NTC

I think you still have a problem. Once the fet turns off, you have the - input heading for zero volts. that will keep the fet off.

Have you considered the version I mentioned that parallels the switch with a resistor, to keep some voltage across the heater and its current sensing resistor even when the switch is off. Of course, you also have to have the other divider across the same switched voltage.

I have simulated this configuration in LTspice, with a heater that has a lot more negative tempco than yours does, and it holds a constant temperature nicely. I used a resistor ten times the heater resistance in parallel with a PNP transistor at the positive rail.

think harder, dude. If you interpret "do you understand dumbfuck" as "do you understand ME" (IOW read it as you referring to yourself as dumbfuck) then Johns comment is, indeed, ironic, and quite funny.

Cheers Terry

You still sound as dumb as a post.

thats a good trick with ASCII text. At least I know how to use a comma.

Cheers Terry

John It's very easy to see that your solution would work. Resistor keeps some current going to the balancing loop so the comparator can operate.

Also, one could physically attach the high-side resistor to the heating tube and so not lose the power from it.

Now, take another look at the circuit I just posted. You say that when the fet turns off, the negative input heads for zero volts. Not so, at least not right away: Mosfet on, cap at essentially zero volts. Then the mosfet turns off, drain goes high. At the instant this happens, the discharged cap looks like an open circuit and when the drain goes high it pulls the negative input high along with it. R4 is magnitudes greater than either the thermistor or Rs, so the negative input is pulled essentially all the way to the positive rail, then the cap discharges through R4. The cap has to discharge from Vcc to a volt or even quite a bit less than that, depending on Rs, so the power mosfet stays turned off for several RC time constants. It's always going to turn back on after this. No matter how hot the thermistor may be. So the limiting factor will be the duty cycle. When the thermistor is hot I'd have a low but not zero duty cycle heating it.

Your solution is probably better. Before building it I'll look at comparator and op-amp data sheets to see what's bothered the least when the common mode slews up and down, during the switching. Thanks, Kell

Right, the VEGETABLE PEOPLE (cue spooky music -- it's Halloween) couldn't seem to figure that out when they were designing their heater. I got in touch with DBK and I'm looking into their liquid flow heaters. The guy on the phone, Steve Smith, said they've sold some to people for the application I'm considering. However, he said the self-regulating temperature of the heating element in them is 180 Centigrade. I said that sounds pretty high. He's going to see if he can get some data and e-mail it to me. Is it really possible that these liquid flow heaters have such a mondo inner temp gradient or so much loss from surface radiation that they need the heating element to get up to 350-360 F?

Not from where I sit.

John Fields a écrit :

Sure. Want an example?

ta1 shi4 ben4 yu2shi4 ta1 bu4 dong3

Cheers Terry

Now, now. Be nice.

Best regards, Spehro Pefhany

The biggest problem I have with this circuit (aside from it not controlling the temperature very well) is that it has a balance of positive and negative feedback applied to the opamp, and exactly what components you have in mind is not clear. I had a hard time picking values that gave the multivibrator operation you describe. Lots of combinations produced some sort of linear control with lots of mosfet heat.

Hi Spehro,

I knew you'd understand that. To be fair, I had to read JF's post a couple of times to figure it out.... but it was funny.

Ni3 you3 mei2you3 han4zi4 ?

Cheers Terry

Wo3 neng2 shi3 yong4 han4zi4? Yi1 dian3.

They're intellectually interesting, and work to thoroughly disambiguate the seriously overloaded phonetics.

Here's a somewhat depressing, if realistic, rant on the language:

Best regards, Spehro Pefhany