My 76,000 rpm pulsed dc motor

[mbrownn]

"What are the numbers of the fets you are using?"

I have used: stw8nb100 latest run, stw11nk100z, the stw50nb20, and the fairchild fqa5n90. The higer the Vdss rating of the mosfet the higher the recovered voltage. In those terms, to date the two best performing have been the stw8nb100 and stw11nk100z. The Rds of each mosfet also plays a major role in terms of speed and current draw. The stw50nb20 has been the clear winner in terms of speed, reaching 76000rpm and some change on 27.2Vdc input.


"It sounds like the 1000v fets have a lower resistance which is what I need."
stw50nb20 200V 50A .047 ohm
stw8nb100 1000V - 1.3 OHM - 7.3A
stw11nk100z 1000V - 1.38 OHM - 8.3A
etc.

Sounds like you need something closer to the stw50nb20 but it's rating is only 200V. It will handle the 30A. Wow you need 30A.

Still need to comment on your post earlier with the link. I printed it and read it just can't intelligently comment on it at the moment. To busy repairing and documenting catastrophe.

Thanks for the info

I am experimenting with a 220v 2kw universal motor which runs at under 10A but when I change it from series wound to shunt wound its current will be three times that to get the same power although the voltage will be cut by thee.

In truth I expect I will have to run the motor at much less power than that to get the effects I want.

[mnsman]

Hopefully up and running again tonight

I think the epoxy should be fully cured by tonight. The coil is ready for a light test run. Last enclosure was a bit on the ugly side not knowing it was going to get so much camera time. New enclosure is shiny and ready for it's photo op.

[quantumuppercut]

hi mnsman,

Can you help me do an experiment? I was trying to study the back flow current in a DC motor. Basically when you connect a battery to a DC motor, it will draw a current, however, this current is not a DC one. I was able to determined that it also flow back to the source but do not have a scope to examine it closely. If you're not too busy, can you help me do a scope shot across a low Ohm resistor in series with the motor? At high speed, I think it has current flow back. Thanks.

[mnsman]

Looking for best POWER MOSFET

To All,

Looking for the best power mosfet. My ideal candidate would have the highest Vds and the lowest Rds. Any input? I want the lowest possible Rds so all of the current will go to the coil. The coil is 1.1 ohm so every bit of reduction in the Rds can be critical to efficiency. The Vds value appears to be critical to the amplitude of the spike created by the collapse of the coil's magnetic field. I have experienced higher amplitudes with higher rated Vds power mosfets. I have used the following mosfets:

stw50nb20 200V - 50A - .047 Ω Best performance in terms of speed and current use. 76K+ rpm @27.2Vdc input (power supply), spike around 320=350Vdc

Fairchild FQA5N90 900V - 5.8A - 2.3 Ω Bigger spikes (never let cap charge past 450Vdc) Slower speed ~46K rpm with same input as stw50nb20

stw8nb100 1000V - 1.3 Ω - 7.3A Biggest spikes 1kV+, Slower speed 71k rpm @ ~39Vdc input (3 bats)

stw11nk100z 1000V - 1.38 Ω - 8.3A Similar to stw8nb100


I need to find the best choice. I'm operating under the presumption that running more than one in parallel will reduce the resistance Rds the same way resistors in parallel would. If this is the case and my Rds is 160 mΩ and I have say 4 in parallel that gives me 40 mΩ Rds which is close to the value of my best performer the stw50nb20 with 47 mΩ. More would theoretically be better. Presuming all of that is correct... so far what I've found to be my best choice for the money is the:

Cree CMF10120D-Silicon Carbide Power MOSFET
Z-FeTTM MOSFET
N-Channel Enhancement Mode

Digikey $20ea
Is my thinking correct?
Is there a better choice?
Jim

[mnsman]

hi mnsman,

Can you help me do an experiment? I was trying to study the back flow current in a DC motor. Basically when you connect a battery to a DC motor, it will draw a current, however, this current is not a DC one. I was able to determined that it also flow back to the source but do not have a scope to examine it closely. If you're not too busy, can you help me do a scope shot across a low Ohm resistor in series with the motor? At high speed, I think it has current flow back. Thanks.

I would certainly do that. But you'll have to wait for me to get back on line. Hopefully tonight I'll get it going again. Will run it slow tonight and work my way up. Don't want to have to replace anything too soon. Do you want the scope ground connected to battery ground? You must be looking for a voltage higher than input. What value resistor? Remember the motor is 1.1 ohm so the resistor would have to be really small like .1 ohm max. Not sure what I have in the way of really small big watt resistors. Now that I think about it I would have to check on what I do have. That could be a problem.
Will do it if I can find a small enough resistor. Nudge me if I forget with holidays coming.

[OrionLightShip]

@mnsman

IFZ44N has a 17.5 milli Ohm Rds on with really fast switching and they are really inexpensive on ebay in multiples. Perhaps a Vdss of 55volts is insufficient for your use.

Sounds like you need a transient protection scheme found here.

from wiki: A transient-voltage-suppression diode may be either unidirectional or bidirectional. A unidirectional device operates as a rectifier in the forward direction like any other avalanche diode, but is made and tested to handle very large peak currents. The popular 1.5KE series allows 1500 W of peak power, for a short time.

You already stated that you need to optoisolate your hall effect so that is taken care of.

Hope this helps

Orion

[quantumuppercut]

I would certainly do that. But you'll have to wait for me to get back on line. Hopefully tonight I'll get it going again. Will run it slow tonight and work my way up. Don't want to have to replace anything too soon. Do you want the scope ground connected to battery ground? You must be looking for a voltage higher than input. What value resistor? Remember the motor is 1.1 ohm so the resistor would have to be really small like .1 ohm max. Not sure what I have in the way of really small big watt resistors. Now that I think about it I would have to check on what I do have. That could be a problem.
Will do it if I can find a small enough resistor. Nudge me if I forget with holidays coming.

Hey,

Thanks in advance. Any time you feel like it. I was using radio shack 3V little motor. Yes, I just want to see if there is actually some current flow back to the battery. The scope does not have to be battery ground. I always see ground as something relative. lol Thanks again, I'll pay you back if I have a chance.

[mnsman]

Back on line

Ran it up to 72K tonight. Short run. Seems fine. Charged cap over 1KV in about 20 sec. and everything held.

OrionLightShip, if your listening I'll post reply to you tomorrow- it's past my bedtime. Been hoping you would chime in. Wanted to ask you about variable capacitance.

Jim

[mbrownn]

To All,

Looking for the best power mosfet. My ideal candidate would have the highest Vds and the lowest Rds. Any input? I want the lowest possible Rds so all of the current will go to the coil. The coil is 1.1 ohm so every bit of reduction in the Rds can be critical to efficiency. The Vds value appears to be critical to the amplitude of the spike created by the collapse of the coil's magnetic field. I have experienced higher amplitudes with higher rated Vds power mosfets. I have used the following mosfets:

stw50nb20 200V - 50A - .047 Ω Best performance in terms of speed and current use. 76K+ rpm @27.2Vdc input (power supply), spike around 320=350Vdc

Fairchild FQA5N90 900V - 5.8A - 2.3 Ω Bigger spikes (never let cap charge past 450Vdc) Slower speed ~46K rpm with same input as stw50nb20

stw8nb100 1000V - 1.3 Ω - 7.3A Biggest spikes 1kV+, Slower speed 71k rpm @ ~39Vdc input (3 bats)

stw11nk100z 1000V - 1.38 Ω - 8.3A Similar to stw8nb100


I need to find the best choice. I'm operating under the presumption that running more than one in parallel will reduce the resistance Rds the same way resistors in parallel would. If this is the case and my Rds is 160 mΩ and I have say 4 in parallel that gives me 40 mΩ Rds which is close to the value of my best performer the stw50nb20 with 47 mΩ. More would theoretically be better. Presuming all of that is correct... so far what I've found to be my best choice for the money is the:

Cree CMF10120D-Silicon Carbide Power MOSFET
Z-FeTTM MOSFET
N-Channel Enhancement Mode

Digikey $20ea
Is my thinking correct?
Is there a better choice?
Jim

I suspect some of the reason for the larger spikes must be the speed that the FET switches off. Lower resistance allows for more current, faster ramp up of current and lower losses

[mbrownn]

hi mnsman,

Can you help me do an experiment? I was trying to study the back flow current in a DC motor. Basically when you connect a battery to a DC motor, it will draw a current, however, this current is not a DC one. I was able to determined that it also flow back to the source but do not have a scope to examine it closely. If you're not too busy, can you help me do a scope shot across a low Ohm resistor in series with the motor? At high speed, I think it has current flow back. Thanks.

Cant give you the scope shot but I will describe it.

It is a dirty signal, you get a ripple given by the change in impedance as the coils in the rotor are switched on and off. As each new pair of commutator segments comes into contact with the brushes the previous pair is shorted and the pair before that reverse their flow. As the coils are never perfectly matched you get slightly differing heights on each wave of the ripple and it is possible to work out what the motor speed is by watching one ripple that might be a little higher than the others.

The pair of coils that reverse their flow cause a radiant spike that is clearly visible on each wave of the ripple. Depending if their is a capacitor across the brushes you also get an oscillation. Add to this the fact you get a less than perfect contact between the brushes and commutator causing more spikes and oscillations and you can imagine what it looks like.

As the motor is loaded the frequency of the ripple drops but the magnitude and duration of the oscillations caused by the spikes increases although the frequency of these oscillations does not change.

Input 12v DC, magnitude of ripple 60mv, magnitude of the spikes up to 1v under load.

The motor used is a 12v DC permanent magnet blower motor, It is old and has visible wear on the commutator, It has a capacitor across the brushes of unstated value.

This test was on a motor with no shunt resistor. The test you want would probably show increased magnitude of the ripple and spikes.

The current is DC but is fluctuating in magnitude so I can understand why you might think it is not DC.

On a Bedini motor you do get a small reversal of current in some situations.

[OrionLightShip]

@mnsman

I'm here, been following your progress for awhile. Need to re-read and get a handle on what you are trying to accomplish though. I'm just fascinated by anything that spins that fast and doesn't explode.

Are you following skycollection on youtube? He has a pretty cool magnetic levitation motor using a simple starship coil to hit 30,000.

Ran it up to 72K tonight. Short run. Seems fine. Charged cap over 1KV in about 20 sec. and everything held.

OrionLightShip, if your listening I'll post reply to you tomorrow- it's past my bedtime. Been hoping you would chime in. Wanted to ask you about variable capacitance.

Jim

[quantumuppercut]

Cant give you the scope shot but I will describe it.

It is a dirty signal, you get a ripple given by the change in impedance as the coils in the rotor are switched on and off. As each new pair of commutator segments comes into contact with the brushes the previous pair is shorted and the pair before that reverse their flow. As the coils are never perfectly matched you get slightly differing heights on each wave of the ripple and it is possible to work out what the motor speed is by watching one ripple that might be a little higher than the others.

The pair of coils that reverse their flow cause a radiant spike that is clearly visible on each wave of the ripple. Depending if their is a capacitor across the brushes you also get an oscillation. Add to this the fact you get a less than perfect contact between the brushes and commutator causing more spikes and oscillations and you can imagine what it looks like.

As the motor is loaded the frequency of the ripple drops but the magnitude and duration of the oscillations caused by the spikes increases although the frequency of these oscillations does not change.

Input 12v DC, magnitude of ripple 60mv, magnitude of the spikes up to 1v under load.

The motor used is a 12v DC permanent magnet blower motor, It is old and has visible wear on the commutator, It has a capacitor across the brushes of unstated value.

This test was on a motor with no shunt resistor. The test you want would probably show increased magnitude of the ripple and spikes.

The current is DC but is fluctuating in magnitude so I can understand why you might think it is not DC.

On a Bedini motor you do get a small reversal of current in some situations.

Thanks mbrownn,

Seems likes a lot of ripples. I guess what I'm trying to find out is can the magnet induced back a higher EMF and the source voltage. This would show a negative value on current although it could be small. Expecting a perfect commutator seems hard. I have connected an LED and it lit up, can only interpret as a reverse current. If the magnet can induced back a higher voltage than source, then in my theory, we can just power a motor and the source gain energy. Kinda hard to imagine. Thank you for your information. Sounds like it's easier to see on a pulse motor than DC motor.

[mbrownn]

Thanks mbrownn,

Seems likes a lot of ripples. I guess what I'm trying to find out is can the magnet induced back a higher EMF and the source voltage. This would show a negative value on current although it could be small. Expecting a perfect commutator seems hard. I have connected an LED and it lit up, can only interpret as a reverse current. If the magnet can induced back a higher voltage than source, then in my theory, we can just power a motor and the source gain energy. Kinda hard to imagine. Thank you for your information. Sounds like it's easier to see on a pulse motor than DC motor.

No, the BEMF will always be lower than the forward EMF but there are tricks to increase the forward EMF by inducing radiant spikes. Try looking up "coil shorting" The radiant spikes produced by coil shorting and coil collapse travel in the same direction as the forward EMF and so assist a motor. It is normal to short out or smooth these spikes, John Bedini showed us how to collect them, I am proposing a way of using them to help power the motor to some degree.

Having said all that, if you have lit an LED wired against forward EMF then there must have been a current flow. How did you wire it in the circuit?

[mnsman]

I suspect some of the reason for the larger spikes must be the speed that the FET switches off. Lower resistance allows for more current, faster ramp up of current and lower losses

Good point. I have to do a comparison of the switching speeds of each of the fets used. The thought had occurred to me at some point but for some reason I have dwelled on the Vdss. The more abrutly the charge across the coil is removed and the more tolerance for a spike at the drain (Vdss) the bigger the spike in return. Makes sense to my little mind.

[mnsman]

@mnsman

IFZ44N has a 17.5 milli Ohm Rds on with really fast switching and they are really inexpensive on ebay in multiples. Perhaps a Vdss of 55volts is insufficient for your use.

Sounds like you need a transient protection scheme found here.

from wiki: A transient-voltage-suppression diode may be either unidirectional or bidirectional. A unidirectional device operates as a rectifier in the forward direction like any other avalanche diode, but is made and tested to handle very large peak currents. The popular 1.5KE series allows 1500 W of peak power, for a short time.

You already stated that you need to optoisolate your hall effect so that is taken care of.

Hope this helps

Orion

Thanks for your input OrionLightShip. The power mosfets I have tested all have a transient-voltage-suppression diode built into them. So I think I'm okay in that regard.

As far as the IFZ44N goes, all of it's specs are good except the Vdss. I presume it would drive the coil even better than my top speed performer the
stw50nb20 200V - 50A - .047 Ω. This is due to the lower Rds. I would probably see more speed with less current but much much lower radiant spikes with collapse of the coil's mag field. Is that good or bad? I'm not sure but I've been operating under the premise that a bigger spike is better.

Periodically I look to see who's lurking in the corner on this thread and I check to see what said person's expertise is. I've noticed you a few times and read a few of your posts regarding variable capacitance. Having said that what is your opinion on adding variable capacitance to my device? This is just a thought exercise at the moment. If you've looked at the animation of my proposed HV experiment with the motor you will have discovered I plan on placing the rotating mag between two oppositely charged plates which is of course a capacitor. P-Motor experiment description and animation - YouTube Since variable capacitance has been associated with OU I'm trying to envision variable capacitance being added to this device. Any thoughts?