New concept evaluation, booted from two spots already and going for the third.

[Ragewizard]

Hi, all

I am an RC airplane hobbiest and I would like to build a test engine stand to test my engines with different props.  I would also like to tune the engines to get the peak performance out of my engines if possible.  Usually to do this, I place one of my motors on a test stand with a prop and start it up and put it through its paces while listening to the sounds.  Now I would like to actually quantify my results with a system that can be connected to my laptop.  Of course I have no idea how to put together a data acquisition system, but why let that stop me.  I thought I would pose a question here to see how much of a hassle it would be.  So below I have listed all my requirements and I hope that someone can direct me to atleast giving me a direction to start picking the components that I would need to achieve my goal.

RC Glow Motor Test Stand Requirements:

1.  Thrust measurement.  The engines can typically produce upto 25 lbs of thrust with a proper prop configuration.

2.  Temperature measurement.  The engines operate at temperatures that are comparable to a typical automobile engine ( -40 F to 400 F)

3.  Tachometer measurement.  The engines can turn as much as 20000 RPMs.  I have a optical tachometer that is suppiled from the hobby shop and works qute well with up to 4 blades on a prop. 

The above are the only 3 measurements that I need to tune the motors to maximize their performance.  In addition, it would be awesome if my data acquisition system wasn't tethered to a plug from a house which would be cause for an extention cord.  Battery powered is the best. 

Of course the final requirement is the data needs to be able to be displayed in a graph on the laptop computer.

Let me add that this is probably one of the most confusing boards ever, is there really that much traffic here to warrant all of the sub titles?  I would also like to add that I'm really just looking for something to buy that isn't too complicated if possible.  I feel that I could learn all of this stuff, but I really have no interest in this field, I'm just looking for something more than what I have for around $1000.  Well maybe $1500, if it goes to around $2000, I would cough up blood but I would probably still go for it.  $2500 would be excessive in my view.

[MattBradley]

Hmm... don't know of any commercial product that will give you just what you want.  But here's my first stab at what you should do.  Others will probably add their own refinements and suggestions.

1) PC (preferably laptop)

2) DAQ card (for the PC) that can measure 3, 4, and 5

3) Load cell (for measuring torque)

4) A way to convert RPM to a voltage or bring it into a counter

5) Thermocouple

Then you need the software to do your test.  Given your budget and lack of interest in writing your own software, you would probably have to get somebody to develop it and give you the executable.  The LabVIEW development license will eat up your entire budget otherwise.

The optical tach probably won't help unless you can interface it to the computer or DAQ card somehow.

Does that help?

[Ragewizard]

Of course it helps, and thank you.  I was checking out the labview price and it seems quite heavy for my requirements.  In addition, I generally base my software learning curve on the cost of the software, and labview seems like it would be a good bit of learning, and probably way more software than I need in terms of capabilities.  It seems like things are not too hobbiest friendly.  Is software my only cost prohibitor?  I was checking out a USB-6008 that someone else suggested and it goes for only $170, and has several analog inputs for my thermo couple, and load cell.  I was planning on busting open the optical tach and highjacking the digital signal coming from the processor into a digital input or signal counter or something already on the 6008 model.  If it doesn't have any digital inputs, then I would highjack the signal coming from the sensor, or find out what kind of sensor it is and get one.  Reading through the signal inputs makes me worried however, all my sensors are giving mV/V and the DAQ thing has +-10V input.  So it looks like I would also want a amplifier of some sort to increase the voltage, and maybe smooth out the signal while it was at it. 

[pallen]

To put together an NI based system including the development software would not be possible given your budget restraints. 

There are other options however.  Tronssen Robotics has a bunch of USB based sensors and the like that might be suitable for what you need.  http://www.trossenrobotics.com 

You might have to get a little creative to integrate the force cells they have.  But they're cheap. 

These boards are divided for good reason.  There's a lot of traffic here. 

Good luck with your project. 

[Dennis_Knutson]

I've done a little RC flying myself but only with electric power. I thought I had seen low cost systems discussed in some of the hobby magazines but I could be mistaken or the available systems may not meet all your requirements. I'll have to search through the old issues I have.

[Spex]

Hi Ragewizard,

I think you do have some good options with NI tools...

First, to address your software concerns, NI has a simple, free software package called LabVIEW SignalExpress LE (the full version costs money) that gives you a quick and easy way to view and log data from our data acquisition devices.  This tool is configuration based so it doesn't require any programming knowledge and has a pretty low learning curve.  With LabVIEW SignalExpress you can create your own dashboard for your DAQ system for monitoring the thermocouple temperature, load cell thrust, and tach frequency at the same time.  You can then manually adjust the engine until you like what you see. 

Second, you need some hardware to bring the above sensor measurements into the computer.  I would recommend an NI bus powered (no extension cord!) USB DAQ card.  You mentioned the USB-6008/9 boards, and that would be a good start, but may not be good enough.  That family of boards is low cost and has analog inputs and an event counter input (up to 5 MHz) that could be used for the tach. The 6008/9 is not generally a good fit for thermocouple measurements because it doesn't have internal gain or high enough resolution to get an accurate thermocouple measurement.  Also, the 6008/9's built in counter cannot perform a hardware timed frequency measurement, so it may be difficult to deduce actual rotor speed from the tach input when using the 6008/9, especially with Signal Express.  I'd recommend upgrading to the USB bus powered M Series DAQ devices.  With M Series, the analog inputs have programable gains to get better precision from a thermocouple and load cell measurement and they have advanced onboard hardware counters so they can perform frequency and period measurements directly on the DAQ card and pass that data to your LabVIEW SignalExpress dashboard.  The lowest cost USB M Series board is the USB-6210.

I'm not an expert on sensor vendors, but I'm sure some time on Google could find some good sources for the load cell and thermocouple.

The rest will be up to you because I'm sure you need to build some test fixture where you mount the engine, etc.

Regarding all the different board sub-titles... NI sells thousands of different products and we try to offer support for all of them on these forums.  Breaking down the boards by different product lines can be confusing if you are not familiar with NI's product lines but helps experts in different fields support according to their expertise.  Sorry you were booted from a couple groups before you found a home on the LabVIEW  board 🙂  If you do decide to use an NI DAQ card with SignalExpress and have any questions as you start to implement your system, you will probably want to post directly to the separate SignalExpress board.  😛

I hope this helps,

[Ragewizard]

This 6210 sounds like it may be what I want.  From reading the spec sheet on it, it sounds like I can basically plug and play with all of the sensors that I need.  It also sounds like it is an over kill with what I need, but if it is the lowest cost unit with USB power ( no electrical plugs), then I think I can hang with it for a copmplete package under or about $1000. 

I was on the phone last night and was given another option with a USB 9219 unit that has 4 analog inputs and the person tells me that I can have a digital tach signal connected as well, but I would need to have the labview software to count the upticks.  So my question is, what is 500 bucks better about a unit that only has 4 analog inputs, and doesn't count the digital signal like the 6210 model.  There must be some advantage to the more expensive unit.

[speleato]

Hello Ragewizard,

The 9219 is a slow module, I would not use it for the digital tach as you will be sure to miss edges.  A counter would be much more appropriate.

The reason the 9219 is expensive is that it is a very flexible module.  It has built in signal conditions for RTD, thermocouples, bridge circuits, voltage, and current measurements.

Sebastian

[Ragewizard]

Thanks to everybody putting up with me as I struggle to get this.  I don't think that there are any stupid questions just stupid people, and sadly I am more towards the latter when it comes to this.

Speleato tells me that 9219 is slow, and looking at the specs it has a sample rate off 100S/s/ch.  I read this as 100 samples per second per channel right?  If so your right, and I guess I should spell out more of a sample speed to capture my prop rotating potentially at 20000 RPM and being a 4 blade prop that would be 1334 times a second.  I would assume that I would atleast need that speed to sample, but I was also reading a little deal on aliasing, which I admit I don't fully grasp yet, but it sounds like I want to sample at least twice my rate of change, which would make it around 2700 S/s/ch.  So the 6210 with 250k S/s would read as 250 thousand samples per second.  I notice that it doesn't have the channel designation behind it, so would I be correct is assuming that I then need to divide that speed by the number of channels?  If so, is it by the number of possible channels, or the number of channels being used?  Or maybe I'm totally off base with my assumptions.

Also, isn't signal conditioning a good thing that I may want?

If I have a load cell that has a 10V excitement but only a 3mV/V output, the max it is going to put out is 30mV.  I would think that if my device can handle a larger input signal, I would want to maximize that voltage and try to eliminate noise while doing it.

I noticed that the 6210 model has 4 input ranges with the smallest being -200mV to 200mV.  At 30mv, I will only be using a small percentage of the range to acquire data, would this then compromise the quality or resolution of the data that I am trying to record?

Once again thanks.

[speleato]

Hello Ragewizard,

You are correct that the 250 kS/s are shared across all the channels you are sampling.  This is not a major concern for the digital tach though, because you would be strongly encouraged to use one of the counters on the 6210 to count the edges.  According to the spec sheet the counter can operate with a clock of up to 80 Mhz, which will easily exceed your requirements.

You are also correct to point out that signal conditioning is a good thing.  In the case of the thermocouple things like cold junction compensation are included in the module and are important to getting an accurate reading.  Similarly the range of the measurement can be important depending on the resolution you need.

One suggestion then would be to look at using two separate products.  Something like the 9211 is a 24bit moduld designed for thermocouple readings and can also read voltages in a +/- 80 mV range.  It is slow, 14 S/s, but I think in your case you are primarlily interested in the steady state response.  The NI 6501 would work to measure the digital tach.  It has a single counter that can handle a signal at up to 5 MHz.  The other digital lines are software timed so are not suitable for any sort of high speed application.  This keeps the cost in the same range as you were looking at before.