MEMM

[PaulLowrance]

There's something about the MEG shots that I don't understand. I don't understand why the actuator current resembles a sine wave if the voltage is a square wave. Shouldn't it be either a saw tooth or an exponentially increasing saw tooth wave? If the core begins to saturate then the permeability decreases, which means di/dt should increase.

The only way it could be like a sine wave is if the resistance is significant. You know, an RL time constant. If the resistance is appreciable then initially the current will increase and eventually will level out due to resistance. You know what I mean?

Why would there be any appreciable actuator resistance? He shows the coil as 1.6 ohms and if the MOSFET is properly turned on it should be 0.04 ohms.

Regards,
Paul Lowrance

[MeggerMan]

Hi Paul,
I am not sure about the replicating the output load resistor, a carbon resistor should be the same what ever you do to it. 
I will just use a large wattage carbon resistor or series of them.
I can see there being a possible issue with a wirewound resistor.

Also JLN's scope shots include the power for the pulse drive circuit so to show that the output efficiency as being better at the higher voltage may be a result of the driver circuit using a fixed amount and this fixed amount becomes less significant to the overall power consumption as the voltage input increases. 
I think therefore the driver circuit needs to be run off a seperate supply which is what I shall do.

I do not think the coil is a 1.6 Ohm resistor, according to the circuit diagram, this is actually a resistor in series with the coil to reduce the current flow through the coil.  I do not understand why this is required and will in effect cause a power loss in the pulse circuit in the form of heat.  I can only think this is there to allow a larger input voltage.
I will take this off the board and therefore it can be added externally as an option.

From all my simulations the drive voltage for the pulse needs to less than a volt, therefore by seperating the drive supply and using my variable output PSU I can get a sub 1Volt pulse into the circuit.

The sine wave shaped drive current is probably a result of the resonance built up in the core.
I think when you first apply power to a transformer, there is a huge current surge until it starts to stablise, probably lasting a couple of cycles.
The coil's capacitance and core saturation can also shape the current.
I know very little about this topic.

Have a read of this:
http://www.butlerwinding.com/elelectronic-transformer/pulse_transformer/index.html

Regards

Rob

[PaulLowrance]

I do not think the coil is a 1.6 Ohm resistor, according to the circuit diagram, this is actually a resistor in series with the coil to reduce the current flow through the coil.

OK, I just thought 1.6 R was just depicting the coils internal resistance.

I'd like to build this circuit in LTSpice and see what happens? Any estimates what the internal actuator coils capacitance is?

Regards,
Paul Lowrance

[PaulLowrance]

I just completed an LTSpice circuit simulation of Naudin's MEGv2.1 and found some very interesting results. I used v2.1 because Naudin shows the actuator currents, which is important. Below are three images of the circuit, the actuator current, and the output.

I found out that the unusual actuator current is due to mutual coupling and not capacitance. Actually capacitance has the opposite effect in that it causes the current to initially surge.

Anyhow, after doing some research on various size inductors and how series & parallel resistance and parallel capacitance changes between inductors I came up with some good estimates on Naudin's MEG. Given wire size and so on I calculated 1.4 ohms for his actuator coil. Naudin has 1.6 ohms. So it's safe to say the extra 1.6 ohms simply represents actuator resistance. Also the 37 ohms is the secondary resistance.

Amazingly I was able to find an actual BUZ11 spice library, so this is very accurate simulation as far as parts.

Initial simulations were showing completely different actuator currents than Naudin's. After getting into the details discovered the secret was mutual coupling coefficients. So I entered appropriate coefficients according to Naudin's MEG core. When they were entered you can see the results are almost exactly to Naudin's MEG with one major difference, the secondary output voltage versus actuator current.   According to simulations, I had to bring the input voltage way down to get 280 mA actuator current, BUT Naudin's output voltage is 3.86 times greater!  According the MCE theory this extra output voltage is caused by what I call Magnetic Momentum. What is very interesting is Naudin's scope shots shows the output as 28.98 W and the input as 2.373 W. That is 12.21 times greater. We know that power is relative to the square of the voltage. If we square 3.86 we get 14.9.  That's very close to Naudin's 12.21 ratio.

I am not saying the above is proof that Naudin was telling the truth and that his scope shots are not faked, but I spent a good amount of time with LTSpice and the MEGv21 circuit, playing with stray capacitance and all. So far it is very interesting and worth investigating.

Attached is a zip file that contains the LTSpice file of the Megv21 and the BUZ11 spice library for anyone who wants to play with this. LTSpice is one of the best simulators. Best of all, it's free, created by Linear Technology, one of the best high end ADC manufacturers in the industry. Note that you will need the place the BUZ11 file in [your hard drive]:\Program Files\LTC\SwCADIII\lib\sub\

There are 3 attached images, which I think you need to click on to enlarge. One of the circuit. Another of the actuator current --> I(L2) purple graph.  Another of the secondary output --> Orange is power, purple is voltage, turquoise is current.


Regards,
Paul Lowrance

[MeggerMan]

Hi Paul,
I do not know how to calculate the capacitance of a coil, there are so many variables that I guess you would need to wind it then test it.

I downloaded the LT Spice and ran your simulation.
What an amazing piece of software this is!

I added an extra 1.6Ohm resistor to the circuit in series with the input coils and it seems to make little difference to the current, which seems odd.

You can see the back emf voltage in the input coil jump to 70V for a very brief period, then it bounces slightly at around 28V until the next pulse.
This means the mosfet would need to be able to cope with around 80 volts across source and drain.

How did you couple L1 and L2 to L3 and L4?

I have looked at the spec on the TL494CN chip again because I was worried that it would not be possible to have a dead time between pulses, so I started at looking at other function generator chips today to see if there was an alternative and the closest I came up with was the Exar XR-2206 chip:
http://www.exar.com/products/XR2206v103.pdf

Looks good but only has a single output so I would need to couple this to a flip-flop and then some output transistors.

So I went back to the TL494CN and I realised that I had mis-read the test wave forms.
It is possible to have a dead time between pulses by using the PWM input.
So its possible to vary the amount of dead time between pulses.

I found a rather good/inexpensive multi-meter for measuring frequency and duty cycle.
Tenma 72-7735
http://mcmb2b.com/cgi-bin/test/72-7735.html
It has an RS232 output too.
They also do an inductance meter 72-8155 but I cannot find a local supplier for it.

Regards

Rob