Pulse motor circuit's and design's

[TinMan]

I would like to brain storm pulse motors as a group.
Mix and match circuit's.Maybe there is a combination that will just all fit together.
I will start by posting one of the most efficient pulse motor circuits i designed.
This circuit will work in a S/S situation,if you remove both diodes across the base/emitter junctions.

[gyula]

Hi tinman,

Trying to understand how your pulse motor in your schematic may work, I can see no battery voltage reaching Q2 transistor collector, is it intentional?  Voltage from Battery 2 goes to collector Q1 via L2 (i.e. lower coil)  and the kick back spike from L2 can go to both batteries via D4 diode and also induces a certain voltage in L1, although it seems not utilized too much from L1 (i.e. from upper coil) for charging because the right hand side wire end of L1 goes via VR1 and R2 (50 Ohm) and D1 (and via R1 and D2) to the common negative so that losses occur in those resistors.
One more question:  does the axis of the bifilar coils face the magnets in a radial direction towards the rotor shaft or the coils are positioned tangentially to the rotor?
Finally: in this schematic as shown, what is the role of transistor Q2? 
Sorry if I ask dumb questions...

Gyula

[TinMan]

Hi tinman,

Trying to understand how your pulse motor in your schematic may work, I can see no battery voltage reaching Q2 transistor collector, is it intentional?  Voltage from Battery 2 goes to collector Q1 via L2 (i.e. lower coil)  and the kick back spike from L2 can go to both batteries via D4 diode and also induces a certain voltage in L1, although it seems not utilized too much from L1 (i.e. from upper coil) for charging because the right hand side wire end of L1 goes via VR1 and R2 (50 Ohm) and D1 (and via R1 and D2) to the common negative so that losses occur in those resistors.
One more question:  does the axis of the bifilar coils face the magnets in a radial direction towards the rotor shaft or the coils are positioned tangentially to the rotor?
Finally: in this schematic as shown, what is the role of transistor Q2? 
Sorry if I ask dumb questions...

Gyula

Hi Gyula
Q2 is not suppose to recieve any battery voltage.To understand how this circuit work's,it would be handy to have the standard SSG circuit next to you.
In the standard SSG ,L1 is always coupled together via (what would be) VR1 R2 and D1. In this circuit Q2 is used to decouple L1.
In this circuit the inductive kickback travels through D4 when Q2 switches off and decouple's L2.Only B1(charge battery)recieve's this charge.At the same time Q2 switches off,and L1 is decoupled.This sends the inductive kickback of L1 through D3 and into B2(run battery).As both windings and wire size in L1 and L2 are the same,the charge that B1 recieve's is determond by the value of R1 and VR1.
Q2 is just used as a switch to complete the circuit of L1(trigger coil)from the base to emitter of Q1,and to allow L1 to become open when L2 dose.
As far as which way the coil should be placed,well it will work either way.But it will perform better with the coil core parallel to the magnet's-as per standard pulse motor configuration.

Here is a replication of the circuit that one of my forum members built.
http://www.youtube.com/watch?v=5HwUuPQn2W8

[exnihiloest]

For an easy power pulsing circuit, I think that the simpler method is to use a MOSFET driver with a non-inverting input, for example the TC4420.
A simple or double RC circuit between the output and the input will allow a possible oscillation and determine the frequency, this loop being also possibly triggered by a controlled switch or a nonlinear element in series for a synchronous operation with a motor.
The TC4420 is cheap, has a small case (8 pins DIL), no need of external components, is able to output 6A with 25ns rise/fall time and is perfect to drive a possible power MOSFET. There is no need to bother with transistors when we see such little marvels...
These chips are optimized for driving capacitive loads, but provided that we stay sufficiently below their maximum ratings, they can also perfectly drive inductive loads.

[TinMan]

For an easy power pulsing circuit, I think that the simpler method is to use a MOSFET driver with a non-inverting input, for example the TC4420.
A simple or double RC circuit between the output and the input will allow a possible oscillation and determine the frequency, this loop being also possibly triggered by a controlled switch or a nonlinear element in series for a synchronous operation with a motor.
The TC4420 is cheap, has a small case (8 pins DIL), no need of external components, is able to output 6A with 25ns rise/fall time and is perfect to drive a possible power MOSFET. There is no need to bother with transistors when we see such little marvels...
These chips are optimized for driving capacitive loads, but provided that we stay sufficiently below their maximum ratings, they can also perfectly drive inductive loads.

Hi exnihiloest
Thank you for your input.I am interested in the type of setup you have talked about.Would the TC4420 drive a pulse motor setup without the mosfet's,providing the ohm's of the coil and battery voltage were correct-in that they dont exceed the 6 amp rating of the TC4420?
My other question would be-will it cope with the inductive kickback from the drive coil?.

Cheers
  Brad

[Allcanadian]

@tinman

Would the TC4420 drive a pulse motor setup without the mosfet's,providing the ohm's of the coil and battery voltage were correct-in that they dont exceed the 6 amp rating of the TC4420?
My other question would be-will it cope with the inductive kickback from the drive coil?.

That would be kind of like hauling hay bales in a ferrari, sure you could do it but it wasn't designed to do it well. The Mosfet driver is designed to charge/discharge the gate capacitance of the mosfet as fast as possible. This makes the mosfet operate more efficiently because the mosfet is a switch which dissipates energy when opening or closing. If the mosfet is not 100% on or 100% off then it acts like a resistance which starts dissipating energy in the form of heat. The faster the mosfet turns fully on or off then the lower the operating temperature and the higher efficiency.

We could think of it like a mechanical relay which starts adding resistance the slower it makes or breaks the circuit which means it generates more heat and becomes more inefficient the slower it switches between the on and off state.

AC

[TinMan]

@tinman
That would be kind of like hauling hay bales in a ferrari, sure you could do it but it wasn't designed to do it well. The Mosfet driver is designed to charge/discharge the gate capacitance of the mosfet as fast as possible. This makes the mosfet operate more efficiently because the mosfet is a switch which dissipates energy when opening or closing. If the mosfet is not 100% on or 100% off then it acts like a resistance which starts dissipating energy in the form of heat. The faster the mosfet turns fully on or off then the lower the operating temperature and the higher efficiency.

We could think of it like a mechanical relay which starts adding resistance the slower it makes or breaks the circuit which means it generates more heat and becomes more inefficient the slower it switches.

AC

Hi AC
Yes i realise that about the mosfets,which is why i didnt want to use them in the system.But if the TC4420 could be triggered by the current produced in L1 in the above circuit,then the mosfets may be a viable option-although they dont realy like inductive kickback circuits.Also the transistor has the ability to switch on lightly or heavy,depending on the current sent to the base-where as the mosfet is either on 100% or off 100%,as you stated.So using mosfets we have to regulate current flow via pulse width/time,where as with a transistor we can regulate via base current.

[gyula]

Hi Brad,

Thanks for the details and the video link.  It was okay for me that Q2 did not receive battery voltage, only its role was not fully clear. Now I understand how it was intended, I have redrawn your schematic to a style which makes it a better understandable schematic for me. If you disagree with it, please tell.

So as you wrote on Q2: "it is used as a switch to complete the circuit of L1 (trigger coil) from the base to the emitter of Q1 and to allow L1 to become open when L2 does". Well I think somehow the switch-on moment for Q2 must come a little bit earlier or at the same moment in time rather than a bit later than that for Q1 to fulfill the expected task a 100%, right? This would need a dual channel scope comparison of the two collector waveforms just around the switch-on moments.

Poppy mentioned in his video that the run coil had 800 turns and the trigger coil had 400, your schematic indicates equal number of turns for the two coils and this latter sounds better for me because less voltage is induced across a coil with less number of turns and this reduces charge power. Poppy also wrote on your forum that the average input current draw was 37.6mA and the output charging current was 11.2mA at one test.  

I think an improvement in the charge power could be had by using a higher Beta transistor type at both Q1 and Q2 because a 2N3055 has low Beta and may need a few mW base-emitter power to switch on and that power is taken away from the induced voltage/current in L1.  A better transistor type is for instance ZTX853    (100V, 4A, see data sheet here: http://www.diodes.com/datasheets/ZTX853.pdf )  and costs about 1 USD at Digikey ( http://www.digikey.com/product-detail/en/ZTX853/ZTX853-ND/92592 ) of course there are similar high speed switching types with high Beta and very low saturation collector-emitter voltage but their price may also be higher. A 2N3055 at Digikey now is 1.82 USD. (A notice with these better transistors: you cannot abuse them as much as you "got used to" abusing a 2N3055 type, they need more attention.)
Further improvement may be had by shunting the trimmer potmeter with a capacitor, this can reduce the AC voltage drop due to the spike current but tests are needed of course.

Would like to know whether you actually managed to measure the efficiency of this setup (output torque/input electric power) or you simply draw this conlusion (one of the most efficient pulse motors you designed) on comparison to other setups how good this one drained and charged batteries?  Just to know what level of efficiency we are to improve.

Thanks,  Gyula

Hi Gyula
Q2 is not suppose to recieve any battery voltage.To understand how this circuit work's,it would be handy to have the standard SSG circuit next to you.
In the standard SSG ,L1 is always coupled together via (what would be) VR1 R2 and D1. In this circuit Q2 is used to decouple L1.
In this circuit the inductive kickback travels through D4 when Q2 switches off and decouple's L2.Only B1(charge battery)recieve's this charge.At the same time Q2 switches off,and L1 is decoupled.This sends the inductive kickback of L1 through D3 and into B2(run battery).As both windings and wire size in L1 and L2 are the same,the charge that B1 recieve's is determond by the value of R1 and VR1.
Q2 is just used as a switch to complete the circuit of L1(trigger coil)from the base to emitter of Q1,and to allow L1 to become open when L2 dose.
As far as which way the coil should be placed,well it will work either way.But it will perform better with the coil core parallel to the magnet's-as per standard pulse motor configuration.

Here is a replication of the circuit that one of my forum members built.
http://www.youtube.com/watch?v=5HwUuPQn2W8

[exnihiloest]

...
Would the TC4420 drive a pulse motor setup without the mosfet's,providing the ohm's of the coil and battery voltage were correct-in that they dont exceed the 6 amp rating of the TC4420?
My other question would be-will it cope with the inductive kickback from the drive coil?.
...

At none time the output is open. Either one or the other of the two output MOSFETs is conducting, so the back emf will return to the power supply, and the chip can support a 1.5A reverse current. A diode can also be connected to protect the circuit.
I don't say it's perfect. I only say that it's an easy solution that can work provided that the current requirement is enough below the maximum ratings. It's the solution that I would choose and try if I needed a power pulsing circuit. For 1 or 2€, I wouldn't hesitate to take the risk to burn one and in this case, to add a buffer with a power MOSFET.
It's not a solution of the state of the art. There are surely dedicated chips for that but not so common and not the same price.

[TinMan]

Hi Brad,

Thanks for the details and the video link.  It was okay for me that Q2 did not receive battery voltage, only its role was not fully clear. Now I understand how it was intended, I have redrawn your schematic to a style which makes it a better understandable schematic for me. If you disagree with it, please tell.

So as you wrote on Q2: "it is used as a switch to complete the circuit of L1 (trigger coil) from the base to the emitter of Q1 and to allow L1 to become open when L2 does". Well I think somehow the switch-on moment for Q2 must come a little bit earlier or at the same moment in time rather than a bit later than that for Q1 to fulfill the expected task a 100%, right? This would need a dual channel scope comparison of the two collector waveforms just around the switch-on moments.

Poppy mentioned in his video that the run coil had 800 turns and the trigger coil had 400, your schematic indicates equal number of turns for the two coils and this latter sounds better for me because less voltage is induced across a coil with less number of turns and this reduces charge power. Poppy also wrote on your forum that the average input current draw was 37.6mA and the output charging current was 11.2mA at one test.  

I think an improvement in the charge power could be had by using a higher Beta transistor type at both Q1 and Q2 because a 2N3055 has low Beta and may need a few mW base-emitter power to switch on and that power is taken away from the induced voltage/current in L1.  A better transistor type is for instance ZTX853    (100V, 4A, see data sheet here: http://www.diodes.com/datasheets/ZTX853.pdf )  and costs about 1 USD at Digikey ( http://www.digikey.com/product-detail/en/ZTX853/ZTX853-ND/92592 ) of course there are similar high speed switching types with high Beta and very low saturation collector-emitter voltage but their price may also be higher. A 2N3055 at Digikey now is 1.82 USD. (A notice with these better transistors: you cannot abuse them as much as you "got used to" abusing a 2N3055 type, they need more attention.)
Further improvement may be had by shunting the trimmer potmeter with a capacitor, this can reduce the AC voltage drop due to the spike current but tests are needed of course.

Would like to know whether you actually managed to measure the efficiency of this setup (output torque/input electric power) or you simply draw this conlusion (one of the most efficient pulse motors you designed) on comparison to other setups how good this one drained and charged batteries?  Just to know what level of efficiency we are to improve.

Thanks,  Gyula

Hi Gyula
Yes,Q2 must start to switch on befor Q1-as Q1 cannot switch on until it recieves current from L1.Poppy and most guy's on my forum just use what they have lying around at the time.A wide variety of coils will work,that is the beauty of the pulse motor that uses transistors.In reguards to the 2n3055-is there any other transistor? lol.The reason we tend to use that transistor,is because it can take some hard knocks.For the beginer,it would be the best to use.But in saying that,i have a bucket that would have close to 100 of dead 2n3055's in it-mostly because of trying to drive them with to higher voltage.

As far as efficiency go's,my window motor with this circuit gave me the best.I never bother trying to measure mechanical output,as it is always very low.I normaly go for P/in and P/out.The window motor would run on 2.5 volt's at 780 micro amps,and return 592 micro amp's at the 2.5 volt's-so around 75.9% efficient.This was using super cap's,and is why ex's post sparked my interest when he was talking about charging caps without loss-as i was basicly doing as he stated by charging the cap with an inductive kickback pulse.I always managed to get back more than half in the charge cap-and that is joule's not volts.

I have made many different circuit's for pulse motor's,but the one above is the most electricaly efficient so far.I have no doubt that a far better circuit could be thought up,but the one above is my bench mark to beat-thus the reason for this post.It would be good to see how far we could take the simple pulse motor as a group.So any ideas you have,please post away.

[forest]

Could I attach my question to this thread ? 

Are you experienced with using mosfet drivers to fully open mosfet as fast as possible , especially TC44XX series ?

In datasheet I found that a low ESR film capacitor is required for decoupling at least 1uF for TC4420 and also a pair of low ESR ceramic capacitors 100nF (0.1uF) between some pins of chip to ground.
I have done a long research and found nothing about such low ESR capacitors :-( Of course I found that thee are MKP impulse capacitors but rated for bigger voltage like 250V and big in size. I even contacted manufacturer of WIMA capacitors but they never heard about low voltage impulse film capacitors for mosfet drivers. Here is the question : how important is it for reliable working of driver and if I can replace those by a tantalum type SMD capacitors , which are also impulse type but polarized ?

Please bear with me I really need your help..... Another question is how costly is to get samples from Microchip of TC4452 (the most powerful driver they have). I heard that they apply processing fee and if that's more then the cost of those parts in my country then it's not worth of trouble. Do you have any commercial email address so you can ask Microchip support about the processing fee ?

The problem is that I try to make Aviso style IGBT = set of mosfets and need to open them fully with a total capacitance of gate in 100nF range :-(


Thank You in advance for any comment....