Electric Motor Secrets

[Peter Lindemann]

Next Test

Goodmorning everyone
Peter, as for the reedswitches, I am sure that the problem was the offcenter comutator, because when I could get only three pulses per revolution, I turned the comutator where the fourth pulse should be, and then if I pressed the comutator towards the reedswitch very hard, the whole thing bent a little bit, just to overcome the offcenter ( a half of milimeter or so) and the reedswitch did switch. Anyway, this is not important anymore, althought the reedswitches are good and easy to operate if you have your comutator right and everyone should start with them.

I did some experimenting. At first I disconnected and completely isolated all three outputwindings from each other and the circuit. I hooked everything up like in previous circuit without the isolated output. Also I put a computer fan near the transistors for cooling purposes. Again, I firstly tried to run the motor with only one transistor. Rhe RPMs were bad - only some 200 or 250 (i did not measure them) also the transistor got very hot fast and the fan could not handle that heat, the amp draw was about 0.8A at 24V (just like when isolated output was used). So I stopped the test.
Now I tried to run the motor with all three transistors. This time The RPMs were rising very fast and stabilized at about 2383 The amp draw was 1,48A at 24V. But the interesting thing is, that when using all three transistors, they did not get hot, only warm, nothing the fan couldn't handle. In the short test (a minute or so) the charging battery went up from 24.8 to 25.4V
Both batteries consist of two 7Ah 12v batteries in series. I should upgrade to bigger batteries, because this amp draw of 1.49A is not healthy for these little the batteries. Also the charging of the output batteries is happening to fast. I should charge and discharge them at C20 rate. But its good for short tests

Thank's
Jetijs

Jetijs,

Excellent. Thanks for going back over everything and getting the motor running on all three transistors. 2383 RPM is pretty good speed for now.

Of course, I still recommend adding heat sinks to the transistors when you can and your idea of bigger batteries is also a good idea.

Now, I would like you to go back and hook up the isolated output on the other three windings. You will find that the second battery charges just as fast with this alternate set-up and that the motor operation is not significantly different. You can leave the current output arrangement in place, but disconnected. That way you can run comparative tests back and forth with each system to see motor speed and charge effectiveness with both ways. If you do go back and forth between these wiring set-ups, MAKE SURE that it is wired correctly each time you make the change or you can blow your transistors.

There is a reason I want you to understand this isolated output system because it makes certain other things possible later.

Great work,

Peter

[Jetijs]

Yes Peter, you are right. I just tried the isolated output circuit again and with three transistors it works just the same. The battery on the output charges just as fast with both circuits and the RPM is the same. But why is that when I am using only one transistor for motor to run, the transistor heats up quickly, but when using three transistors, they get barely warm althought the total current draw has risen?
Thanks,
Jetijs.

[Peter Lindemann]

Making Progress!

Yes Peter, you are right. I just tried the isolated output circuit again and with three transistors it works just the same. The battery on the output charges just as fast with both circuits and the RPM is the same. But why is that when I am using only one transistor for motor to run, the transistor heats up quickly, but when using three transistors, they get barely warm althought the total current draw has risen?
Thanks,
Jetijs.

Jetijs,

Very good. I'm glad the motor is stable in the new configuration and is operating the same with both output circuits.

The reason your transistors get hot when only one is hooked up is probably related to a combination of ON-TIME per cycle in relation to heat dissipation rate. When only one transistor is used, the magnetic field strength across the gap is probably under 400 ampere-turns. This turns the rotor slowly, so the transistors are in the ON-TIME window for a long time, drawing maximum current. With no heat sink, the heat builds up quickly. When all three transistors are used, the motor can speed up to the rise-time maximum quickly so the time the maximum current is across the transistors is very low. Your amp meter is sampling at a certain rate and since the RATIO of ON-TIME to OFF-TIME is always the same for the motor, no matter what speed it is going, it looks the same to the meter but is very different in the transistor.

Now, I'd like you to connect your scope probe across your amp meter. It has a low resistance shunt inside that it uses to produce the meter reading. We can use this also to look at the current on the scope. Then you will have both the AVERAGE current reading on the meter and the current wave form information on the scope, generated from the same source. The meter reading should be equivalent to the area under the curve on the scope in relation to the full cycle time. From this we can determine what the true peak current is that the transistor is switching.

The wave form should be a straight ramp up with a vertical drop-off. It will also have a flat top at start up which will get shorter as the motor speeds up. At full speed, the flat top may (or may not) disappear completely leaving the simple ramp or "saw-tooth" wave.

Let me know what you find.

Peter

[Jetijs]

Peter, I made the scope shots.
Here's where my amp meter is hooked up:


And this is how the current waveform looks like:


This is at full speed, drawing 1.48 amps at 24V.
And here's how the waveform looks like when I load the motor down:


As the motor is gaining speed again, the flat top of the wave gets shorter and shorter

[Peter Lindemann]

Excellent

Jetijs,

Very good. So, my predictions were pretty close.

From the scope shot it looks like we have this information:

Inductive rise-time = 1.5 divisions
Total ON-TIME at top speed = 3 divisions
Maximum current = 13 divisions
Total wave length at top speed = 17 divisions

Current wave-form at top speed occupies at least 75% of the area defined by 3 time divisions and 13 amplitude divisions. This area is also defined by the ammeter reading of 1.48 amps.

So, with the settings on your scope, we should be able to quantify these readings, even though we do NOT know the value of the resistor we are looking across. Also, we can cross-check your tachometer and scope time-base since the RPM should equal the frequency of the wave-form divided by four, since there are 4 pulses per revolution.

After reporting on this data, please move the ammeter to the negative side of the output (between the negative terminal of the battery and the group connection to the three output coils) and run the test again. This will show us how much energy we are recovering. In terms of current, it looks like you are getting about 1/3rd back, or about .5 amps, but at a higher voltage. So this time, please report the voltages of both batteries as well so we can compare WATTS in and out, not just current.

Great work!

Peter

[lighty]

Also, we can cross-check your tachometer and scope time-base since the RPM should equal the frequency of the wave-form divided by four, since there are 4 pulses per revolution.

Um, shouldn't that be (WaveFreq/4)*60? I mean we're trying to calculate RPM not RPS.

[Peter Lindemann]

Yes

Um, shouldn't that be (WaveFreq/4)*60? I mean we're trying to calculate RPM not RPS.

Lighty,

Of course, you are right. This time it was me "assuming" Jetijs (and everyone else) could figure out the relationship between frequency and RPM. Thanks for clarifying!!

Peter

[Peter Lindemann]

OFF TOPIC: Humorous Interlude

Everyone,

Here is a great "hi-tech" laugh. Enjoy!!!

Peter

Retro Encabulator

[gabriolaman]

haha i remember seeint thats a while back i dunno what the hell hes sayinglol

[Ted Ewert]

Everyone,

Here is a great "hi-tech" laugh. Enjoy!!!

Peter

Retro Encabulator

Now I see where Beardon gets his material.

[Jetijs]

Ok Peter
Here are my new results. I made a new scope shot just like the previous one, just this time I adjusted it so that the wave aligns to the grid for easier calculations. So Here is the shot with scope leads across the ampmeter on the primery negative side:



The amp draw decreases slowly while motor is running as the speed increases and it stabilizes at 1.46A at 22.9V after about two minutes of motor run time.
So next I put the ampmeter between the common negative of the coil and negative terminal of the charging battery. I put the scope leads across the ampmeter and got this waveform:



There are longer spikes, but in this picture they are hard to see. Here's a another picture so you can see the spikes better:
http://www.emuprim.lv/bildez/thumbs/...9__medium_.jpg

The readings are as follows:
Current draw from the primary battery is 1,46A at 22.9V. Current on the amp meter on the charging battery is 0,17A and the voltage on the charging battery is 25.6V. So when we calculate watts, we get 33,43W getting IN and 4,35W getting OUT. That would be about 13% that we are recovering, right?
Now, in order to recover more, we need to adjust everything so that the current waveform on the primary does not get that flat top, but is cut just before that so that the waveform would look like a straight line ( / ) and occupies only about 50% of the pulse area on the scope. This illustrates what I mean:


The red area represents the current going IN and the yellow is what we can recover. The longer the top flat line, the bigger the difference in the IN/OUT ratio. We need the waveform to be just like in the first illustration with the 50:50 ratio for maximum recovery. Am I right?
So what do we need to do, to get rid of that flat line? Reduce some more turns of the coil? Or should I better make the ON time a little bit shorter on my commutator wheel?
Thanks,
Jetijs

[Peter Lindemann]

Totally Correct

Jetijs,

Everything you reported and interpreted from the data is correct.

At this point, I would leave everything where it is electrically, and cut a new rotor piece with extremely close tolerances. This will allow the motor to produce more torque for the same electrical input and to speed up with the timing you have so it will run in the window you describe. You are absolutely correct, all you can have back is the Yellow Triangle, so the best system will only put in the matching Red Triangle before shutting off. We just have to get the maximum torque for the input.

In the meantime, you can play around with advancing or retarding the timing to get the maximum speed out of the rotor you have and watch how the power input/output ratio changes. You can also look at the output recoverable from the primary windings directly using the other output circuit. It should be slightly more than on the isolated windings, but very close.

Great work!!

Peter

[Jetijs]

Ok, thanks Peter
I will order a new set of rotor plates from my laser cutting guy. In the mean time I will do some experimenting and post all the results and observations. Also I still have not changed the magnetic bolts in my rotor with non magnetic ones. And in closer inspection I noticed that the gaps in my timing wheel are a little too big, because the impulse turns off just a little bit after the middle of the 30 degree turn. So adjusting the gaps and changing the rotor bolts may be enough to gain the speed needed to get the correct waveform Also, If I use the previous circuit without the isolated output, I have three strands of wire left unused. Maybe I could add another transistor in my circuit and use four strands of wire as the power coils? That also should increase the speed.
Thanks,
Jetijs

[lighty]

Also, If I use the previous circuit without the isolated output, I have three strands of wire left unused.

You could also use isolated strands to charge the capacitor and then discharge it to the main power battery. Just a thought.

[Jetijs]

Yes, lighty
I was thinkink about that too, that way I would not have to use two battery sets. Also there is some radiant charge in these BEMF impulses that are going into the charging batteries, so there is also the battery conditioning process happening, just like with Bedini SSG. This means that if I swap these batteries around for some cycles, the battery capacity should increase to some degree.