Self Built Dynamometer accuracy

[Diablo_niterider]

hey guys,

i wanted some inputs regarding the accuracy of self built dynos from any teams that have built their own dynos,

i came across a tech paper from rit on their self built dyno , any inputs from them would be great cause we are thinking on similar lines

we are pursuading the university to buy us a dyno and incase that does not happen build our own

it would be great to have inputs from you guys

thanks in advance

[Diablo_niterider]

hey guys,

i wanted some inputs regarding the accuracy of self built dynos from any teams that have built their own dynos,

i came across a tech paper from rit on their self built dyno , any inputs from them would be great cause we are thinking on similar lines

we are pursuading the university to buy us a dyno and incase that does not happen build our own

it would be great to have inputs from you guys

thanks in advance

[Wesley]

We've found our dyno (using a Land and Sea load cell) to be accurate to the competition dynos to 2-3HP,

We've just finished building our new dyno to replace the aging one we had for the F4i.

Just make sure you have some sort of data acq, because logging data by hand is a real pain.

[PSUAlum06]

I don't think accuracy is nearly as important as precision.

[Steve O]

I have to agree 100%. Ever noticed how different your numbers are going between a mustang dyno and dynojet? The numbers you get out of your dyno do not matter as long as they are consistent. Honestly I don't care if my dyno says I am getting 15 hp or 150 hp as long; if my car is faster than your car does it matter? A dyno is a tool for you to quantify the changes you make to your vehicle so make sure your numbers are repeatable and you should be all set.

[Wesley]

Different words, same meaning.

If it isn't precise, it isn't accurate.

[Jevon]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Wesley:
Different words, same meaning. </div></BLOCKQUOTE>
Sorry to not add to the discussion but I wanted to point out that accuracy and precision are not the same. Accuracy and Precision (http://en.wikipedia.org/wiki/Accuracy)

[Diablo_niterider]

yes definately accurate ad precise is what i am looking for.

@ wesley - can youshare a few detsils of the dyno you guys built i.e time reqd , cost, type ,problems , anything would be of great help

we are planning for an water cooled absorpton type engine dyno and buying the available data aquisition system and motor control unit.

let me know what you guys think about it

[Superfast Matt McCoy]

The Oklahoma dyno is not a good role model for young and impressionable dynos. It's basically just a land and sea water brake/computer, some electronics, and an engine all welded together with some haste and after-the-fact reinforcement for load paths. It's very loud, very hot, and looks like it was slowly assembled buy four generations of engine team leaders who never talked to each other about it.

A dyno is just a brake, like a very inefficient water pump, and a strain gauge attached to a torque arm. The higher the strain, the higher the torque. If you can log the strain and RPM, you've got a dyno. Assuming you have a limited budget, a dyno computer is an unnecessary luxury if you have a simple data acq and a spreadsheet program.

As far as repeatability, get a strain gauge that doesn't have a lot of thermal creep and spend some time on your cooling system to ensure consistent engine coolant temperature. If you can maintain some consistency with ambient temperature and humidity, that would be ideal. This could be as easy as testing during the early evening and adjusting the time appropriately for different day's conditions.

You can get 80% of the way to a perfect dyno for very cheap if you are careful about the important things. I recently visited a major auto manufacturers racing engine test facility and I was blown away by the millions and millions of dollars that were spent on specialty equipment and facility upgrades just to get repeatability under 1 hp. if you're within 3 hp, I'd call it good.

[Wesley]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Jevon:
Sorry to not add to the discussion but I wanted to point out that accuracy and precision are not the same. Accuracy and Precision (http://en.wikipedia.org/wiki/Accuracy) </div></BLOCKQUOTE>

Yeah, if all Blips are Bloops, are all Bloops Blips and all that. SAT/ACT semantics.

From a technical point, I agree. But you have to answer him in the context of his own question and intent.


And Matt, we've built a new dyno.

[VFR750R]

we've spent alot of money to get 1hp repeatability, but that' roughly .1%. that's .1hp on a 100hp engine. The major factors you must work with is a driveline that doesn't take power and doesn't change. That pretty much rules out chains/belts. Direct drive is the best, and you may be able to hook the absorber directly on your transmission output shaft, if not, a small driveshaft with universals works fine.
Water and oil temp; you'll have to have some sort of control anyways, otherwise you'll overheat. I'd use a open cooling system heat exchanger as been described on this website in other posts.
Air temp, humidity, barometer. Record all three at the inlet to your engine (preferrably real-time while making runs, but not necessary, and use the SAEJ1349 correction factor. This is the current industry standard. This will allow you to compare tests from different days.
You should also consider
-Airflow over the engine, think really big fan
this keeps electronics, headers cool, and makes stuff last a lot longer, it will also be a source of repeatability keeping BOTH manifolds the same temp every run.
-Reasonable exhaust that will allow your neighbors to not call the police everytime you run. Big industrial mufflers, or car mufflers.
-air to the engine should come from away from the engine for best repeatibility, you don't want hot air off the exhaust or 'exhaust' going in the intake.
Lastly, come up with a 'procedure' where you start with the same water and oil temps, do the same number of runs/pulls. decide on ramps/steady state. Always do everything the same way for every test. You may alter your procedure after you see how the engine behaves.

Make sure you test the dynos repeatability BEFORE you start doing testing. Don't assume it will run the next day like it did today. A bad dyno isn't much better then no dyno.

[Superfast Matt McCoy]

Keep in mind that J1349 is only valid for corrections of less than 3%. I imagine it's acceptable to be beyond that for your testing as long as all your testing ends up in roughly the same correction range. We tried using that correction in the mile high city with some... impressive numbers.

Also, I'm glad to hear no more Sooners are risking life and limb while dyno testing.

[VFR750R]

Most areas are much closer to the 72F, 29.34"hg of 1349 then the 60F, 29.92"hg of j607/STP. luckily we're so close that our corrections are with +-1% of 1.000. And all of that is baro, we have a dyno air system that controls air temp and humidity (one of those expensive things that gets us to .1%)

what did you guys use instead?

[Superfast Matt McCoy]

J1394 is the way to go, no doubt. For FSAE teams Dyno testing in the garage, there's something to be said for minimizing the variation in ambient conditions. It's something I first noticed when testing in "If you don't like the weather, wait a few minutes" Oklahoma. There was more consistency in Denver, but the magnitude of the correction made it clear that there is a limit to it's ability to accurately correct.

I can't say how much you'll be off if you correct 5%, but when you get to 18% the numbers get pretty far off. Like stupidly optimistic.

[VFR750R]

Did you guys ever work on making a 'Denver' correction factor? Basically a correction for how an engine would run at Denver's altitude and you just happen to be running in denver. At least you would have limited correction, even if its way low for how the engine would run at competition.

[Superfast Matt McCoy]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by VFR750R:
Did you guys ever work on making a 'Denver' correction factor? Basically a correction for how an engine would run at Denver's altitude and you just happen to be running in denver. At least you would have limited correction, even if its way low for how the engine would run at competition. </div></BLOCKQUOTE>

My Denver testing experience was with Furniture Row Racing, the NASCAR team. I bet with an FSAE car, a team would be capable of predicting the proper tuning correction for their fuel map including the power curve shift from the change in intake pressure wave velocity. With the intake design in a Cup car where the carb is sitting on top in the middle, the pressure waves of each cylinder and the interaction of the cylinders makes the problem much more multidimensional. In any case, they shut down the engine department and started leasing engines before we got that far.

[Mikey Antonakakis]

Have any teams built an inertial dyno? Our team just inherited an eddy current dyno from the mechanical engineering department, but it is about 70 years old and literally takes up half of our shop. It is going to need a lot of work to get it going again, and I just don't think it's feasible. Also, just the generator weighs over 3000 pounds, so there is virtually no portability.
I know inertial dynos have been used pretty successfully in karting with motors that make about 20 hp. http://www1.cedar-rapids.net/tdkmotor/
I'm thinking since there isn't a huge motor to overheat or a water brake to overheat, it should give pretty consistent results. It will probably end up being much less expensive than an eddy-current or water-brake dyno, and pretty easy to fabricate. Data acquisition would consist of a hall-effect sensor, then some simple conversions in the computer to get a torque and hp graph if the moment of inertia of the flywheel is known.

[Drew Price]

I think that it is a fairly feasible idea, but there are a couple things that need serious consideration for an inertial dyno setup.

1) Cost. You will need something very heavy. Steel prices are quite high. Shipping is also pretty pricey with current gas prices, so making or buying a flywheel heavy enough to keep diameter to a reasonable size so that you have a hope of packaging it might actually be comparable to saving for a small water brake by the time you make shields, get suitable bearings, etc.

2) Safety. Your team will have to discuss the feasibility of safe operation. This will be a very heavy, very quickly moving object. Any imbalance during it's manufacture or operation from bearings wearing out or failing, axle not exactly centered, parts made slightly out of round, mounts fatiguing, etc. can be a serious safety hazard. Many high powered race vehicles require a scatter shield or blanket around the flywheel housing.

One of our guys worked in CAT's engine calibration lab for the last 6 months. They operate the equipment from another room, and the engines operate in basically a bomb shelter. One of the test labs has tracks along the floor, up one wall, and onto the ceiling where a harmonic balancer came loose from the crank of an engine being tuned and walked it's way all around the room before stopping. I'm sure it took some equipment out with it.

Just make sure you don't overlook anything if you decide to do this. If the bearings fail or rotor becomes unbalanced in a water brake, you probably spill a lot of water, and it starts making horrible noises, but doesn't try to make it's way outside of the building.

IMHO.

Best,
Drew

[Mikey Antonakakis]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Drew Price:
I think that it is a fairly feasible idea, but there are a couple things that need serious consideration for an inertial dyno setup.

1) Cost. You will need something very heavy. Steel prices are quite high. Shipping is also pretty pricey with current gas prices, so making or buying a flywheel heavy enough to keep diameter to a reasonable size so that you have a hope of packaging it might actually be comparable to saving for a small water brake by the time you make shields, get suitable bearings, etc.

2) Safety. Your team will have to discuss the feasibility of safe operation. This will be a very heavy, very quickly moving object. Any imbalance during it's manufacture or operation from bearings wearing out or failing, axle not exactly centered, parts made slightly out of round, mounts fatiguing, etc. can be a serious safety hazard. Many high powered race vehicles require a scatter shield or blanket around the flywheel housing.

One of our guys worked in CAT's engine calibration lab for the last 6 months. They operate the equipment from another room, and the engines operate in basically a bomb shelter. One of the test labs has tracks along the floor, up one wall, and onto the ceiling where a harmonic balancer came loose from the crank of an engine being tuned and walked it's way all around the room before stopping. I'm sure it took some equipment out with it.

Just make sure you don't overlook anything if you decide to do this. If the bearings fail or rotor becomes unbalanced in a water brake, you probably spill a lot of water, and it starts making horrible noises, but doesn't try to make it's way outside of the building.

IMHO.

Best,
Drew </div></BLOCKQUOTE>

I feel you there. Luckily, our engines will probably never make over 100hp. I just did a bunch of calculations, and a very bad approximation for stress (on the safe side though) and I have a factor of safety of at least 2 using a mild steel disk that weighs 200kg, has a radius of 306mm, a thickness of 95mm, and would spin at 340rad/s at an engine speed of 15000rpm with a 4.633 final drive ratio. This would give us a 7 second pull, which would simulate the load in second gear pretty well. But these calculations will definitely not be anywhere NEAR enough justification to call this thing safe. But size-wise, it seems feasible.

[VFR750R]

Should be more feasible then water brake, although you'll still have to have a 'brake' of some sort to slow that sucker back down. If you put 100hp into it over 7 seconds; that's alot of kinetic energy.

Consider the limited kind of data you'll be able to get too, sweeps are really nice but unless you run a turbo engine there's nothing to learn from them you couldn't learn doing steady state. But steady state will show you much more resolution, and allow you to tune maps faster ect. Also consider a 'baker fan', i've seen them used to tune tractor carbs at fairs/tractor pulls. And with 6 gears in the transmission, it would be easy to get 6 spread out data points.
Safety cannot be understated, for any dyno. I've seen driveshafts come apart at speed and do things you couldn't fathom to HEAVY steel shields. I've also heard a story of a flywheel not tightened that sawed through an aluminum bellhousing, and through 4 sheets of drywall into another room.