RomRaider

There is no need in better sampling.

You just need to use LeastSquares 5th-9th order appoximation for RPM vs time instead of "tricky filtering" for 3000RPM-FuelCutRPM region.

After that TQ\HP may be got very consistent.

That's what I'm doing with the XYTrendline Class in RR and the order is controlled by the user on the screen.
Both sample interval and RPM (or Speed) are interpolated as both are noisy. Even sample time varies about 10% while logging. As you state, high sample rate data didn't make much difference. I tried samples at a rate of 65msecs up to 200msecs and the quicker rate only allowed you to go a couple higher in the order, like 11 to 13 before the trace gets all ripply again.

The lower the order the smoother the curves are. But for curves follow the reality you may need to limit lower and upper RPMs when LSM rises the residue for low order approximation.

I am just a purist as you could see before..

It will definitely give better results, even if it is only slightly better, but that's not really what I was getting at.

The other benefit to bumping the baudrate would be the ability to log other parameters while running the dyno without losing a decent sampling rate. Whenever I take dyno plots from full logs, the results aren't quite as good as the bare minimum for dyno. This will allow us to overlay critical parameters on the graph from the same pull with no spreadsheet. As much as I like the spreadsheet, it is tedious while roadtuning.

If I understand correctly, more samples will also allow a bigger RPM range of the interpolated curve, but I don't think that's a big issue.

The problem also gets worse with a high power car. A 4th gear pull in a 500whp car won't have nearly as many samples as a stock car.

Looks good.

Any chance of putting in a toggle for displaying kW and Nm?

Also is there anyway to save it?

One future feature request would be to add a small digital display panel below or above the dyno graph to display upto 4 selected parameters - or hardset them so it at least monitors FBKC, FLKC and IAM.

Only reason is so you at least have an idea if you get any knock and can quickly stop the WOT pull

Leslie


Edit: If you use the same "Side Tab" type setup as the dashboard does, you could incorporate the parameter list, then you could select what parameters you want to view above the dyno panel. Here is a quick mock up I did:

Yes I do plan on making the parameters user defined, i.e.: metric and imperial. It shouldn't be difficult to include SI units as well.

Save? If you mean the graphical display, just right click on it and press Save as.
If you are referring to the raw data, then all you need to do is log to a file the same parameters as the Dyno records. That would be Engine Speed (RPM) for a manual and Engine Speed and Vehicle Speed for an automatic transmission. The Dyno also watches Throttle position such that it only captures WOT data.
So on the logger just set these parameters and press the Record Data on the Dyno tab and when ready, if you're set up to use the Defrost button to record a log, press that too.

Great idea! (looks like you did it already )
I had thought of recording other values. Initially, I was going to do AFR and Boost pressure...but, the more parameters you add the slower the sample rate. For the current items which are just the RPM and Throttle (and defrost switch), the sample rate is 97msec (about 10 samples/sec). The best I've seen is 65msec (about 15 samples/sec) with just RPM (and defrost switch).
Adding four more parameters drops you to about 6.5 samples/sec.
The situation would be much better if we could increase the baud rate. Going to 9600 baud would allow you to record three more items (with 1 byte return values) at 15 samples/sec.

One enhancement to the current Logging code would be to drop the Defrost switch query of the ECU when the user de-selects that option in the Logger menu. Currently the switch is always checked regardless the user's preference.

I'll put this on the list for the future, k?

Leslie, why not just use the graph or gauge display? No real need to watch the dyno in realtime.

I think time would be better spent implementing parameter overlays on the dyno graph, and improving the alert/warning system.

Wow, awesome work! Can't wait to see this integrated into an official beta release once you are happy with it.

Let me know how you go...

Any way to run existing logs?

At the moment, no. But it is something I'd like to add.

New test version attached to post topic39-34.html
Now at rd-v3 which does imperial and metric calculations.
Select units before recording or change units after recoding and click "interpolate".
I may have to switch to dual Y-axis to increase the grid area used by the power curve.

New test version attached to post topic39-34.html
Now at rd-v4 which has dual axis. This allows the metric curves to fill the graph for better resolution.

How are the colours? I'm interested to know if anyone with red/green colour-blindness has any problems seeing the graphs.

Maybe have an option to use balck and white color scheme. Also, make power line solid and torque line dashed/dotted so it is easier to differentiate between the two (regardless of color scheme).

Also, another feature request. Can you have it so that when you press the "record data" button, it doesn't start recording data until the first 100% throttle is recorded (you could indicate actual WOT recording with by a red dot - virtual led)?

This will ensure it only captures WOT data and provides the most accurate log for interpolation.

Leslie