DAQ

[Rami]

our team planning to use the EDGE ecu from PE....as we see from the datasheet that the ecu has built in data logging up to 1 MB... i wonder if i add sensors like (wheel speed sensor,accelerometer, vehicle speed,another sensor for brake system performance)....would this lead us to add an additional DAQ unit to store this data...and send the info in a telemetry system for analyzing only...or by using the 1 MB memory could be enough??...our team is first year...so don't be harsh on me http://fsae.com/groupee_common/emoticons/icon_smile.gif

[NickFavazzo]

I'm not the electrical guy on the team however I know a little, you may run out of inputs with the ecu, also (we use MoTec) if you want to use maths and micros the loggers can have that capability. 1Mb turns out to be more than you expect, an enduros worth at least [depending on logging rate and how many channels etc]

[Rami]

hi nick,
thank you for your information,but we have 8 aux analogue inputs and 6 digital inputs so i think that we had a lot of space for placing many sensors...we run the PE ecu because its very cheap in comparison with motec so we avoided in the first place using motec due to the cost....beside the Ethernet communication that is included in the EDGE is really helpful in the telemetry http://fsae.com/groupee_common/emoticons/icon_smile.gif

[NickFavazzo]

Keep in mind what sensors you use and what you want from the DAQ, you have 4 tyres for wheel speeds, steering angle, G-sensors, engine sensors, extra engine sensors(that you can use to help you tune and then remove them), gear position. heaps of sensors.

Define what you want from your DAQ then find out how many inputs you have left, when you have done so, look at how often you want the data saved, I imagine the manual would give you an approximation on how much you can store too.

[Rami]

as for now we did not get the ECU so i think we don`t have the capability to have the manual....here is what i think about the hole situation and please please help me if i am wrong or not.....Ethernet connection in pe ecu is very good as we are going to make a telemetry...the data will be saved in the ecu and the unit can send the data to dashboard through a can bus and we use the Ethernet port in the ecu to connect it to a PCB and antenna that can send the data wireless to a receiver... the receiver will be a laptop but our main problem is the software interface...we think about the ECU software but i think that's not true...we still for now can`t get any information how the capacity of the data...as we still in design phase and we don't have any hardware to test http://fsae.com/groupee_common/emoticons/icon_frown.gif...so please tell me if iam wrong or not

[Ashmeet]

@rami, I'm also researching about the same ECU. The Edge System Datasheet at this page http:/ /www.pe-ltd.com/ downloads.htm says 1MB of datalogging, which translates to around Approx 5hrs of data with 10 channels at 20 Hz. Now its up to you to figure out how much you want from it.

Ashmeet Singh Sidhu
Captain
IIT Roorkee Motorsports

[B Lewis @ PE Engine Management]

Hi Guys,

Ashmeet, your correct about amount of data. There is some book keeping in the data also but 10 channels at 20 Hz is about 310 min of data.

Please feel free to ask questions directly.

[Cardriverx]

We use the edge ECU (Pitt).

The internal memory will be fine for a runs worth of data, maybe a couple hours. If you do not mind taking the data off the ECU that frequent, then it is fine.

Also keep in mind that if you want GPS, at the moment you will need to buy a separate DAQ unit anyway.

[Rami]

Originally posted by Cardriverx:
We use the edge ECU (Pitt).

The internal memory will be fine for a runs worth of data, maybe a couple hours. If you do not mind taking the data off the ECU that frequent, then it is fine.

Also keep in mind that if you want GPS, at the moment you will need to buy a separate DAQ unit anyway.

The team is the first team of our school so i would like to ask you if i can take the data out form the ECU, we can save it,right? because we build the formula student branch and upcoming teams in the next years can start from our data that we saved.

also i would like to know if you mean by GPS that you send the data on board to a receiver that can make the team evaluate the performance of the vehicle during the event (telemetry),then why i should buy external DAQ???

[Chapo]

I think that this fits in here, it was mentioned in the previous few posts, what is peoples opinions on the use of GPS in FSAE?

To my knowledge most GPS used for racing has a positional accuracy of approx 5m and a update rate of 5-10hz (based off the MoTec system and general commercially avaliable units). This seems to be too large for plotting a decent FSAE track map due to the sheer volume of close packed corners.
Of course if you were to use D-GPS thats a different story, but a system like that tends to cost a lot more and requires more setup. Or the GPS could be use to correct INS drift, but again that is more of an advanced use and costs a lot to set up.

Where I am going with this: Have people been using GPS with their DAQ systems? Has it been compared to a G and wheel speed sensor trace?
How well does it work? Lastly, is it useful or did it just make a good intergration project that now makes people say "ooooo GPS"?

Chappy

[Rex Chan]

Our 2010 car ran a Race Technology DL1 system (5Hz GPS and internal accelerometers). We did not have wheel speed sensors. Track maps were made using the DL1 Analysis program, which used GPS and accelerometer data. Track maps were sometimes very good, and sometimes bad (succesive laps overlaid were not the same shape), but you could tell slaloms.

This year, the 2011 car has run so far with 5HZ GPS (MoTeC GPS-G1), and wheel speed sensors, but no acclerometers (logging done in the M400). Thus, the track map is purely off GPS, and shows pretty similar behaviour to the DL1: sometimes the map is great, other times there is a shift; but you can still tell slaloms, even without correction from accelerometers. However, the addition of wheel speeds has shown that GPS speed sucks - it lags actual wheel speed, doesn't always get it right, and is just plain wrong most of the time. Maybe its something we're doing wrong, but after seeing the wheel speed traces vs GPS, I would not GPS speed to tell you how fast you're going at any pint in time (it's just too unreliable). BTW: we're using hall effects triggered off steel trigger plates bolted to the hubs.

Oops: now that I went to get a screenshot in i2 of wheel speeds vs GPS speed, the data looks very good. It seems as though some of our data has very good GPS to hall effect, and others do not:

GPS vs hall effect wheel speed: GOOD (http://www.facebook.com/photo.php?fbid=10150358234308036&set=a.10150358234268036.355263.559588035&type=3)
GPS vs hall effect wheel speed: BAD (http://www.facebook.com/photo.php?fbid=10150358234358036&set=a.10150358234268036.355263.559588035&type=3)

[Rex Chan]

One point I forgot: we use GPS to make track maps, so that we can tell where stuff happened on track, so even a crappy map will still be ok, as long as we can tell where the car is.

And we don't use the speed of the car for much, so the inaccuracy of the GPS speed doesn't matter. However, I'm an engine guy, so my use of data may be very different compared to those using it for driver performance stuff.

[Cardriverx]

Originally posted by Rami:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Cardriverx:
We use the edge ECU (Pitt).

The internal memory will be fine for a runs worth of data, maybe a couple hours. If you do not mind taking the data off the ECU that frequent, then it is fine.

Also keep in mind that if you want GPS, at the moment you will need to buy a separate DAQ unit anyway.

The team is the first team of our school so i would like to ask you if i can take the data out form the ECU, we can save it,right? because we build the formula student branch and upcoming teams in the next years can start from our data that we saved.

also i would like to know if you mean by GPS that you send the data on board to a receiver that can make the team evaluate the performance of the vehicle during the event (telemetry),then why i should buy external DAQ??? </div></BLOCKQUOTE>



You can take the data out of the ecu. All you do is hook it to the computer, open the software, and download the data from the ECU to your computer.


By GPS I mean Global Positioning System, not live telemetry. I was just saying that the PE ECU does not have GPS at the moment.

[Cardriverx]

Originally posted by Chapo:
I think that this fits in here, it was mentioned in the previous few posts, what is peoples opinions on the use of GPS in FSAE?

To my knowledge most GPS used for racing has a positional accuracy of approx 5m and a update rate of 5-10hz (based off the MoTec system and general commercially avaliable units). This seems to be too large for plotting a decent FSAE track map due to the sheer volume of close packed corners.
Of course if you were to use D-GPS thats a different story, but a system like that tends to cost a lot more and requires more setup. Or the GPS could be use to correct INS drift, but again that is more of an advanced use and costs a lot to set up.

Where I am going with this: Have people been using GPS with their DAQ systems? Has it been compared to a G and wheel speed sensor trace?
How well does it work? Lastly, is it useful or did it just make a good intergration project that now makes people say "ooooo GPS"?

Chappy



We just bought a DL1 MK3 with the 20hz GPS upgrade (no changes to actual hardware, but it uses the raw data from the GPS to calculate position on the computer at a rate of 20hz, it is supposed to be more accurate).


We hope that with the open (to the sky) skidpad that we use for testing and with the built in accelerometers, that we will be able to get good enough data for accurate driver training. If we can, I think it would be a great asset to have.

[whiltebeitel]

However, the addition of wheel speeds has shown that GPS speed sucks - it lags actual wheel speed, doesn't always get it right, and is just plain wrong most of the time. Maybe its something we're doing wrong, but after seeing the wheel speed traces vs GPS, I would not GPS speed to tell you how fast you're going at any pint in time (it's just too unreliable). BTW: we're using hall effects triggered off steel trigger plates bolted to the hubs.

Oops: now that I went to get a screenshot in i2 of wheel speeds vs GPS speed, the data looks very good. It seems as though some of our data has very good GPS to hall effect, and others do not:

GPS vs hall effect wheel speed: GOOD (http://www.facebook.com/photo.php?fbid=10150358234308036&set=a.10150358234268036.355263.559588035&type=3)
GPS vs hall effect wheel speed: BAD (http://www.facebook.com/photo.php?fbid=10150358234358036&set=a.10150358234268036.355263.559588035&type=3)

Could the error be coming from the changing loaded radius of the tire? V=Rloaded*RPM

[Rami]

Originally posted by Cardriverx:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Rami:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Cardriverx:
We use the edge ECU (Pitt).

The internal memory will be fine for a runs worth of data, maybe a couple hours. If you do not mind taking the data off the ECU that frequent, then it is fine.

Also keep in mind that if you want GPS, at the moment you will need to buy a separate DAQ unit anyway.

The team is the first team of our school so i would like to ask you if i can take the data out form the ECU, we can save it,right? because we build the formula student branch and upcoming teams in the next years can start from our data that we saved.

also i would like to know if you mean by GPS that you send the data on board to a receiver that can make the team evaluate the performance of the vehicle during the event (telemetry),then why i should buy external DAQ??? </div></BLOCKQUOTE>



You can take the data out of the ecu. All you do is hook it to the computer, open the software, and download the data from the ECU to your computer.


By GPS I mean Global Positioning System, not live telemetry. I was just saying that the PE ECU does not have GPS at the moment. </div></BLOCKQUOTE>

I still don`t know why do every one buy a GPS system where we can use a wireless Ethernet communication that can be build easily using EDGE or any other engine management system.

[Zac]

Originally posted by Rex Chan:
Our 2010 car ran a Race Technology DL1 system (5Hz GPS and internal accelerometers). We did not have wheel speed sensors. Track maps were made using the DL1 Analysis program, which used GPS and accelerometer data. Track maps were sometimes very good, and sometimes bad (succesive laps overlaid were not the same shape), but you could tell slaloms.

This year, the 2011 car has run so far with 5HZ GPS (MoTeC GPS-G1), and wheel speed sensors, but no acclerometers (logging done in the M400). Thus, the track map is purely off GPS, and shows pretty similar behaviour to the DL1: sometimes the map is great, other times there is a shift; but you can still tell slaloms, even without correction from accelerometers. However, the addition of wheel speeds has shown that GPS speed sucks - it lags actual wheel speed, doesn't always get it right, and is just plain wrong most of the time. Maybe its something we're doing wrong, but after seeing the wheel speed traces vs GPS, I would not GPS speed to tell you how fast you're going at any pint in time (it's just too unreliable). BTW: we're using hall effects triggered off steel trigger plates bolted to the hubs.

Oops: now that I went to get a screenshot in i2 of wheel speeds vs GPS speed, the data looks very good. It seems as though some of our data has very good GPS to hall effect, and others do not:

GPS vs hall effect wheel speed: GOOD (http://www.facebook.com/photo.php?fbid=10150358234308036&set=a.10150358234268036.355263.559588035&type=3)
GPS vs hall effect wheel speed: BAD (http://www.facebook.com/photo.php?fbid=10150358234358036&set=a.10150358234268036.355263.559588035&type=3)

The quality of your GPS velocity is dependent on logging rate and the number of satellites that you're picking up. You likely need to log at a rate faster than 5 Hz.

[Chris @ Motec]

Most GPS units accept multi-path signals because they wish to lock on quickly and/or under difficult situations. This creates very poor positional data, though speed and heading are usually still good. The MoTeC GPS-G1 / Garmin 5 Hz unit will accept multi-path signals. It is differential or DGPS from SBAS/WAAS corrections, not from a base station. 5 Hz is not enough for speed, but plenty for positional data. Yes there is a 0.25 second offset in the signal so you will have to \use the GPS correction functionality in i2 with a manual offset of 250msec.

We have replaced our Garmin 5Hz unit ($300) with a cheaper / newer model which provides ever so slightly better positional data, nicer speed data and all at a cost of $180. This newer model updates at 10 Hz which should help the speed traces to look better.

The MoTeC GPS BL which is either 10 or 20 Hz does not accept multi-path signals. It's accuracy can be estimated from satellite count, where a count of 9-10 satellites will be in practice 6" accuracy. More than enough for FSAE cars. When the satellite count is below 6 then accuracy is horrible with drift as you've seen in other units.

The cheaper GPS units which accept multi-path signals mean that you can not use satellite count as an estimation of accuracy because those signals create error! Most high end systems do not use base stations anymore. They use SBAS/WAAS and futher improvements from subscription based services in the L1/L2 frequencies such as RT-20 corrections. All of our units are raw GPS calculations, they are not estimated with an accelerometer.

[Rex Chan]

MoTeC recommend you log the GPS-G1 (5Hz) at 20Hz. Due to space limitations in the M400, we are currently logging at 10Hz.

[Chapo]

Good to hear from a Motec rep on this topic, I have seen what you guys can do with your track maps through I2 (pro) via the FSAE course's you guys run (worth the money if I say so my self).

With the speed thing rex, how does the unit you use calculate speed? I know my GPS watch (probably not the same level of chip as you ould use in a car) calculates speed off a 5-10 second average. I would assume that the chips used for this purpose would have a much shorter delta T. Though it could still be a posiblility.

I know that a masters project on our car is looking at using a D-GPS (base station) corrected INS system, but to my knowledge there has been a delay in physical testing (for work reasons/equipment failure) If anyone else is using a system like this it would be good to hear what they say about the accuracy.

[Test Driver]

VBox 100Hz GPS, priced accordingly.