OT: rolling road plot experts needed...updated inital post

[Gridgway]

So if you had an actual rolling road power run plotted against what the engine should be doing, what would you conclude?

 

here or this might be better here

 

The headline in bhp terms is peak 206 bhp versus the 210 it should be. I think the torque tells a very different story.

 

ETA: I have now uploaded a scan of the actual plot. It looks a little different to my excel rendering as I have cut off at 7k rpm.

 

Also: what effect does the ambient temperature that you enter into the RR? I assume it's a correcting factor. How much difference does it make? I presume only a few percentage points?

 

Graham

 

Edited by - gridgway on 2 Feb 2009 20:12:35

[Paul Deslandes]

Graham, it would be interesting to have a bit more info about the engine and the tests. I would guess the engines a 2 litre Duratech or a Vx with a peak torque in the 180s.

 

The only reason the calculated peak power is anything like what's quoted is because torque gets somewhere close at near peak rpm. Below that its so far off as to make me wonder whether the measurements are right. The measured torque curve looks a bit smooth for a real engine but that's maybe because I'm used to looking at my Vx curves which are a lot more wavey!

 

Could be all manner of things wrong, strangled induction, tiny exhaust.....

Tell us more....

 

 

[Paul Deslandes]

looking at it again it almost looks as though the plot could be flywheel torque in blue and rear wheel bhp in red!

 

 

[Billyboy]

Hope it isn't a new R400 No I see its not... 😬

 

Edited by - Billyboy on 2 Feb 2009 10:17:37

[Gridgway]

Thanks all. It's a 2.7l engine (not in a caterham!) which is currently running lean (no idea what it was doing at the time of the RR plot, but possibly the same). It's running mechanical fuel injection and throttle bodies. It looks like the known fuelling problem (ie it's know from exhaust analysis to be weak) is causing the torque hole.

 

Or could there be other things that might be wrong? Cams and timing, valves, rings?

 

The plot is smooth as I just mocked it up in excel. The real one is quite jittery. I'll scan it and post it later.

 

Graham

[dannylt]

Why do you think there's a torque hole? Are you comparing it with something you expect that we can't see? If the engine is that close to quoted peak output, I wouldn't worry, rolling road errors in general can be far greater than this. Correction for temperature is quite significant, air pressure (in the UK) less so.

 

Most rolling roads will plot lambda on the same chart. And if it ever looks dangerously lean they would abort the run anyhow.

 

I assume it's a an old Porsche engine or something, surely it's best to ask a specialist?

[Ian Macquarie]

It’s a while since I looked seriously at my rolling road plots. However I seem to recall that when I was at the rolling road, the engine speed was not recorded by a sensor on the engine but was calculated by putting in a figure for mph per 1000 rpm (or something similar which had the same effect.)

The rolling road can measure power accurately based on what it has to counteract (absorb) however it then calculates engine torque based on what speed it “thinks” the engine is running at.

If it “knows” the power is a certain number and “thinks” the engine is at a low rpm it reports a high torque. On the other hand if it “knows” the same power and “thinks” the rpm is high, it will report a low torque.

I notice that in your graph of expected torque and measured torque the measured torque peak is at 5900 and the expected torque peak is at 5100 rpm, If the “engine speed factor” mentioned above was adjusted to bring the torque peak down to 5100 rpm then the torque would rise to pretty much your expected figure.

 

Ian

 

 

[Gridgway]

Thanks for the thoughts. To respond:

 

Danny:

The reason I think there is a torque hole is by comparing the actual with the expected in the excel plot. This also matches observed behavior - lean running, running like sh!te in the mid range and generally less than impressive "go".The ambient temp on the RR was set to 32 degrees C, when I expect it was actually more like 15. What effect would that have? Cause the RR to over-read (be generous) or under-read (be conservative)? Lambda was not being measured in any way as far as I know. It is having specialist care, I have the plot so thought I would try to match observed behaviour with the plot and maybe have fun fitting causes to the observations in the plot (effect).

 

Ian:

It could well be that the set up of the RR caused the peak torque to occur too high in the rev range and the magnitude of the torque to be too low as you say. It'll get an RR run as soon as we have finished our remedial activities to see where we have got.

 

Graham

[Gridgway]

Ian, just a re-track...the RR measures torque, so if it gets the engine revs wrong, the torque curve will move left/right but will not change in magnitude. So in this case where the RR shows the peak torque at revs which are actually too high for the engine, if we move it down to 5100, we get 162bhp! Now that feels much more like what it does!

 

Graham

[Gridgway]

And going for the record number of posts on one's own thread...

 

I have effectively moved the plot of the rr run down by 1000 revs (as that was easiest) and posted the pic in the album. Now to me that makes a lot more sense. A curve which kind of follows the characteristics of the engine but just "not up to it" is a better result.

 

The cause could then be something like the wrong fuelling across the range.

 

Of course, it could just be straws desperately being clutched at whilst I wait for the result of the expert's ministrations!

 

Graham

[Ian Macquarie]

Graham,

/bottle of wine warning

The rolling road probably does measure torque – at its rollers. Unless it has an accurate value for the gearing ratio between engine and wheels (including road wheel to roller ratio) then it can not accurately calculate engine torque unless it measures engine speed directly using a sensor. However if it measures torque at its rollers and "knows" the speed of its rollers (which is likely) then it CAN calculate power accurately.

// bottle of wine warning

 

Ian

 

 

[Gridgway]

but...(my bottle is pretty much finished!)...

 

It measures torque in ftlb up the rev (of the RR machine) range as well as transmission losses in ftlb down the way. It adds the two together and produces a graph of torque up the side against a linear scale along the bottom. That produces the shape and magnitude of the graph. The operator just adjusts the scale along the x-axis to suit the car.

 

Graham

PS I can't wait to get it back on a RR observed by me!

[Ian Macquarie]

but...(Actually I’m on the second! – and in danger of getting my posting record changed)...

 

”It measures torque in ftlb up the rev (of the RR machine) range as well as transmission losses in ftlb down the way. It adds the two together and produces a graph of torque up the side against a linear scale along the bottom. That produces the shape”

I agree up to this point.

“and magnitude”

I disagree here

“of the graph. The operator just adjusts the scale along the x-axis to suit the car.”

When the operator adjusts the scale along the x-axis, he should also adjust the scale up the y axis, as the power is the product of torque and (rotational) speed.

 

Ian

 

 

 

[Gridgway]

yep, ok, (I've moved on the the Genepi now...)

 

The plot is torque only, i have steered clear of the power plot as being meaningless!

 

Graham