Micro CoGen.

OK, so, my guess in the last post was actually right. This alternator is designed to produce around 12-15V, and the electronics package contains a Boost Converter which pushes which up to 50+ to charge the battery.

I was confused by the lack of large inductances necessary for this type of converter, but someone on another forum pointed out that it's using the motor coils (stator) themselves as the inductance. I didn't know you could do this.

Once you get this, everything makes sense. The five phases come into five daughterboards which each contain a FET bridge (two large FETs bonded directly to the substrate) with one of the FETs connected directly to Ground via a precision shunt / current measuring component. The idea is to short the winding directly to ground initially in the AC cycle, let the current start to build up in the inductor/coil, and then turn the FET off; the resulting flyback voltage provides the higher output. This is being done under control of the central processing module (see the photo above). And this in turn responds to CanBus commands on the four-pin control interface.

Honestly hacking CanBus commands is beyond me and I'm after a simpler solution in any event, although this is a pretty impressive and clever machine. So I'm going to try rewiring the stator coils arrangement and seeing if I can get more voltage out of it that way, and just leave the electronics out of it. Conceivably, I could build my own boost converter - they can be extremely efficient - but that's Plan-B for now.

So things have moved on a bit. Some good things, some bad things, kind of at a pause right now.

You cannot view this attachment.

Things got off to a bad start when I decided I needed to remove the pulley from the alternator to fit a different size one. Not having a large-enough impact wrench, I gave it to a guy at the local garage/auto workshop, forgetting to tell him the main nut was a left-had thread. I heard a loud snap and a sheepish engineer appeared, holding the alternator in two pieces. No repairing that, so I had to buy another alternator (it turns out Suzuki use exactly the same device, and it's cheaper because there's less demand than for Ford).

The alternators are nicely made - SKF bearing at the rear and the main one says "made in germany" so is likely to be good quality.

I have set up a test rig, driving the alternator using an old washing-machine motor with a speed controller so I can vary the rpm.  I'm using a bench power supply to feed the field coils/rotor, so I can adjust the current, and running the output through a five-phase bridge rectifier (just two three-phase units in parallel) and watching the AC output with an oscilloscope and the DC output with a meter. It's open-circuit output - no load - but I reckoned it would be representative.

I do indeed get very clean DC out of it. However, I'm finding I need to get the RPM up *really* high. Here's a graph at 3600 rpm, showing how output voltage varies with field current :

You cannot view this attachment.

Current in Amps is along the bottom (x-axis) and you can see we've pretty much reached saturation by 10 amps - no useful increase beyond that.

So let's fix the current at 10A and then wind up the speed :

You cannot view this attachment.

Nice linear relationship, BUT we're at 5,000 rpm before hitting 24 volts, which means to get to 48 volts (or a bit beyond it, if we want to charge batteries) we would need to be at around 11,000 rpm. Hmm.

Looking at the car this unit comes from, the crank pulley looks roughly 2 x the size of the alternator pulley, so in order to charge the batteries, the engine would need to be doing over 5,000 rpm. That doesn't seem right ! How often will it be doing that ? The idea is it does this under braking, capturing waste energy, and you're not going to braking at 5,000 rpm unless you're on a race circuit.

So, where are we ? I could probably get this useful by running it really, really fast, but I don't want to do that as 11K rpm sounds excessive even if the device is rated for that. And I don't see how in it's intended application (the car) it can do anything useful.

Is it possible the big chunk of management electronics I took off this is doing something very clever like boosting lower voltages up to enough to charge a 48v battery ? Some sort of boost converter? I am next going to try wiring it back into the circuit and trying things, although I have a feeling it may need some sort of control input to tell it to go into charge mode.

Any thoughts appreciated, as usual, and I will let you know what transpires...
Richard

Reassembled without the Management unit, and with the backplate brought out so each phase can be led out, and a connection into the sliprings.

You cannot view this attachment.

Now awaiting the arrival of some large bridge rectifiers. I had planned to use an old washing machine motor plus speed controller to build a test rig that would enable me to spin it up at different speeds and put a little electrical load on it, but the pulleys have different groove sizes :-( Rethink required.

I paid about USD120 for it. It's about five years old so has some rust etc, but that would clean up and it looks good. If I was going to use it long-term I would probably replace the bearings with some good ones, and get some spare brushes.

Keith, that's an interesting pulley ! Must put a lot of load on the front bearing. But I think that's the 12V alternator from the same car - it has just one large terminal so must use frame as ground, whereas I'm getting the feeling the 48V subsystem is run completely isolated from the main 12V circuit, so my 48V alternator has two large terminals, like this :
You cannot view this attachment.

If I put the part number off your listing into a search I get 14V units.

I've done some more checking and actually now think this is a delta (or pentagon!)-wound unit - I can find no star-point. The coils for each phase are wound with a group of four wires, and it's the start and end of that skein (so 4+4=8 wire) that I can see bring brought out to the terminals. With four wires in the skein for any given winding, the resistance is so low that my simple multimeter can't resolve the difference in resistance between different points on the outputs.

I am going to try running it up and seeing what happens. I should be able to use a couple of three-phase bridge-rectifier units, right ? Just only part-using the second one. All you're doing as you add phases is to add more diode-pairs, so I can link the outputs of the two units in parallel, and attach each phase to one of the inputs, leaving one unused. I think that should work OK.

Thanks for the feedback !

So I've done some more digging and maybe answered some of my own questions.

There are 80 slots, and 16 poles on the rotor, which gives this alternator a pole pitch of 5. Since there are 5 connections coming out, this seems to be a 5-phase alternator (!)

Each phase would then have eight coils (80 divided by five (phases) and then by two (sides to a coil)), and since each connection crimp has eight wires into it, it seems those eight coils for a phase are connected in Parallel, not in Series as seems more common. I've no idea why you would do this but it must put the emphasis on current not voltage ? And the more phases you have, the smoother the operation in motor mode, and possibly the more torque on startup ?

The fact that I find similar resistance between all pairs of connectors then suggests that this is a star-wound alternator and there is the mother of all star points (with 5*8 or 40 wires joining up) somewhere. Given that number, the star point could be more a star "ring" running around the other end of the stator. I will take the back (slip-ring end) of the casing off and look.

Still very interested in peoples' thoughts on the suitability of this for an inverter/generator setup. It looks well made and solid.

As part of building an inverter-generator setup I am looking into 48V alternators.

I realised that cars increasingly have such things built in to provide a "mini-eco" function. They effectively have an entirely separate 48V electrical system comprising a 48V starter/alternator, and a 48V battery. The alternator mode puts energy into the battery during braking, and the battery can use the "motor" mode of the device to both provide the stop/start function and also provide a bit of extra power when accelerating.

So, these are pretty interesting devices and available relatively cheaply on the 'used' market, and I figured I would see if I could use one as the basis of my generator. I bought a "salvaged but servicable" one from a Ford Focus (a 1-litre petrol car) which is a 48V 150A unit - it was cheap enough I reckoned if I broke it while dismantling, I would at least have learned something. I found a slightly strange setup and I'm wondering if people here can shed some light on it. Obviously things will be complicated by the Starter functionality, and it might even be possible to use that, but let's see.

First weird thing is the stator has 5, (yes five) connections. It's densely packed, 80 slots :

You cannot view this attachment.

The five leads come out onto the back and are labelled U, V, W, X, Y. There is no Z like I was expecting (well actually I was hoping for just 3, but you can't have everything). They do not seem to be isolated pairs, and there is about half an ohm between every combination / pair of terminals. Here's the backplate :

You cannot view this attachment.

The wires are led through fat copper tracks to spade-type terminals which are spot-welded onto tags from the electronics package, in the visible groups of 2, 2, 1. This is turns out is so they can be led into the regulator/manager in the right places. The regulator package is large! and heavily potted/sealed so was a devil to get apart. It looks like this :

You cannot view this attachment.

Forgive the fragments of plastic from my ham-fisted and exasperated attempts to open things up. There was a LOT of potting, spot-welding, and sharp-edged connectors and my pain levels got to a point where the red mist descended and the hammer and screwdriver came out. I was reflecting that I didn't really want the control package, just the main body of the device, so I wasn't concerned about re-using it.

You can see five small boards with golden components on them - this is where the five stator leads come in. The prominent gold components are precision resistors, so something is monitoring current very precisely. I can't see the rectifiers but I think they're potted-in up the top of the picture.

The rotor has a small magnet attached to the end, which I suspect is driving a hall-effect sensor in this package so the system knows where the rotor is, rotationally - it needs this for the Motor mode, I think ? Because it has to generate / synthesize a waveform to move the rotor around based on where it is relative to the stator coils. Some controllers do this by sensing the change in inductance as the rotor moves past stator coils, but this appears to use a sensor. I guess that gives it more control.

The control harness comes in at the top-right, it's a six-way connector but only four of them are used, you can just see four little loops of wire coming onto the central processor board.

So it is WAY more complex than I need, but if I could figure out the stator windings I could perhaps use it. The sliprings/rotor connections are very accessible. So I'm posting this (a) because people might like to see inside these things and (b) to ask if anyone knows what the five stator tags signify and whether it's possible to three-phase rectify this.

I might try tearing-down the manager a bit more to try and find the heavyweight rectifiers, and there must be some chunky switches in there (FETs or a relay maybe?) to switch between Motor and Alternator modes.

Thoughts appreciated !
Richard

I am kicking off a project to build a generator. It will be a smallish Kubota diesel driving a chunky 48V alternator, which can put energy into my 48V batteries which in turn feed an inverter. I'm going for this approach because the inverter will be my main power source - good output, clean sine wave, versatile - and most of my power needs are intermittent, so having a genny run all the time is not a good fit. Plus, this way is a lot less fussy about alternator speeds.

I think I'm looking at a small Kubota engine like the supermini range, Z482 etc, but Kubota seem to advertise some of these - and their larger 3-cylinder brothers - as "1800 rpm" versions like this :
https://global.engine.kubota.co.jp/en/products/detail/126/
They have derated it in the spec to 5kW / 4.5HP, but is this engine "special" in any other way ? What have they done to it, I wonder, to make it a "1800 rpm" version. Am I going to get different results getting a "normal" 482 and just running it slower ?

I want to end up with a slow-running engine for noise and reliability, so 1800 is kind of a goal for me, and 4kW would be a good-size output... just trying to understand the range here.

Anyone have any thoughts ?
Thanks !
Richard