How to quickly turn on/off N-FET switching high side of 55V?

Spoil sport!

I don't think that would work for this application - it'll be subject to 10s of gs of acceleration in all directions. I don't think relays can handle that, but I guess there could be some special purpose ones out there. Also, I'm vertically constrained - I have about 2x8x8cm for this guy - and most relays (of course not all, but most) are pretty tall.

-Michael

An alternative to IR's photovoltaic driver is Vishays LH1262CAC .

They also make a new one with higher Isc but I cant find it anywhere. This one vo1263aa .

The LH1262 is available at Newark in through hole, they did have over

IR has some good literature here (you have to send away for it though)

Dioncs also makes them with pretty good drive current but again none of the regular distributors seem to stock it.

I like all sorts of optocouplers, dc to GHz, LED, laser, silicon, GaAs. I like SSR's too, which could be used here for beefier drive.

Michael later noted that the load is a motor, occasionally switched, so the fet will be turning on into inductance. Piece of cake, dissipation-wise.

Mouser has them starting around $6 or so. Given the fets, motors, power supplies, all that, another 6 or 8 bucks is invisible in the noise.

It's not a painful little problem. It's not a problem at all. Why don't you be positive about suggestions, rather than combative and defensive? Nobody posts fully-engineered designs here, well, except JF maybe.

But, since you're only turning

Frowned upon by who?

John

Hi Winfield - I really like that IRF photovoltaic isolator chip. I saw those a while back, but had completely forgotten about them.

Can you tell me, would this circuit be faster (or better in some way?) than the circuit John Fields posted? For some odd reason it sits better than me, I guess because using isolated parts always seems like the easy/lazy way out :). I guess I'm a bit of a masochist in that way...

-Michael

OK, keep the dc/dc converter, but drive the fet gate with a tiny spdt relay. That ain't bad. Gate rise/fall times will be insane, so add a series gate resistor.

John

I didn't take "painful" as derogatory; I just understood Win's comment to mean the DC-DC itself still needed designing, which would involve some additional effort.

I like the photo ideas and dc-dc both. All photo & semi is sexy somehow.

For more frequent switching rates than our OP needs, I was going to suggest directly driving the big FET with a rectified fly-back pulse & turn-off with an opto or some such might be fun.

Oops, piece'o'cake on the rising edge. Payback on the falling edge.

How would you like being criticized by the author of The Art Of Electronics?

But in a situation like this, big-buck power parts and probably one-off, it's easier to buy a dc/dc for $6 or so.

The dc/dc could *be* the gate driver, with a dirty trick for speedup maybe.

After the first flyback, you could just bang it now and then to keep the gate charged.

A second transformer could turn on a smaller npn or fet to discharge the gate, plus a resistor across the zener.

I've combined transformers (for high current) with a wimpy pv isolator (for dc).

John

On the Hubble Telescope platform support thingy I did years ago, I used transformers with a DC-restore circuit to drive the HexFETs yet allow PWMing.

Likewise on the satellite launch spinner for Honeywell Satellite Systems Division.

...Jim Thompson

I have a great book, The Hubble Wars by Chaisson. Whan they couldn't get it to focus, they had excuses then meetings and then investigations and then screaming. At one meeting, when they finally figured out that the mirror was ground wrong, one of the optics designers stepped out into the corridor and vomited.

An amateur telescope maker, with a candle and a straight-edge, could have spotted it.

John

I stand corrected, I should have said "John likes optoisolators" I was thining of the many sketches you've posted using them for various unusual things, like HV amplifiers. Do you remember those?

Yes, although turn-off protection may be in order.

Yes.

Hah, now just who's being combative and defensive?

I often do, because it's fun to work things out, and more often than not something changes as one goes from a nice-sounding concept to a circuit with specific part values. You know that experience well, I'm sure.

I'm also more likely to work out the details and post them if it's something I'm interested in for my own purposes, and then the usenet archive becomes something of a notebook. I'm continually surprised when I do a Google search on some exotic topic, seeking information, and right away up pops a post I made 7 years ago, or whatever, with all kinds of detail worked out. I'm surprised first of all because I'd forgotten working on the topic, and second sometimes because the post looks to be much better than anything I could have written. The implications of that are a little bit disturbing.

That, clearly I hope, was a joke. If a subtle, wry joke is made, isn't it wrecked by adding a smily?

However, the IR and Vishay opto-voltage parts are truly cool and useful, especially for switching huge MOSFETs that aren't available in the small SSR IC packages, but they do suffer from excessively-slow turn on and turn off, e.g., 5uA into 2000pF, etc. Combining one of those with your opto-isolator idea was appealing to me, so I pursued it in detail. I would have thought this was paying you a compliment. But if you don't see it that way, I take it back, by all means, and ask for your forgiveness.

Hi Winfield - a couple more questions for you: What is the purpose of the resistor across the bottom optoisolator? Also, you mentioned that the IRF1407 is a large die mosfet. How can you tell this? The reason I ask is that I am interested in using a FET with a smaller on resistance, but I want to make sure I know what to be looking for. Also, why not only use one inverter and drive the B optoisolator directly? I'm assuming it's because you want the A optoisolator to switch before the B optoisolator, but why?

By the way - wonderful weather today, right? I'm very tempted to just pull up a couch and sleep at work tonight.

Thanks so much for all your (and everybody else's) help, I really appreciate it.

-Michael

Well, that's John's part, but I can answer for him. That's just to insure the MOSFET is off, and stays off, when the 5V logic supply is off.

It's necessary to supply a reasonably-high current both to turn ON and to turn OFF the MOSFET's gate. So you have to pull both directions. In John's idea, either A or B is always on.

I may not have a choice. The cars are stuck on Land Blvd outside, and I wouldn't be able to even get out onto the street. People are out of their cars, cleaning snow off their windows, etc., and it's coming down hard. I was going to leave over an hour ago, but it's been stalled like that the whole time.

Well, we're enjoying ourselves. You're getting a half dozen good ways to do it, so have fun. But if this is for a serious real-world high- production project, and not just a one-off, the maybe you'd better get some real help involved in the decision making, instead of all of us usenet playing-around wannabies! :-)

Oops, the cars are moving again outside, so I'd better get going.

Nobody will forgive you until 3e is in the bookstores.

John

Is there one being worked on? I hadn't heard of it, though it would seem that it is due!

-Michael

Ah - that makes good sense.

I understand the necessity of needing to be able to quickly sync or source current into our out of the gate, but it looks like the circuit has a delay added in between A and B in the form of the inverter. I mean, couldn't the switching circuit look like below?:

TTL | | | ------+----|>o-----+ FOD617D (2) inverter optoisolator

With the way you have it drawn, it seems to me that there must be some reason to switch A before B, but I'm not sure what that would be, especially considering that the delay between switches would be so small.

It's not for a production project - just a research project, so no fear :)

Lucky! I'm still stuck.

-Michael

I seem to recall I first heard that Win was already "working on it" sometime shortly after I'd purchased the 2nd edition as an undergraduate in the not-quite-mid-'90s... :-)

I get the impression that Win had a lot more spare time when he and Paul were working on the 1st and 2nd versions than he does today... victim of his own success, I guess!

and here is another idea for application of the IR and Vishay opto-voltage parts :-

Reverse Bias supply for a detector diode !

Jure Z.

Good idea. Has anybody verified their internal high-side charge pump?