e-Buggy: Alternator Conversion

widodo

100 mW
Joined
Feb 17, 2010
Messages
39
Just Join today, but already found so many interesting information for my current project

It's a 2-seater buggy, previously powered by 5.5HP brigg&staton Engine, but now powered by BOSCH 80A converted alternator running in sensorless mode.
this video was taken from the early stage of the project, notice that it was powered by turnigy 2600mAH hanging down on the seat rather than the car battery in the car



As the plan is to use the buggy for 2 children and my self, I'm using a two stage chain gearing giving up a total gear ratio of 21:1. So far it was able to climb 25 degree ramp as shown in the video below.


I'm currently using HK 150 car brushless ESC running in sensorless mode, and a brushed ESC for the rotor supply, both ESC controlled by AVR8535 processor, which gets an input from joystick functioning as throttle, and a slider potensiometer as gear selector. For a very powerfull start up, I'm powering up the rotor up to 7V, and decreasing it to 3V in 'Auto' gear setting, an a fix voltage supply for 1st, 2nd, 3rd and 4th electronic gearing. It is now run with 70AH car battery which proves more than enough for 5KM rides with only 1.5 voltage drops.

A very usefull nature of this kind of motor (compare to the permanent magnet one) is that you could have a torque/RPM on demand. Boost the rotor voltage up for the highest torque but with the lowest rpm (I could go as low as 70RPM :D , for a sensorless motor this quite amazing), and go as high as 7000+rpm with 12V supply (3V rotor supply) :D
 
Nice.

It has been discussed, but this may be the first real example of using a brush controller for field-weakening.
 
That's verrrry cool. 8)

Are you running the motor on 12v?

Do you know what the current draw is?
How much current does the field winding take?

I think you could increase the voltage to 24v or higher with the existing windings for more power.
 
Thanks Guys,

But first, please forgive shall there be any typo or grammatical error as English is not our main language here in Indonesia.

@Fethcer: Yes for the cost efficiency shake, I'm running in 12V. Having a motor that you could program in term of Kv, actually removing the requirement to go higher in Voltage supply. Nevertheless, as the torque is linear to the rotor flux, hence linear to rotor voltage supply to certain degree, but the downside is that it is also linear to the BEMF which in turn limits the RPM, going up higher in motor (Stator) supply will gives larger RPM band/per given rotor flux as it will take twice the rpm before the BEMF reach the stator supply level. But as I could easily decrease the rotor flux, I choosed to just decrease it, hence decrease torque (which I don't really that big after the buggy start to moves) hence, decreased BEMF and Increased the RPM.
My rotor got around 3 Ohms resistance, given max 12V supply, the theorytical is only 4Amps, I think its saturated there as I don't see noticeable increase around 3.5-4 V supply. Using Clamp meter, the maximum amps draw is around 40Amps, that when I goes up the ramp.

@Grindz145: I'm Controlling the both ESC with Atmel AVRMega 8535 (cost less that USD90 including LCD, programmer cable and 2 IO expansion), you cold easily use other AVR as long as you got the build in Analog to Digital Converter (ADC). For beter understanding of how the element interact with each other, I had install an 2 lines LCD just to monitor throttle signal, rotor RPM, Battery Voltage and rotor Voltage
utf-8BSU1HMDAxMjEtMjAxMDAxMjgtMDcxN.jpg

All the number, except for the RPM are still in its 10-bit Decimal value with a 5V reference. Hence, I used some voltage devider and op-amp to bring the measured value down to 0 - 5V. I used to work with assembler until I found BASCOM AVR for programming, which only need around 30 instruction to make a those thing happened.

Further during the project, I also found that sony Play station joystick is really suitable for the throttle control, hence used it on this buggy as throttle and active-braking control
utf-8BSU1HMDAxMjUtMjAxMDAxMzAtMTUyN.jpg


Sorry for the focus issue as the photo was taken with my BB. :oops:

I'm currently thinking to put the Hall inside my motor and thanks to the forum already have some ideas on how to do it. Furthermore,I found out that brushless computer fans is a source very affordable hall sensor. Inside you cold find 276ICs which is two open collector magnetic switch able to detect both north and south, I used this for my rpm sensor :D
 
Awesome. So its basically excite the stator coils and run. You didn't break it out into 3 phase then?

Do you commutate each phase of stator winding?

This is so awesome :mrgreen:
 
Thanks,

Well alternator is basicly a three phase generator/motor with a controlable rotor flux through a slip ring (i.e. carbon brush), so what I did was removing the diode plane and rewired the cable into a "delta"
utf-8BSU1HMDAxMDctMjAxMDAxMTctMTcxM.jpg

In this picture I used ND 30Amps alternator using the default "Star" wire configuration with less power (the Delta provides 1.6 times of power compares to the star if I wasn't mistaken)

Once you power-up the rotor through the slip-ring, it become a normal brushless in-runner motor, so you could spin it off using normal RC brushless ESC, just make sure that you use the large one as the current for start up is quite big. I always used at least 100Amps rated ESC, which is now is quite affordable through the online stores in HK :D .
 
Excellent thread, wimodo.
 
Great work! I love the active field weakening. Very cool!

-Luke
 
Field weakening is blowing my mind right now... Minimizing BEMF makes so much sense. I have to do this too! Thanks Widodo! I wonder what the breakdown voltage is of these....100V? :twisted:
 
Yup, I did found out it was really blow my mind, hence caused of many moonlighting work... :D

If I may, I would suggest to use this test arrangement to understand all interaction beter
utf-8BSU1HMDAxMDYtMjAxMDAxMTctMTcxM.jpg


I used a cheap Brushless ESC, a Brushed ESC and two servo tester/driver, see how "weakening the rotor" will increase the rpm, you could also feel the torque by your own hand, it will helps a lot

Anyway, would love to help if you want to start yours, in another forum I'm currently helping out a guy in sweeden for his 6 Yo son's buggy
 
Could there be any benefit to combining field weakening with permanent magnets? eg have both coils and permanent magnets on the rotor, so the coils are only used when higher torque is needed, but at cruise the field coils are simply turned off for greater efficiency.
 
You'd have to map out the magnetic fields of the magnets vs the coils so that they don't interfere with each other. Could take quite a bit of experimentation.
 
I must say " Bagus" keep up the nice work ! i like the micro controller setup and readings. This is gonna be a nice transport in indonesia!
 
ha ha, good bahasa MrKang :D , Thanks

@Jhon in CR: I think have seen that, they event activate the coils to fight out the magnet and I'm agree with amberwolf that its quite challanging in term of mechanical and electrical.

On the otherhand, Other that its weight, electromagnetic rotor only consumes 1-2 Amps, that would translate only to 24Watts, but you get the ability to have a "regulated" charging of regenerative braking so you could charge the battery directly without the possibility to blown it off. Not mentioning the ability for free-wheeling, in large brushless motor, its very high torque in brushless motor impossed the need of one way bearing on the gear, at least that what we do in RC helicopter otherwise the plastic gear tooth will evaporates on the first throttle. :shock:
 
widodo,

very cool work! Bascom does make life simpler I have to agree!

I can't let my son see this or he will end up with a powered pull wagon (now that he finally realizes daddy makes odd vehicles and devices for fun and profit, he figured out I could make his little red wagon go without me pushing)... he even said I could "remote control" it like the RC10 we have to keep him "safe".

Right now my excuse is the hubs are too big (direct drive motors) and too expensive for a toy (Astros) but if he realizes we could power his wagon with a belt a pulley and an old car alternator... lets just say my car won't start the next time I go to use it.

Thanks.

-Mike
 
John in CR said:
Could there be any benefit to combining field weakening with permanent magnets? eg have both coils and permanent magnets on the rotor, so the coils are only used when higher torque is needed, but at cruise the field coils are simply turned off for greater efficiency.

Yes, I think that would be quite doable. Permanent magnets could be glued into the crack between the pole pieces to give them some permanent magnetism. Add coil current to increase the field. You'd just need to check which side is north and which side is south when the field current is applied to make sure it's polarized the same way as the permanent magnets.
 
Hows this going? :) alternator engine is coool :)
 
Any updates? Also did you say you post on another forum? I like the budget approach. It looks really promising.
 
Apologies for not being able to updates as our fiscal year end process really tied me down.

Anyway, our eBuggy has been our main mean of transportation for the morning walk every saturday and sunday. But being a sensorless means that there are huge of possible improvement from the performance side.

Hence, started up my Enhancement project for a Sensored Alternator conversion with Integrated motor controller.

The Idea is to go for a sensored commutation mode with budget motor controller which has a build in rotor controller. To make thing more interested, put an initial goals of:
1. Low Cost sensor and should use the rotor magnetic field rather than go for optical sensor wheel /separate sensor. The consequences would be that this will be a fix non adjustable sensor.
2. The Controller should be easily constructed, cheap and with High Amps (at least 100A) and Hi Voltage (at least 30V) Half Bridge Driver. The Driver should be easily upgraded into IGBT

To push the research cost down, I'm using small Nippon Denso 35A converted alternator (USD20) and use plain old TIP2955/3055 transistor in the development phase. The Savvy driver will only be implemented if the Transistor model already working flawlessly.

Project BLOG:

1. Sensor & Installation: The First attempt is to utilized AN276 hall sensor IC from salvaged Computer fans. I do have the experience with this IC for my RPM meter and it works ok. The only challenge is as this IC gots 4 pins so its wider than normal hall sensor. Trying to install this near the end or rotor as in the picture below
utf-8BSU1HMDAxNjYtMjAxMDAyMjItMDc0N.jpg


It has been found though the sensor work flawlessly, but the magnetic field was not. Due to the arrangement of the Claw Pole, if the sensor was put too near to the end of one pole, there isn't enough opposite flux to toggle it as it should be. Hence the only solution is to have the hall sensor installed exactly in the middle of the stator where both claw size is exactly the same. But it will means that I could not use the wide sensor for this and need to make a grove in the stator tooth.
Hence, ordered HONEYWELL S&C SS40A (SGD 1.98 each) from farnell singapore, had a machine shop make the grove and voila... the sensor work as expected (six commutation phase).

2. Controller logic
Implements the sensor input into 2 external interrupt channel and 1 Analog Comparator channel, as I need to detect toggle event in every sensor. So far the Sensor & Interrupt works ok (with output only to LCD screen), I'm currently working on the Transistor driver when the work held me down. Should have the POC model in a week or two. Atmel's application notes helps a lot especially on the commutation logic. Still looking out for additional PWM output for the rotor controller without huge processing expense as if approached through normal port.

3. Controller Driver
Rather than going into a standard push-pull on each Phase driver which requires N & P Channel MOSFET, I'm planning to used IR2101 ICs (SGD 4.66 each, one in every phase output) to drives 4xIRF3007 (75V/80Amps N-Channel Mosfet giving a total of 160Amps for each phase, its a clearance sale item hence it only cost SGD 2.24 each). With this approach, I don't have to use the P-Channel mosfet which was more expensive do to its more complex manufacturing process. Nevertheless, the boostraping technique enabled by the IR2101 is quite complex hence to isolate the problem, this will only be implemented once the transistor Model works ok. So far, the IR2101 documentations shows that MOSFET-IGBT conversion is just plug and play with a special attention only on Gate Voltage threshold.

Will share the progress later on if anybody interested :D
 
There are some way way better FET options out there. IRFP4468PBF. It would save you money, reduce complexity, and make for a much cooler running controller.

We've dubbed this style of direct controller, "cavemen controller." Its been discussed on here a few times, but I think you will be the first to have actually made one. :) Very Cool! Make sure you use a FET driver with anti-pass-through logic, or just wiggling the rotor forward, then backward a bit at the wrong spot will result in instant plasma ball from where your FETs used to be. Lol
 
Back
Top