PWMing a solenoid for RPM control

[madlabs]

Hi All,

I got my new alternator/gas engine going, it's running great. Pulled 165 amps (@12V) out of it briefley and ran it for 1/2 an hour at 120 amps. Gonna be good to have this beast handy as the solar power heads south for the winter. One odd thing, I found I had to alter the PWM frequency of the field coil to get smooth output. I've used 10kHz for every other alt, but this one likes 7kHz. Whatever.

Which brings us to the questions. The generator I modified to accept the alt has idle at no load feature. The engine idles along until there is a load and then a solenoid is released and the engine kicks up to full RPM. I'd like to take advantage of this feature but improve it. I'd like to be able to at least somewhat match the RPM to the load. My thought is to try PWMing the solenoid so that instead of just being either idle or full throttle it can range inbetween. My first concern is that might burn up the solenoid, although I can't really see why. If that happens though, I can switch to a servo.The second issue I see is that matching the RPM to *exactly* what needed will make for a lot of hunting or a fancier program then I want to or am able to wriite. So this is the scheme I am thinking of, and I was hoping to see if y'all think it will work.

Fire up the gen, run full throttle until I drop below say 100 amps output. Lower the throttle and readjust the field coil PWM duty. Run like that until we get below 80 amps, kick the throttle down, readjust the field coil PWM. Keep going until we get below 1.5 amps a batt and shutdown. You get the idea, lower the RPM in steps and not try to vary it dynamically.
Whaddya think?

[madlabs]

Ouch. Just after I wrote the above, I may have thought of a deal breaker. The engine has a mechanical govenor and that would likely need to be adjusted as well. That sounds like more trouble than it's worth.

Tell me I'm wrong, please. I was having such fun thinking of the fuel savings...

Jonathan

[Flux]

I don't think you can do what you want with a solenoid. Linear solenoids are rare, the one you have won't work that way. It will either be pulled or released and I doubt that pwm will keep it in any other position any more than you could do it with a variable voltage supply.

To do what you want I suggest you use a geared motor ( car wiper motor would probably do fine) and use the geared output to control the tension on the engine governor speeder spring. Each position of the motor will result in a different engine speed. You could probably use a current signal from the alternator charging current ( or possibly alternator excitation volts) to set the best engine speed for easy running at the chosen current with good fuel economy. You will probably need a servo pot connected to the motor output shaft to reference against your control signal.

Should work , I have done similar things for other reasons including removing the speed droop of the engine governor.

Flux

[madlabs]

Flux,

Thanks for the advice. IIRC I have a motor from a rotisserie that ought the be strong enough, or I can certainly round up a windshield wiper motor. Being a redneck I have access to lots of dead cars. 

So in your scheme, I would have a servo control the throttle and the wiper motor adjusting the governor?

Thanks!

Jonathan

[rossw]

I don't think you can do what you want with a solenoid. Linear solenoids are rare, the one you have won't work that way. It will either be pulled or released and I doubt that pwm will keep it in any other position any more than you could do it with a variable voltage supply.

I know you said "rare" not "impossible"... my generator setup uses exactly this setup.

Feedback via magnetic pickup, counts the 112 teeth on the ring-gear as it goes around, through controller, which uses PWM output to a "linear solenoid" that is coupled to the throttle.


The solenoid spring return (out) and current to actuate (pull in).

The mechanical clunkyness with the bits of metal on the throttle arm was during early times, I was terrified things would go beserk and run the engine at full throttle - this was just a mechanical stop to prevent more than a "comfortable" throttle setting while I "got confidence" with it all.



The speed sensor. Since the alternator is belt-coupled to the engine, this gives me engine RPM and via small calculation, output Hz


Showing engine RPM fluctuation over time. The spike in the middle is the application of 2.5kW additional load.

[fabricator]

Well what about pics of the alternator mounted to it? That's the money shot

[rossw]

Well what about pics of the alternator mounted to it? That's the money shot

The discussion was only about the control of the engine.... but since you ask (remember, these are during construction so things are still a mess)



Alternator in the frame, ready to lift the engine in place. It's a 15KVA, 3phase, 15000 RPM machine ("Syncro", made in Italy). It has electronic AVR.
I rewired it for double-delta to give me a decent single-phase output.




Drive is a decent toothed belt. The tensioners needed work once it was all done. The belt being so long tends to slap a fair bit without them.

[fabricator]

That's a really nice looking setup, damn good idea to mount the head under the engine to keep the foot print small, what happened to the Leece Nevill idea?

[rossw]

what happened to the Leece Nevill idea?

Huh?

[fabricator]

Maybe I got threads confused, I thought originally this was the engine that was gonna get the Leece Nevill.

[JW]

Fab;

I have nothing to loose(AS A FORMER ADMIN) by telling you "admin" allows you to create more problems than you solve with your contributions...

YOU SUCK....

jw

[rossw]

Maybe I got threads confused, I thought originally this was the engine that was gonna get the Leece Nevill.

Nup, this one was always going to be the Sincro HB4.
(Deleted link that doesn't work)

Not shown in those photos (but the corner rests are there) - is a large drip-tray that goes under the engine, over the alternator for those inevitable drips, spills etc - when changing oil or filters etc.

I also put an extra oil reservoir/cooler there - holds an extra 5 litres of oil. It only needs an oil change about every 500 hours, although I have pushed it to 1200 hours once - oil test shows there's little difference between 400 hrs and 1200 hrs run (it runs on pure propane). I just like to change it a little more often so I can keep an eye on it.

[fabricator]

Nope I can't access that site.

[fabricator]

Fab;

I have nothing to loose(AS A FORMER ADMIN) by telling you "admin" allows you to create more problems than you solve with your contributions...

YOU SUCK....

jw

OK.

[kurt]

jw mind your business any more of your useful thoughts and i will put you in the read only club for a month or two. fabracator it is good to see you and rossw having a reasonable conversation finaly. i hope it will continue.

by the way rossw i could not get to that pdf either without my user/pass to your home website. 

[birdhouse]

ross-
that system is pretty bad!  i remember when you first posted about it.  by chance do you have any KWH/gallon of LPG figures?  just curious more than anything.

and yes, i'm glad you and fab can have cordial conversations!

adam

[rossw]

that system is pretty bad!  i remember when you first posted about it.  by chance do you have any KWH/gallon of LPG figures?  just curious more than anything.

I have done some calculations, and enough to know that if I was doing it again, I'd do it differently.

Although my reasons for doing it the way I did were (and remain) sound, the reality is that the engine is just way oversize.

Fortunately, because I use almost all the heat from it - salvage heat from exhaust, and from the block, it all goes into hot water for either domestic hot water, or heating the house, I don't consider the "waste" to be really "waste" at all, because it's all used. Probably more efficient than most gas space heaters or even gas water heaters.

As to consumption, I just got a load of gas delivered a while ago. It was 6 months give or take a couple of days. 1256 litres. Given that represents *ALL* our external energy, and it's been over a cold and wet winter, thats under 7 litres/day. (1.5 gallons)

With winter here, we've had a lot more (office) lighting than usual, and several more computers than usual, and the wife has been running the incubator all winter for a number of eggs and abandoned chicks, we've been averaging close to 22 kWh demand per day, and getting barely 3kWh from sun some days.

If we said the average was 15 kWh/day, and 1.5g/day, thats 10 kWh/gallon. I think thats an overstatement, in reality its probably closer to half that. (The same gas tank also does all our cooking and runs the clothes dryer) Does 5 kWh/gallon sound in the right order of magnitude? (given most of the energy is actually given off as heat to help keep us warm, thats probably not so bad)

[kevbo]

Model airplane servos are really really cheap.  They need a 5V power supply and a .5 to 1.5 msec ttl pulse to set the position (usually +/- 45 degree travel with 1ms is centered). Repetition rate of the pulses is not critical, 10-100 Hz works fine with most of them.  Not too bad to generate this with analog multivibrators, or pwm circuits, and trivial with any sort of microcontroller.

You can either govern electronically by connecting the servo "hard"  to the throttle blade, or you leave the mechanical governor, and link the servo via spring.

[madlabs]

Haha, away at work for a couple days and the the thread wandered. This project is using a Leece-Neville and is running well. Kevbo, I have lots of servos and have used them with a uC many times. My controller is a DIY job and was running my last lawnmower/alt project, so all I had to do was hook it up to my new charger.

I'm gonna get around to playing with the project today, thanks for the help!

Jonathan