It seems like I should be able to calculate my RWHP from a nice 3rd gear datalogit run, is this true and if so how can I do it?

 

This has been batted around a few times. I made a post about a month ago where I did it with injector duty, but I was informed that it wasn't a very accurate measure.

-Scott

 

There is definitely a way to do this as proven by the guy who has offered to make the Data Log Labs systems available for PFCs as he has for other ECUs (Miatas etc) but so far there has been only a limited interest from PFC users.

I am still trying to get a working math model based on acceleration (based on time, distance and speed) which is then manipulated with a constant based on weight.

But as I said numerous times, "if I could have done math, I would have been a doctor instead of a lawyer!!"

 

Well I have a GForcePro, I guess I could just use that :-)

 

The problem with accelerometers has also been discussed here. G-Force and Pocket Dyno are excellent tools for lots of things, but accurately estimating RWHP is not one of them. Because of individual differences in each run, tire spin, bog, etc. you can wind up with wildly different HP estimates for the same car on the same day at the same test spot. Using an accelerometer to determine whether a mod or adjustment gives a better or worse HP result is just not going to work.

 

jeff48, I spent a couple hours one day fooling with this. Amazing what I've forgotten since college....Unfortunately I just had a half throttle run to work with, but I swear this should work reasonably. You have time, RPM, and speed, so from that data, torque and horsepower should be able to be derived. I got some graphs, but I have no idea of the accuracy. They seemed reasonable. This technique would not account for rolling and drag losses, but should still be valuable for evaluating mods. FWIW, I used a lot of info from this website....
http://www.newcovenant.com/speedcraf...horsepower.htm
I'd be interested to hear if anyone can confirm any validity of datalogit=>horsepower conversions using this or a similar technique.

 

Tailhappy

Wow looks like I have a steep learning curve to deal with. I will make a couple of 2000-8100 RPM runs in 2nd gear and send them to you also. (I know that will cut the RWHP estimate down by percentage points, but I have too many tickets to risk a 3rd or 4th gear WOT run).
Thanks for picking up the work on this concept.

Jeff

 

That'd be cool. Or maybe I could just post the spreadsheet of what I have so far. I'm suspicious I've still screwed something up with it....

Last time I had trouble getting the file by our firewall though

 

We'll see if it works this time....

This was just a lazy acceleration from a log that I had. I'm confident there's still mistakes in here, but at least it might be a starting point. Unfortunately I ran out of time that day and haven't had a chance to look at it since.

 

Taking a look at the spreadsheet now. I am adding a tire size calculator on the face of the sheet. Then, using Skotx's Datalogit clipping utility I will put some numbers through the spreadsheet and let you know what happens.

 

Wow cool, you guys are great. I'll grab some data from my run and post it in a bit.

 

there appears to be a bug with both the website and your spreadsheet...regardless of tire size the horsepower does not change....need to factor in torque change due to tire size...also I compared these numbers (with the tire size factor/error included) to a dyno log....results....log 295hp vs. 335hp calculated

 

Cool, yeah I was sure there still had to be problems with it since I didn't get a chance to check over it again. I wondered about the whole tire size thing too, but the website seemed to know what they were talking about so I just went with it. I still say you should be able to use common physics equations to do the calculations. Such as a = dv/dt, and F=ma. I just kept screwing up my units when I tried to go that route.....

 

after revisiting the standard formulas or is it formulae...I believe their calculations are correct...but the calculated results don't correlate closely to 2 dyno logs run on separate dynos (neither the datalog files or the logs from the dyno.)..also doesn't reflect 40-80 mile pavement runs... approx 10% errors

 

Anyone else have some dyno logs to compare to? If we could just apply a 10% correction factor or something along those lines, it still might be a worthwhile spreadsheet for rough-estimates.

 

Here is a log, I haven't had time to prune it though.
http://nopro.com/~jedon/Jedon_Log_20030620_1234.zip

 

The math in Tailhappy's spreadsheet looks good to me. Changing tire size will not affect the calculated HP, because it increases torque by the same percentage that it decreases RPM. Calculated power should stay the same unless either weight or acceleration changes.

 

 

OK guys.

I have run a number of tests with the data I have and somehow I keep getting different results (between 175 and 210 HP). I am attaching a text file with two (non-top-end) runs on it. See what you guys make of it. My test weight is 3100 (me and a full tank of gas).

 

Ok, I thought I’d give this horsepower calculation another shot. I used the logs from the dyno shop. I averaged a 10 mph range around peak horsepower (also a flat part of the curve). Then I converted the acceleration for that range into g’s, plugged it into the equations and came within a few horsepower of what was measured by the dyno.(provided my car weighs 3000 lbs; touring model, owners manual states 3210). I also checked another log from a different shop and came up with similar results. However when I used Datalogs of the dyno runs the calculated horsepower was off by quite a bit from the measured value.
I’ve noticed that the Datalog sampling varies depending on how much you’re monitoring, for example if you have a couple of watches, (I use AFR, and knock) on as well the map watch you get samples around .2 s, whereas if you activate just the monitor, it samples closer to .1s. I believe this variation contributes significantly to the error in calculated horsepower. In addition, the Datalogit s/w varies from one version to the next, if you use 1.07, sampling can occur in less than .05s.
So is there any benefit to any of this, who knows? What I desire is the ability to calculate horsepower using Datalogs exclusively, to that end after my last dyno session, I logged several 40-80mph 3rd gear runs, I calculated the acceleration from 40-80mph plugged it into the horsepower formula and using 110 mph the calculated figure was within a few horse of the dyno. The use of 110mph was somewhat experimental, really a fudge factor or correction factor, although coincidentally 80*1.391 is about 111mph, where 1.391 is the 3rd gear ratio. Perhaps some others, who know what their horsepower is, can perform this little experiment to either discredit or to add credence. Please remember to use version 1.07 and do not activate any additional watches.

 

Sounds pretty good. 3210 is the GVWR, so you should be a lot lighter.

Right, because the Dynojet doesn't provide the same load as the car does on the road, so the datalog has different acceleration times. If that doesn't make sense, think about what would happen if you added 500lbs to your car. The Dynojet would never know, but you'd sure feel it on the road.

I suspect that 40-80 is too wide a range to calculate a single number accurately, though you could generate a nice curve across it. The 1.391 correction was no coincidence; it matches your 3rd-gear RPM to the 4th-gear run on the dyno.

Good work.

 

Thanks for the comments Eggie. I'd need to clarify that my statement:

"However when I used Datalogs of the dyno runs the calculated horsepower was off by quite a bit from the measured value." was referring to logs I had taken with the Datalogit while on the dyno.

Hence my speculation, that the sampling by the Datalogit is causing the error....since I believe the Datalogit logs and the dyno operator's logs should have the same acceleration times...but they don't. Later, I will verify via the horsepower calculator what a 0.1s error relates to in horsepower error.

I don't know how accurate 40-80 mph times can be in predicting horsepower due to the road factors you mentioned...I was hoping that perhaps we could use the horsepower formulae to get within 10 hp...but we need others to datalog. Maybe just taking a one second interval (around peak horse) instead of 40-80 would lead to more accurate results. The problem I had with this method was using the correct mph for the calculations, the correct speed (mph) may not be in the range you're recording. Maybe it will be 1.391 times the avg. speed ???

 

Probably the best way to do this is 4th gear, 1:1 ratio in the driveline, but with my peak HP @ nearly 8,000rpm, that's pretty damn fast in 4th... ~140ish

But by stipulating a 3rd gear pull, we could also work in an air resistance factor since we would know the speed of the car (from the logs, and the coef of drag is constant and published), oh, AND the top of 3rd isn't go-directly-to-jail-don't-pass-go-don't-collect-$200 speeds

K

 

The dyno software simply calculates HP/TQ based on the acceleration of the roller.

If you knew the weight of the dyno roller (which you can get from the dyno guys software), you would plug that in as your car weight and get nearly as accurate numbers on the dyno as the dyno software does. This would be a good way to tweak your calcs, but like I mentioned above, no air speed drag on the dyno like you get on the street which is what we are really working towards.

As a side note, my Blitz PowerMeter has spit out very accurate numbers. Unlike the g-force based accelerameter devices (g-tech/pro), the PowerMeter taps onto your ECU RPM signal and calculates HP based on RPM delta/time in 3rd and 4th gear, so the crazy gtech 1/4 mile pulls are not needed, just reset the unit, put the car in 3rd or 4th, and wind it out. Powermeter then displays a "graph" of your HP curve and peak HP, unfortunately I haven't found a way to extract HP@RPM from it yet.

K

 

The G-Tech Pro DOES read RPM. It does this through the power connection. You calibrate it a couple of times and it does a good job.