I really don't want worms everywhere, just out of curiosity what do some of the mainstream dyno technologies use for guesstimating the "At the Fly" output from the "At the Wheels" measurement?

I've searched around, seen some comments that suggest a static ratio and others which suggest it's something like +X static BHP and THEN a ratio... if that makes sense.

My car has been on two different dyno dynamics dyno's and the ATF value given was 127% and 129% respectively. My ATW readout was almost 30bhp less on the second one though which suggests they may use something a little more complicated than a flat percentage increase.

I'm assuming the settings would be the same across all same-brand Dynos, or is it something configurable by the operator?

It's a shame really that ATF figures are the universal benchmark for performance, considering they appear to involve a great deal of guess work and assumptions!

 

I'd give little credence to dyno bhp figures and rely on the bum dyno and how the car drives.

 

Agreed 100%

I still want to know though :P

 

You are an analytical bloke 😃

 

It just seems to be a figure which (rightly or wrongly) is a benchmark for almost everything we do to our cars and from what I know so far... It seems to be such a "loose" statistic.

Does transmission loss suffer diminishing returns or is it a flat proportion? Learning how the dyno technicians/equipment deals with it would give me at least a clue!

 

Calling Shaun.... Mr. Fennings 👍

 

Dyno Dynamics RR's use a flat percentage when run in shootout mode - the car should have been run in shootout mode on both DD RR's (21.8% iirc for Scoobs on the older software).

Other RR's like MAHA and Dastek use coast down loses to measure the "drag".

At the end of the day they all read much a muchness if set-up and operated correctly.

Post the graphs up and we maybe able to give you more of a "clue".


Although I'm just about to go out on the ****, so doubt I answer back again tonight.

 

PMSL

 

ATW is the only real comparison. It's hard to put a fixed value on it as the age of the oil in the gearbox mixed in with the condition of the drive train can all attribute to this % loss ATF.

 

You've probably missed me saying this before, so just in case I'll say it again.
You can't do that either as all dynos even measure the wheel figure very differently.
We've had one customers car ran on 3 dynos in 5 days.
There was a 80bhp discrepancy across the board in the wheel figures and only a 20bhp discrepancy in calculated flywheel hp.

 

ATF figures from chassis dynos are pointless comparisons really as the Dyno didn't measure that figure. It measured the power at the tyres.
(not the wheels you will note - thats different again)

A bigger issue for all Dyno fans though is correction, which, interestingly enough we are discussing over on another forum and I have been educating folk on how open to abuse Dyno figures really are and made a video to illustrate the point. I didn't know if there was such a topic over here so I didn't bother replicating the topic on Scoobynet for fear of seeing tumbleweeds blowing by ... Lol

 

Thanks guys, useful info.

My dyno experiences so far have been:

Dynodynamics: 219bhp ATW and 284.6 ATF = 130%
Dynodynamics: 248bhp ATW and 315.6 ATF = 127%
Hub Dyno (dynapack?): 282bhp ATW and ??? ATF.

I looked at some other graph printouts from other people all across the power range and dyno dynamics always seem to hover between 125-130%.

Basically my car was at three different stages of tune/hardware at these dyno runs, so the differences if ATW don't concern me between the three. You can see that the percentages I worked out across the two (different) dynodynamics runs is slightly different too so that suggests that they DON'T operate a flat percentage increase.

The final run was on a hub dyno, and the software only provided an ATW measurement of 282bhp. The technician asked if I was bothered about an ATW readout and said "not particularly" but he basically got a calculator out and tapped in 282 x 1.15 to give me 324.3 ATW - by his admission it was a conservative conversion.

130% is quite different from 115%, but the 115% used was on a hub/chassis dyno so I'm not sure whether there's a science behind the difference.

I'm just curious, the car feels faster so I'm a happy boy - I still like to know how things work though.

 

Doesn't that mean that the dyno isn't calibrated? Or do the dynos use the correction factor back to the fly as the calibration?

 

No, what it means is due to the larger/smaller contact patch, different methods of tie down, different dyno braking methods ie inertia or eddy current they all throw out differing wheel figures too.

 

The problem is, without all the correction data, your figures above are pretty meaningless with regards to actual accuracy.
Unless you have data to prove that air temps and ambient pressure were somewhere near correct, you have no real idea why the Dyno said what it did.

The flywheel figure is obviously based initially on the only accurately measured figure, the one at the tyres, and if that is incorrect due to fudged correction, the flywheel figure will be an even larger percentage incorrect than the flywheel figure guesstimate.

I should have the first of the 2013 shoot out beta test softwares installed on my Dyno next week as the factory want some feedback on it before they release it so I will see if I can figure out what they use for flywheel figures but I am confident it is NOT a simple Percentage.

 

Thanks Stu, that was pretty much what I was after.

I'm well aware that even ATW measurements are subject to far too many variables to be properly accurate. The main benefit for me is a simple arbitrary scale to see whether a specific change has resulted in an improvement or not.

I've now found a hub dyno locally which I'll be using rather than traditional "rolling roads" where possible to rule out at least a few of those variables.

I can feel climatic differences with my "butt dyno" so I'm well aware that a hot day will change the performance of my car, but that still is irrelevant with regards to this thread.

I was interested in how the "ATF" estimates are produced from an "ATW" measurement, regardless of how accurate or not it may be.

 

dont forget that also in dyno scenarios the car is not actually moving so your engine will be running hotter and be making less power than it would out on the road.

even with a good fan setup most cannot match the airflow of actual road driving. imagine walking near a dyno fan that was able to blow air at the equivalent speeds of a 3rd or 4th gear pull.. try putting your head out the window at those speeds, dyno fan doesnt even come close lol

 

Not all dyno's are that bad. A dyno Cells effectiveness can be measured quite easily by graphing the cell temps during a few back to back runs. If the cell doesn't get any hotter then the cooling is fine, assuming the cell temps are taken within 2m of the engine bay and are measured BEHIND it, not in front of it. (Assuming correct cell airflow direction). The biggest issue with cells as far as a dyno's accuracy is concerned is lack of cell airflow (not engine) and lack of exhaust gas extraction. Either of these deficiencies will equate to inaccurate results and make the whole process pretty pointless which is, for me, one of the biggest surprises when I visit installers with such setups. They spend so much on the Dyno, but nothing on the cell to make it work correctly, accurately and, most importantly for tuning, repeatable.

 

A good Dyno will be within 2% run after run and as long as you get the all important correction data and saw the strapping tension yourself you can be reasonably sure your power gains are indeed accurate. You can always ask for at the wheels and of course non corrected data if you want to be absolutely sure, but the Uncorrected data is absolutely dependent on the quality of the cell itself and NOT the chassis dyno of course.

 

engine dyno is only way to know for sure

different dyno produce different results, transmission losses are just a guess.

So much can affect the result then its just realy a guide rather than a exact figure, i;d be mroe concerned with the AFR figures from it

 

Why is an engine dyno "only way to know for sure"?

It's the only way to know for sure assuming that it's accurate (which is another bone of contention) and that you like to know how much power you have with none of your ancillary parts connected, which in many maybe pretty pointless.

Transmission losses are not "just a guess" with some of the dyno's. This is all part of the process of measuring coast down. OK... I know this fires off other "discussion", but it's far from guess work with a number of dyno set-ups.

When all said and done a dyno is a tool and if used correctly, can be a very good tuning tool. I put way more emphasis on real performance and if I measure decreases in acceleration times in-gear, I'm a happy chappy knowing that tuning has had a positive effect on the black stuff.

Stu,
Is your version of DD dyno measuring coast down losses. If it doesn't how can it not (sensibly) only apply a fixed % loss factor in shootout mode (you appear to suggest the latest version doesn't)?

 

Its a tricky one to answer specifically as they wont release any specific information about it regardless of how much I ask, but from the comparisons ive done, their flywheel power guesstimate depends on the original HP output, which to be fair is an absolutely sensible criteria.

To illustrate my thoughts on this, lets say we have a 4wd road car that makes 400bhp at the wheels and we suggest it has a 20% loss through the transmission so we will assume it to be 500bhp ATF.

If we now tune that engine and manage a peak power output at the wheels of 1000bhp. Are we honestly going to suggest that its 1250bhp because we lost 250bhp through its transmission?

If so, can anyone tell me where that 250bhp actually went because the fact of the matter is, we cannot destroy energy, so what did we convert 250BHP of energy to? Before you say heat, just go and work out how much heat that is...

 

They are indeed not just a "Guess". However, they don't measure dynamic drivetrain losses, only static, and if those two items where the same, why does you transmission get a lot hotter under acceleration than it does coasting to a halt? Are the losses not obviously more under power?

More food for thought, if we take a car that's very common for rear brake drag, your old Sierra RS500 cossie is a very good example... what effect are those dragging brakes going to do to the flywheel estimate that the dyno is going to produce after measuring the exaggerated deceleration rate of the drivetrain? Why else do you often see a tired and frustrated dyno operator pop just "1 click" on the handbrake during rundown? LOL

People sat in the boot add a similar "increased deceleration time" factor by adding 300kg of weight to the back axle and dyno hub bearings.

The list goes on - best bet, just ask for at the wheels figures and all correction factors used.
If anyones interested in learning about correction factors, give me a nod, either on this topic or a new one.

 

Stu,
Using some form of equation based on original BHP makes some logical sense - although it's still a flawed method regardless. It has already been remarked upon, that higher BHP cars are perhaps being over egged using the DD fixed percentage shootout mode of the older software.

I believe that DD have now moved away from DIN correction in favour of SAE - is that correct or just hearsay? Not that it should matter too much, but us UK'ers have been using DIN correction since Pontius was a Pilate, so it's yet another confusion factor.

In many ways I can totally understand why many of the existing "old software" DD operators don't want to "upgrade".

One thing I have been really impressed with, in regards to say the Surrey Rolling Road DD set-up, is that of repeatability. It's nothing short of outstanding that I have taken my car there, several months between runs, and figures were within 1-2bhp (at the same level of tune). THAT'S the kind of results (regardless of what the actual figures were) that I find beneficial to me, especially when making staged modifications when you want to see proper net gains.

 

They left DIN about 2 software updates ago so at least 2 years now mate. They favour their own ATMC2 and recommend that all DD Operators use that.


Totally agree mate. Without repeatability its just an expensive power run machine and pretty damn useless as a development tool. I am still at work trying to find another few BHP from a Ferrari Race car and due to having 8750rpm of lag (I.E: Its normally aspirated) it would be impossible to be sure of gains without rock solid repeatability once we get down to the last few horses on race fuel.

 

If your interested in the difference the correction standards make, here is some data from mine. (Copied and pasted info that I wrote on another forum)

Ok,
As most of you hopefully now know, a Rolling Road of any form simply measures power at the tyres (Not wheels, that's different)


That measurement is then corrected to a known standard that takes into account:

• Atmospheric pressure
• Ambient air temperature
• Humidity
• Intake air temperature

The problem here is that there are quite a few standards out there today, and the latest 2013 Dyno Dynamics Dynotech machines offer correction to the following standards.

• ATC
• ATMC1
• ATMC2
• SAEJ95

Or of course we can read it totally uncorrected and not really care about the powers accuracy against industry standards at all. (It will still read gains and losses, but the figures will be somewhat meaningless units with regards absolute accuracy to OEM level)

So, taking the run from the video where we placed the air probe into a cup of tea, we have the following data.

Atmospheric pressure = 1024mb
Ambient air temperature = 26.1 deg c
Humidity = 58.48%
Intake air temperature = 68 deg C. (The temperature of my cup of tea.)

The runs then produce the following power graphs depending on the correction standards chosen.

Uncorrected = 395.7bhp (No correction added - pure measured power)



Corrected to ATMC1 = 424.8bhp (Adds 29.1bhp)


Corrected to SAEJ95 = 426.9bhp (Adds 31.2bhp)


Corrected to ATMC2 = 429.4bhp (Adds 33.7bhp)


Corrected to ATC = 430.8bhp (Adds 35.1bhp)

 

What info are you trying to get form the dyno? how much power the engine can produce, so unless you know the amount of air going in you'll never know.

As far as transmission loss, its an educated guess, but thats about it.

 

Stu,
Do you know what the net difference is between how much correction is applied by using ATMC2 vs DIN?

Tidgy,
I don't understand what you're referring to. I know exactly how much air is going into my engine..... it's measured by the MAF and on a dyno vs the road, those airflow logs are very close when using a good dyno facility.

We could keep going round in circles here when it comes to anything rolling road power measurement related - i.e. in the transmission loss area for instance. The same can be said for WHP (I mean tyre HP lol). Different types of RR and of course set-up can give different readings. Hence why I come back to my "testing acceleration times" as being a good way of measuring gains in the real world.

I still personally feel that an engine dyno is somewhat as floored (but for different reasons) as rolling roads. They are a tool for tuning and regardless of what type, should be seen as that only. Unfortunately we're in a world that loves the "figures" - I'm no different than most others in that arena.

To me (personally) it makes sense to use what the majority of others use, so at least you have some form of comparison across the board (regardless of how accurate the figures really are). This is why sites such as Surrey Rolling Road, ScoobyClinic and Engine Tuner are probably the Subaru "standard" venues to compare graphs within the UK imo (all virtually the same set-up and appear to provide equally comparable results). If it's utter bull**** at least it's a level playing field of bull****. It's the reason why I'll travel 200mile round trip to "just" get a dyno run.

 

Morning old boy.
No idea mate, but if you have the DIN algorithm we can certainly work it out on a spreadsheet. (or I know a man who can over on PF)

 

The bold, Very close, is what i mean shaun.

For most folks/situations then rr dyno is more than enough, but my comment is about being completley accurate.

A experiment is only as good as the control, and on a dyno you dont have that much control over airflow etc, unles syou go full wind tunnel etc.

I agree your unlikley to get the same results around differning dyno's, too many variables