Accelerometers

[ajb]

Hey guys,
I'm doing some preliminary research on DAQ systems. Right now I'm interested in the DL1 Data Logger, which comes with either a 2g or a 6g accelerometer. Our car should corner at about 1.5g's. However, I wouldn't be surprised if their were spikes above 2g's. Can anyone with DAQ experience tell me if they regularly see spikes in the data above 2g's? Thanks alot!

[TMichaels]

Hi ajb,
are you only talking about lateral acceleration or are you also thinking about measuring longitudinal and vertical acceleration?

Regards,

Tobias

[ajb]

Tobias, I am interested in measuring only lateral and longitudinal.

Andrew

[jrickert]

It really depends on the bandwidth of the accelerometer and how you mount them. The acelerometers I have dealt with have bandwidth into the hundreds of hz. They pick up chassis vibration just as well as car motion. As an example you might be pulling 1g but be picking up 1.5g of vibration. This would max out a 2g unit. If you have suffecient analog to digital resolution to resolve the 6g unit accurately I would go with that. Accelerometers usually need some form of low pass filtering either electrical or software. Make sure to mount it where there is the least potential for rattling. We use a 3g and have not noticed any adverse effects but I think this may be on the edge unless you mount it to a pillow or find one that is naturally low bandwidth.

[exFSAE]

+1 to mounting with some sort of inherent vibration damping.

Also really is good to mount the thing as close to the CG as possible, and you won't get fake spikes of lateral force on turn-in.

[Rex]

Big question is how do you plan to use the data. The DL1 analysis program is really slick, and I'm pretty sure it does quite a bit of data cleanup and interpolation for you behind the scenes - I've never had a problem with the 2G unit I've used, but then again I'm not looking at the raw data, just the overall results through the analysis software. I don't know the details, but I'm sure one of the guys from Race Technology could tell you about how their software filters noise/vibration/etc before it gives you the analysis results.

So if you're just using the DL1's analysis suite to look at your g-g circle and to time laps, I probably wouldn't worry about the 2G setup. Only if you plan to use the raw data do the above comments really become critical. But then of course, if you just want raw accelerometer data without the cool analysis software, you could get that much cheaper than a DL1!

[TMichaels]

I don't know the details, but I'm sure one of the guys from Race Technology could tell you about how their software filters noise/vibration/etc before it gives you the analysis results.

@Rex:
You should know how your data is preprocessed otherwise you know nothing about the restrictions of your analysis.

Regards,

Tobias

[Rex]

@Rex:
You should know how your data is preprocessed otherwise you know nothing about the restrictions of your analysis.

Generally agreed - I think we're probably on the same page here. Maybe it's not coming through clearly in my prior post, but my point is that the ONLY way you don't care how the data is processed is for the most basic functionality of the unit - i.e. like I use it for. My application is used on an older FSAE car just for driver training, so as long as the data processing is consistent (i.e. gives repeatable relative results between different runs/drivers) it suits our needs perfectly. Certainly for this application I don't need to know the details of how it works, so long as it works consistently (which it does).

But if you're going to be trying to use the DL1 data to tune the car (or really anything else requiring a high degree of precision) I'm sure you would want to get in touch with the manufacturer and understand what is going on behind the scenes before acquired data is presented to the user. I can imagine a team getting roasted in design review because they present pretty graphs from slick DAQ software without fully understanding how those results were generated!

I'd just hate to see the guy spend the extra money on a 6g accelerometer in a DL1 if all he's doing is timing laps or comparing drivers, when I know the 2g works fine for this purpose.

[ajb]

Thanks everyone for really informative responses! We would like to do some extensive analysis, but I would not be surprised if this year we don't do much more than analyze the basic signals (rpm, throttle position, lat and long accel., speed, etc). I only say that because I would assume that testing (something we've never had time to do) will take much longer than expected, just like the design and manufacturing phases always do.

While it's possible that we won't be able to do much more than analyze the basic signals, I would view purchasing the DL1 as an investment for future years. What do you guys think? Thanks again for your input.