PWM modification

[alpha-dog]

I do have a MX067 pwm modified to use a P-channel mosfet that is working.
Basically the P-channel fet switches the positive voltage from the battery. I've only found one mosfet that is a good substitute for the N-channel fet we have been using.
The problem was finding a P-channel fet with the same Vgs(th) gate threshold signal as the N-channel fets use that we find in our pwm's. Part # STP80PF55 has the same gate threshold signal requirements and can be used in our pwm's. The gate ( pin 1 ) and drain ( pin 2) connect the same way as the N-channel fets. The source ( pin 3 ) of the P-channel will connect to the positive battery post and of course ther N-channel source ( pin 3 ) connects to the negative post ( ground ).
How I modified the MX067 ( MX068 uses same schematic ) follows:
1) I cut the ground run from the driver portion of the pwm to where the ground spade terminal is.
2) I cut the run between L+ and positive spade terminals.
3) I attaached a wire from L+ spade to the ground run returning to the driver circuit.
4) I removed the flyback diode - D1.
5) I removed the N-channel mosfets and installed in thier place a P-channel mosfet.
How I connected the pwm to the HHO electrolyser:
1) I connected a positive voltage from the ignition circuit to the + spade terminal on the pwm.
2) I connected ground to the L+ spade terminal on the pwm.
3) L- still goes to the HHO electrolyser.
4) The ground spade connector now connects to the positive post of the battery ( or the HHO relay )

I do not know if this will work better than N-chanel pwm's. I believe it will work better. When I was testing it my resovior and electrolyser were leaking. I'm in the process of making repaires. One thing that I did notice is sthat the higher PRF I dialed in the current would drop. This is with a 60% duty cycle.
I should know more this weekend.

[tomisnt]

Thanks for your contribution alpha. I will play with one soon.

[alpha-dog]

I ran a few test today comparing two MX067 pwm's. One has N-channel mosfets and the other has P-channel mosfets. My little test electrolyser only produces .5Lpm at 4.7 amps with the N-channel. Although the gains were small the P-channel pwm out preformed the N-channel pwm by maybe as much as 10% with the same current draw at a PRF of 100 hz. This is after a few hours of running the electrolyser.
Russ

[Painless]

So, you're saying that the p MOSFET PWM produced more HHO for the same watt power?

I've been thinking about the stored power within a cell when its not receiving power from the PWM. Imagine a PWM that controls both the p and n sides with synchronised MOSFET's, yet, during the off cycle it connects positive to negative allowing to cell to power itself.

[alpha-dog]

So, you're saying that the p MOSFET PWM produced more HHO for the same watt power?

I've been thinking about the stored power within a cell when its not receiving power from the PWM. Imagine a PWM that controls both the p and n sides with synchronised MOSFET's, yet, during the off cycle it connects positive to negative allowing to cell to power itself.

Yes!! I used the same electrolyser, KOH solution, ect, ect. Also, after the test were done with the standard pwm I immediately tested the P-channel pwm. So the water temperature was the same. I made a mark on my HHO flow meter at water line with the standard pwm and retested 3 times verifying the mark. The P-channel consistently moved the 1.5 liter bottle up 5/16 inch higher at the same amperage ( 4.7 amps ) and PRF (100hz ).
Also, that is what I have been saying about SDC current. Using the mosfets that I've found STP80NF55 and STP80PF55 working at the same Vgs(th) gate threshold signal. When one conducts the other is off. You could use an inverter to produce a gate\not signal. With a gate and gate\not signal and these mosfets provides the switching needed for SDC current in an electro-magnetic cell.
Russ

[alpha-dog]

I got my flow meter in yesterday and was able to do some better test on a modified P-channel pwm. ( I'll be doing the current limit mod next ).

At:
PRF = 4.761Khz
Duty Cycle = 80%
Amps = 8.2 amps
Lpm = 500 ml
volt = 10.8
MMW = 5.65


One thing very interesting I discovered is that the higher the PRF the less current used. MMW dropped a little bit ( 50 ml ) from DC current, but Pulsed dc current dropped from 13 amps to 8.2 amps. Using this low of a voltage should give bad results but the pulsed dc is very efficient.

[Gforce]

Keep in mind if you are measuring the voltage with a meter you will get the average. The voltage is actually getting switched on and off with the PWM.
When it is on the voltage is higher. If you were to look at it with an Oscilloscope you would see and measure the peak voltage.

13.8V......___ ..___....___
Ave -----|--|--|--|--|--|------
________|...|__|...|_|...|______

The thing I am pondering is it more efficient to use a variable voltage rather than a PWM. This can be done by adding an inductor and capacitance.

[alpha-dog]

Keep in mind if you are measuring the voltage with a meter you will get the average. The voltage is actually getting switched on and off with the PWM.
When it is on the voltage is higher. If you were to look at it with an Oscilloscope you would see and measure the peak voltage.

13.8V......___ ..___....___
Ave -----|--|--|--|--|--|------
________|...|__|...|_|...|______

The thing I am pondering is it more efficient to use a variable voltage rather than a PWM. This can be done by adding an inductor and capacitance.

I use an o'scope. Also, I measure DC before the pwm. That's why I can measure PRF and duty cycle. I'm thinking that pulsed dc allows the stored energy to be used. That's just a theory of course, but there is a marked increase at the same current draw between the N-channel ( common pwm ) pwm and the one I have modified for P-channel mosfets. I'm going to make current limiting work next, then it'll be ready for my truck.

[Q-Hack!]

Keep in mind if you are measuring the voltage with a meter you will get the average. The voltage is actually getting switched on and off with the PWM.
When it is on the voltage is higher. If you were to look at it with an Oscilloscope you would see and measure the peak voltage.

13.8V......___ ..___....___
Ave -----|--|--|--|--|--|------
________|...|__|...|_|...|______

The thing I am pondering is it more efficient to use a variable voltage rather than a PWM. This can be done by adding an inductor and capacitance.

Isn't ASCII art a PIA...


13.8V....___....___....___
Ave ----|---|--|---|--|---|------
________|...|__|...|__|...|______


It works a lot better if you set the font to a standard width like "Courier New". Those variable width fonts like "Times New Roman" will be different for people on Mac's and Linux boxes.

[alpha-dog]

Thought I'd let you know that the current control comparator works fine. Look for a schematic.