PDA

View Full Version : The Science of Dyno Testing



volvotuning
Friday 3rd February 2006, 05:15
I have written an article about dyno testing. For anyone who wonders why people argue about accuracy of figures etc, have a read of this. It's quite long, and a bit complex in places but I've tried to keep it as simple as possible.

Here is the link -

www.dyno-tuning.com\dyno-tuning\dynoscience.pdf

Adam.

JUDGENINJA
Friday 3rd February 2006, 05:50
Nice ...not overly complicated article...now it's late so go to bed...!!!!

Andrew
Friday 3rd February 2006, 09:24
and the correct URL is:

http://www.dyno-tuning.com/dyno-tuning/dynoscience.pdf

Well done Adam.

Vikingxl
Friday 3rd February 2006, 09:53
Cheers Adam!

volvotuning
Friday 3rd February 2006, 10:28
If anyone does read it (without falling asleep) and would like any clarification about any point discussed, please let me know so I can include it in the document.

Adam.

Mrsmopp
Friday 3rd February 2006, 10:44
Blimey - beauty and brains LMSO!!

Useful info, makes for interesting reading (and i am still awake!!)

Thanks

x

Engineer
Friday 3rd February 2006, 10:48
Should make it a sticky as it would help new and old members alike who may miss it this time.

weasel
Saturday 4th February 2006, 01:13
interesting - thanks.

siamblue
Saturday 4th February 2006, 03:07
Cracking read Adam, very interesting :)

Gary

volvotuning
Saturday 4th February 2006, 13:04
I've made a a few ammendments, thanks to feedback from Chris Kay!

1. Spelling mistakes removed (due to writing at 4am LOL).
2. Chassis Dyno description paragraph ammended.
3. Explanation why measuring drag loss on coast down will never give 100% accurate engine power.

Thanks!

Adam.

aeroresh
Saturday 4th February 2006, 15:06
Very interesting and I didnt fall asleep.

Who dares to post up the link on VOC then? :bricks:

mikezero
Sunday 5th February 2006, 10:07
Found this info on the scooby forum, very interesting reading!

http://www.dentistmapped.com/roadtuning.htm

volvotuning
Sunday 5th February 2006, 11:44
I do agree with most of the stuff said on there, but there is one point which does need to be addressed - ROAD MAPPING!!!!

Road Mapping is for race cars and NOT for road cars. As it states, the track mapping is the final testing stage, which I also agree with.

Mapping on the road from scratch is generally a bad idea, and anyone who says it's a good idea will normally say so because they don't have a dyno. Mapping on the road is very danegrous for a start, and requires speed of 40-130 mph on a T5, which is quite obviously impossble to do legally in the UK on the public roads. Also, on the road it's not possible to hold a fixed rpm if you need to do static load mapping. I've mapped on the road before and it's not something I'm comfortable with, mainly from the safety aspect.

In summary, you should always map on a dyno and test on the road. Not the other way round.

Adam.

craig
Sunday 5th February 2006, 19:21
I have written an article about dyno testing. For anyone who wonders why people argue about accuracy of figures etc, have a read of this. It's quite long, and a bit complex in places but I've tried to keep it as simple as possible.

Here is the link -

www.dyno-tuning.com\dyno-tuning\dynoscience.pdf

Adam.


ya i no !!!!

pault5
Sunday 5th February 2006, 19:41
Very interesting and I didnt fall asleep.

Who dares to post up the link on VOC then? :bricks:



Go on I dare you :remybussi

volvotuning
Sunday 19th February 2006, 23:27
If anyone is interested in what sort of fan you need to give a turbo car half a chance on the dyno, here is the one we use -

http://www.dyno-tuning.com/dyno-tuning/fanside.jpg

And before the clever ones comment, yes it's parked outsided to show you a comparison of size against a Pug 106. :)

Even this thing struggles in the peak of summer to keep intake charge temps down.

Anyone who has been to us knows that we have two fans - a very small one, and the large one shown above. We use the small one when doing some diagnostics work where we are not looking for best performance figures. But when we are trying to get the maximum performance out of the car, we always use the large one.

Adam.

GuyD5
Sunday 19th February 2006, 23:59
Missed this the first time round, thanks for putting the picture up and bumping the thread Adam ... makes interesting reading and explains a lot of what goes on with the dyno testing ... thanks again mate :)

Geezer
Wednesday 23rd August 2006, 08:24
Excellent read…thanks for that Adam.

Not having ever seen a Dyno in action, your article was well worth the time taken to read and no I didn’t fall asleep.

I have a couple of questions however.

With regard to “Operator Issues” and more specifically “Incorrect Coast Down”, you mention that during coast down on a manual car, the operator must dip the clutch at the end of the run to disengage the clutch from the engine to allow the Dyno to measure the power back to the clutch or the flywheel. Some operators put the car into neutral, which has the effect of lowering drag losses, thereby reducing power figures when convenient.

Question: What is the procedure when you have a car equipped with an Automatic transmission on the Dyno? The same as above? What is the nett. effect of what happens?

Question: We have all seen Dyno charts that have all these squiggly lines all over the place, but how do you read them, and how can you tell if they have been “fudged”?

Can you please write an article on this as well at some stage?

volvotuning
Wednesday 23rd August 2006, 09:20
Excellent read…thanks for that Adam.

Not having ever seen a Dyno in action, your article was well worth the time taken to read and no I didn’t fall asleep.

I have a couple of questions however.

With regard to “Operator Issues” and more specifically “Incorrect Coast Down”, you mention that during coast down on a manual car, the operator must dip the clutch at the end of the run to disengage the clutch from the engine to allow the Dyno to measure the power back to the clutch or the flywheel. Some operators put the car into neutral, which has the effect of lowering drag losses, thereby reducing power figures when convenient.

Question: What is the procedure when you have a car equipped with an Automatic transmission on the Dyno? The same as above? What is the nett. effect of what happens?

Question: We have all seen Dyno charts that have all these squiggly lines all over the place, but how do you read them, and how can you tell if they have been “fudged”?

Can you please write an article on this as well at some stage?

Good questions!

I will add the answers asap to the document. Interpreting a dyno chart is sometimes quite tricky, especially since many dyno systems present the info in different ways. Auto testing is a little more complex, and does sometimes present the dyno operator with some problems, so I will explain this in detail too. Give me a few days :)

Adam.

TwistedSanity
Wednesday 23rd August 2006, 18:15
interesting read Adam, one point i must have missed in the text(which i feel is the most important) is dyno calibration, how and how often do you calibrate?, ive seen some muppets put there mums saxo on rollers before and calibrate from the factory specs :-), how do you do yours?
surely none of the text is valid if the calibration is incorrect

volvotuning
Wednesday 23rd August 2006, 21:01
interesting read Adam, one point i must have missed in the text(which i feel is the most important) is dyno calibration, how and how often do you calibrate?, ive seen some muppets put there mums saxo on rollers before and calibrate from the factory specs :-), how do you do yours?
surely none of the text is valid if the calibration is incorrect

Yes you are right in that the accuracy of the "science" of dyno testing hinges on the correct calibration of the dyno itself, just like any other scientific instrument. Of course, my text assumes that all dynos are correctly calibrated, but since assumption is the mother of all f**kups (as Hamish keeps saying), I do really need to explain calibration. I'll briefly explain it here, but will add a more detailed explanation in the article. Certain uses for dynos don't need calibration, and I will explain that too.

Here goes...

Every dyno needs to be calibrated so that the power and torque measurements that it displays are indeed true, precise and consistent, and not under-reading/over-reading. Let's not confuse incorrect calibration with apparent under-reading and overreading of power because all the factors mentioned in the article can affect the power readings. Therefore, calibration cannot be determined by testing a car and comparing to factory specs!!! :) Nor can calibration be confirmed by putting a car on a different dyno and seeing if the results are the same. That would be like comparing two watches to see if they are both telling the same time, when they both may be 3 minutes slow. Calibration itself therefore needs to be precise and correct.

The method of calibration of the dyno depends on the method of testing with the dyno. For example, our dyno can use one of two modes for measuring power and torque - 1) Inertia and 2) Torque Cell. This means that our dyno needs to be calibrated seperately for each method.

Inertia calibration - this subject is indirectly mentioned in the article. Remember the equation Force = Mass x Acceleration, where the mass of the roller is constant, so by measuring the acceleration you can calculate the force and therefore the power. (The formula is a bit more complex when applied to a rotating mass, but the principle is the same). This is a fundamental law of physics which cannot be changed. But the best bit is that the same formula automatically calibrates the dyno! The mass of the dyno (including retarder) is precisely measured at the factory. These values are stored in the dyno software and cannot be changed. So the dyno is calibrated from the factory for Inertia testing and this will always remain correct because the mass of the roller does not change. Ie, for inertia testing, the Dynojet does not require any calibration. I must add that for this calibration to be correct, the weight of the entire rotating mass must be measured extremely accurately, and even the bearing drag must be accurately measured. For this reason, a dyno which was not designed to be an inertia dyno from the factory will probably never ever give accurate results.

Torque Cell Calibration - This is the method which most dynos equipped with a retarder will use. Torque is measured by a torque cell. If you apply a precisely known amount of torque, then you can calibrate the torque cell to give you the exact same torque reading as the true torque you are applying. This is commonly done by applying a known mass to the end of a known length bar (calibration arm), which is attached to the retarder. This rotational for exerts a force on the torque cell which enables the calibration to be made. Unfortunately, this is where all such dynos can be incorrectly calibrated, be it on purpose or by accident. By applying slightly more or less weight on the calibration arm, or by using a slightly longer or shorter calibration arm, then the calibration will not be correct.

This is where our Dynojet has a big advantage - we can "confirm" the torque cell calibration by comparing the dyno test results from a torque cell run with the results from an inertia run. The power and torque curves will be exactly the same in inertia mode as in the torque cell mode if the torque cell is calibrated correctly. This is how we can prove that our torque cell is calibrated correctly, since the inertia calibration is always 100% correct. We use a normally aspirated car for this test, because a turbo car can often behave differently with repeated runs due to rising charge temps.

So what does this mean in terms of the tuner and the customer? Well, a good tuner has no interest in having a badly calibrated dyno for the simple reason that this would prevent them from making accurate measurments themselves. We tune cars for people who taking racing very seriously, where 1 or 2 bhp extra can be enough of a difference to win or lose a race. If our dyno was not accurate, then we would simply not be able to do our job properly. In terms of what you the customer sees, well that really does depend on the dyno and tuner you are using. How do you know if a torque cell dyno is calibrated correctly? The honest answer is that you don't! That is something that you learn about with the reputation and credentials of the tuning shop you are using. Does it matter? That depends on how accurate you need the results to be.

Adam.

volvotuning
Wednesday 23rd August 2006, 23:07
Question: What is the procedure when you have a car equipped with an Automatic transmission on the Dyno? The same as above? What is the nett. effect of what happens?


You can't disengage the clutch on an auto like you can on a manual, so in an auto you can only put the car into neutral. Obviously this means that you are only measuring coast down losses back to the neutral gear, so you can't really measure the complete transmission losses. Autos also pose another problem since they always want to kick down at full throttle. This makes it very difficult to dyno an auto below approx 3500 rpm. Torque converter slippage adds even more problems!



Question: We have all seen Dyno charts that have all these squiggly lines all over the place, but how do you read them, and how can you tell if they have been “fudged”?


In a nutshell, you can't really tell if the dyno run has been fudged because the dyno software will plot what it measures. So fudging the results can only really be achieved by not measuring the data correctly in the first place, or by the factors that influence power and torque as already mentioned in the article. It's possible that the dyno operator can manually edit the raw values recorded by the dyno software in order to give a different graph, but this is something which we cannot do anyway because our dyno data is encrypted.

One useful tip - by measuring only wheel power you can eliminate any fudging caused by incorrect drag loss measurment when trying to calculate engine power.

The "squiggliness" (if there is such a word) of the graphs depends on how fast the data sampling is and how much of this detail is included in the graph. For example, if you choose to plot the graph at 250 rpm intervals, you will have a very smooth line, compared to one that is plotted at say every 1 rpm. A smooth line can therefore hide potential running problems such as ignition retard due to knocking, but this isn't really a fudge and is often used for presentation purposes to show overall characteristics.

As for interpretation of charts, I need to do this as a seperate topic and need some pics to demonstrate. Need time for this as it's quite a detailed subject, especially with different dyno software available.

Adam.

BlackBeast
Wednesday 23rd August 2006, 23:31
You can use my graphs for that if you want to Adam!?! :) :) ;)