Check this out !

http://www.cmsraceperformance.com.au/roaddyno.html

I am ordering mine on monday, can't wait....

I was going to build my own with an accelerometer and a labjack data logger ! stuff that I can but it far alot less and the software is top notch.

 

Very very very interesting!
Must do some more research on this product!

crispeed
87 Rx-7 TII
9.20@150mph
2600lbs

 

i read the article, to start with it seemed to be talking about flywheel HP - as one would assume seen as how it wants numbers like diff ratio, gear ratio, tire roll out etc etc etc, then it gives you the rwhp and you can add any number you want for the driveline hp loss.

seems a little odd to me

 

No No son, you are missing the main thing...

The factors you see are to take into account drive/tire loss, if you want true RWHP you do not put in any loss.

Accelerometers are the true accurate way to do things, the only thing you have to be carefull of is to not get wheel spin, i.e test in 3rd gear. Also go to a weigh bridge and get a "real" weight for you're car as tested.

It is easy enough to get the CD (coeff of drag) and frontal area for the type of car you are testing, then input in the values to get a true graph of RWHP.

Need also to test on a flat road....I trust this method over all dyno's except inertia types which do not suffer from calibration and heat related accuracy issues.

The system is kick ***

 

How much? I didn't see a price anywhere. Would you input a rotary as a two or four stroke? I just skimmed over the page. I am rather interested in this, but lack your knowledge

 

2 stroke.

The price is $365 Australian

 

Damn, thats about $175 USD?

 

I'm getting one. Thanks RICE.

 

yeah yeah yeah i understand all that stuff, its covered in like 5th form physics.

i must have missed seeing the piece which said about leaving out the diff etc etc etc.

Its surely is a good price and in a few months will probably look at getting one myself. that coupled with my mates FJO wideband and we should be able to tweak his GTR and my '7

 

Hi Rice,

Long time no chat!

I actually tested one of these a while ago. I tested it on a few cars and found it a bit inconsistent.

It is also completely useless in a car that wheelspins. The instructions were also a bit strange the way they said that runs need to start from idle or very low rpm.

The software and the download capabilities were also a little weird, and sometimes when it looked like it was recording (light flashing correctly) it wasn't recording. The plastic clip that is used to read the ignition also melted while I was testing it.

The unit is actually based on a "home brew" unit from the US that has been around for years.

Maybe they've fixed a few things since then.

 

looks good enough for me for the $

 

Actually the price is in US. So its $365 (US). The price does include shipping though.

 

No, this is off his site

https://www.paypal.com/cgi-bin/webscr

It's in US inc shipping.

a777, yeah it will be inaccurate if you have wheel spin, as is anything... a mate of mine went on a dyno dynamics and got less power than he expected due to wheel spin, the funny thing is that the operator did not pick up on this ! After doing a simple calc based on roller speed and what rpm the engine was doing during the run I calculated that the engine was turning 30% faster than what the rollers said it was !

In regards to the testing in the lower speed or rpm, I think this is irrelevant for us, It is applied for cars to limit the time spent accelerating to get the most accurate readings, I recommend a total time of say 8 to 12 seconds to get the most resolution. The longer time you spend the more the inbuild inaccuracy in the acceleromoter adds up (it is VERY VERY accurate, but over long times it gets worse).

So long as the CD, Weight, and Gear ratio, DIff ratio, Tire size are correct the unit will give you a result to with in +- 1 to 2% (Better than ANY accuracy between two different runs ona chassis dyno, let alone different ones (same type but different operators)!) If you use it for the time I recommended. Also pick a nice flat piece of road (I have been using the same one for 5 years for my Autronic data logger) to get accurate comparative data over the years.

For the money I will get the system, If it has any issues then I will resort to building my own (as I originally intended to do).

 

why on earth would they put the price of $365US on the webpage of an australian shop??????

Its gotta be in AUS $$ or else EVERYTHING in that shops catalogue would be in US$ and it isn't

 

http://www.charm.net/~mchaney/roaddyno/order.htm

 

its because everybody loves america . that sounds like a better g-tech, let us know how it turns out rice

 

OK READ THIS !

I take this back, I made a mistake I thought it was an accelerometer. It is not from what I understand now.

When I did an initial search I found his other product which was a dual axis accelerometer, then followed the link to this product.

It purley works on recording spark output ! and claulating everything else form this and the data a person send into the software !

THE MACHINE I WAN TO MAKE, relies on an accelerometer, all you will need to do is take a coil output for rpm and weigh the ca as tested that's it. My system requires independant speed measurement (I have not finalized on the type yet).

Basically how it works is that you accel in 3rd say from 2000rpm to 9000rpm, all through this range the G level is recored, then you wack the car in neutral and record the G to slow back down to the original speed you started at (again need flat raod). Then use basic physics to derive power and torque.

There will be no guessing about axel ratio, gear ratio tire diameter CD factor, frontal area etc...ALL you will need to have is a flat road need not be flat just consistent for the run and traction (test in 3rd or 4th).

I think I will develope my own as it will be totaly accurate (+-1%) and will suffer from less variables, still may buy the other one for a bit of fun

 

sorry for spelling mistakes !

 

I have bounced this idea around A LOT. I've got the circuit all figured out / processor etc. The only thing that's stopped me was the lack of a market (ie it's hardly worth it for just myself).

The best way to set up is to mount a DUAL channel accel. in the y-z plane. This allows the software to calculate out the 'tilt' of the car under hard acceleration. You can flag a failed run from wheelspin by measuring engine accel vs actual car accel which would seriously reduce the variation in HP / 1/4 calculations.

So would you be interested?

-pete

 

I'm down ! saves me making one

 

Hmm... thats an interesting concept... in truth all you'd really NEED is RPM input, 'cos you can calculate acceleration from that after you input some known values (gear ratio, tire diamater, weight). You could then calculate how much flywheel horsepower by coasting down in-gear and adding the absolute values from each measurement.

Actually, this is the method I used to calculate the torque curve on one of my cars It's a little clunky but it does work well enough for curiosity's sake, when using a stopwatch... but if you have a computer measuring it all...

 

No you cannot, because coasting down in gear takes into account engine friction.

It must be done out of gear as I stated, this will give you the aero and mechanical load placed on the car via the chassis.........that is why you need to measure the G's in this situation. What you measure here you add to the acceleration G's to get the total G force required.

You need this for every speed point you did in any one gear then you can calculate out TRUE RWHP, the rpm pick up is only for translation of the road speed to the rpm of the engine (IN THE SYSTEM I WAN TO MAKE OR BUY). You only use the TRUE road speed to determine the RWHP required to abtain that amount of G.

Hence no need for tire measurements (unless the speed sensor needs this input) i.e pittot tube type would not need any such input, no need for diff ratio, no need for gear ratio, no need for CD or frontal area.

Just need speed, corresponding total G's and corresponding RPM. The decel G force I suggest should be calced on the same stretch of road that the acceleration was done on to get totally accurate readings.

Not this same method was used and recommended in "Turbochargers" by Hugh MacInnes. The only difference now days is that we have the ability with modern electronics and sensors to do a whole dyno curve in one run with excellent resolution.

See page 41 of his book for full description of the theory, It is called the Dick Griffin system, note Inertia type dyno's work on the same principle and they are the MOST accurate ever devised.

 

The only thing you need with this system is a accurate weight input, and a consistent piece of road, and good traction, if it spins in 2nd then try 3rd, if this no work, try 4th !.

I just need someone to write the software for me could use something like a labjack to obtain the data from the accelerometer, rpm pick up, and the speed sensor.

If not just input data values into Excell and make you're own graphs after a bit of sorting of the data.

 

JAMECO sells 2-axis accelerometers, and we've used them with a MoTeC ADL to datalog G forces (acceleration) and cornering loads.&nbsp Should be $100 a piece...

Yep, the MoTeC backs out speed (and gear shifts) by inputing a ratio of MPH to RPM - by just keeping track of RPM, it know how fast you're going.&nbsp The accelerometer is accurate enough to detect tire slip!



-Ted

 

You're right, forgot about engine friction Oh well, you could just use an arbitrary percentage for driveline losses and a rough guesstimate of air losses based on calculayted vehicle speed, Cd, and frontal area...

or you could just get an accelerometer and be done with it It's a nice way to think outside of the box though.