Data Acquisition Tech Discussion

Al Seim

Well-known member
There - I created a real data acq tech discussion thread.

Greg - have you used that dual channel Bosch sensor? It looks intriguing. It does appear to have an unusual taper seal seat near the nose, but looks as though it might work with a regular copper washer.

I agree with your sensor routing philosophy. If the ECU already has or can use the data, wire the sensor to the ECU and send the data to the display or logger via CAN or RS232. If not, wire the sensor directly to the display or logger.

Table switching is EXTREMELY useful if racing with a sanctioning body that uses chassis dyno results for classing... :twisted:
 
Al Seim":3q9gp4as said:
There - I created a real data acq tech discussion thread.

<3

Greg - have you used that dual channel Bosch sensor? It looks intriguing. It does appear to have an unusual taper seal seat near the nose, but looks as though it might work with a regular copper washer.
I've installed it, and I've fired up the engine with it to verify data. But I have not really "used it" yet.

And I did The Bad Thing that everyone hates: I re-tapped the back of the block deeper to 1/8-27 NPT and Teflon-taped the M10x1 sensor to screw in. I do know all the warnings, and I do know I done bad, but I done bad before and gotten away with it.

I'm actually more worried about direct-mounting the sensor to the block versus screwing in the M10 to a 1/8 NPT. It's something I really never, ever do as I'm always worried about fatigue failure. But the damned thing only weighs just over an ounce so I'm less worried than if it was, for example, one of those two-pole VDO pressure sensors. I will be inspecting it often.

Megasquirt-->Dash2Pro is the first time I've even used CAN (well, except in an OE vehicle.) I did it more out of laziness than anything, allowing me to not have to wire up tach, TPS, and coolant temp to the dash display/logger. It was actually fairly straightforward, really just two wires between the Dash and the MS. Looks like the MS will also send AFR, volts, fuel injector pulsewidth and (most importantly) boost to the dash (I'm guessing it's just MAP relative to ambient), so that may be useful to log as well. However, as noted, I'll be using a cheap Adroid tablet as a MS tuning/logging device, connecting via Bluetooth. If I wanted to get fancy I could mount the tablet to the dash somehow and see pretty much any parameters that the MS sees, input or output.

And, related to our prior emails conversations, I tapped an output from the MS to drop the radiator fan relay based on coolant temp, so that was useful.

I'm using direct-burial Ethernet Cat6 to route the data (I have the most part of a 1000-foot spool left over from wiring the house to the detached garage). One of the twisted pair is being used for CAN H/L, one pair is for sensor 5V ref and ground from the dash, one pair for the analog inputs from pressure and temperature, and one of the remaining wires for the rad fan relay solenoid. The only real downside to using Cat6 is that it's solid wire, so I need to watch out for fatigue failures. But nothing going through there is mission-critical, and I'm providing the monkey a manual override switch for the rad fan.

All sounds simple but it took a SHADload of time poring over manuals and documentation to figure it out. I wish I could put purple LED illumination on it or something to show it off. But in the end I'll be happy if it all works as designed... - GA
 
It's extremely unpredictable but we've seen sporadic issues with the internal diaphragm rupturing on pressure sensors - probably from vibration, although pressure pulsation could also be a culprit. It seems very application-specific, with location on engine AND sensor orientation probably both issues.

We had a low volume car manufacturer customer who was seeing a LOT of this type of failure until they remote mounted the sensor and problems stopped. OTOH I've had a pressure sender (same type that had failures for this customer) on my VW HP engine for years.

I'll pretend that I did not read your description of thread sealing on the Bosch sensor. OTOH some of the VW oil filter mounts have a 1/8BSPT sensor port right next to an M10x1 sensor port, so probably half the cars out there have been graunched at least once. And I'll confess that "back in the day" before I knew what BSPT was I'm sure I put more than one NPT fitting into a BSPT hole.
 
Al Seim":2okjhzsi said:
And I'll confess that "back in the day" before I knew what BSPT was I'm sure I put more than one NPT fitting into a BSPT hole.
I did the same in my 240SX.... in about 2007 when I put an oil pressure gauge in it. sucker has been there >13 years and about 60k miles, 35k of which have been track miles. still ain't fell out yet!

Another nice option for CAN is the easy logging of most of your engine and chassis data right along with the actual driving data. you can easily correlate things like engine temp-->power loss-->laptime increase through a race when you have all of the data right there and can see the acceleration difference over the length of a race.

The body/traction control/ABS module in my FRS logs all four wheel speeds, "average speed", brake pressure, yaw, etc as part of the factory nanny systems. traction control gets turned off, but the computers are still running. All that comes across to my dash via the CAN connection, so I can watch things like wheelspin on corner exit, the ABS engaging and seeing each wheelspeed wiggle in a braking zone. cool stuff. of course I'm not doing any of that while on track, but after action reports tell the tale of my poor driving. lots of cool stuff to look at in the modern cars, should you care to look and digest what it's telling you. of course, the kicker is what to DO with all of that data once you have it.
 
Yep - I've used the wrong fitting in the past as well. It can be done.

So the cool Bosch "double sensor" is a nice and clean approach, but one of the big reasons I'm using the sensors I use now is that I can get all of them at any auto parts store close to any track. GM and Ford temp, pressure, MAF sensors are plenty affordable, and most ECUs are able to work with them. Log the data in the ECU, port it to the dash by CAN and log it there too if you like.

One thing I'm adding this year is a 2 axis accelerometer + yaw sensor that feeds the ECU. Whittle was able to decode the CAN protocol that Hyundai-Kia uses and we can get the sensors from the boneyard for like $25. Also looking at how to use a string pot for a steering angle sensor. It's helpful to see what the chassis is doing when looking at the ECU logs. AND with Mega Log Viewer you can actually open both the data logger and ecu data and merge them to look at the whole picture. You can find pretty cool insights by making 3d maps of different variable combinations.

I'd like to do more, but I know that I just don't have the bandwidth/time at the track, what with all the distractions of prepping the car mechanically and then driving it. Hopefully one of my kids gets into the data side of it and we can take it a bit further. Until then, I'm putting more on it than I can reasonably use at the track unless Andy is there with me to focus on that stuff.
.
.
.
I still can't get over how accessible this stuff has become. I mean the products that Al and his competitors are offering lowly club racers like us used to be only in the realm of the Penske's, Gnassi's etc. of the world. It is really cool.
 
chois":11tb7x8i said:
I still can't get over how accessible this stuff has become.
Totally agree. I first started getting into data in the early-00s with the Race Technology DL-1 Mk1 that I thought was the cat's meow*, and it has quickly grown to the point of data saturation...and I find myself sometimes between sessions reverting back to just checking lap times.

I was reading through your plans and agreeing with how cool it was but thinking "you ain't got no time for that!" but it sounds like you're grooming a data acq team so there's that... ;)

*I hid my old Mk2 DL1 in the interior of the historics 914 and it auto-starts and stops. I run oil pressure and temp through it, as well as AFR (dual Dellortos) and CHT. This way I can pull out the data card and check for any issues that may be occurring...plus, Al sold me a NOS legacy DASH3 really cheap that I have peeking through a hole in the dash panel to flash at me for anything out of spec...
 
chois":38z2ayyd said:
...I'd like to do more, but I know that I just don't have the bandwidth/time at the track, what with all the distractions of prepping the car mechanically and then driving it. Hopefully one of my kids gets into the data side of it and we can take it a bit further.

That's how it works between my Dad and I :D . I learned all the tech/data bits and do the laptop tuning/data analysis and he's the engineer.

Put the mega log viewer in front of your oldest and see if he picks it up, he might just start digging into it and really enjoy it. I grew up on modern technology and got a good grasp on this stuff when I was a teenager - and kids nowadays are lightyears ahead of where I was when I was their age.
 
My first data logger was a VHS camera pointed at the dash. The dash had a WOT light and a brake light in addition to the usual gauges.

After running data for a few drivers in recent years, I have found the most important points IMHO are brake pressure and timing, and the most important data point.. Steering wheel angle vs location on track.
A certain open wheel series puts up all of the driver data to compare. My driver was losing a tiny bit of exit speed. VS the # 1 and 2 drivers. Looked at the data , flat down about the same point on track with very close to the same speed at apex. The difference was my driver had to put in more wheel past the apex due to having slower wheel speed at at the turn in points.
Wasnt in the data , only on video .
 
Protech Racing":hf1233i5 said:
My first data logger was a VHS camera pointed at the dash. The dash had a WOT light and a brake light in addition to the usual gauges.

After running data for a few drivers in recent years, I have found the most important points IMHO are brake pressure and timing, and the most important data point.. Steering wheel angle vs location on track.
A certain open wheel series puts up all of the driver data to compare. My driver was losing a tiny bit of exit speed. VS the # 1 and 2 drivers. Looked at the data , flat down about the same point on track with very close to the same speed at apex. The difference was my driver had to put in more wheel past the apex due to having slower wheel speed at at the turn in points.
Wasnt in the data , only on video .

Just FYI, the AIM system I have in my car logs all that (steering angle, individual wheel speeds, throttle pos, brake pressure, position, etc). benefits of modern ECUs with a high speed CAN bus that watch all that stuff already. I just plug the AIM into the CAN line and tell it which addresses to watch. voila.
 
Brett W":1aat6p6f said:
Is there not an OE steering angle sensor we can appropriate?
Most cars since about 2003 that have some sort of "stability control" will have a steering angle sensor. That said it's usually a steering wheel angle sensor, not one on each front corner of the car. having one on each spindle might be handy if you're fiddling with ackerman changes and want to see actual angles on each side.
 
Modern steering wheel sensors are integrated into the column. It's certainly possibly to adapt one, but a simple string potentiometer on the rack/pitman arm would be easier.
 
The single largest discrepancy between the fast and pretty fast drivers has been turn in wheel speed . This assumes very good overall precision on track .
 
Protech Racing":2e28g30y said:
The single largest discrepancy between the fast and pretty fast drivers has been turn in wheel speed . This assumes very good overall precision on track .

Just curious, does that mean slow, med, fast, or jerk the wheel to pitch the car?
(I noticed the faster I got with my car, the slower I found myself turning into a corner since I was driving smoother and pitching it less. With FWD, that technique could be different.)
 
The wheel speed that results in the largest arc . Fastest wheel speed at the turn in .
Small difference from FWD to RWD and power out put. More power goes best with extending the straights, i.e. more wheel before the apex and opening up post apex.
 
Protech Racing":3jprdzz7 said:
The wheel speed that results in the largest arc . Fastest wheel speed at the turn in .
Small difference from FWD to RWD and power out put. More power goes best with extending the straights, i.e. more wheel before the apex and opening up post apex.
ahh gotcha. first time I read your point about wheel speed, I thought you were referring to the rate of steering wheel input at corner entry, not the actual vehicle's wheel speeds. (since we had been discussing steering wheel angle sensors just a couple lines above).
 
Back
Top