Mosfet Heating Circuits

[Harvey]

Fine Tuning the Gate

Place a 2 Ohm section ( a little less than 30cm?) of Nichrome wire (same as used to build the resistor) in series with the Gate Pot.

Make a flexible jumper of 16 AWG wire with an alligator clip on one end - it should be just a bit longer than the Nichrome wire to allow the Alligator clip to be turned 90° and still reach the end of the wire. Secure the other end to the nichrome wire by a suitable method. You now have an infinitely variable 2 Ohm precision resistor in series with the gate pot.

Use the Gate pot for course adjustment, and then slide the alligator clip along the nichrome to fine tune. It would be best to start with the clip in the center of the nichrome wire so you can either increase or decrease the resistance as required from that point.

This method has not yet been tested on this circuit, so please document your results for other readers to benefit from. I have used this method for calibrating high current shunts in test equipment, only in my case I used copper wire due to the high currents involved. So I know the method is a very accurate means of arriving at the desired resistance.

Cheers!

Harvey

[ashtweth]

Gad thanks for posting and helping great to see you back man and thanks Harvey for helping.

[gadh]

Hi Harvey

Timebase: upper - 200v/div. lower - gate: 5-10v/div, shunt: 500mv/div.
(depend on image, i do not recall which of them now. but the gate amplitude was close to 10 since the battery was 9v and the resistance was low)

Wire length - as in Glen's diagram (90cm or so)
wire Gauge - i think ~14AWG (1.5-2mm in diameter, suits for house power 16amps)

see also my next post - reply to the next post of yours.

[gadh]

(for Admin: i tried to edit my last post but the changes were not saved ! so i repost here)
Hi Harvey

Thanks 'ashtweth' for your encouragement ! (was is your name ? not clear here...) wish me luck...

Timebase: upper - 200v/div. lower - gate: 5-10v/div, shunt: 500mv/div.
(depend on image, i do not recall which of them now. but the gate amplitude was close to 10 since the battery was 9v and the resistance was low)

Wire length - as in Glen's diagram (90cm or so)
wire Gauge - i think ~14AWG (1.5-2mm in diameter, suits for house power 16amps)

About the 2ohms wire - why my solution does not suite ? (10ohms in parallel with the gate pot.) i think gate value of > 10ohms is irrelevant to the test anyway.

About the battery measurements - i also did a discharge/recharge with my charger and i saw that for Pb batteries the error margin was 5% (1205 dis., 1255 chg.), and for NiMh (9v battery) it was about the same when doing both actions in the same day. when recharging and leaving the battery for 1 day, it loses about 5-10% of its charge , plus more 1% per day. (for Pb its about 3% per month, so i'de ignore the losses for this type). But also i'm going to test with a car headlamp as you suggested.

IIRC you did not answer my question about connecting the 555 to 12v battery and seeing no resonance at all. do you have an idea on this too ?

Thanks
Gad

[Harvey]

(for Admin: i tried to edit my last post but the changes were not saved ! so i repost here)
Hi Harvey

Thanks 'ashtweth' for your encouragement ! (was is your name ? not clear here...) wish me luck...

Timebase: upper - 200v/div. lower - gate: 5-10v/div, shunt: 500mv/div.
(depend on image, i do not recall which of them now. but the gate amplitude was close to 10 since the battery was 9v and the resistance was low)

Wire length - as in Glen's diagram (90cm or so)
wire Gauge - i think ~14AWG (1.5-2mm in diameter, suits for house power 16amps)

About the 2ohms wire - why my solution does not suite ? (10ohms in parallel with the gate pot.) i think gate value of > 10ohms is irrelevant to the test anyway.

About the battery measurements - i also did a discharge/recharge with my charger and i saw that for Pb batteries the error margin was 5% (1205 dis., 1255 chg.), and for NiMh (9v battery) it was about the same when doing both actions in the same day. when recharging and leaving the battery for 1 day, it loses about 5-10% of its charge , plus more 1% per day. (for Pb its about 3% per month, so i'de ignore the losses for this type). But also i'm going to test with a car headlamp as you suggested.

IIRC you did not answer my question about connecting the 555 to 12v battery and seeing no resonance at all. do you have an idea on this too ?

Thanks
Gad

Hi Gad,

I have found that edits requiring more than a single character will not work unless you use the 'Advanced' button.

Regarding the 'Timebase', this is the Horizontal Deflection Time/div setting (see Item 29 on pages 2-7 & 2-8 in this manual:Tektronix 465b Manual)

Thank you for the wire info - this may have played an important role in Glen's tests due to the inductive reactance of the wire at his frequencies. We may have to calibrate this length to give similar results at your frequencies. We will address this later if necessary.

The reason that the 2 Ohm wire is helpful, is because it is a smooth surface that the clip can be slid along. The POT is not a smooth surface, it is a coiled wire and the wiper tends to jump from one coil to the next. Therefore, the closest resolution you can obtain is in increments of a single coil circumference inside the pot. For example, if the pot as a 1mm coil diameter, the best resolution you can obtain is the resistance of 3.14mm worth of wire in the pot. Depending on the resistance in mm that could be several tenths of an ohm. This is what makes it hard to adjust. So the smooth surface fixes that problem.

The headlamp test in my opinion is a good way to test the battery discharge and recharge values.

Also, in addition to the test I outlined, a fully charged battery should be able to have the headlamp attached with no appreciable reduction in voltage for at least several seconds after attaching it. Testing a battery under load in this way is one of the best ways to ensure the voltage being read is not just a 'fluffy charge' of static voltage on the terminals and plates. The 'fluffy charge' will be dissipated almost instantly after attaching the headlamp. I have personally witnessed as but as an instantaneous 5V drop by this method on 800aH truck batteries that were thought to be fully recharged but were not.

I did address the 12V supply to the 555 problem briefly in post #175 4.b (http://www.energeticforum.com/induct...tml#post117908) Since Aaron, Glen and myself have all obtained the aperiodic oscillation using the 12V timer supply, it would seem that the lack of this oscillation in your setup is unique to your arrangement for some reason. I would need to take some time and analyze all the differences involved to pinpoint the exact cause. I hope it is something simple like a disconnected wire, but if not then we will have to ascertain why.

[gadh]

Regarding the 'Timebase', this is the Horizontal Deflection Time/div setting (see Item 29 on pages 2-7 & 2-8 in this manual:Tektronix 465b Manual)

I know what is timebase, but obviously i was confused... Ok, so the timebase is usually 5us/div, and in the 'zoom' images its 2us. which gives us 140khz i think in the last test. my overall range of oscillations , disregarding the load amplitude, just playing with the freq. - can vary between 80khz and 300khz max.

I did address the 12V supply to the 555 problem briefly in post #175 4.b (http://www.energeticforum.com/induct...tml#post117908) Since Aaron, Glen and myself have all obtained the aperiodic oscillation using the 12V timer supply, it would seem that the lack of this oscillation in your setup is unique to your arrangement for some reason. I would need to take some time and analyze all the differences involved to pinpoint the exact cause. I hope it is something simple like a disconnected wire, but if not then we will have to ascertain why.

maybe i was misunderstood or i did not understand you correctly - i think all of you used another 12v battery for the 555 (as i used 9v for it) so we have 3 batteries overall - 2 for load, 1 for 555. what i'm trying to do is to connect the 555 to one the existing 12v batt. of the load so i would use only 2 batt. overall. is that what you Aaron and Glen did ?

Thanks,
Gad

[Harvey]

Hi Gad,

For image #3 one full cycle takes up ~17µs if your setting was 5µs/div. According to my calculation that is around 58.8 kHz. Thanks

As regards the batteries, Aaron and I both used the same batteries that the load had been connected to while Glen used a completely separate 12V battery. I used the configuration shown in my earlier post with the LED's, and I don't recall precisely how Aaron connected his. Aaron and I both did tests at 12V only as well.

One thing I did find rather curious in some of my tests was that the thermal radiation from the resistor was different across it's surface and was voltage polarity dependent. This means that swapping wires and resistor position did not change the orientation, the heat remained hotter at one end than the other. I don't recall now which end was the hot side, the MOSFET or the B+ end, but it was rather consistent that the gradient was there. I think I made a post about it when I did the test, or maybe it was in a video - I can't remember, too long ago. But it was enough to call into question the IR method of temperature measurement as the gradient was several degrees across a one inch long aluminum housed resistor.

[gadh]

For image #3 one full cycle takes up ~17µs if your setting was 5µs/div. According to my calculation that is around 58.8 kHz. Thanks

I think in this case the timebase was 2us so its 142khz... i never got to so low freq. and i always calc. the freq. when i get the oscillation and record it in my detailed XLS file.

One thing I did find rather curious in some of my tests was that the thermal radiation from the resistor was different across it's surface and was voltage polarity dependent. This means that swapping wires and resistor position did not change the orientation, the heat remained hotter at one end than the other. I don't recall now which end was the hot side, the MOSFET or the B+ end, but it was rather consistent that the gradient was there. I think I made a post about it when I did the test, or maybe it was in a video - I can't remember, too long ago. But it was enough to call into question the IR method of temperature measurement as the gradient was several degrees across a one inch long aluminum housed resistor.

in my case this does not matter, since i sink the load resistor and my thermometer together in a sealed thermos filled with distilled water

I now recall seeing some rust powder in the water since i use regular metal connectors to the load resistor wires so i assume the rust is from them. can it affect the water conductivity and the test ? i thought on measuring the water resistance but i got weird results before (when no rust was found) - between 200kohm and 20mohms. what is the min. value for our experiment ?

[Harvey]

I think in this case the timebase was 2us so its 142khz... i never got to so low freq. and i always calc. the freq. when i get the oscillation and record it in my detailed XLS file.


in my case this does not matter, since i sink the load resistor and my thermometer together in a sealed thermos filled with distilled water

I now recall seeing some rust powder in the water since i use regular metal connectors to the load resistor wires so i assume the rust is from them. can it affect the water conductivity and the test ? i thought on measuring the water resistance but i got weird results before (when no rust was found) - between 200kohm and 20mohms. what is the min. value for our experiment ?

The water should be acting as a dielectric, not a conductor. The ferrous rust in the water is only an issue if the quantity is great enough and the particulate size is small enough to form a colloidal suspension facilitating a current path. If this is your concern, you could place a small magnet near the bottom to collect the particles away from resistor. It is doubtful a few connectors could supply anything near that needed to change the water resistance. It is more likely that air suspended salts are being dissolved in the water if you are reading a reduction in resistance.

When taking those types of readings, ensure that you do not touch the probes as a 20M meter can measure your skin resistance and thus skew the reading. Any voltage differentials in the water can also give odd resistance readings. If your water was conductive, but measured a 10M value, the resistive current ratio would still be 1:1M so your error margin would no doubt swallow that up.

But, using clean sealed water for each test would probably be advisable and good scientific practice.

[gadh]

my digital charger accuracy

Hi harvey.
i measured its accuracy by using a car lamp (rear lamp) - 21w/12v rated.
results: see attached screen capture from my Excel.
In short:
used MAH: 1271
charged MAH: 1151
accuracy: 90% (means -10% from expected)

I assume the lamp took less than 21w, since in all my tests the charging process used about 5-10% more energy than used

My conclusion: maximum of +10% /-10% accuracy.

[Guruji]

Water heater

Hi guys did anyone manage to use this circuit with a filament heater?
Sorry for Keeping asking this question but something tells me that this circuit can be used to high amp heaters.
Thanks.

[Harvey]

Hi guys did anyone manage to use this circuit with a filament heater?
Sorry for Keeping asking this question but something tells me that this circuit can be used to high amp heaters.
Thanks.

Actually Guruji,

This circuit is better suited to a high voltage filament requiring less than 6A continuous (or 20A pulsed) current. The MOSFET can withstand up to 1000V on an inductive spike.

So let's say you wanted to produce 2000°F. You can do that with only 8.5W of power (see http://www.energeticforum.com/inductive-resistor/5359-mosfet-heating-circuits-2.html#post91200) using a one foot length of 40AWG N8 filament wound tight around a 1/32 arbor diameter and 0.36A of current. The wire would have a resistance of 65.7 ohms so a little less than 24V would do it. Therefore, you could drive 16 of these in parallel for a total of about 136W with this circuit with all 16 producing an IR temperature reading of 2000°F in free air under DC conditions. You could then multiply your wire length by up to 40 times and multiply your voltage by that same factor if you like and still be within the limits of the MOSFET, so you could drive 640 of these elements in a 40 x 16 array with 960 VDC to produce 5.4kW of heat at 2000°C and you would still be below the maximum ratings of the MOSFET at 5.7A

Of course that would be one massive heater with 640 feet of wire all wound up in it and the batteries would drain nearly as fast as your car battery would if you left your headlights on, but there would be no doubt as regards the heat produced

[Guruji]

16 circuits

Hi Harvey are you saying 16 circuits? I have built the one 12v battery circuit the negative wave one. So you're saying if I put more than 12v input would produce more? And what about the 555 timer it's only 15v rated I think.
Thanks

[Harvey]

Hi Harvey are you saying 16 circuits? I have built the one 12v battery circuit the negative wave one. So you're saying if I put more than 12v input would produce more? And what about the 555 timer it's only 15v rated I think.
Thanks

I'm saying that if you used a heating element as I have described, that 16 of them in parallel could be driven from a single MOSFET without exceeding its current rating.

I then added, that if you wanted to, you could place 40 of those parallel circuits in series with each other for a total of 640 heating elements and then increase the supply voltage and still not exceed the MOSFET drain voltage rating.

So it while it may seem hard to believe that a single IRFPG50 could drive a 5.4kW heater, I have just explained a matrix that would function that way.

One MOSFET circuit, 640 resistive heating elements, 5.4kW of heat.

Of course the control circuit for the MOSFET would have to properly powered, but if you will recall, Glen used a separate battery for his timer. And not too many of us have eighty 12V batteries laying around that we can series together for the 960V supply.

Also, in a real world application, special attention would need to be given to the BEMF spike to ensure that avalanche did not occur on the Drain pin if it were to exceed 1kV.

The purpose of my post was to show that this circuit is not a high amperage circuit, but instead is a high voltage circuit.

[Guruji]

Circuit

Hi Harvey as I am understanding you rightly so it's impossible to use this circuit for a normal heater element right?
Thanks for response.