[CZD8T]

How do people here feel about using datalogs to create a virtual dyno? I've seen it done in the past and it's claimed to be within spitting range of an actual dyno run as long as you can control the environment. (flat road, foot down, not bumpy or curvy, etc.)

I have a OBDlink MX+ bluetooth adapter and got it set up to datalog TPS, RPM, MAP, and AFR to see what it would look like in the V dyno and I think it did pretty well. It seems to fall roughly in line with most expected stock runs. I have a catless front pipe and a catback which should skew the results much.

I might try and go this route to see if its predictable and comparable to others who have put cars on actual dynos with mods such as headers, tuning etc. This is with the dynojet option selected at correction factor 1.09 if that means anything to anyone.

 

[TxDivemaster]

Trying to understand your stats as a noob...

Weight difference was that your estimate l based on you in it or actual?

The rpms:hp - would it have been any different @ redline?

That's 207 to the wheels with your mods, has me curious now what the actual is to the wheels bone stock.

 

[CZD8T]

So it takes into account a good number of variables, a lot of which would normally be calculated by a dyno operator anyways I think. I looked up the official weight for a limited manual, and when I process the log there is a spot to add driver weight, so I weighed myself and filled that box in. You add the gear ratios and final drive, then tell it what gear you did the run in.

It's entirely possible that I may have made more peak, I was pretty sure I had hit redline but it may have been an error on my part/datalogging. I may try for some different roads to compare. I did do 2 runs and they were virtually the same so it did seem repeatable.

I don't understand the correction factor myself, I'd have to look into that some more.

 

[Soldmyfiatforthis]

Curious about this, what did you need hardware wise to do this? I've done an actual dyno so I could compare with baselines I've already set

 

[CZD8T]

Curious about this, what did you need hardware wise to do this? I've done an actual dyno so I could compare with baselines I've already set

I have a Bluetooth scan dongle, that can also do datalogging among other things. I bought it back when I had a ford GT350 to enable some gauge cluster features.

It's the OBDlink MX+ scantool with Bluetooth, and it has a free companion app I downloaded on my android phone with some datalogging features. It's not super intuitive and clunky to use but it does actually work. I had to modify the file to remove the header it puts in automatically so the virtual dyno program can use it.

If you google virtual dyno it'll most likely be the first hit you get, it's a pretty popular program. Also I think that technically any sort of tool that can read the sae PIDs and calculated PIDs should work if you can get an app or program to datalog. I do think it needs to have a pretty high polling rate if it's over Bluetooth.

 

[TRU-BOOST]

HP seems about spot on. The peak torque looks a bit unrealistic on the high side. But honestly its about as accurate as most of the "professional" dyno numbers you see on the internet.

 

[formerMiata]

I think it can give you an idea of the power a car has, but probably isn’t very useful for testing mods or tuning because it’s very difficult to control the variables at play. You will get different results as the wind shifts, etc.

 

[CZD8T]

I tried a different road today, and it really highlights how sensitive this ends up being. Both were flat, and straight. The biggest difference I can think of is that the red run I had my windows down, and the temperature was significantly higher. The blue run was windows up, lower temperature. Seems like the best way to do it would be to find a straight abandoned road so you could have actual repeatable results.

$200 dyno pulls are looking vaguely more tempting these days.

 

[TRU-BOOST]

That TQ curve in the red definitely isnt right. The blue one at least resembles what everyone sees on a stock header. I agree that if you stick to the same section of road and use similar conditions the results should be close enough to see if certain mods have any impact. You have to get rid of as many variables as possible though. Using only sensor data it will never be spot on, but good enough for an idea of whats going on.