How can I design a homemade dynamometer for small engines?

[Ayrity]

hey all, as you may or may not know, I am playing around with some different engine designs for a project, and I am using small engines, under 5 HP, and I was wondering if anyone had a suggestion about how to get a reading so I could determine what gave me more power vs what could hurt performance? I was thinking something like make a frame, bolt in the engine and a large electric motor, run a 1:1 ratio pulley and put a volt meter on the electric motor then i could try to calculate watts and convert to horsepower. I suppose I would have to get a strobe to get rpm as well. any other bright ideas? thanks

 

[Cliff_J]

I'm assuming you would like the least expensive? I'd think a couple stripes on the flywheel and a RPM sensor from Tower Hobbies or similar model aircraft source would be a very inexpensive way to get RPM. That or you could use a magnetic or some other pickup to get a pulsed output but would be more work than a paint job.

I'd suggest finding some sort of load cell, a water pump or electric motor tied to a big load (like big resistor wires from space heaters) and then simply measure force at a known distance to figure out torque. It could be beam deflection, a weight scale, or even a load cell. If the electric motor case is free to pivot on a second set of bearings concentric with its shaft bearings, and you have a 1 foot long arm to prevent it from pivoting on this second set of bearnings, now you can measure torque easily with a scale.

Then its a simple scalar equation to find HP, if you're in ft-lbs of torque, its RPM * ft-lbs/5252 = HP.

The biggest question: How do you vary the load so you can hold the motor at an RPM and open the throttle completely? With resistance wires you could probably work out some sort of variable resistance, maybe you could use on old stove (cheap at a garage sale) or something like it.

If you just measured the watts, you'd need to factor in the efficiency of the electric motor. It wouldn't hurt your precision so each change will still have its relative merits, but your accuracy would be hurt. Measuring torque is the method used by many engine dynos and many of them are the water pump variety (they vary the amount of water introduced) but the really accurate dynos are eddy-current types.

 

[FredGarvin]

Put an alternator on the engine and vary a resistive load on the alternator via a load bank of resistors.

 

[berkeman]

The simplest dyno is the DynoJet style, where you go from idle to full throttle on the engine, and measure the d(RPM)/dt at a load that has a good-size moment of inertia (like a weighted drum). You only need to have a photodetector and data acq arrangement and do the math to calculate & plot the horsepower & torque versus RPM. The DynoJet style graph is the one typically shown in motorcycle magazines, for example. The constant load & steady RPM dynos are harder to make and maintain, although at the 5HP level, dumping the energy won't be as bad as at 100HP.

 

[Ayrity]

cliff: I understand your idea for getting torque, and its a good one, simple. what size electric motor would I need? and would I need to know how much resistance I was putting out on the electric motor or wuld i know that from what the scale told me in how much torque i was makeing? correct me if I am wrong, but when the engine is not accelerating (or decelerating) then the torque being created by the engine is equal to the load the electric motor has on it, am i right?

berkeman: How would you suggest I measure the derivative of the rpm? how would you make a simple (since you say it is the simplest) version of the dynojet model?

 

[berkeman]

Just use a photoreflector or photointerrupter and put a zebra stripe or series of holes along the perimeter of one side of the weighted drum. Use a DIO data acq card (or even just do something with the parallel port) with your PC to pick up the on-off transitions and time them.

 

[Cliff_J]

Ayrity - you'll need an electric motor of higher horsepower than the motor you wish to test. And like I said, an old stove provides 4 variable controls for resistance that are designed to handle higher power levels and you have 4 heat devices and could easily add more in parallel to offset the voltage difference and get the resistance low enough to allow enough current to flow.

An alternator might work, but you'd need a self-exciting alternator or at least one with enough residual magentism to be self-exciting, otherwise they need an external power source. And it would need to be a large one, an old 60A variety only outputs just under 1hp so even if its efficiency is poor its a stretch to be a full 5hp load.

Yes, if you had the engine under test connected to the electric motor as its load, and RPM was steady, then its an equal torque on either unit. So the engine is fixed (for ease you just bolt it down) and the load unit is pivoting. If you use a 1ft long torque arm and measure the force needed at the end of it to prevent rotation of the load unit, that's your torque.

An inertia dyno like a dynojet is really just a super large flywheel. The flywheel's ability to store rotational inertia is expressed as a term "moment of inertia" or to avoid confusion the term is sometimes specified as "rotational moment of inertia" to avoid confusion with bending. Think of a cantalope and a bicycle tire - spin them up and the bicycle tire will resist being stopped or started in motion more than the fruit. Even if they had the same weight on a scale, the distance that weight is from the pivot point makes a big difference. Grab one end of a hammer and swing, now grab the other end, same idea. Its not just the mass, but where that mass is in relation to the center of rotation.

Since it takes a lot of energy to spin up a big flywheel, you can derive the force needed to spin up that flywheel with respect to time and therefore the power as well.

This works in the opposite too, that's why racecars will have a lightened flywheel, its not just to lighten the car but to also reduce the horsepower needed just to spin up the engine and instead allow that horsepower to be transmitted to the ground.

There's benefits to both methods too long to list here. Controlling the test conditions is going to be the most important, it seems the correction factors used and the sensors are the biggest complaint you'll find, especially with the inertia dynos. They are so easy to use, those details are overlooked and when 1% or 2% gains are being marketed the lack of tight control over variables really makes the results questionable.

But if you have an accurate means of determining barametric pressure (an old altimeter retired from service from a Cessna or something) and a wet/dry bulb thermometer to measure temp and humidity, and a SAE correction factor to STP then you could measure reasonably precise data for not a lot of money.

 

[Ayrity]

why does the electric motor have to be more powerful than my gas engine? can't I just hook up a variable resistor (or your stove idea) to any size electric motor (granted it should be a bit beefy), maybe like a starter motor from a car? how much power do those have usually? and as long as it has enough load on it, it will provide the needed resistance... maybe I am just not thinking clearly I am sorry for my confusion.

I was thinking maybe it would be easier to use the water pump idea, use a water pump from a tank of water to a hose with a variable nozzle on it, and make the opening for the nozzle smaller to inclrease the load. Knowing the amount of water pressure isn't important, if i understand correctly the only important things are to get the ft*lb of torque and the RPM, the load is just used as a tool to get those 2 data points. I think this way would be easier, and cheaper (snag an old pump from a junkyard or something)

 

[Cliff_J]

Its just that, you'd need it to be beefy enough. If you have only a 1HP electric motor, that's not going to be much of a load for a 5HP motor.

The water pump is likely going to be cheaper and easier, it is the popular dyno load for engine dynos.

Yes, if you know torque and RPM you know power. Even big automotive companies plot out 500rpm increments where they measure the points. With an eddy current setup or an inertial dyno you can apply fancy math and fast sample rates to get even better data. But most just smooth the curve and call it a day, pretty easy to do in Excel once you do it once.

 

[Ayrity]

cool, thanks a lot for all your help :)

 

[brewnog]

I think the electrical dyno would be easier to do. A 5hp motor isn't all that huge, and it's dead easy to make a load bank out of a stove or couple of electric heating elements, and just switch them in as necessary in known increments. It's exactly how the ones we use on our engine testbeds work, except they go up to about 2500kW. We've not used the old water dynos for years. You could even use 100W light bulbs to give an instant visual display.

A friend at work has just built one, on a slightly smaller scale, for an electric racing car we're building. Biggest difficulty he found (and he just used lightbulbs as his load cell, which is the reason!) was effective cooling.

 

[Ayrity]

hmm, where would i get a large enough electric motor? it looks like 5hp electric motors are kinda expensive from what I am finding...

do you think i could in place of a 5hp motor use something smaller, like a 1hp motor and just use a pulley ratio between the gas engine and electric motor, then divide by the ratio when doing my calculations for the gas engine based on the data collected?

 

[Cliff_J]

Ayrity - you could overpower the electric motor and only suffer from shortened life of brushes and so on. Maybe not an issue if plentiful and cheap.

If you take a 12V motor and apply 24V to it, you get roughly 2x the RPM and 4x the power (since power is volts*current) so you'd want at least 2.25:1 gearing at 5x the power level if the RPMs were matched. But if the electric motor is rated at a different no-load RPM, then you'd want to factor that in as well to your gearing. Maybe just a set of gearing options you could use to adjust to the RPM range and power level would be good in addition to varying the resistance seen by the motor.

 

[Ayrity]

by saying "when the rpms awere matched" do you mean when i have a motor that has a lower HP rating but can endure high enough rpms? and i thought that the electric motor was not all that important other than to provide resistance in a controlable way isn't that correct? so let's say I am using a 1 horsepower electric motor and an engine rated at 3HP, and i also assume I have as much resistance as I need. I should use the larger pulley on the electric motor, and a smaller pulley wheel on the gas engine correct? and the only reason for the ratio in the pulley system is to make sure I don't blow up the electric motor right? now i can measure RPM right off of the engine pulley with a laser tach, but as far as correcting for torque, which I am measuring off of the electric motor with the arm off of the freely rotating housing, so for example if i used a 3:1 ratio and I get 20ft*lb of torque, then the gas engine is really making about 6.67ft*lbs (ok so this probably isn't a realistic number for that small engine haha but just for an example).

my point is that i don't need to bring the electric motor to exactly the power rating of my gas motor for this to work, it can be "bigger" (or thought to be bigger by way of gearing). i think that's right...

is all this math correct? thanks for the help, higher powered electric motors are expensive, even on ebay, not to mention shipping for these heavy things!

 

[brewnog]

All looks alright to me, having just woken up. Don't forget to think about losses in your transmission, (belt slip, friction etc) and to include some kind of generator efficiency into your calculations.

 

[Ayrity]

alrighty thank you very much. some other non-morning checks would be appreciated

 

[Cliff_J]

Its the reverse - you over-rev the electric motor just so you can get enough resistance.

Take a handheld drill with a 2-speed transmission, spin the chuck to spin the electric motor in low and high gear. The low gear (maybe 3:1 motor:chuck) offers a lot more resistance because your hand needs to spin that electric motor faster.

So you need to spin the eletric motor faster, it'll get the smaller pulley. Then if it shows 2ft-lb of torque and you're on a 3:1 ratio, then you have 6ft-lb of torque at the gasoline motor.

Generator efficiency isn't important, its just a load cell there to provide resistance and it doesn't matter if the resistance is from electrical or heat generation.

The belt efficiency should be 97% or so if you keep enough belt-pulley contact to minimize slippage. Its generally bad practice to drive a smaller pulley from a larger one, chains are only maybe 90% efficient (throwing numbers off more) so with the smaller pulley on the electric motor you may need to devise a way to put in an idler pulley very close to help the belt wrap around the pulley and make more contact to avoid slippage.