I had an idea about building a gas powered quad. It would be powered by some of the new small 4-stroke engines like the honda gx-35.
I know what you're thinking... gas engines don't respond fast enough. But, everyone bases this on using the throttle and carb to control the power output. My idea is that in addition to the throttle you could use a microprocessor controlled CDI ignition to control the power.
With a microprocessor ignition you could immediately cut or reduce power output by skipping a spark for super quick power reduction and you could also retard the spark for a smoother, but still quick reduction.
At say 3000 RPM (typical) you could reduce power within 1/3000 of a sec by skipping a spark, or every other spark for a quarter sec or something of that nature. The throttle could also participate, but the spark control would be immediate.
I'm curious how a gas quad like this would perform. A Honda GX-35 puts out about 1.5 hp, and many sources say it easily turns a 20x8 prop at around 3000 RPM.
Would that be too large for practical use in typical applications? Each engine would be about 1.5 hp @ around 8 pounds (very lightly modified). That would be 6 hp @ ~32 pounds. I've also read that when the flywheel/magneto (using CDI) and clutch are stripped they can come in under 6 lbs. That would be ~6 hp @ 24 lbs.
I'd like to know if this is new territory or if someone has already tried this and what their results are. I haven't been able to find anything of the sort, so maybe it's a truly new idea.
@Jake: I just looked at the GX35 specs. In continuous duty, I'd call that a one horsepower motor. But since everyone tends to show peak, that's about 5 lbs/hp.
You'll get no debate from me regarding 4-strokes being preferable to 2-strokes, in all aspects except power-to-weight ratio. Rotax rates their 2-stroke aircraft engines at 300 hours TBO, while the 4-strokes get around 2000 hours.
Remember that Honda has a price point to hit here, especially in light of the wave of really, really cheap powerplants coming out of China.
Let's look at a hobby-class 2-stroke:
It's relatively expensive, but it's also probably due to low volume and high margin. To be fair, let's use its shipping weight at 8.6 pounds and claimed 10 horsepower; that's a 0.86 ratio, or nearly six times better than the Honda in this key metric.
Once it can be shown that there's a market for this, we'll see even more of them.
How NOT to do it :--
-)An engine with alternator and then electric motor because efficiency would be so terrible. The more the devices between the source of energy and the destination of energy, the more the losses because all the devices would all individually take away bits of efficiency leading to much lower use-able energy by the electric motor which doesn't even have a 100% efficiency on it's own.
How to DO it :-
-) 2-stroke (perfect power to weight ratio), with disadvantages too though (vibration, pollution & noise).
-)Direct drive system from the engine to the rotor(reduces loss) with variable pitch.
-)Since the whole system would be much heavier than conventional electric qauds, it will be less sensitive to gust/other sources of instability due to big inertia so the speed of the digital servo will be adequate.
Thank you, Kevin. I don't really relish a debate; my mission is to educate. Yet when I walk through the math and show independent references for my assertions, there is still dissent. Rather than debate, which is counter-productive, I merely lay out my position with basic mathematics and references to other sources when I make claims of technical validity. The chips fall where they may.
Virtually everyone here recognizes that electric motors have come a very LONG WAY in power-to-weight ratio over the last decade, mostly due to MOSFET control advances and better permanent magnet materials. That this forum even exists is testament to that fact. I went so far in another thread as to lay out the basic foundational logic for a motor-generator device as a technically sound solution to the whole energy density dilemma in electric flight. The technical argument doesn't even require the conversion efficiency an exotic wave-disk engine, although the APU in the link you provided sports a rather astounding density of 305 watts/pound (not counting fuel weight). In my holistic solution analysis, I assumed 10% Joule conversion for a garden-variety two-stroke. So, then, perhaps the basic question (or hang up) for some must be...
What exactly IS the difference between a motor and a generator?
It is obvious that a motor converts electrical power to mechanical power, and some of these hobby-class PM poly-phase AC motors (there really is no such thing as a brushless DC motor...thank the MOSFETs for the the DC to AC conversion) indisputably do it at greater than 80% efficiency. Well, the fact is, a generator IS a motor which is simply employed to make the power conversion in the opposite direction. Is a generator less efficient than a motor? Yes, it is. But why? And by how much?
I didn't spend much time searching before I found this academic paper from Belgium precisely focused on collecting efficiency data on using motors as generators. They're discussing AC induction motors, which have their own losses due to creating the magnetic field (stator windings) so "our permanent magnet mileage my vary" (better). Suffices to say, they measured an approximate worst-case 8% difference in efficiency when a motor was used as a generator. The best case difference in their "high efficiency" class was a percentage point or two. I quote from the paper's conclusions page, "...typical characteristics for high efficiency machines or machines with a high power rating, have comparable efficiencies in both generator and motor mode..."
Kevin, I'll go one step further and say that off-the-shelf equipment IS there now, just waiting for someone to do the integration work.
When putting power into an electric motor the MOFSETs switch the power rapidly to different coils to follow the movement of the permanent magnets in the rotor. Hopefully I'm understanding that part correctly at least.
Am I correct in thinking you just need a simple diode rectifier circuit to extract power from the motor? Or do you still need some sort of electronically controlled switching? I haven't seen any of the actual circuits used in regenerative braking, power generation, or the like.
I've been looking around at DC motors to use for a generator in my system. I only really want to generate enough power for the electronics (RX, telemetry, video TX, etc.). I'd like to put out 4W of video power on the ATV band, a 1W telemetry system, and of course running the servos and APM.
An outrunner brushless motor might work well for this requirement. I'm also considering working on a few brushed DC motors I have on hand, but I think that reliability and maintenance might be a factor there. They would be pretty cheap, mostly salvaged in fact. However, with such cheap outrunners out there it might not make much sense.
Any ideas about efficiency, control electronics, etc.? Would a higher kv motor generate more total power or just more voltage for an alternator?
You are correct that a simple diode rectifier is sufficient to extract power from a motor. A 3-phase diode bridge is what you'll need for a brushless DC motor. This will generate a rail whose voltage will be roughly proportional to the rpm (minus 2 diode voltage drops).
You can of course get more fancy and do this with FETs or SCRs, which is how larger converters work, and although more efficient this is an order of magnitude more complex, and if you get the timing wrong - poof.
The higher the Kv of the motor, the lower the voltage, as Kv is rpm per volt. You will need a low Kv motor.
Thanks for the info, and good catch on the Kv.
Unless someone has more ideas I'll put together a simple 3-phase diode bridge and scope the output as a start towards figuring out a filtering and regulating circuit.
Jake, I would take the approach of determining your sustained power requirements and first selecting a suitable gas engine which can produce about 50% more than that. There are far fewer suitable gassers available, while there's a huge variety of electric motors at your disposal.
Aside from the obvious importance of light weight, look at the engine's RPM vs. torque curve. You'll have to play around with this, but perhaps setting an RPM point just shy of maximum torque would work well. Naturally, implementing an off-the-shelf engine in a use-case where it would sustain maximum power for a long duration is just asking for reliability and heat dissipation issues.
Once you have an operating RPM point set for your engine, selecting a KV rating (with the caveat that the resultant voltage will be peak, not RMS) for the electric motor/generator is a cinch, again slightly over-sizing for reliability.
One thing I have learned the hard way from running 15 feet of wire from the batteries to my motors, inductance is NOT your friend. In fact, I propose the adoption of this essential platitude for electric vehicle design: inductance is evil - capacitors are your friend. Do not be afraid of putting lots of big lytics bypassed with MLCCs to stabilize your supply rail. Yes, big capacitors are expensive, which is why almost all ESCs have cheap, small, generally inadequate ones with "make your wires short" diatribes in the owner's manual.
(pardon a little bit of digression)
Case in point: Paul Pound's railing against hobby-class ESCs in his various papers over slew rate limiting code imposing restrictions in response time (the code is there to protect the cheap ESC MOSFETs from the even cheaper capacitors), and his complaints about transient supply rail collapse, are all due to a lack of sufficient capacitance on the supply rail nearest the MOSFET drains. The fact is, virtually all hobby-class ESCs rely heavily on using the LiPoly cells' capacitance for inductance mitigation and shunting of the switching ripple current. They get away with this in the vast majority of cases because most people just plug-and-play. Well, LiPoly cells really aren't made for high ripple current shunting because, while they do have capacitance, they're not capacitors. (rant mode off)
Here's an entertaining little primer on supply rail inductance and MOSFETs:
Since you're presumably going to be powering a load directly from this generator, don't skimp on the output capacitors, as an unstable supply rail is a very bad thing.
You are a font of wisdom Brad, thanks again.
I checked out some 3-phase rectifiers at mouser and they are expensive! Apparently auto alternators use the same type of 3-phase rectifiers. So problem solved! I can pull alternator parts out all day long for free!
The article you linked to led me to think that it wouldn't be that difficult to incorporate a drive circuit, and that almost any drive circuit could be used without too much modification for regen. So I got to thinking that perhaps with the right gearing I could also use the alternator as an electric starter!
I'm looking at using a cheap Chinese clone engine, but the real GX-35 has a compression letoff for easy starting. If I can get something working I'll probably use the real GX-35s down the road. Electric start would be a real nice safety feature.
You don't need specific phase rectifiers, just use 2 standard full wave bridge rectifiers and ignore 1 of the AC side inputs and obviously connect the two dc sides together.
You can get a full wave , high voltage or amperage heatsinked full wave bridge block at Radioshack. Again, each block has 2 AC inputs and then the DC + and -. 3 wires, 4 ac inputs, ignore one of the inputs.
I've always wanted to do this. How i imagined getting around the throttle lag was to have a central engine running 4 of these: http://www.hobbyking.com/hobbyking/store/catalog/14923.jpg. Except instead of 3 servos running the rotor independently, the 3 servos would only move the rotor up and down adding or removing trust, for each of the 4 props/rotors.
Just a thought.
Robert, I've seen a quad using constant speed props in this very venue. There was a video and a paper associated with it as an experiment from some university. A bit of searching should pull it up.
I just had the idea of using 4 separate gas powered helicopters attached to a central core, then you could use the individual collective servos to control the pitch of each of the 4 rotors. Obviously this would be more expensive than using a single central gas engine, but it is a thought none the less.
With a special cnc gear box mounted to a central engine, you could then use the helicopter tail boom idea with servos to control each rotor's pitch.