Just to give you a background, I was once employed in the automotive electronics industry but have now semi-retired and occupy my time working on the electrical systems of narrow boats on the UK canal network. Not much money in it but it is a lovely way of life.

I often get involved in wiring up boats with smallish vintage diesel engines that are preferred in historic or replica historic boats. A lot of these engines were produced without electrical starting systems and it can be difficult to retro fit them because there is no ring gear on the flywheel or mounting for the starter.

One thing that is often installed along with the engine is a large alternator. This is necessary to supply the power to recharge the battery banks that then feed an inverter for mains power. An alternator that is often used is capable of producing 160A DC and is a three-phase stator exited by a 16-pole rotor. The rotor is magnetised by a single DC bobbin coil and is brushless. So that the alternator can be driven to maximum it's speed of 8000rpm by an engine that will only do 1200rpm, it is belt driven with a 1:6.7 pulley ratio.

It occurred to me that it ought to be possible to drive the alternator as a motor by injecting a three phase signal into the rotor windings whilst supplying the rotor directly and thus turn the engine sufficiently fast to start it. At a guess the engine would need to be turned at about 200rpm to start it. This would require the alternator to be driven at 1334rpm and form the phase/pole ratio; the frequency of the signal would have to be 178Hz. I am quite prepared to be corrected on these calculations though! As a guesstimate, I if the output of the alternator is 160A then the locked rotor current would probably be about 200A.

I am guessing again that what I need in the way of control is a switch which when made starts the process at a low frequency and slowly ramps it up to full speed over a couple of seconds and then maintains it at that speed. When the switch was released the process would be stopped. There are a couple of other technical issues such as how to integrate this with the alternators original purpose and some sort of interlock to prevent the starting system being engaged when the engine was already running but these can be sorted out later if I can get the thing motoring.

I have been looking for a BLDC controller/driver IC that has analogue or simple switch inputs, as narrow boats are best kept simple and their owners abhor too much technology but I have not had much success.

I would be grateful for any pointers that you may have...

Regards
Arnot

 

Yes a rotating field on the stator from 3-phase AC with DC on the rotor should produce a motor. I'm not at all sure it would have the torque to turn over a cold engine, and the belt would be an issue, but let’s suppose. You wouldn't want to do it very long because the fan isn't very effective at low rpm. And I would expect that you would need a higher voltage probably 30 or 40 V RMS at 200 A. An inverter to do this would be a pretty special item, but feasible. You would really have to want to do this because; it'd probably be much more expensive in both cost and weight than the small PM starters available today. The car companies are already doing something like this where the starting and alternator functions are integrated with the flywheel. Because of the diameter advantage and OTS nature of their system, you might consider adapting that, which would be even more compact, if you run a flywheel in these applications.

Regards,
cww

 

This sort of thing has been done on an experimental basis in automobiles, but not with standard components. The experiments were for automobiles which needed high output alternators to drive things like active suspension and electronically controlled valves. The combined starter/alternator was much bigger than a standard alternator, but smaller than a standard starter motor. They opened all the electronically controlled valves at once to take the compression off the engine until it was up to starting speed, and then started the engine. This sort of option is unlikely to be available to you, so you will probably need a fairly high starting torque.

I'm not familiar with narrow boat engines, but even if you were able to modify the alternator to act as a motor, would it have enough torque to start the engine? If you used a larger alternator/motor could you deliver enough torque through the belt drive?

A lot of standard BLDC IC designs are intended for 90 volts or higher. If you are operating at 12 volts, then the number of diode drops in the power circuit becomes a lot more important than it does at 90 volts. Also, at lower voltages you are dealing with higher currents.

I can't recommend any ICs for your application, but I hope this has been of some help

 

Thanks for the assurance about motoring an alternator. There is only 12v available and bringing in a higher voltage would probably negate any benefits the system might have.

Again, changing the flywheel is a non starter (pun intended) the whole idea of vintage engines is to maintain originality and I suspect the cost of a bespoke flywheel would be horrendous.

Whilst there are plenty of starters available, fitting a ring gear to an old flywheel is a problem because they are often rusted on. As well as this there often is no obvious place to put a starter motor without moving a lot of other engine bits around, on modern engines, the starter mounting is cast as part of the main cylinder block or bell housing.

Torque may well be a problem but if the alternator is capable of supplying in the region of 2Kw I am hopeful that the motor output should be in the same order and that should be sufficient. Using a polyvee belt and suitable pulleys I think that the belt drive problem can be overcome.

I have discovered a number of BLDC motor control IC's but they all seem to be intended for integration into a larger system with some form of serial control and for my application this is just too involved. It seems that some of the little electric bicycles use something like the functionality I want but I can't seem to find just what chips they use.

Alternatively, Atmel and Microchip both have microcontrollers that can be configured as a controller and then interfaced to an external bridge. This may be the way to go but I had hoped not to have to re-invent this particular wheel.

IR do a variety of suitable FETS which I have sued in the past and with a bit of a parallel array I am sure the RDSon can be got down to a reasonable level.

It was just an idea anyway...

Regards
Arnot

 

You might find more success with using a positively arcane piece of technology from the annals of 'the automotive history' in using a generator instead-of an alternator. Personally, I'm in-favour of using whatever works for the application at-hand, even-if that technology is considered antiquated in this modern age. In-truth, a generator can-be used as a motor, just-as an electric-motor could-be used as a generator. If you do some searching online you can find some decently powerful 12VDC motors, theoretically, you might-have better luck trying-to incorporate a powerful DC motor as a generator and using-it as a generator when it's not firing the engine over.

I've in-fact been planning something similar to what you've been trying, and was searching online for some information referencing the idea of using alternators for a starter, but there's nothing wrong with generators. Heck, you might even-be able to convince your customers it's worth-it because it's an older technology, instead-of some 'modern technological doo-dad'.

 

It's fairly easy to use an alternator as a motor. But you have to run it on three phase AC, bypassing the diodes as they convert AC to DC but won't obviously do the inverse. And you have to have a source of DC for the rotor excitation, bypassing the regulator, which is of no value to the plan. After this it's a pretty standard three phase motor with characteristics similar to a PM rotor. You can vary the torque with the rotor excitation. If you leave the diodes in place, the output terminal can be used for the rotor excitation with suitable means to control the current. But even a 100 amp alternator is less than a 2HP motor and is questionable for starter service unless you overdrive it. This might be possible for intermittent starter service. And actually, there was quite a bit wrong with low voltage, high current generators. That why alternators are universal. And when it's cheap enough, the starter will be AC as well. Brushes
are of the old world.

Regards
cww

 

I have heard some hybrid vehicles are using a combined alternator starter motor system named Belt Alternator Starter or BAS for short, so it may be close to an off the shelf part. Some companies are Valeo SA makes Valeo Starter Alternator Reversible System which generates electrical energy and is capable of functioning as a conventional starter motor and GM's mild hybrid belt-alternator-starter system.

 

RC models use a component called an ESC with DC input, 3-phase AC output and PWM speed control signal to power permanent mag, brushless motors. HobbyKing is selling an ESC they advertise as 190 amps for about $30. In your case you still need a variable 12v dc to power the rotating field and a simple timer (555) circuit to provide the 5v PWM control signal. Search Wiki, Youtube and rcgroups.com for lots of information. When you get something along those lines working share with my email as I don't often check out Yahoo. charlescrail * sbcglobal.net (Los Angeles)

 

I have never tried this but have made alternators into motors by exciting them with 3 phase. getting enough torque will be a problem. is it possible to engage compression release on the engine, get it spinning, then re engage the compression. my experience with diesels is they start quickly if there is enough rpm to produce the required heating of the air on compression. if they are turned too slowly the air cools as it is compressed. holding the exhaust valve open should do it. the three phase drive voltage can be obtained with a stepper motor drive with additional power fet transistors to handle the current.