It's been talked about for awhile. Some have even taken kart systems and used them on their RC cars. Soon you will no longer have to resort to those extremes.

Eagle Tree Systems is in the final stages of releasing a data acquisition system for RC cars. I got a test unit to help them debug it and while I haven't actually got it installed and running. I can give you a few preview pictures.

The test unit (which doesn't even have a case yet) included the following:
Y cables for the steering and throttle servo (for position recording)
RPM sensor
Current sensor
Voltage sensor
Temperature sensor

Weight (no case) including all of the above sensors is 1.5 oz.
Size (no case) 2.75" x 1.25" (about the size of a normal ESC)

Not sure what other options will be available, but it looks like there is room for 8 more channels of data!

I plan on installing it today and giving it a test run as soon as the rain stops... never fails, just when you need decent weather, you get rain!!!

We plan on offering this system for sale and will have a dedicated forum for questions, tips, etc.

Below are a few pics to wet your appitite. I'll post more info as I get further in my testing.

Will it be out by Christmas? I need something else to put on my list,

I guess it depends on how the testing goes.

I'd rather see info like side G loads and slip angle info in helping me get my chassis dialed in. It would be great to see the difference between sets of springs/tires/bodies/wings etc and then tie it to the RPMs

How much does it weigh? Do you download to a laptop?
Tom

First, the unit I have a first production prototype, it doesn't even have a case yet. The software is still beta, so what I now have is subject to change.

Not yet sure of the full options that will be available but looking at the raw data I see places for X & Y G-Force. The data is downloaded to a computer. The present software allows playback in a "real time" display. While that may not be optimal display for closely inspecting the info, the software they have for airplanes does allow for graphing. You can also import the raw data to Excel.

Weight (specs in the 1st message) is 1.5 oz as shown.

Since it is raining here I haven't had a chance to run the car outside. I did run it in my livingroom but the 8 x 6 foot of space available didn't really tell me much except that it worked.

I am attaching a graph I did in Excel for some of the data on this test run. The RPM & Amp Draw use the left scale, the Pack Voltage, Receiver Voltage and Motor Temp. uses the right scale. This was captured at 8 samples per second so you can see I have about 25 seconds of data here.

Another note on the data that can be collected. It will record any "glitches" you get in either your steering servo or your ESC... good for finding out if a receiver voltage drop is causing control problems.

can you give us a ball park figure on what the street price is likely to be on something like this?

I'm drooling already.... This is a gadget guy RC racers dream... and that's exactly what I am...

The current street price for the aircraft system is $150.00. I suspect the car system will be priced very close to this. Since I don't know for sure we can guess that it will be close to the aircraft system since the functions are pretty much the same. They normally include a couple sensors, my guess is these could be the 2 servo Y cables and the RPM sensor.

Also... another goodie, they are working on making it a lap counter also. An external IR beam will trigger an onboard IR sensor to record when your car passes the beam. I suspect it could be used with multiple beams to get split times on different parts of the track.

The nice thing is, is that the system is modulor. You can buy the main unit now and buy other sensors as you need them... they just plug in and you configure the system to collect whatever data you want. Want RPM, Volts and Amps now... great, just get those sensors. Want G-Force or Lap Counting later... no problem, just get those sensors and plug them in... reconfigure the unit to collect that data and you are set to go.

Sounds like a fair price to me... (well less would be better... but anyway)...

IR beam trigerd lap/segment time... COOL.

For me Amp, volt, and RPM would be great. It might be really cool to have Volts at battery and Volts at motor... however this may not be all that usefull, since the voltage at the motor is plused and that may make it somewhat hard to interpet in any meaningfull way.

Put me on the list of buyers...

To take motor voltage, you could just connect the voltage sensor wire to the motor instead of the battery... I haven't tried it so I have no idea how well it works... will have to try that.

Considering the present day motor dynos (single purpose units) are priced at 500 bucks or so, $150 seems like a bargin to me... ;)

It certainly seems as though this is a “serious effort” and not just a toy. Where the motor information would be nice to have, I think I would be more interested in chassis information. I’m not sure just how they would go about that but things like shock and front-end movement would be good to know. ANYTHING that would show what is really going on when the rubber meats the road!

This is pretty interesting, I really dont see a need for it for RC cars atleast for me. My interest is in it because I do data acq for a living on military aircraft, but my stuff typically is in the 100 lb range but that is for 100's even 1000's of parameters. Hank if the company needs any technical suggestions for data acq as it could relate to RC cars direct them my way.

This is designed off of a already manufacturered and marketed unit that is built for RC cars. Considering the "guts" are pretty much the same and only the layout was changed to make it smaller, the only real R&D is for the software. For what I would consider a "low end unit" (under $200.00) it does a lot.

I'd like to see 4x the memory, higher sample rates and more sensors. But then would there be a market for something that cost twice as much?

Right now I have my car wired for RPM, voltage, current, motor temperature, steering and ESC. As it is, the wiring is a rats nest. I suppose it could be made a little neater but having many more sensors then this can end up being a real pain to wire and maintain.

Pat, I'll contact them and see if they require any assistance.

I agree with Mclin, the more we know about what is REALLY HAPPENING on the track the better off we are,this may just completly change the way we think about set ups.Not only that but think what it will do for manufactors when it comes to new designs.

The major addition of the new system is amps. They have an area on the board for acceleration so I'm assuming that it's just not in there yet.

Once you have amps and RPM, it's off to the dyno. You can also see the changes in gearing which is a must in 12th. I'll be running it a little more this weekend.

Ok, once they get the load sensors G force and all that stuff, then transmit it to a remote seat that I sit in and it gives me real time feedback so I can feel the car break loose in the corner and correct for it... :)

For those that are testing the system, could you please contact me via email. I'd like to have a private area where we can discuss the system and what is being found. Also, we can discuss how best to address some of the improvements that are needed.

Here is a picture of the case that the unit will be enclosed in. I understand the production label is slightly different then the one shown in this picture.

What are the servo connections for?

It records the position of the steering servo and throttle esc/servos.

Just an update, the latest version of the software allows real time recording of data while hooked up to the computer.

This could lead to all kinds of interesting options.... how about a full chassis dyno that records motor RPM, Amp draw, MPH and Voltage? This could make dedicated motor dynos obsolete!

Has a price been determined yet? Curious as to how much this is going to cost, software included.

Thanks

On tower hobbies website they have the airplane version for $144.95

The price is not set yet so saying what i think it would cost could give a false impression. BTW, what do you think it is worth and what, out of the above sensors, should be included in that price?

They really do seem to be doing their homework on this one. Looks great. I do have one concern though and that is the type of computer it will take to use it. The reason I say that is, most of us use a pretty cheap lap top to carry around to the race track. Maybe USB adaptors could be an "option" or at least put it in the directions as to where to get them (Radio Shack part numbers and such).

Just a thought. I for one am VERY interested in this thing.

Keep in mind guys that this is a "set up tool" and it's not legal to race with it in your car. But, that is why it is being made.............SETUP!

As for what sensors?.................All of them, don't go half way now!

Put me on your list Hank!

Didn't read one of your post above before I posted the last. ON THE TRACK, LIVE INFORMATION!! This COULD do away with dynos and achieve better information!! IF it could send back AMP, RPM and VOLTAGE, and they could figure some type of "time factor" into it, then TORQUE could be calculated also.

DANG! Put me on your list TWICE!!! and UNDERLINE it!! LOL

It should run fine on most any system that can run Windows 98SE or later. I have a 233Mhz laptop that I will be trying the software on.

For the most part the computer does nothing but download the data from the recorder and store it for further use. The software does have a "payback" mode where you can see the results just as if it was in your car, but I find this is not as useful as the graphing feature that will be added.

The "real time" part of the software is not completed yet, so I have no idea how that works yet. Non the less, I will be testing on my laptop so we can see if it works on one with that slow of a processor and will report how well it works on that.

USB to Serial port adaptors are available. I'll look up some RS numbers but I'll bring this up to them and I'm sure they can find a supplier so they can offer them too. Also, there are USB PCMCIA cards available for laptops. But a minimum of Windows 98SE will be needed for proper USB support in most cases.

With the way laptop prices are dropping it's almost worthwhile to buy a new or used one. New Dell laptops can be had for $600.00 and good used ones are available for $300 or less

I see no reason why it couldn't be used during racing. It isn't "active" in any way, it does nothing to "control" your car, it is passive in that all it does is collect data.

Note on possible new feature: At this time the fastest collection rate is 10 hz or 10 readings per second. They are going to try and boost that up to 20 hz. This would be most useful for oval races as this will cut down the total collection time to somewhere around 60 seconds but will provide detailed information for those fast 3 second lap times we see on some short oval tracks.

Yea this is defintly going on my list of must have's sign me up Hank.

10 Hz, that really means that we can only really see what is going on with any hopes of accurately recreating the signal would be at a frequency of 2Hz (Nyquist theory).

A motor dyno would have to sample at approximately 3500 Hz to be of much use.

Data is so easy to misinterpret and the data acq is so easy to take for granted, I guess thats why I have a job though.

Hank, I haven't bothered to look up the rule because it is just not that important to me LOL but if I remember it right, there may be some questions about it. Like I said though, I just don't care, all I would want it for is tuning anyway.

Does this use the cars battery for power or does it have its own power source? What is the voltage required?

Lateral G's would be very useful. Then you could compare steer input to lateral G input. This would tell you if you have a over or under steer situation.

I wanted to post this interesting throttle curve that I found when the data recorder was hooked up to my GM ESC in a pan car. The curve shows about 3 sec of an acceleration curve. The car was off the ground and the wheels stopped. I fully gunned the throttle for a short time to come up close to full speed and then let off the throttle.

The ESC has a "traction control" system turned on where it is suppose to limit the current for a short time to help eliminate wheels spin when you quickly hit full throttle. I noticed when hitting the throttle that the motor would sort of stall for a short time after I hit the throttle and then go to full speed… this was strange, I never noticed it before.

Looking at the graph we can clearly see that the RPM is actually limited for about 1.5 seconds before full power and RPM is allowed. Since it doesn't appear that the current is actually limited, the ESC must be internally limiting the amount of throttle being applied during that time.

Also notice the small amount of "regeneration" when I let off the throttle.

Opps, forgot to attach the graphic, here it is.

Hank,

Somewhere up there you asked what sensors we thought should be included in the price for $150 or so... Well... From my perspective, Voltage, Current, and RPM are essential for electric car system components.... I'd like to see all three in a package deal, even if it meant paying slightly more then the $150 price. I'd not really like to have to buy extra sensors just to get what I would consider essential for electric RC car info.

I believe that there will be some type of "base" unit with general sensors (servos, RPM) that can be used by all kinds of vehicles, then packages available. Since the needs for electric and nitro are different, it makes sense to do it that way. But I ask so we can see what people think.

Everyone's feedback is being read by the manufacturer so keep it coming.

10 Hz, that really means that we can only really see what is going on with any hopes of accurately recreating the signal would be at a frequency of 2Hz (Nyquist theory).

A motor dyno would have to sample at approximately 3500 Hz to be of much use.

Data is so easy to misinterpret and the data acq is so easy to take for granted, I guess thats why I have a job though.


Pat, Since they read the messages here they wanted to comment on your post. To quote their email:

The recorder does sample motor current/voltage at a much, much higher rate than 10Hz, but it averages the values it records and reports that average once each logging period (10 averages/second for 10hz).

I've had a couple questions via email about mounting the RPM sensor. In the attached example the sensor is mounted to the rear pod plate with servo tape and the two magnets are super-glued to the diff hub.

The recorder does sample motor current/voltage at a much, much higher rate than 10Hz, but it averages the values it records and reports that average once each logging period (10 averages/second for 10hz).

For all intensive purposes that is no different than sampling at 10 Hz. The averaging does smear out noise, but it also smears out your signal.

Hi Pat,

The purpose of the current/voltage acquisition feature is to provide logging ahmmeter/voltmeter functionality in an actual race setting, rather than just on a bench.

To get accurate results, we have to take tons of almost continuous samples over the logging period (1/10th second for example) and average these together for each log. This of course is required due to the sinusoidal nature of motor current draw. If we didn't do this, and only sampled once every 1/10th second, we'd get random numbers corresponding to whatever the current draw was at that instant.

As it is, the values logged now agree closely with units such as the Astro Whattmeter. The Whattmeter seems to average over a much longer period however, like 1-2 seconds rather than 1/10th second.

As you point out, the unit is not designed to log at the rate necessary to plot the actual sine (or other) waves of motor current draw that occur as the motor rotates thru a single revolution. I'm not precisely sure how this level of data would actually be used for tuning, but I'm interested in learning! Let me know your thoughts on this and if there are different data that we can consider logging in the future.

By the way, thanks to you all for providing all this great feedback! We're paying close attention to it and it's very helpful.

Bill Parry
Eagle Tree Systems

Hi Bill

Its great to hear from you. Just want you to know that Im not bashing your product in any way, im always amazed how small these things can be made.

I think these very short "abnomalities" can actually be very useful in tuning a motor. But that is just conjecture on how and I wont get into that.

My main reason for thinking a10Hz sampling rate isnt enough is that when coming out of a corner a motor accelerates a car to atleast 80% of its top speed in about 2-3 tenths of a second. This is pretty much what makes a difference between a slow car and a fast car. If the phenomonem happens this fast I think that this is the most important thing to capture for data analysis.

Off topic but possibly interest to you, I recently did a job where 35 strain gages needed to capture the peak strain caused by gun fire. The pulse took place over 1 millisecond. Typically it would be sampled at 10x this, but since we wanted to capture the peak within 99% of its full value I sampled 25x every millisecond. I think uping that to 35 or so would have helped but I had run out of bandwidth.

Hi Pat,

Definitely not considered bashing in any way. Great feedback!

The recorder should be able to go up to 20hz. Do you think this data rate would get you the info you need?

Thx,

Bill

It would definately help, is bandwidth or frequency response of the sensor the limiting factor? Personally I would give up the other functions for a 100Hz sample rate with a record time of 60 seconds, I would also settle for 10 bit resolution too.

I know next to nothing about data acuasition, not even how to spell it... However I'm with Pat in the sense that I really would be interested in having at least the option of seeing everything as fast as possible , even if it meant a realtively short collection period was possible... 60 seconds worth of data, in very fine detail would be very nice for some interesting things... Heck even 30 seconds worth of data would be cool, if it meant getting raw data that was not filtered and such.

I wouldn't want this to be the only mode available, but it certianly could prove to be very interesting at times.

I'll see how fast we can get this puppy to go!

Bill

Man! I'm almost ready to PAY to help test this thing LOL

I am posting attaching a screenshot as an example of what can be displayed. You can configure the display to show a number items that are recorded and show then in graphic or digital format.

When you first load in the data (as shown here), the display shows the maximum and minimum values for the complete run. For example, in this case we can see the maximum RPM, max Amp Draw, minimum Pack Voltage, etc. Not all the data is here as an incomplete data file is being used.

More example coming as I do some more testing.

Hey I really like the look of the playback software.

Still there is nothing like looking at the raw data files for info though. :)

didn't see anything to do with G loading or the rpm difference between L & R or F & R to determine wheel spin or slip.

Graphing is being added to the software but is not included yet. You can look at the saved raw data now in Excel. There is a spot in the raw data to store G Force but I suspect the G-Force sensor development is not complete yet. There is only one RPM sensor so you can not get wheelspin.

The data holders in the raw data files now include (I would guess this is still subject to change): Steering Throttle RPM Speed Rec-Bat*10 Temp1*10 Temp2*10 Amps*100 PackVolt*10 GForceX GForceY ThermoA ThermoB IsEvent EventError EventData Timestamp

[QUOTE=patcollins]10 Hz, that really means that we can only really see what is going on with any hopes of accurately recreating the signal would be at a frequency of 2Hz (Nyquist theory).


Actually the Nyquist Theory states that the sampling frequency needs to be greater than twice the maximum frequncy in order to remove any aliasing from the signal. That would mean 5 Hz resolution at best, not 2 Hz.

Hank, do they have any intentions of trying to beam the data wirelessly. Being able to watch the data in real time while the car is on the track would be very cool.

Also, do they have a website. I do DAQ for a living and would be interested in learning more about what they are doing

SS