Setup Drawing

[RogerAS]

Hey Everybody,

Dinges, I don't have a digital camera, (I know and I do web design) so I thought I night try this:

If this is too small I can send larger via private email.

Anyway, this is how I have things right now. Many small parts are not shown, like fuses on batteries, volt meters in house, AC distribution box etc.

So here ya go.

RogerAS

[dinges]

Hi Roger,

Looks very nice, more or less the standard setup, I think. Bit of solar, bit of wind, gas generator backup.

Was primarily interested in a picture of your gas genny, but there's no hurry. The design for mine is more or less beginning to evolve in my mind. I notice you've used a 5" pulley on the engine and a 2.5" one on the alternator. Rheostat control is a bit crude in my opinion, but works fine, and a great solution for the electronically-challenged people

You seem to have a limited RE system (still, better than mine, I have none...), so either you have very limited power-hunger or are you still partially on the grid?

Maybe with a bit more wind (say, 1kW) you could really reduce the need for the Honda genny as a backup.

Peter.

[RogerAS]

Hey,

No, we are absolutely 100% off the grid. Yes, we are very starved for power at times, like now when it's so feakin' hot. Our long term goal is to have 1.5Kw of solar and about the same for wind. I'd also like about 2 times the battery bank, and maybe a real inverter, or 2. If that was in place I don't think the engine would be an issue at all.

I've tried the homepower mark8 design a couple times and I must be doing something way wrong 'cause it keeps blowing the power transistor after a few hours. I can't figure it out, so I stay with what works.

I've often wondered if I couldn't just use the rheostat and a 555 to time pulses to a big mosFet. I never have tried though. Maybe this winter.

I'll work on gettin' a scan of a real photo (or photos) as soon as I can.

See ya,

RogerAS

[hiker]

your going to kill that honda-lug a engine down and they wear out quick..

try at least a one to one with that 5 hp--i would  go with a slightly bigger on the alt and smaller on the engine..and make sure its turning in the right direction or you will burn out the brushes for the field coil...built a number of car alt setups over the years--got the idea from a 1960"s popular mechanics.................good luck.

[commanda]

I've tried the homepower mark8 design a couple times and I must be doing something way wrong 'cause it keeps blowing the power transistor after a few hours.

Sounds like it's overheating. Lick your finger and touch the case of the transistor. If it sizzles it's too hot.

I'd also be fitting a diode in reverse parallel with the transistor (Kathode to emitter) to catch the back emf from the field when the transistor switches off.

Amanda

[dinges]

Overheating could definitely be an issue.

Have you tried to use a different transistor, if only from a different brand? I understand that lately a lot of bogus parts show up, made in the People's Republic of China. Like 2N3055s that are way below specs and won't survive long. Maybe the same happens with MJE2955s. Perhaps try a completely different supplier from your usual one.

That back-EMF diode is a good idea, will immediately add it to the schematic.

[RogerAS]

Hiker,

Look real close at the image agian. I have a rheostat between the supply side of the positive going to the field and the field connection on the alternator. This engine has several hundred hours on it and it's still real strong. I can apply full load to the rotor, or just a hint. Yep, I know about the fan setup and all that. My sig file (notice change from bunny ref) tells all. This ain't my first rodeo! I can switch the field off and shut down the engine from in the house. Nice when it's nasty out. Too bad it doesn't have electric start.

I used to have a Kabota EL300 one lung that was never right after a rebuild by a non Kabota mechanic. I miss it, but wish I had put all that money into PV's and good magnets/wire. When that engine was good I could start, control throttle, engage charge and shut down from the house. Now I have to go out and get everything going manually.

Try the factory pulley and a rheostat on the alternator. It works. Of course you need a non regulated type alternator.

BTW, I can push 80 amps out of the Honda/Alt if I want to, but don't!

Others,

Could I replace the output transistor with a big FET and eliminate the little 2N2222A altogether? I've already tried a big diode, and that didn't help. Could be bad parts in bad results out! I have been know to use old salvage parts.:-)

RogerAS

[ghurd]

"a diode in reverse parallel with the transistor"

What happens to the back EMF after that, if the switch is opened? (instead of Q2)

Should the switch also have a reverse diode?

(Just seems like a lot for the caps and LEDs to take)

G-

[commanda]

Good point. Diode should go to other side of switch.

Amanda

[Flux]

In the Homepower circuit D2 is the freewheel diode and should catch any inductive energy from the field.

I can't see that adding more diodes anywhere else is going to help.

The transistor will need to be well heat sinked, it is partly switched but with the high base drive resistor (100 ohms) and the 0.1uF capacitor slowing the base drive it will be little better than linear as far as the transistor is concerned.

The 0.1 uf capacitor is almost certainly to stop rf oscillation and it may not be effective with some transistors, sustained rf oscillation kills things quicker than anything else. The position of the capacitor may be critical on layout.

An RC snubber of .22uf and 10 ohms from Q2 collector to ground may do the job better but without a scope the only indication of oscillation will be heating of the transistor.

An older hometaxial transistor will be much tougher than the present epitaxials and as Peter said, some transistors are of doubtful quality to put it mildly.

It should be an easy life in that circuit if everything is right, but without a scope it is difficult to find things that wreck the job.

Flux

[ghurd]

Can someone tell me why D2 needs the specified heat sink?  The from RF?

I didn't expect this to cycle fast or sharp enough to get D2 hot, partly because of Q2 being sort of linear.

G-

[Flux]

I can't see that D2 needs any heat sink. Not sure what the suggested D2 is but any 5A diode should do.

The intended switching frequency is not the issue, that won't cause any problems with anything.

Parasitic oscillation at frequencies above 10kHz is what cooks transistors when they are near their cut off frequency. This will have no effect on D2.

Flux

[americanreman]

1:1 is the way to go, we had this discussion in another thread, you can still run the motor up to 3600 rpm if you want, but 1200 is about right for a nice charge 1:1.

Saves fuel and wear and tear on both the motor and the batteries.

With engines you have no load speed and loaded speed, people just don't seem to understand, I hear about stalling and trouble getting the rpms right, in short, go to 1500 rpm with the throttle at no load and adjust the field until you drop to about 1300 rpm and you have a good running machine. Don't lug it too much.

I built another one today and experimenting with some led voltmeter schematics, instead of turning a led on at a set voltage, I am trying to send power through a different rheostat connected in place of the led, figue 3 or 4 rheostats that are used at different battery voltages and can be adjusted and fine tuned with the final circuit shutting down the engine when a set voltage is reached. Obviously the point is to control amps in and not boil the batteries.

[dinges]

To be honest, there's a few things I don't understand of this circuit.

First of all, like youse said, D2. Needs to be heatsinked? Why, it's only there to suppress voltage spikes, I'd think.

Secondly, that 100ohm/10W resistor. Why 10W? Seems like a lot of overkill for something that sees little more than the base drive for Q2.

Thirdly. Do you guys still agree on the necessity of an extra diode over Q2? Seems a bit redundant, because of D2 (in a previous response I didn't realize that D2 was already present). I don't understand Ghurd's & Amanda's discussion, of it having to go over the switch as well.

Have thought about using a MOSFET as the switcher for the rotor field. However, the usual N-channel (I think) don't seem to work, it have to be P-channels. And then there's that depletion vs. enrichment thing. In other words, me & designing with MOSFETs is not a good combination. Will stick with the original circuit, unless anyone has a good suggestion on how & what MOSFET to use.

Personally I'd add a few 10n caps here and there (over B-E and C-E of Q2, for example) just to be on the safe side. With 100n & 100R at the base of Q2, switching will be slow anyway, as Flux correctly stated. Maybe a bit of ferrite here and there too, shouldn't hurt.

And the obligatory fuse, of course.

[Flux]

D2 will be maintaining the field current when the transistor is off. At best it will be carrying an average current of 2A. No normal diode will need heatsinking, I have no idea what the suggested diode is.

There is no point in extra diodes anywhere else, D2 maintains continuous current and any transient spikes during commutation will be too fast and low energy to bother the transistor.

If you change the circuit to use a mosfet you will need a high side driver and you will have to switch it hard and fast, mosfets don't like the semi- linear mode.

If you do that you will need a fast diode for D2 and likely a snubber across the mosfet.

I have done similar things many times for exciters with AVR's and at this low power it is just as easy to do it linear and not have any risk of switching spikes. Efficiency is no issue, you are just automating the manual field resistor.

[JW]

"D2 will be maintaining the field current when the transistor is off. At best it will be carrying an average current of 2A. No normal diode will need heatsinking, I have no idea what the suggested diode is."

 No wonder you say this Flux.

 The field winding can't be that high in impeadence. Hence closing cycle, no matter how sharp, can only generate so much back emf. Obviously less than 2 ampere's, unless its fairly large(the reluctor).

 Once, when I was working with an Infinity (Nissan/Japan) Mastertech- on a 1988 ford P/U truck( in 1992-3). The A/C(air-conditioning) compressor clutch would cause the 10amp fuse to burn-out that protected/power'e'rd the electro-magnetic circuit, everytime it was turned off. I just was visiting over the summer(summer job) from University when I was about 19. Anyways the tech came up to me, and said, how about using some of that Detroit education you got last year. I replied what do you got? I have a A/C comperessor fuse that keeps blowing in this truck. I have checked the coil impeadence the way the manual says to check it(observing polarity) and nothing should be wrong, but I can put fuses in it all day long, and every time I shut it off, the a 10 amp fuse will burn up. But if I put in a 30 fuse its ok. I then asked him then did you check the in4001 diode? He said what diode. I then told him to go to radio shack and get one and replace the one thats there. Later he came back and said I found the diode in the clutch coil.

 I wonder about the size of the reluctor in this situation(in this posting the one that the D2 works with), but even more interestingly how a properly sized Zeiner diode may or may not change the circuit.

JW