SS, 2 is theory, but ever look at a sine wave? If you sample twice the frequency of it you could actually end up hitting the zero crossings and think you have no signal what so ever. In the real world 5x the filter cutoff frequency is the minimum acceptable to avoid "folding" effects of your data. 10x is more of an accepted value.

The only thing 2x is really good for is if you only want the power spectral density of the data and dont care about the actual amplitude of the data or signal recreation. Thats something they dont teach in any class, but you can find tons of published papers on the subject matter, hopefully one by me soon on maximum amplitude capture.

Tons of good publications can be found through www.telemetry.org

I don't see wireless anytime in the near future... for many reasons. Their web site is http://www.eagletreesystems.com

Thanks for the info Hank.

If you decide that testing that unit is too much work for you I am sure I could help you out with that :)

I love the idea of this and I am sure the release version will be sweet.

Pat - lets see - gunfire testing in Maryland. I bet you are somewhere near Aberdeen :) I am a DAQ guy myslef, but most of what I do is 0-20 kHz noise acquisition - accels and microphones for measuring automobile service brake noise.

I wonder if the logger could be configured to run at different sample rates within the 10-20 hz scan rate. For example, temperatures change very little in .1 seconds, but amp draw does. It would be interesting if you could run a sub-scan at a higher rate (50-100 Hz), but only 1 or 2 channels and run the rest at a slower rate (.1 Hz).

Just thinking out loud
SS

Not Aberdeen, thats Army, I work for the Navy about 2 hours south of Aberdeen.

Pat,

My brother is at Aberdeen with the Army so I assumed. Sounds like you do some interesting work.

SS

Small world eh, I graduated college with several people that work at Aberdeen. It is an interesting job, never the same thing twice however the way us "support" people are treated leaves alot to be desired.

billpa

You had mentioned averaging data into 1/10th second samples to get out the higher speed stuff. Has any thought been put into using a presample filter for the current sensor? I think a good filter would help with obtaining good usable data for this. Not sure on how small it could be made but a 6 pole Butterworth is fairly simple and does the job quite nicely.

Hi Pat,

Tell me more about what the additional filtering would provide you. We currently already do the averaging in firmware to get the average current (and voltage) over the sampling period. The results we get on both current and voltage (and hence AH and Watts) closely matches what traditional instruments like the Whattmeter record.

Thanks!

Bill

What a presample filter can do is can get rid of higher frequency signals (its a low pass filter) that you can't properly recreate anyway due to your sample rate. Lets say the highest frequnecy of interest would be 100 Hz, I would design a filter so that the signal would be 99% there at this 100 Hz, for a 6 pole butterworth filter the fc would be 1.4 times your sample rate or in this case 140 Hz. Then I would sample atleast 5x fc to accurately recreate the signal.

These filters are often refered to as anti aliasing filters, they really help clean out noise that otherwise wouldnt do anything except affect what you are trying to measure in a bad way.

The hard part about DAQ isnt getting information, its knowing what information that you are not getting. Presample filtering helps in this respect.

Personally I would get rid of the averaging feature and take data at whatever rate you can and put a filter in there, that way you wouldn't have to worry about it and with an electronic filter such as this there is alot less error in what is actually measured.

It would probably add a bit to cost but would result in a much better unit, one that I would be more apt to trust the data from.

Dang!! Just think, last month all I was interested in was getting a good stop watch!!!!! LOL

I haven't seen any erronious data yet. Don't forget guys, with the extra sample rate, you're going to get overwhelmed with data.

Just as an example, for a track like Cleveland...The running line is 330 feet or so and the lap times were 10-12 seconds. with the sample rate at 10 HZ, that's 100 to 120 data points at between 3 and 3.5 feet average (5 or so on the straight)(3000 points plus in 12th). With the exception of that "thing" going onto the straight, you could capture the entire layout easily and draw it from the data.

Sample any faster, and you'll lose the end of the run. You end up missing how much the motor and batteries drop off and that's very important. Even with more memory, the data might not fit into Excel easily.

I'm currently cutting out laps at the beginning, middle and end of the run.

Don't forget us oval guys... 3 scond laps...

Or worse yet... a drag racer...

All I need for many purposes, is a few good laps worth of data... If I need more laps, and there's a option to slow the collection rate down and I can get a whole 4 minutes, I'm good.

A 220 ft track at 3 seconds and 10 Hz is just over 7 feet apart on the track and 30 datapoints per lap. I can see where it would be nice for oval racing to have a little faster sample rate but there is plenty of data in a 4 minute acquisition to pull out what you need.

The one thing that would be nice would be to have a marker in the data for the start of each lap (or at least close). It's pretty interesting to see how the two ends of the track are different.

I can "look" really fast with one of these. Looks are important to us slow guys!

Since im mostly a drag race only guy will this thing work with say 10 or 17 cells without blowing up the voltage and amp readings.... i remember the victor iq tracker it didnt start recording till after 3 sec into the run no good for drag racing we need something that will record right at the pull of the trigger .. let me know want to put it in my record setting cars to find out whats going on inside......albie

The unit draws its voltage from the receiver. Since you most likely use a 4 or 5 cell reciever and don't feed 12 to 20 volts into your receiver you shouldn't have any problems. The unit starts recording as soon as you turn it on, so it will catch before the beginning of the run and even after the run is over.

Now, I am not sure how high of a pack voltage it can read, Bill should be able to answer that. If nothing else you should be able to put a voltage divider into the pack voltage circuit to chop the voltage the unit reads in half.

From what I understand the amp sensor is rated 100 amps but will handle 120 amp spikes.

Seems to be some off-topic action slipping into the dynothread.
sorry.

http://www.hobbytalk.com/bbs1/showthread.php?p=628996

Hi,

Re max current/voltage, he system will measure voltage up to 50 volts and current up to 100A. Nothing bad would happen with higher current spikes.

Thx,

Bill Parry
Eagle Tree Systems

Actually, when I posted on the other thread, I was more interested in how people would be using the data with their dyno's.

Also, I'm working on a new spreadsheet for the new logger that I will post sometime after the new year. I was thinking about automating things a little but I'm pretty much too lazy to write the macro's. You guys can get an idea of what's going on from the old data that I posted (it used the Flight Data Recorder).

At full throttle you are measuring motor and pack voltage (there's basically no resistance in the circuit). At part throttle, you are measuring pack voltage only but if you're at part throttle you have more than enough power anyways.

One last thing, the "cockpit" playback in the program is very useful for finding your place on the track and pausing the run to record times for a spreadsheet. Also, you can use it to quickly compare two runs by switching back and forth between files.

Fred, That's fine, just wasn't sure if those in the Oval area had seen this thread. Anyways, as you can see, I've open a whole new area for the system. Carry on... ;)

One last thing, the "cockpit" playback in the program is very useful for finding your place on the track and pausing the run to record times for a spreadsheet. Also, you can use it to quickly compare two runs by switching back and forth between files.


Fred,
Do I understand correctly, that the excell sheets as well as the Cockpit will be available.
Not one OR the other.

Man that souns like the best of 2 worlds then.


But comparing 1 graph to oneother by pasting #1 over #2 in the same graph, will not be possible?!

thnx

I haven't seen any erronious data yet.


Not sampling fast enough will not cause erronious data, but it can make you think the shape of the data is different than what it really is. Remember data acq is taking discrete points not the entire signal. Only knowing these discrete points forces you to interpolate what the data looks like in between them.

Imagine if your system only took data before you started and after you quit, it would read 0, but that would in effect be correct. To find out whats in between you have to sample faster and faster until you find a "good point".

In my field we generally know what the maximum frequency of interest is because we know the natural frequency of a vibration mode and we will often capture 2nd, 3rd, 4th harmonics etc.

I have taken data as fast as 10,000 samples per second and still needed faster to see what happens "inbetween".

Wait I got the perfect example, ever watch an old cowboy and indian movie? Ever see the wagon wheels look like they are going backwards? This is because movies are actually discrete images taken at a time interval. It turns out that the capture rate of the camera wasnt fast enough to show a true signal of the wagon wheel and the data was aliased. I know this is a difficult concept to grasp for someone not in the field, it took me a while to believe it when I first read about it but that example should help.

PizzaDude, Graphing will not be a feature when the software is first released. It will be added in a future release. Graphing is now being done by us using Excel. So by using Excel it is possible to import or cut and paste two or more data sets if that is what you want to do.

I wish I knew Excel better so I could make a "snappy" spreadsheet to distribute but about all I know how to do is import the data into Excel and create a graph.

The data is saved to your computer in a space delimited format so it can be manipulated by the user however they want.

Sounds great ( I keep repeating myself)

Well even the excell as it was shown shows a lot clear info!


Questions keep popping to mind.I hope you don't mind me asking!

How long is the recording time of this data?
Is the recording time expendable?
Can it be used troughout a days testing and at home be read-out by the PC?
Wil it be represented as for example 5 different runs?
Which can be saved with each individual setup?

As some of the pics shown as well as reading this thread, should be pretty close to release.
Will there be a first batch (proto or early production) first be available for members of Hobbytalk?

Well, that's it for now!

Pizza

The Excel sheet was somethign that I was working on using the old logger. The data from the new logger is pretty much the same. The only downside to using Excel is that you have to pick out the data that you want to graph from well over 1000 lines. It's really not that hard if you have a template, but it is a little time consuming.

Pat, I understand the problems with sample rates. Work with them every day. Two things. First, You can get a pretty good idea of what's going on from the data with the current sample rate. Also, Bill is saying that there is averageing over the .1 second sample time. This will act like a filter filter for what is probably a very dirty signal. You will still see the initial spike and have a good idea of it's shape.

Just an FYI. It takes a couple of tenths of a second to pull the trigger and for the speedo to react. I'm guessing from the data that this ramp up reduces the current spike somewhat.

As for the logger questions, It is not currently expandable but it has plenty of memory to record several runs and then take it home. The data remains stored without power. You have to keep track of what you ran on each run. I can get over 8 minutes if I turn some of the extra stuff off.

As Fred has stated, I have also found the information given very useful even if it is at 10hz. I guess we could go to the extreme of saying we would have to gather info at least at 180Khz (20K RPM * 3 armature segments * triple samples per segment) to get the kind of detail some are asking about, but that's not likely to happen. Anything less then that is filtering

I think most will find the info very useful. As shown in the graph segment here > http://www.hobbytalk.com/bbs1/attachment.php?attachmentid=6709 and as Fred brought up, I would think most never thought that it takes a couple tenths to pull the trigger fully... that is something no dyno duplicates.

will there be a way to get the logger upgraded to show more amps.. in drag racing 100amps isnt seen to estimated 30 ft after the launch so there is alot of data gonna be lost from the begining of the run till the motor winds and the amp drops under 100. also say i want 2 rpm sensors can this be done as i would like to see how far the car carrys the front wheels .. i think this is great and in a drag car can provide alot of useful data especially as the amp draw can be used with the dyno... albie

Im not saying it has to be that fast. Hank your math is off by the way you forgot to divide by 60 to convert minutes to seconds.

A couple tenths to pull the trigger and for the speedo to react, man could I get a holeshot on you.

Oh yeah... divide by 60. Even so, at 3K hz just the electronics to count and store that amount of info for one channel would cost more then this whole unit costs.

Wait till you get one... you might be surprised at what the data shows you... ;)

If you want to get an idea of how quickly you can pull the trigger, get out the old stop watch and see how quickly you can start and stop it with your left finger. Thing is, the trigger throw is much farther than the button on the stop watch. We saw the same results logging at 100 Hz on our Pace Sci logger. I've yet to find relavent data that's missed at 10Hz and I have used loggers that sample much higher.

You can view a preview of my "public" review of the unit. You can check it out at http://www.hobbytalk.com/datarecorder/

All,

Excel macros is a good portion of what I do for work. We have a variety of functions and subroutines that do most any kind of data import and graphing you would want. If you list your requirements with a hunk of sample data I would be happy to do the macros for free.

Post the data, the requirements, and give me a couple of days and I will get it done for you.

Let me know
SS

The new software that just came out tonight has graphing built-in. So, for the most part, using Excel will no longer be needed.

Although some may find using Excel may be easier.... YMMV.

All,

Excel macros is a good portion of what I do for work. We have a variety of functions and subroutines that do most any kind of data import and graphing you would want. If you list your requirements with a hunk of sample data I would be happy to do the macros for free.

Post the data, the requirements, and give me a couple of days and I will get it done for you.

Let me know
SS


Squirrel,
I am interested for sure.
A nice idea could be if you can combine the CDR-data (CarDataRecorder) and laptimes combined.
Laptimes fi being imported from Maylaps.com.
Or imported from the Orion lapcounter software....

Hank,
Will Eagletree be devolloping an integrated lapcounter in the near future?

PizzaDude, It may be possible to make a lap counter. I am looking into it now.

Hank this is the best thing since the motor dyno... i hooked the rpm sensor today in my drag tc3 which i use to test my drag motors and breakin my drag batteries... tc3 hit top speed of 61 mph and 65000 rpm.......... i double checked all the parameters and the motor in the car is my #2 drag motor with my #3 6 cell drag 2400s ..............cant wait for the g force meter

what would it take to add a "Flash Media" or memory stick type card to the data recorder for additional memory?

Albie, Let us know how it works out in your drag car.

SW, It would be pretty much impossible to add memory. Why would you need more then 6+ minutes of recording?

Ok, I think I saw it posted but is there a mode which allows me to plug the car into the computer while on a stand and watch in real time??? This is my deciding factor....

hank i recieved the new sensor today and all is working great.. only thing is the amps meter says high of 17.86 but the car was drawing 117.86 which the graphing shows..

vtl.. yes the recorder can go live and run on a stand and you can see things in real time.. albie

Cool... Been lurking since I somehow found the link to the recorder and to here... I toasted a P2K2 this weekend and would have to seen why, I'm gathering this would have told me why if I had it last week...

I have just got mine :thumbsup: now its time to start playing LOL

Thanks Hank

Scott J

Yes, there is a real time mode available.

If someone would post a small chunk of Car Data Recorder data and some Lap Counter data I will make a macro to inport and graph it.

Let me know or PM me to take this discussion to email
SS

Can someone with the car data recorder give me some info:

I'm interested in knowing if you take a stadium truck like a XXX-T and a stock motor and 3000 or 3300 batts and nail the throttle from a stand-still and/or out of a turn how low does the battery voltage go? how many amps is the motor drawing?

How about wide open on the straight...just before breaking for the turn, what is the battery voltage and how many amps is the motor drawing?

Thanks for your help.

Jerrit

I've seen somewhere around 40/45 amps and a drop in voltage for about 1.6 volts on startup. Amp draw at the end of a long stright is about 14 amps with a voltage drop of about .6 This is on a track with fairly low traction.

albie, Could you report the problem to billpa@comcast.net I haven't drawn that high an amp rate so I never noticed it before... most likely just needs a tweak to the software.

hank iv talked to bill about this already.. he is a great person to deal with.. i took the car out yesterday and made a few runs the car pulled 118.2 amps the voltage was 8.3 until the amps hit then the car battery dropped to 4.78 volts this was for about 2-3 tenths of a sec...then the amps leveled out to about 70 and the voltage went up to 6.67. this little unit is very interesting and just seeing how the voltage drops is amazing.. also when the amps hit and the voltage drops the reciever voltage drops also on the cyclone .. im gonna put either a stock motor or a 19 t in the car and play with that... dont like putting wear on my drag motors... albie

You mean you aren't running one of our receiver packs in your drag car???? ;)