[ODaily]

So I recently became interested in converting an automotive alternator into a DC motor. Google was helpful, but I had to get the info by bits and pieces from various places. I decided that I'm going to go ahead and convert an alternator, AND document it from start to "magic smoke". I really just want to explore the idea, learn some, and leave behind a record for the next poor schmuck who decides to try this. So.........

The givens:

I'll be using a 3G (3rd generation) Ford 130 amp alternator. Because, A: I have one (free), and B, I really like the big heavy mounting lugs on the case. If I do mount it to something, then that'll be super handy.

My intent:

To see what and how it shakes out. The basic plan is to duplicate what I've seen done before, and then try to step it up from there in terms of output. At least until magic smoke occurs.

So stage one will basically be a physical modification of the Alternator into Alter-motor, and getting that moving. I have already ordered a RC type Speed Controller (HobbyKing WP SC8). It's rated at 120 amps, and 2-4 lipo batteries. That's a really poor (IMHO) way of rating voltage, it basically works out to 14.4 volts max. As I'll start with a 12 volt lead acid battery, that should be fine. Upgrades may be necessary in the future. First it's gotta work. 120 amps at 12 volts is 1440 watts. Roughly. Real world doesn't always work that well. We'll see.

 

[ODaily]

Will be back soon. Have pics, but having problems with the place I used to use to host. I'll have it figured out soon.

 

[ODaily]

Yup, alot like that. I see alot of finished or at least finished on the benchtop projects. I've seen a few go-karts and such. Not much on the process.

Here's the front of the conversion candidate. Its from a 2003 F-150 with 5.4L motor.

Here's the back of the donor.

 

[ODaily]

First step, remove the brush and voltage regulator assembly from the back. It's 4 torx bit screw, don't yank the two in the middle (one of which is under plastic cap).

Bottom side of brush assembly. Note the uneven wear on brushes. When reusing used brushes, probably a good idea to keep most of the rotation in the same direction.

 

[ODaily]

Removing the middle two screws will separate the brush holder from the regulator. Then there are 3 places on the back side of the regulator where there are bits of plastic that come through the aluminum plate and are smashed to keep the regulator on the plate. Like plastic rivets. Cut them off flush to the surface. You should be able to pry the aluminum plate and plastic housing apart now, forever ruining it's ability to be an OEM regulator.

In between you'll find the circuitry attached to the aluminum plate. Which I heated with a propane torch until I could pop it off the glue/plate. The Plastic housing also has a number of metal connections that were snapped upon separation and a bunch of non drying (silicone ???) goop. I removed the goop as best I could and trimmed back the connectors to prevent any possible shorts. I also took the opportunity to grind away the OEM electrical connector. Be careful, one of the pins is electrically connected to the brushes, which we want to keep isolated. Make sure no contact can happen.

This is what I ended up with.

I reassembled the pieces and then cut (dremel) 2 small notches next to the screw holes. This should allow me to attach electrical connectors / wires to the two screws, each of which is connected to the brushes. All else stays isolated.

 

[ODaily]

Went back to the alternator, removed the red main battery wire connector and the nuts under it, and discarded them. The three screws that hold the case together were next, then I gently tapped / pried the rear case off.

The white half circle is the rectifier, and it's riveted to the aluminum housing underneath it, and soldered to the windings. To remove it, I heated the winding connections with propane torch, while prying up on the aluminum housing. The solder points gave one by one, until it was free.

 

[ODaily]

Next I removed the rectifier assembly from the aluminum piece underneath it. I drilled out the rivets and pried it off. The diodes have soldered on wires that can be cut / unsoldered / or just ripped off. After that discard the rectifier and use a hammer to drive out the remaining diodes in the casing. They're just pressed in. Picture shows a before and after. (I have a spare).

My bearings are good, so I have no need to go deeper. It's not hard, and I'd do it if there was any question regarding the bearings. But, I'm confident they're good.

 

[ODaily]

Tonight I was going to add a terminal block and run wire to each of my legs. I know I have multiple terminal blocks from various sources, but I can't find any of them. Anywhere. Typical me.

Tomorrow's shopping list:

Terminal block
Woven cloth insulator sleeve (to cover bare wires)
Maybe a sprocket. I might weld one onto the pulley as I'm concerned about belt slippage if I really lay into it. Will have to have a large enough center hole to accommodate the socket / nut. I think I can get one at the local Farm and Home store.


There is one thing I ought to mention about this specific alternator. Most alternators that I'm familiar with have 3 leads to the field wiring. 3 legs, 3 leads, makes sense. The other end of each of the coils is connected together deep in the spaghetti nest of wiring, and you don't really see it. Sometimes that center connection is brought up out of the coils as a 4th lead.

This alternator has 6 leads. I almost rejected this style alternator for another as I didn't understand it, and couldn't find any real info on how it was wired. Until I found a patent application on google that just happened to contain both an exploded view and electrical diagram with the connections numbered / labeled the same in both.

You can read it here. https://patents.google.com/patent/US6528911

If you hold the field so that the 6 leads are arranged with 2 on the left, and 4 on the right, as pictured above, then number them 1-2-3-4-5-6, from left to right, then 1, 3,and 4 should be tied together as the common field connection, and 2, 5, and 6 are the leads that will be powered by the speed controller.

 

[ODaily]

Ended up stripping apart an old power supply to get the 6-32 screws I needed to mount the terminal block, but it's in there. I also soldered the leads to some solid copper wire and ran them to the block.

I wanted to run a twisted wire to double up the size. It didn't come out as well as I'd hoped, so the center connections are bare twisted copper, and the leads are insulated wire.

Not alot of room in there. The squarish looking gap in the plate at 10-11 O'clock is where the brushes go, and the area at 12 O'clock actually has the rear housing touching giving zero head room.

I'm gonna wait on running the leads out, until I get the controller I ordered.

I also found a grommet to go in the rear housing hole where I'm going to bring the wires out, and soldered ends onto a connector for the brushes and got it ready to go.

Looks like I'm waiting on UPS.

 

[ODaily]

The controller showed up, had to go get some connectors for it. They expect you to use one of their motors, so it only comes with one side of the connectors. Anywho.

I made some leads and brought all 3 legs, and the center connection out of the case. I had to give up on the grommet I used. In fact, I had to cut a section out of the case to give the wires enough space to align the case during install. They ended up where the grommet used to be, but still needed the room to get there. Mostly because I hard mounted them to terminal block so close to where they exit. No give. Also got the brushes re-installed and check for lack of continuity between wiring and the case.

 

[ODaily]

Picked up a sprocket, but it's too small by far.

I'll go back and get another soon, but take the pulley with me.

In other news.
I'd been thinking of making a crude dyno to test the motor as I worked through it. Then it occurred to me that pushing my oversized self down the road would make a much funner test. I can evaluate it by top speed reached. With that in mind, I went looking for a go-cart frame, but came up with a stripped Polaris 2wd 4-wheeler frame instead. There's just enough left to meet my needs. Steering, brakes, rear axle, but No seat or body. Ought to be gloriously hack looking. Still at a buddy's house, but I'll drag it home soon enough.

 

[ODaily]

i'm thinking this alter-design will stall at a very low torque and not be effective or efficient as a motor.

No doubt about that. At least as it currently is. I should be able to get something useful out of it by upping the volts later. Mostly a learning experience.

My original dyno thought was to mount a disc brake rotor and spindle/brake caliper so that they both could rotate freely. Then add an arm from the caliper to extend out and rest on a digital weight scale. With the motor running to turn the disc rotor, you could use the brake to almost stall the motor.

If the arm was 1ft long, then the weight on the scale in lbs (assume set to zero with arm at rest on it) would be the ft/lbs of torque. If you checked the rpm at the same time, then horsepower is only a little math away. Take a bunch of measurements across an RPM range and you could graph the results like any real dyno.

But land speed of a big guy on a 4-wheeler frame seems like more fun.

 

[ODaily]

Finally got around to hooking everything up. Needed a way to provide the ESC with a throttle signal, and I just happened to have an unused Arduino Mega laying around, so I fixed that up, and now I have this.

Tried to take it for a spin (yes, it was deliberate ), but kept having problems. It will rotate, but only in a narrow and kinda weird part throttle band. Very torquey, slow RPM. I forgot to plug in the armature winding at one point and got an entirely different response. Tried to spin, but at very low RPM and throttle.

I think that I might be wildly over-powering the armature winding. I did some measurements and got 12.6 volts, at 5.12 amps. The most effective, and adjustable method would probably be to use the Arduino to PWM the power via a transistor.

Before I go there, I'm gonna try the easy way. Math says that to cut the volts in half, I'd need a 1.171875 ohm resistor at 30.72 watts. Note, I probably fudged that a bit, please have mercy. That's not standard issue for sure, but I live in farm country. I can grab the resistor commonly used to adapt old 6 volt ignition systems to 12 volts. It won't be perfect, but should let me know if I'm chasing the right problem.

Tomorrow then.

 

[ODaily]

Messing with the armature current has gotten me thinking. I see people who talk about dropping the current at high RPM to increase the RPM. Apparently with a corresponding drop in low RPM torque.

This seems counter-intuitive to me, but unless people are faking videos, then I've seen it work.

Questions this has given me:

Is the increase in RPM only valid for a free-running no load application?

Is this a case of just finding a "correct" current that will provide both torque and RPM? Or, will these two goals require conflicting current levels?

If this does require adjustment to get the low end torque and the high end RPM, then what is it a function of? Do I need to reduce armature current according to RPM, motor load (measured by amps drawn?), or some other factor?

Would this be a linear relationship, or graph as some kind of curve?

Good questions, I think. But not for tonight. First I gotta get it running well and consistent, then I can start throwing variations at it.

 

[ODaily]

Just kept messing with it trying to figure it out. What I noticed was this pattern.



1. Turn Controller on. (May or may not calibrate range, makes no difference.)

2. LED's are off (supposed to be when at neutral throttle)

3. Can give light throttle in forward direction, motor rotates. Red LED on (normal).

4. If I exceed approx 1/3 throttle, motor stops and LED remains lit.

5. Returning to neutral causes LED to go out, and seems to reset the problem.

6. Repeating a 2nd time, same results, except now I can't reset it, and LED stays on permanently.

Turn it off / back on. Then everything above repeats. Running it in reverse causes red LED, but nothing else. Manual says that solid red LED with no rotation means bad signal.

So.... I went out and got a servo tester. Basically a box with a knob that's used to signal servos to move / and or signal ESC's to run.

Motor will now reliably run in the forward direction. Not much speed. I can also get the green "full throttle" LED now.

I got it to run in reverse twice. Not sure how. Noticeably reduced speed and power. I can get the red LED in reverse, even when not running, and I never get the full throttle green LED in reverse.


I have the programming card, and everything is set per the instructions with forward / reverse rotation. I'm thinking it may just be a flaky ESC. Gonna keep playing with it. Maybe inspiration will strike.

I should also probably mention that I put an automotive points style distributor resistor on the armature electro-mag. They had two, a 2.0 ohm, and a 2.8 ohm. I got the 2.0 ohm one, and measured somthing around 7ish volts at 3something amps. I don't think this helps the running at all, but there's no need to suck down 5 amps right now either.

 

[ODaily]

I almost grabbed an alternator like that for this project. Looks like the innards are much more accessible. Nice.

I've messed with it some more, but not much improvement. It's torquey enough that it ALMOST flips the unsecured Altermotor over on the bench top when accelerated quickly from a stop. Rpm's seemed low. Wanted a way to check RPM and found a strobe app for my phone. With a white mark on the shaft, it appears steady at approximately 15.8 hertz. Google says that's just a touch under 1000 rpm, which is about what it looks like.

So it looks like I got more torque than I expected but much less RPM. Not that I had much idea what to expect. I had wanted to mount and test it, even at a crawl, on the 4wheeler, but I don't think there's enough power to bother yet.

Looks like it's time to move to the next plan. While plenty of RC controller projects are out there, the more successful looking go-cart projects seem to be using an E-bike controller. I need to procure one. A buddy has a junked e-bike that I may try to rob, or I'll have to go E-bay and wait for it to ship from China.

That brings up the question "which one?". Looks like the standard voltages are 24/36/48 with most controllers supporting multiple. Most wattages are 250 watt. I do see some up in the 800 watt range. Probably what I'll try to get. Also got to read the fine print for those that will work without the Hall sensor. It's alot to dig through, especially as specs on e-bay tend to be "flexible".

I did find this interesting write up on the controllers if anyone is interested in the internals. http://www.etotheipiplusone.net/?p=1688

 

[ODaily]

How much does your 4 wheeler weigh when loaded? How fast do you want to travel, and how quickly or slowly do you want to accelerate up to speed?
What size battery pack will you use? From that you can calculate your power requirement.

I'm actually working in reverse of that. Seems silly, I know. Unless you think of this as a motor project only. I just need a testbed and while I could have used a waterpump, big fan, friction brake, etc..., the 4wheeler seems like a fun idea.

Good find on the sensorless controller article, interesting to read.

Thanks, man. Don't know if you noticed but there's several articles there. In the next one he revamps if from 350 watts to 1800 watts. Good read.

Were you testing a 30amp esc on it? Because 30amps won't push that alternator very hard, especially only at 12v power supplies.

120 amp. It's a HobbyKing WP-SC8

It's interesting stuff..must tell the wife I need to get one lol.

Good luck. Might wanna butter her up real good first!

 

[ODaily]

Finally got a controller actually ordered. 36 volt 800 watt ebike controller from china / ebay. Looks like one seller is moving most of them, so that's the one I went with. Shipping is high, but ironically the cost plus shipping is almost exactly what I paid for the RC controller.

Now I just gotta wait for it to arrive, from Guangzhou, China.

 

[ODaily]

You plan on upgrading the controller with higher voltage caps and higher power fets?

Not right away, but the plan has always been to head for the release of the "magic smoke" in pursuit of "Just how much can I get out of this?".

 

[ODaily]

I HATE waiting for international shipping, but the controller finally showed up today.

So, of course, I took it apart first thing.

General impressions,

63 volt caps. There's some headroom there.

12 FETs total. YPI8323C, Google had nuthin on em.

Dual Shunt wires. Interesting

Lots and lots of wires. Some will be pointless. A haircut is in order.

Flat copper bar / strip has been used to reinforce the traces across the FETs similar to how I see modders upgrading them with copper wire. BUT, the traces are not reinforced from the power wires to the FETs. Kinda defeats the purpose, I think. Fixable though.

More spare space than I expected. Going to shove an extra cap in there.

Wires are thin and meager. Anything that carries a load will need replaced / upgraded.


In other news; I really gotta get on the stick now and source some more batteries. Gonna stay lead / acid for awhile longer, and don't want to buy new.