current into it is not a big deal. When power is restored it the time I'm worried about. But I don't know if it matters that the input is connected to the battery for 40 us.

il

ne

Yes, I have that impression.

In that case I won't bother responding to your other points except to point out that your schematics show AO6407 FETs, the data sheet for which is trivial to find. Under Absolute Maximum Ratings Vgs is listed as +/- 8V Maximum. If you want intermittent failures after production has started, just hit it with 12V.

e:

as

nt

it

Ok, my bad. I thought I had checked that but I got something wrong. The p arts on the board are something else, these were picked for the simulation because they are in the LTspice library. I get tired of fighting the tool to get third party models to work.

How about using two diodes? Too easy I guess.

That's ok, I've made worse design errors. I think resistors lower than 10K and a circuit similar to the one on that web site will switch fast enough if you pay attention to resistor power dissipation and Vgs max. R14 and R15 may not be needed or can be a lot lower than 10K. Try a simulation with 4.7K and

Yes a couple of diodes producing Vmain would stop Vin and Vbatt ever meeting each other.

e:

ue

ion

if

may

and

Element-14 in Australia (it used to be Farnell /Newark) has 306 different p arts in stock that fit CD40**. Everything in the first sheet is labelled a "best seller", so it does seem to be still in production. Most if it will probably go into legacy products, but I certainly ran into design problems all though my career where logic that could run slowly from an up to 18V su pply was just the solution I needed. Unlike the 555 which seems to persist because some designers are lazy and unimaginative, there weren't any better alternatives.

Yes I agree it's useful if you want a slow (by today's standards) system working directly on 12V or more. I managed never to design a 555 into anything. I've seen some strange circuits using a 555 including SMPS PWM controller.

e:

ue

ion

if

may

and

D3 is to prevent the comparator from being damaged by reverse power. The g uy designing this board added a part to protect everything from high voltag e and reverse voltage. He never saw an LT part he didn't like and the rest of the board is already rated for +40 volts, just in case! So a simple di ode should protect the comparator and everything else is good already I thi nk.

What web site? The pages you've pointed to don't actually give much inform ation. They are very general.

e:

hem

it

ing

Yes, they will do that. But they won't allow the rest of the circuit to wo rk and the reason why has already been covered. It's also very obvious if you look at what the circuit is doing.

In fact, the diodes are already there for the low current always on path. They work like a champ!

Actually, the predecessor circuit to this one had a FET based diode pair to combine the input and battery circuits. But that was with a higher voltag e input and now we are using a 12 volt input. Use diodes and the battery w ill always be powering the circuit until it runs down.

e:

ote:

wde

de

it

ssue

hit

ation

ool

gh if

K and

gic

parts in stock that fit CD40**. Everything in the first sheet is labelled a "best seller", so it does seem to be still in production. Most if it will probably go into legacy products, but I certainly ran into design problems all though my career where logic that could run slowly from an up to 18V s upply was just the solution I needed. Unlike the 555 which seems to persist because some designers are lazy and unimaginative, there weren't any bette r alternatives.

In this circuit the biggest problem I have is that some of the uses require the input to be higher than the supply which a FET can do well but a logic gate has issues with. Negative excursions are also a problem.

Me either.

If it's the only tool you've ever bothered to master, you can do lots of different things with a 555. It's been about thirty years since this was a useful skill.

I think some of us are having trouble fully understanding your system topology. A block diagram may or may not help. In any case, why does it matter if both switches are on for say 50uS, what is that going to hurt? If it could possibly hurt anything at all then add a diode so it can't.

Also I don't see how any harm could come to the op amp working as a comparator (some of them don't like that) if D3 is replaced by a short circuit. If D3 is really needed then MBRS140 seems overkill.

Your TLV333 will probably be fine as a comparator but in cases where it isn't, the advice is clearly to use a comparator if you need a comparator:

te:

ts

Another LTspice c*ck up. I didn't intend to use an opamp. It was supposed to be an open collector comparator. I copied the part from another design and didn't notice it was not a comparator. Thanks for that one. That exp lains a lot with the slow edge rate I'm seeing.

The harm is if the power input is reversed. The other components in the ci rcuit will be fine, but the op amp^H^H^H^H^H^H comparator would be damaged.

The design review is coming up. We'll see what others think of the LT orie nted design.

te:

ts

so

to

ge

Why would you use diodes and switches??? I don't know if 40 us of shoot th rough is an issue. I thought I made that clear in the original description s.

The input powers a buck/boost charger for the battery and powers the switch selecting the battery or the input power. Is that what you were missing? You talked about running the entire system off the charging circuit so I t hought you must have understood that, no?

I likely wouldn't think LTspice for this kind of circuit. If there was any doubt about correct operation then by now I'd have soldered the parts together and tested it in a lab, or equivalent parts with suitable legs.

If there's a risk of the power input being reversed I'd probably put a series shottky diode in the power line. Reverse power might damage more than your comparator.