Self running coil?

[gyulasun]

Hi Mark,

here are the scope shots. First is with 50 resistor and second is without resistor. I do agree that it makes the pulses square but it takes more current then without the resistor.

Both were adjusted to minimum voltage to trigger gate.

Any idea why this is ... this is the problem I have when I connect the 555. It keeps the pulses square and does not give as good of results compared to using the wavetek with rounded pulses.

Luc

Hi Luc,

The near sinusoid waveform with the spikes across the 100 Ohm resistor comes from the fact that you use a 10:1 probe backwards i.e. its output goes to the signal generator and its input tip (which includes a series 9 MOhm resistor as usual) goes to the gate of the MOSFET.
This means that this series 9 MOhm makes the gate high impedance and this way the gate can easily pick up the stray E field from the toroid coil (the sinusoid voltage also has the same freq like the signal generator output, 33.7 kHz), or if it does not pickup the the E field, then this sinusoid voltage may also come from the drain side via the C_rss drain gate capacitance and it is as big as the scope shows because the gate has a high impedance.

How to solve?  Could use a piece of 50 Ohm coax cable directly from the signal generator input to the gate, a cable length of 0.5-1 meter long.

EDIT  Maybe the 50 Ohm low impedance the gate now will "see" from the signal gen output would mean you will find similar NOT as good results like you find with the 555.

You would have to 'tinker' again for achieving the same good sweet point if you change to coax cable.

Thanks for answering my earlier questions.

Re on you ebay MOSFET types: they have very good low value threshold voltages, a .5 to 1V advantage wrt the 2SK... type, maybe this can compensate for the moderately high 700-800pF input capacitance. If you are careful with the voltages, the P channel type can be "handled".
(The third type is out of question indeed because of its 4.3nF input cap.)
So the first two types are worth trying, hopefully they have good prices.

rgds,  Gyula

PS Or you use an 1:1 probe, not a 10:1 from the signal gen? Cannot recall.

[mscoffman]

Hi Mark,

here are the scope shots. First is with 50 resistor and second is without resistor. I do agree that it makes the pulses square but it takes more current then without the resistor.

Both were adjusted to minimum voltage to trigger gate.

Any idea why this is ... this is the problem I have when I connect the 555. It keeps the pulses square and does not give as good of results compared to using the wavetek with rounded pulses.

Luc

Excellent scope pic's. Well both the signal and the noise has to flow
through the terminating resistor and the signal is winning power wise
with the resistor. Notice how the signal *Transistions* are much more
rapid and much squarer. The Fourier decomposition to sinewaves
would show much greater high frequency action with the squarer wave. 

The other thing is that clearly your coil seems to be resonating at
the frequency you are driving it at. This is seen in the one to one
lower power sinewave base seen in the gate current. The pulses
are due to coil inductance. The sinewave is due to coil re sonace and
you are *not* seeing it in the generator voltage with the t-resistor.
The coil power is beginning to overpower the transistor. And that
is beginning to effect your signal generator leads as noise in those
drooping high levels. With the cleaner signal you can tell that
the resonance sinewave is not a function of drive voltage.

On the NE555 how are you powering it? It may make sense to use
an external ~5volt dc supply for the NE555 then a voltage divider
to set the voltage. for comparison to the SigGen. You need to look
at the *exact upper and lower gate voltage levels* If you just want to
convert the signal to be *more* sinewave like and slower use an RC filter.
Hopefully you have already tried that gate series variable resistor. ie VrCiss =
RC when used with the NE555. In this circuit you won't necessarily get any
operational browny points for using clean digital signal level drives. It's
nicer to look at though. RC filter = series resistor then cap to ground.
put the scope on the actual gate. If you still come up with the same
answers then perhaps OU is based more in coil resonance and less in
the high speed inductive kick?

So:

a) set up variable gate resistor, tune with t-resistor vs tune with remote
powered NE555. Should be same. Also *maybe* same now as witout Vr and
t-resistor.

b) Set up RC filter, tune with t-resistor vs tune with remote powered NE555
Should be the same. Also should be same now without RC filter and
without t-resistor.

c) NE555 under remote power vs NE555 under bulk or pickup power with
voltage set exactly the same - difference in operation? = difference of
power drain of NE555 on operation of circuit.

:S:MarkSCoffman

[gotoluc]

Hi Luc,

The near sinusoid waveform with the spikes across the 100 Ohm resistor comes from the fact that you use a 10:1 probe backwards i.e. its output goes to the signal generator and its input tip (which includes a series 9 MOhm resistor as usual) goes to the gate of the MOSFET.
This means that this series 9 MOhm makes the gate high impedance and this way the gate can easily pick up the stray E field from the toroid coil (the sinusoid voltage also has the same freq like the signal generator output, 33.7 kHz), or if it does not pickup the the E field, then this sinusoid voltage may also come from the drain side via the C_rss drain gate capacitance and it is as big as the scope shows because the gate has a high impedance.

How to solve?  Could use a piece of 50 Ohm coax cable directly from the signal generator input to the gate, a cable length of 0.5-1 meter long.

EDIT  Maybe the 50 Ohm low impedance the gate now will "see" from the signal gen output would mean you will find similar NOT as good results like you find with the 555.

You would have to 'tinker' again for achieving the same good sweet point if you change to coax cable.

Thanks for answering my earlier questions.

Re on you ebay MOSFET types: they have very good low value threshold voltages, a .5 to 1V advantage wrt the 2SK... type, maybe this can compensate for the moderately high 700-800pF input capacitance. If you are careful with the voltages, the P channel type can be "handled".
(The third type is out of question indeed because of its 4.3nF input cap.)
So the first two types are worth trying, hopefully they have good prices.

rgds,  Gyula

PS Or you use an 1:1 probe, not a 10:1 from the signal gen? Cannot recall.

Hi Gyula,

thanks for the details. The scope probe I use on the Generator output has both X1 and X10. I only use it on the X1 setting which has a 220 Ohm resistance from what I can measure.

In the first scope shot of the post in question: http://www.overunity.com/index.php?topic=8892.msg234503#msg234503  The scope probe is not used. I have my alligator clips (50cm or 20" long) connected at the 50 Ohm resistor across the Hot and Ground of the BNC output connector of the Wavetek and still have the sine wave. So I don't think it's coming from the probe.

Thanks for your opinion on the ebay mosfet's

Luc

[gyulasun]

Luc,  it is interesting that with the triangle waves the waveform across the 100 Ohm looks like correct. 

Then the sinusoid wave is still a mistery... Maybe it is picked up by simple induction from the main toroid coil?

Gyula

[mscoffman]

Hi Luc,

The near sinusoid waveform with the spikes across the 100 Ohm resistor comes from the fact that you use a 10:1 probe backwards i.e. its output goes to the signal generator and its input tip (which includes a series 9 MOhm resistor as usual) goes to the gate of the MOSFET.
This means that this series 9 MOhm makes the gate high impedance and this way the gate can easily pick up the stray E field from the toroid coil (the sinusoid voltage also has the same freq like the signal generator output, 33.7 kHz), or if it does not pickup the the E field, then this sinusoid voltage may also come from the drain side via the C_rss drain gate capacitance and it is as big as the scope shows because the gate has a high impedance.

How to solve?  Could use a piece of 50 Ohm coax cable directly from the signal generator input to the gate, a cable length of 0.5-1 meter long.

EDIT  Maybe the 50 Ohm low impedance the gate now will "see" from the signal gen output would mean you will find similar NOT as good results like you find with the 555.

You would have to 'tinker' again for achieving the same good sweet point if you change to coax cable.

Thanks for answering my earlier questions.

Re on you ebay MOSFET types: they have very good low value threshold voltages, a .5 to 1V advantage wrt the 2SK... type, maybe this can compensate for the moderately high 700-800pF input capacitance. If you are careful with the voltages, the P channel type can be "handled".
(The third type is out of question indeed because of its 4.3nF input cap.)
So the first two types are worth trying, hopefully they have good prices.

rgds,  Gyula

PS Or you use an 1:1 probe, not a 10:1 from the signal gen? Cannot recall.

@gyulason,

I can't believe there is either a 1meg or 9meg R in series with the signal
generator as it wouldn't drive into a 50Ohms resistor at all....Measure
with DVM ohmmeter. Scope probes are generally 20Ohms impedance coax
at their highest frequency range so unless we are running at many MHz
the difference 20 vs 50Ohms won't matter. 20Ohms=RC || RL + dcR
So 50ohm video coax for a lab would be nice but not necessary at
these low frequencies. The t-resistor is desirable for Sig Gen.

---

@gotoluc

Well, try a sinewave then. Maybe OU doesn't like inductive kicks?
A sinewave is it's own Fourier Decomposition. It doesn't have any
additional HF signal components or steps. Resonance is a sinewave.
NE555 + RC filter ~= waveform. There is an Intersil cmos IC
somewhat like the NE555 that outputs a sinewave and squarewave
too. I think there is an app note that shows triangle wave generation.
It takes extra energy to generate a squarewave.

---

Man this looks close! You should boost the resistor up on pickup coil
to get 4.2Vdc like 6Kohm - perhaps insert the full diode bridge across
the pickup coil.

:S:MarkSCoffman