The Rotary Transformer Project.

[TinMan]

Depending on the motor type, sometimes a load improves efficiency. There is a sweet spot. Motors are sometimes designed to be run most efficiently with a certain slip angle, which does not occur except with the right loading.
Sorry for stating this if it is already known to you.

Which class and type of motor are you using?
see: http://en.wikipedia.org/wiki/Electric_motor#Specifying_an_electric_motor

Sorry ION,i should have mentioned this in the post or video.
Thanks for the link,but i am quite at home with electric motor configuration's
The house fan motor is a Squirrel-cage induction motor.
The rotary transformer started out as a universal motor or series wound motor,with a 2 pole stator.

[TinMan]

tinman,

Why is the fan blade so small in this test? Where is the original fan blade for the high efficiency (HE) fan? Is it a lot larger in diameter?

Assuming that the HE fan is running at 86% efficiency in your test might be an error. As ION mentioned, you may need to have the correct designed-for load on it before it can attain that 86%.

I would suggest another test with the original HE fan blade on both motors.

Do not be confused with size here,as both fan blades almost do the same amount of work.
The small fan has 7 blades at a very steep angle,and is designed to shif a large amount of air over a small area(it is from a thermo fan setup)
The large fan has 3 blade's of a shallow angle,that is designed to shift air at a lower speed over a larger area.

But to be fair and in the name of reserch,i have carried out the same test with the large fan blade on both devices.
The one thing to remember when you watch the video is that the efficiency of the rotary transformer is lower than that of the last test.
This is due to the fact that i have to run the rotary transformer in reverse (switch polarity)so as the fan blade runs in the same direction as the house fan motor dose.
The brush timeing is not set right to run in this direction,nor can i move the brushes around enough in the right direction to get the timeing right for for this direction of rotation.
This is because the housing bearing blocks are in the way.
But we have run the test reguardless

I will ofcourse post the link to the video as soon as it is uploaded.

[TinMan]

Here is the same test,but carried out with the large fan as Poynt99 suggested.
The efficiency of the rotary transformer is down,as i had to reverse the rotation direction-so as to rotate in the same direction as the house fan motor.
The brush timeing is not set to run in the reverse direction-something to keep in mind.
http://www.youtube.com/watch?v=szpJ97M58G4

[Chet K]

Its my job to ask the dumb questions
sorry

When you say "had to reverse rotation"
do you mean the motor rotation is backwards of your previous run to accommadate the fan?
why can't you flip the blade around and leave the motor rotation and timing issue aside?

Nice weld job ,you have a good eye for balance!
and your movie making is superb ... nice clear and fun to watch [even scary with the weld test at speed scenario  ]

Either you have uncovered some Gross negligence in your country's motor rating system
or you have something here that could get a nice prize [Nobel even......]

BTW
It has come to our attention that folks are harvesting money from the efforts of researchers on these forums
Kickstarter and similar orgs  with the best of intentions are paying people for concepts or prototypes they showcase and some nefarious types are stealing and posting things that are not theirs.for financial gain .
of course some will breathe a sigh of relief as they go the patent route ........
however this "WILL" be dealt with at the root..........

make an announcement at anyplace you post this publicly ,
However
I suspect a very strong Push back is heading towards the orgs that unknowingly subsidize these crooks or men of low moral standards.

thx
Chet

[TinMan]

Its my job to ask the dumb questions
sorry

When you say "had to reverse rotation"
do you mean the motor rotation is backwards of your previous run to accommadate the fan?
why can't you flip the blade around and leave the motor rotation and timing issue aside?

Nice weld job ,you have a good eye for balance!
and your movie making is superb ... nice clear and fun to watch [even scary with the weld test at speed scenario  ]

Either you have uncovered some Gross negligence in your country's motor rating system
or you have something here that could get a nice prize [Nobel even......]

BTW
It has come to our attention that folks are harvesting money from the efforts of researchers on these forums
Kickstarter and similar orgs  with the best of intentions are paying people for concepts or prototypes they showcase and some nefarious types are stealing and posting things that are not theirs.for financial gain .
of course some will breathe a sigh of relief as they go the patent route ........
however this "WILL" be dealt with at the root..........

make an announcement at anyplace you post this publicly ,
However
I suspect a very strong Push back is heading towards the orgs that unknowingly subsidize these crooks or men of low moral standards.

thx
Chet

Hi Ramset
I could indeed turn the fan around-so as the leading edge was the same as the house fan setup.
But i reversed the direction insted so as the fan is pushing the air away from the motor-to keep in line with the house fan.
Remember that by turning the fan around ,dosnt change the direction of air flow,but only the leading edge of the fan blade.

Now come's the process of elliminating error's that could be within the test.
As the fan is no where near cavitational speed-we can eliminate that from any error's that may be being made here.

The two that stand out are the DMM reading's,and the fact that(as you stated) the efficiency rateing of the fan motor is way out.

So now i have to get the second winding on,and start the process of eliminating error's.
Then ofcourse the prony brake test to confirm numbers we have seen so far.

[PhysicsProf]

Here is the same test,but carried out with the large fan as Poynt99 suggested.
The efficiency of the rotary transformer is down,as i had to reverse the rotation direction-so as to rotate in the same direction as the house fan motor.
The brush timeing is not set to run in the reverse direction-something to keep in mind.
http://www.youtube.com/watch?v=szpJ97M58G4

Excellent work, Tinman.  Thanks for taking these measurements and showing results.

Are you still running on 24V, rectified AC?  (22.22V per your latest vid.)
If so, how sure are you that Pinput = Vin * Ain?  RMS measurements? 

[TinMan]

Excellent work, Tinman.  Thanks for taking these measurements and showing results.

Are you still running on 24V, rectified AC?  (22.22V per your latest vid.)
If so, how sure are you that Pinput = Vin * Ain?  RMS measurements?  

Hi PhysicsProf
My scope says that the p/p voltage is about 38 volt's,and the DMM says RMS voltage is 22.22 volts.
As far as amp's go's,i have found in other test useing both low and high frequency's that the DMM dose a preaty good job at averaging the current.
I think i have a smaller 24 volt transformer some where ,which i would think wouldnt have as many losses as the large transformer.
This way i could get a bit closer to actual watts useing the watt meter.
What would be handy right now is a variac transformer,as 24 volts may not be the best voltage for this setup.
Might even have a go at makeing my own.

Much more testing to do yet,and i still have yet to wind the second coil on the stator-which will only increase efficiency!! hopefuly !!

And yes-im still useing rectified ac at 24 volt's.

[PhysicsProf]

  OK - thanks for the response and for your cutting-edge (IMO) work.

   I take it you need pulsed-DC?  but could you possibly use a super-cap for the input energy?  then E = 1/2CV^2 makes the Ein measurement straightforward. 
People are using 555-circuits to go from battery to pulsed DC...

[TinMan]

  OK - thanks for the response and for your cutting-edge (IMO) work.

   I take it you need pulsed-DC?  but could you possibly use a super-cap for the input energy?  then E = 1/2CV^2 makes the Ein measurement straightforward. 
People are using 555-circuits to go from battery to pulsed DC...

At the moment i am useing the rectified AC,this give you a half wave AC or a direct altinating current if you want to call it that.
Meaning that the direction of current flow is single,but the current altinates from 0 to 22.22 volts in this case.
I will be trying a DC pulse setup today,useing my PWM to run the motor.
This will allow for accurate input measurements,and also tell us if the system like pulsed DC.

As far as super cap's go-i have non that are any where near enough to the capacity or voltage needed to do that.

[Chet K]

TM
So a control of sorts would be a good goal, If we can access the fan at a local "China mart"and do a run at another location to establish efficiency or better yet the blade?

Who here has 220- 240 at 50 hertz ?
itsu?
Poynt?
AC ?
Mike ?
Groundloop?

?
I am quite sure you have solutions to verify the "Work" end. t would be nice to pick up a 15 dollar fan and just do a quick run.
With out calibration a pony brake and such still has holes in the data ,A good control works wonders .

Just bouncing ideas around to validate ........
Thx
Chet

[TinMan]

TM
So a control of sorts would be a good goal, If we can access the fan at a local "China mart"and do a run at another location to establish efficiency or better yet the blade?

Who here has 220- 240 at 50 hertz ?
itsu?
Poynt?
AC ?
Mike ?
Groundloop?

?
I am quite sure you have solutions to verify the "Work" end. t would be nice to pick up a 15 dollar fan and just do a quick run.
With out calibration a pony brake and such still has holes in the data ,A good control works wonders .

Just bouncing ideas around to validate ........
Thx
Chet

I would like to know a good way to measure the efficiency of a motor that is accurate,as like you said-the prony brake test can give faulse measurement's.
Short of makeing a wind tunnel with preasure and flow gauges,im not sure?
I can only go on the star rateing of the fan as an efficiency figure for the house fan motor.
I tested the DMM's useing a 23 ohm 10 watt resistor today-they are spot on.
This can mean only two thing's
1-the efficiency rateing of the fan motor is way out
2-we have something worth looking into some more.

Here are some calculations from the last video.
But we are going to drop the efficiency of the house fan down to 80% insted of the 86% it should be-this is just to be on the safe side or the low side of efficiency.

House fan
Input is 33.34 watt's
RPM is 1107
Efficiency rateing on the motor 80%
1107/80 x 100= 1383.75.
1383.75 is the RPMs needed for that motor to be operating at 100% efficiency at the input wattage
This is confirmed by-1383.75 x 80% =1107
So 100% efficiency is 41.50 RPM per watt
This is done by dividing 1383.75 by 33.34 watt's.

Results from the rotary transformer in the last test.
watt consumption-28.57
RPM-1585
so RPM per watt is 1585/28.57=55.47.

As i now know that the DMM's are reading correctly,and the fan's output is linear to rpm(as we are no where near cavitational speed)my first asumption must be that the house fan's efficiency rating is way out.

[ion]

You asked how to measure the efficiency of a motor that is accurate. I covered this a while back.

First step is to take two identical PM motor/ / generators, shafts coupled.

Run a voltage into the first and a  load resistor on the second.

Compute the power in vs power out vs. rpm. Do this for various power input and various load resistors and at different speeds.

You have now approximately half the losses in one motor and half the losses in the generator.

You will have a load line family of curves for future reference.

Now you can use one of the PM motor/generators as a load for your device under test (DUT) since you now know very accurately it's efficiency as a generator. It's output and efficiency is now "calibrated" from the prior steps. It will be a shaft load for your DUT.

This takes a bit of work in the beginning, but is worth it as a tool for future use.

We can get into detail of the finer points if this interests you e.g. how to properly load the generator.

[Chet K]

I Like this idea !!
And I love a good Cheap quality Tool,
Especially when we can collect good quality Data that is completely valid!![and Calibrated too].

@ION
Thanks for taking the time to help and Comment !!
@TM
Thank you for being so Darn good at what you do,and also for your "brutal Honesty" when it comes to manning up about
what is actually taking place on your bench!!

There are some wonderful possibilities in a group that can have the ability the will and the tools to really work thru a
project!
Thx
Chet

[TinMan]

You asked how to measure the efficiency of a motor that is accurate. I covered this a while back.

First step is to take two identical PM motor/ / generators, shafts coupled.

Run a voltage into the first and a  load resistor on the second.

Compute the power in vs power out vs. rpm. Do this for various power input and various load resistors and at different speeds.

You have now approximately half the losses in one motor and half the losses in the generator.

You will have a load line family of curves for future reference.

Now you can use one of the PM motor/generators as a load for your device under test (DUT) since you now know very accurately it's efficiency as a generator. It's output and efficiency is now "calibrated" from the prior steps. It will be a shaft load for your DUT.

This takes a bit of work in the beginning, but is worth it as a tool for future use.

We can get into detail of the finer points if this interests you e.g. how to properly load the generator.

Hi ION
I do indeed have 3 x 24 volt PM motors that are identical.
The motors are 400 watt rated-out of electric scooters.
They are a brushed PM motor.
So i would be very interested in makeing one a test bed for load tests.

My immediate question would be this
Lets say that our generator is 80% efficient,and we have it coupled to the motor unloaded.
The motor is drawing say 20 watts to run the setup.
If we pull a 8 watt load from the generator,should this reflect as a 9.6 watt increase in power consumption on the prime mover?
This would make sence to me when looking at a gasoline generator.
Drawing say 3kw would need say a 4kw gas engine to do so.
Any load drawn from the generator requires a higher amount of power input from the prime mover.

So in theroy,any load i draw from the generator should reflect on the prime mover !+ the 20% loss!

[ion]

Hi ION
I do indeed have 3 x 24 volt PM motors that are identical.
The motors are 400 watt rated-out of electric scooters.
They are a brushed PM motor.
So i would be very interested in makeing one a test bed for load tests.

My immediate question would be this
Lets say that our generator is 80% efficient,and we have it coupled to the motor unloaded.
The motor is drawing say 20 watts to run the setup.
If we pull a 8 watt load from the generator,should this reflect as a 9.6 watt increase in power consumption on the prime mover?
This would make sence to me when looking at a gasoline generator.
Drawing say 3kw would need say a 4kw gas engine to do so.
Any load drawn from the generator requires a higher amount of power input from the prime mover.

So in theroy,any load i draw from the generator should reflect on the prime mover !+ the 20% loss!

Correct, you will always need more power from your prime mover because of the losses in your generator.

Once you can quantify the losses of your scooter motor, you can hook it to any motor of similar size and use the scooter motor in generator mode as a variable brake with the advantage that you now can have a measurement of power (Voltage  squared over the load resistance in ohms equals power in watts). You will also know the losses in the scooter motor so any additional loss is from the driving motor.

First step, lets hook two of those scooter motors together, run one from your power supply, load the other with a known value resistor and we can begin to find out the efficiency of these motors.

The power available in your load resistor can tell you much regarding the efficiency of your system.

Total power input= heating loss in driving motor+ Heating loss in braking motor + heating loss (power) in resistor.

Since a 24 volt 400 Watt motor will draw approximately 16.67 Amps, our load resistor will need to be:

At 400 Watts , 1.44 ohms
At 200 watts , 2.88 ohms
At 100 Watts, 5.76 ohms
At 50  Watts, 11.52 ohms
At 25 Watts,  23.04 ohms
Maybe start with a 10 ohm resistor of 60 watt capacity, it will make calculations easier.
Run your power supply up slowly and make power input readings at  6 volt intervals until you reach 24 volts.
Also record voltages across the 10 ohm resistor at each interval.
If possible record RPM's at each interval, it will come in handy later.
This is the calibration / quantification stage.
After this test we will see we have losses in each scooter motor, but can calculate the efficiency of each.

As an example lets say we are putting from our power supply into the drive motor 50 Watts or 24 volts at 2.083 amps.

But on our 10 ohm load resistor we are getting 20 Volts. This is equal to 40 Watts, so since the motors are identical we can roughly assume to be losing 5 Watts in each motor or 10% of our input power in each motor. Our motors are each then around 90% efficient at this loading.

Efficiency will change for higher loading, but we will have the data if we do the test.

[TinMan]

Correct, you will always need more power from your prime mover because of the losses in your generator.

Once you can quantify the losses of your scooter motor, you can hook it to any motor of similar size and use the scooter motor in generator mode as a variable brake with the advantage that you now can have a measurement of power (Voltage  squared over the load resistance in ohms equals power in watts). You will also know the losses in the scooter motor so any additional loss is from the driving motor.

First step, lets hook two of those scooter motors together, run one from your power supply, load the other with a known value resistor and we can begin to find out the efficiency of these motors.

The power available in your load resistor can tell you much regarding the efficiency of your system.

Total power input= heating loss in driving motor+ Heating loss in braking motor + heating loss (power) in resistor.

Since a 24 volt 400 Watt motor will draw approximately 16.67 Amps, our load resistor will need to be:

At 400 Watts , 1.44 ohms
At 200 watts , 2.88 ohms
At 100 Watts, 5.76 ohms
At 50  Watts, 11.52 ohms
At 25 Watts,  23.04 ohms
Maybe start with a 10 ohm resistor of 60 watt capacity, it will make calculations easier.
Run your power supply up slowly and make power input readings at  6 volt intervals until you reach 24 volts.
Also record voltages across the 10 ohm resistor at each interval.
If possible record RPM's at each interval, it will come in handy later.
This is the calibration / quantification stage.
After this test we will see we have losses in each scooter motor, but can calculate the efficiency of each.

As an example lets say we are putting from our power supply into the drive motor 50 Watts or 24 volts at 2.083 amps.

But on our 10 ohm load resistor we are getting 20 Volts. This is equal to 40 Watts, so since the motors are identical we can roughly assume to be losing 5 Watts in each motor or 10% of our input power in each motor. Our motors are each then around 90% efficient at this loading.

Efficiency will change for higher loading, but we will have the data if we do the test.

Ok well my first problem is i dont have a variable power supply any more-as i fried it some time ago.
I do have a 20 amp PWM that i could put a large smoothing cap on the output.
As i live in the country,it means a 4 hour trip (there and back) to pick up a power supply.

I am building one at the moment,in the form of a low voltage variac.
I will start to build a test bed to bolt the two motors on,so as we can start these measurements.
But i must ask you this.
If i have the motor hooked to the generator,and draw6.5 watts from the generator-but only get an increase current draw to the prime mover of 4.8 watt's-what dose this mean?

Thanks for your help ION,and to all.

[TinMan]

I have now included a PWM into the system.
It's frequency is 2KHz.
It is running from the 24 volt half wave AC.
So we now have created a 100Hz half wave AC that has 20 pulses in each wave.
Then we also have the comutator switching aswell.
In the video you will see that i have finaly been able to dump the high voltage spikes ontop of the pulsed AC wave,and almost eliminated all the pulse current within the AC wave.
I am yet to check the efficiency of the system with a load on the shaft,but will be doing that tomorow.
http://www.youtube.com/watch?v=X163_IilwHk

[TinMan]

Correct, you will always need more power from your prime mover because of the losses in your generator.

Once you can quantify the losses of your scooter motor, you can hook it to any motor of similar size and use the scooter motor in generator mode as a variable brake with the advantage that you now can have a measurement of power (Voltage  squared over the load resistance in ohms equals power in watts). You will also know the losses in the scooter motor so any additional loss is from the driving motor.

First step, lets hook two of those scooter motors together, run one from your power supply, load the other with a known value resistor and we can begin to find out the efficiency of these motors.

The power available in your load resistor can tell you much regarding the efficiency of your system.

Total power input= heating loss in driving motor+ Heating loss in braking motor + heating loss (power) in resistor.

Since a 24 volt 400 Watt motor will draw approximately 16.67 Amps, our load resistor will need to be:

At 400 Watts , 1.44 ohms
At 200 watts , 2.88 ohms
At 100 Watts, 5.76 ohms
At 50  Watts, 11.52 ohms
At 25 Watts,  23.04 ohms
Maybe start with a 10 ohm resistor of 60 watt capacity, it will make calculations easier.
Run your power supply up slowly and make power input readings at  6 volt intervals until you reach 24 volts.
Also record voltages across the 10 ohm resistor at each interval.
If possible record RPM's at each interval, it will come in handy later.
This is the calibration / quantification stage.
After this test we will see we have losses in each scooter motor, but can calculate the efficiency of each.

As an example lets say we are putting from our power supply into the drive motor 50 Watts or 24 volts at 2.083 amps.

But on our 10 ohm load resistor we are getting 20 Volts. This is equal to 40 Watts, so since the motors are identical we can roughly assume to be losing 5 Watts in each motor or 10% of our input power in each motor. Our motors are each then around 90% efficient at this loading.

Efficiency will change for higher loading, but we will have the data if we do the test.

Hi ION
I have orded the drive coupling's,but all i could get was blank's-as they dont make one's with bore's less tha 12mm here-And i need 8mm.
So i will have to pop them in the lathe and bore the 8mm hole-but no problem there.
I have most of the resistors needed,but are only 10 watt resistor's-i am unable to find anyone who supply's the 60 watt resistors?-but im still looking.

In reguards to the DMM's on the input,i have checked there accuracy by removing the rotary transformer and placing a 23.6 ohm resistor across the output.
The meters read 22.3 volts and .942 amp,s
Ohms law says the resistence should be 23.67ohms.
This tells me that the DMM's are very close to being correct.
Is there another way to check to confirm this?

I also have the PWM in the system,so we can now go from 0 to 20 volt's-this PWM wont go any higher that 95% pulse width.
But this should enable us to get the measurements we need -i hope?

Brad

[ion]

The values  I specified don't have to be exact, as long as you can accurately measure what you have on hand. Then E^2/R does the rest.

They do require thermal stability so best to grossly oversize them so there is not too much heating and drift.

I fashion what I need from low drift Manganin wire.  Nichrome can be used but it does drift with temperature.

50 watt resistors are standard. If you test at full power, you will need a 400 watt load.

Improvise with what you have on hand, series / parallel combinations.

Once we get started we can determine a course for future tests with various loads etc.

[TinMan]

This is the first graph plot for the standart motor and generator setup.
The P/in watts is measured as follows.
Amp meter between FWBR(from transformer)and PWM. PWM frequency is 2kHz
Volt meter is across FWBR-set to DC-pulsed DC setting.Frequency is 100 HTz
DMM readings were confirmed useing a 23.6 ohm 20 watt resistor on P/out of the PWM-and are within .03 watts,using ohms law.

P/out is measured using the following method.
Negative from generator to negative on 1000uf 40 volt cap.
Positive of generator-through amp meter-to positive on cap.
In this test ,a 47ohm 10 watt resistor is placed across the cap as the load.
Volt meter is also across the cap.

The next test graph will be taken using a 23.6 ohm resister as the load.

[ion]

Good work and nice graph Tinman ,

Overall efficiency seems rather low roughly 20 to 24 % depending on whether idle power is subtracted.

I suspect the rotational losses will increase slightly as RPM's are increased, so maybe a chart of power draw with no load on the generator at higher RPM's will reveal where the low efficiency at low loading is coming from.

As you start to approach 400 Watts, the power needed to satisfy rotational friction loss will be a smaller percentage of power required, so overall efficiency should go up.

So right now it takes approx. 5.7 Watts/2 in each motor just to spin up to 500 RPM, or roughly 2.85 Watts per motor.

Here is a drawing of how I would do it, Amp and Voltmeter post capacitors to prevent false readings.

[TinMan]

Hi ION
Problem is,i cant have a cap on the input-as my other motor relies on the rectified wave form to opperate at it's best.
So i was trying to keep the power supply and meters all the same with both test setup's.
This way,even if the DMM's are out slightly-the error would remain the same in both test's.
This would mean that the graphing would still show a correct result between the two setup's-as all meter positions would remain the same.

We wont be going any where near 400 watt's,as the power supply setup i have wont allow it.
Also my rotary transformer wont do that many RPM with the power at the level we have.

I will plot a graph today with no load on the generator,and post here when done.

Here is how i have it set up at the moment-have modiffied your schematic.

[Magluvin]

Hi ION
Problem is,i cant have a cap on the input-as my other motor relies on the rectified wave form to opperate at it's best.

Ahh.    So if you have just smooth DC for input, there would be no 'transformer' action from the armature to the stator secondaries?  Very nice.    

Mags

[Magluvin]

And, if no transformer action, no special effects.  

Mags

[TinMan]

And, if no transformer action, no special effects.  

Mags

You got it Mag-no magnetc reversal of the secondary coil = no attraction to the next rotor field approaching the stator core.