Rolling Roads - Good And Bad

[PumaRacing 1]

I'm not a fan of rolling roads. Back in the good old days when most of you weren't born and I wasn't going grey they were ok because they only gave wheel bhp figures, which is all they actually measure accurately, and flywheel figures were left to your own best guess which is actually just as good as the figure the modern ones come up with anyway. However, pressure from punters wanting to know flywheel figures led to people like AMD and others designing computer systems which added coastdown loss figures to the wheel figures and gradually the whole thing spiralled out of control until nowadays no one has a clue what the hell their engine actually produces because it varies so much depending on where you go to.

 

Finally however there's a chink of light at the end of the dark tunnel. Some weeks ago I had a long and fascinating chat with Gerry Gaffney who runs Dastek UK and markets the Dastek rolling road system which is installed at several places across the country. Gerry used to work for Toyota motorsport and has a wealth of experience of engine dyno bhp figures versus chassis dyno wheel bhp figures for the same engine and used this when they were designing the Dastek system and trying to get it as accurate as possible.

 

To help eliminate wheel slip they made the rollers with fluted serrations that the tyres grip onto properly. Other systems just have polished steel rollers which I've long said is a recipe for disaster when you're trying to put a lot of power down. The internal calibration and calculations the system does when measuring coastdown losses lead to much lower, and more accurate, figures than many other systems give. The worst is probably Powerstation's system which adds about 50% to the wheel bhp figure but there are plenty of others which bear no resemblance to reality.

 

I was even more impressed when Gerry freely admitted and agreed that coastdown losses with the throttle closed actually have nothing to do with 'under power' transmission losses as I've long said on my website. However he designed his system to give the best agreement to true transmission losses as it was possible to get by measuring something completely different. On very powerful engines, 400 bhp plus, those losses can be as little as 8 or 10 percent.

 

I've gradually been acquiring data from cars on Dastek rollers over the last year or so and I'm convinced they are more accurate and repeatable than anything else in the UK. This doesn't mean that the operator will set your car up any better than anyone else of course or in fact even have a clue what he's doing but at least at the end of the day you'll get an accurate bhp figure for your engine.

 

I'm not quite to the point of taking a Dastek flywheel figure as being anything like correct compared to an actual engine dyno flywheel figure but they beat any other rolling road system hands down.

[Dino 3]

I've used Dastek's rollers in the past based on their claims that theirs were the most accurate in the country. To get a comparison I went to Sanspeed in Kent a few weeks later and their rollers showed an identical bhp figure and curve.

 

Compare that to Peter Baldwins rollers which were over 30bhp out.

 

An interesting point you have made above Dave and I am glad that someone has pointed out WHY Dasteks rollers are more accurate than most others as I'd never known the reasons until now so thanks

 

Shame they are in bloody Fife!!! At least I know Sanspeeds rollers give the same results and they are only 30mins from me

[DaveK 0]

where abouts in fife are they btw? i really need to take a trip there in the future.

 

they got a website maybe?

[ablister 31]

dalgety bay

[Jonmurgie 2]

Interesting stuff, and seeing as you mention PowerStation by name I shall pass your thoughts onto them

 

Here's an interesting story about the PS rollers that happened recently. Most of you will have heard of Prodrive, they are basically the approved 'tuner' for Subaru's and are approved by the manufacturer (i.e. PPP packs). So anyhow, they have been working on their MY06 WRX/STi tuning packages and during that development then engine work has been all done on an Engine Dyno to offer an indisputable power figure. So with this work done they then come over to PowerStation/Litchfield Imports to chat about who knows what, but while there they ask for their WRX & STi to be run on the Dyno. So, you all ask, what was the outcome... well the flywheel figures given on the PowerStation Dyno were +/- 3bhp from the figures that Prodrive had produced with those engines on an Engine Dyno... interesting stuff hey...

 

OK, I don't know what the wheel & 'drag' figures were and I'd guess they were as stupid as ever, but as I've personally always felt the calculated flywheel figures do seem pretty spot on

 

Just thought I'd pass that info on for those who completely dismiss PS.

[PumaRacing 1]

Interesting stuff, and seeing as you mention PowerStation by name I shall pass your thoughts onto them

 

Knock yourself out. While you're there you can ask them a couple of things. You've said previously that their dyno is regularly calibrated. Calibration of rolling roads can only be done with the wheel bhp figure and as that clearly bears little resemblance to reality I'm somewhat curious as to how this 'calibration' gets done.

 

Secondly. On this recent thread

 

http://forum.205gtidrivers.com/index.php?s...opic=57112&st=0

 

The power graph

 

http://www.mdracing.co.uk/S616KDF.pdf

 

shows peak power at 7k rpm if you go by the flywheel figure and 6k rpm on the wheel bhp curve. The one engine can't have two different power peaks that far apart and two completely differently shaped power curves so which is correct or neither?

 

It's pretty clear what the reason is though. The drag figure, which rises exponentially, is such a huge percentage of the wheel bhp, 56%, that it distorts the curves. Whether or not the flywheel figure bears some similarity to the truth it's inescapable that this dyno operates in a very strange way. I'm afraid that then begs the question does it even operate consistently?

 

As an engineer, when you see something about a machine that just isn't right it's hard to trust any part of its operation until you've found the source of the problem. I want a rolling road to give an accurate wheel figure because that's all it's truly measuring and a coastdown loss that bears some resemblance to the real transmission loss. There are plenty out there that do so the easiest thing with one that clearly doesn't is just avoid it.

 

Gerry Gaffney at Dastek doesn't just market his own make of machine. He buys, sells and reconditions other makes of rolling road and he'll also tell you that any type of roller, from the humblest old Clayton water brake to the modern Sun and Dastek ones should be capable of generating exactly the same wheel bhp figure as any other roller if they're properly calibrated.

 

What you can't escape is that the wheel bhp figure for a specific vehicle in a particular gear at a given rpm is a scientifically measurable quantity. 550 foot pounds per second per bhp. It isn't some fudge figure that only makes sense when the transmission losses get added back. It's the one single thing that any rolling road should be capable of giving the same figure for as any other.

 

If the Powerstation rollers can't measure this properly and can't even tell you what the true power peak rpm is to within 1000 rpm then you can preach their virtues until you're blue in the face but as far as I'm concerned they're just plain wrong.

 

When I buy a ruler or a micrometer I want it to measure things accurately. I don't want it to come with a sticker that says I have to add 30% to whatever it reads to get the real figure. I'll just send it back and avoid that brand in future.

[Jonmurgie 2]

So your basically saying it was pure luck that the Prodrive cars made the same power they had on an ENGINE DYNO on the POWERSTATION DYNO? I'm not sure if it was clear but it wasn't 1 car, it was 2 sepertate cars and I'll post the actual figures when I get into work on Monday if you like

 

Also, I'm certainly not going to argue about the wheel/drag figures on the MAHA dyno as they are clearly odd, but I am inclined to believe the flywheel figure going on the number of cars I've seen on their rollers and especially on the Prodrive results of a few weeks ago...

 

Good read all this though... and I don't remember saying they have the dyno calibrated regularly as to be honest I don't know that, I do know it was calibrated when fitted last December.

[PumaRacing 1]

One of the things you have to bear in mind about engine dyno and chassis dyno power figures is that the engine won't necessarily make the same power once it's installed in the car. The coolant and oil temperatures, exhaust system used, ancilliaries such as power steering and water pumps, inlet air flow and temperatures, proximity of the engine to surrounding bulkheads and wings can all affect the installed power. Unless great pains are taken to control all these factors, and generally only the major OE car manufacturers do this, then the figures just can't be compared.

 

What tends to happen is the installation losses make a high reading rolling road look similar to an accurate engine dyno but in truth they are both measuring different things.

 

Back in 50s and 60s the American car manufacturers used to rate their engines based on gross engine dyno power figures which were obtained without exhaust systems, air filters and ancilliaries. What you actually got once the engine was installed in the car was about 30% less. Now that situation has long gone thankfully but the basic issues are still valid and without knowing every detail of a particular set of figures there isn't really any point trying to draw conclusions from them.

[Batfink 201]

my view is if you have a problem dont fix the results to negate the problem, fix the problem to start with.

 

Dave as you cant be too far from G-Force in Aylesbury, what do you think of their dyno.

Seems to put most 205 mi16's on tb's around 146bhp at the wheels

though their flywheel estimates suck in shootout mode their wheels figure seems acurate enough with different engines with the same modifications

[PumaRacing 1]

Funny old coincidences you get in life eh? The only contact I've had with G-Force is a couple of emails and finally phoning them up yesterday to tell them that if they continue to ignore my emails asking them to remove from their website the article on dynos which they've blatantly copied from mine they'll find a damages claim for copyright infringment slapped on them in my local county court. Other than that I have no knowledge of their business or any figures from their rollers.

 

To be fair to the bloke who owns it I'm told he's hardly ever there and the person who put my article on their website would also have been the one who ignored my email and anyway he's now left them.

 

PS - 146 at the wheels would seem spot on for that level of tune. It would equate back to about 175 flywheel which looks emminently reasonable.

[petert 550]

To help eliminate wheel slip they made the rollers with fluted serrations that the tyres grip onto properly.

 

I just assumed all modern dynos had serrated rollers. DynoMax and DynoDynamics are probably the most common over here. I don't know of anyone who talks in flywheel figures. Once you've driven out the showroom door, everyone talks in hp at the wheels. Flywheel figures from a rolling road seems to be a UK thing.

 

The DynoDynamics setup does have the ability to produce a flywheel figure, but once you see how it's done you'd never worry about. Basically the operator enters a "fudge factor" until the desired numbers come up on the screen. DynoDymanics have tried to take bull out of rolling road figures by introducing "Shootout Mode". The operator then has limited input, meaning dyno's 1000Km away should produce comparable figures.

Edited April 8, 2006 by petert

[rich_w 0]

Jon, I don't suppose they compared the bhp and tourqe graph's? I think that would have been interesting.

[base-1 17]

G-Force use Dyno Dynamics rollers, which are very accurate for wheel figures (although no more than any RR should be I suppose!)

 

The write-up on their website is different to the last time I looked, much less old-looking but basically the same words. I suspect the guy referred to that has left is Chris? G-Force are a useless bunch of clueless fools these days - couldn't even get anywhere near accurate flywheel figures despite us telling the formula off the Puma website, and seemed to just stick in their usual 4WD loss percentage

 

Chris seems to be the only guy who knew anything (and was honest about what he knew and didn't know, which I still think is great, no bulls*it etc), and he now has his "own" place in Silverstone and is very well known and respected - GF's loss!

Edited April 8, 2006 by base-1

[DrSeuss 0]

oddly enough it seems that powerstations choice of maha rollers is the cause of the inaccuracy. I was reading a thread on a civic forum on a totally unrelated set of rollers (same make, different country).

 

The graphs posted showed realistic hp gains, but the coast down losses were once again, laughable.

[smckeown 1]

Jon, I don't suppose they compared the bhp and tourqe graph's? I think that would have been interesting.

 

If you mean why they are not crossing at 5252, look more closely, they are actually correct

[PumaRacing 1]

G-Force use Dyno Dynamics rollers, which are very accurate for wheel figures (although no more than any RR should be I suppose!)

 

The write-up on their website is different to the last time I looked, much less old-looking but basically the same words. I suspect the guy referred to that has left is Chris? G-Force are a useless bunch of clueless fools these days - couldn't even get anywhere near accurate flywheel figures despite us telling the formula off the Puma website, and seemed to just stick in their usual 4WD loss percentage

 

Chris seems to be the only guy who knew anything (and was honest about what he knew and didn't know, which I still think is great, no bulls*it etc), and he now has his "own" place in Silverstone and is very well known and respected - GF's loss!

 

I didn't ask and wasn't told the name of the person who had left but was told that he had set up on his own somewhere and that there had been some sort of falling out between the parties. I really wasn't interested in the details. Maybe I'd better keep an eye out though in case he puts my work back up again on his new website.

[PumaRacing 1]

http://bbs.scoobynet.co.uk/showthread.php?t=469358

 

I came across this last night. Down towards the bottom of the page it shows two back to back tests between Powerstation's rollers and WRC's. The peak power is similar in both cases but the torque figures (i.e. low rpm power figures) are 10% or more lower on the Powerstation graphs. The reason is actually quite simple to fathom.

 

The Powerstation rollers read very low at the wheels (about 20% low from my calculations) as we already know but that should just be a matter of calibration and somewhere inside the machine there will be a dial to do just that. To compensate for this it appears they have somehow altered the coastdown calibration but how this is done on that make of dyno I have no idea.

 

But here's the rub. Coastdown losses rise exponentially. If they are set to somehow fudge the low wheel figure back up to a somewhat correct flywheel figure at high rpm then the flywheel figure at all lower rpms will remain too low because not enough losses will be being added back to the wheel figure. What this does is distort the shape of the power curve, change the rpm at which peak power is shown and give an incorrect torque figure, in fact a curve which reads too low at all rpms other than peak.

 

The only way to correct this is raise the wheel figure across the board by the appropriate percentage and reduce the coastdown losses to more normal ones. You just can't correct a linear error which will be the same percentage at all rpms on the wheel bhp curve with an exponential loss correction which rises sharply with roller speed.

 

The conclusion I'm afraid is that this probably has nothing to do with the make of roller which ought to operate in the same way as any other and be capable of accurate wheel bhp figures if correctly calibrated. It's purely down to whoever set this machine up getting the wheel bhp calibration wrong and fudging it with the loss calibration.

 

I can't believe that the people who run this machine are unaware of its calibration problems. The fact that they haven't done anything about it is what I don't understand. However I'm fanatical about exact scientific accuracy in all things I do and most people don't give a toss I suppose. Their choice though and mine to avoid using anyone who can't be bothered to get things right.

[Bonzai 2]

i have a question which is slightly OT sorry

 

i had my previous car rolling roaded, a corsa 1.5 td, but the computer didn't knock up a torque figure for it, only power. Why would this be?

 

norm power - 87 bhp

engine output - 85 bhp

wheel output - 66.5 bhp

drag output 18.5 bhp

[PumaRacing 1]

Different systems print out different things. If you want torque figures work them out. The equations are on my website.

[Bonzai 2]

i dont think i can, the power graph shows power against wheel speed, not rpm. and i was led to believe petrol cars got torque readouts at the RR station, just not the dirty denzils

[base-1 17]

Dave - this is WRC Tech's website, the bit about their rollers http://www.wrc-tech.co.uk/services-rolling-road.htm

[PumaRacing 1]

Dave - this is WRC Tech's website, the bit about their rollers http://www.wrc-tech.co.uk/services-rolling-road.htm

 

I see they've copied my bloody work as well

 

http://www.wrc-tech.co.uk/services-rolling-road-faq.htm

 

Lifted straight off my website from the 3rd power and torque article just like G-Force did. WTF is wrong with companies this size who ought to be aware of copyright issues completely ignoring them as well as all the copyright notices on my website.

[Jonmurgie 2]

I've had issues with the WRC-Tech lot stealing web content before... I ended up having to send a rather snotty letter to the MD and it finally got removed... you lot think PS are bad, they aren't ANYTHING compared to the WRC-Tech bunch of cowboys!

[base-1 17]

WRC-Tech aren't cowboys, they do cars not websites that's all!

[Jonmurgie 2]

Well I'm sure we all have stories to tell....