Engine dyno hardware specifications

[Speedworx]

Hi All,

I've recently started a new project.

I currently have an analogue dyno (Go Power water-break with hydraulic load cells) set up, and I want to convert it to digital.

I definitely can't afford Go Power's software for this, so my goal is to build it myself (with extras).

Having no previous experience with LabView or with pressure sensors etc, I was hoping to find out what hardware would be most recommended, based on running these sensors: ??

• 8K type thermocouple 
• 3J type thermocouple 
• 4 pressure transducers (combustion/torque/intake pressure/exhaust pressure)
• 1 RPM pick-up
• 2 wide-band air/fuel sensors (which need outputs for pre-warming)
• 1 encoder (720 degrees)

Additionally with the hardware, I would like to be able to add-in some sensors later on if needed.

I've done some research, and have heard that it's best to sample 3 times faster than ideal. Is this correct??

And if this is true, my dyno has the capabilities to run intermittently @ 12 thousand RPM, so that would mean a sampling rate of 600?

I recently bought 2 new LabView books, which have given me a bit more insight, and from them I've constructed a few VI's. I'm sure I'll have more questions on this subject once I start building with LabView.

As I said, I've no previous experience in this area, so any feedback would be greatly appreciated.

Thanks in advance,

Josh

[esmaham]

Hi Josh,

Here are some good links to get you started with LabVIEW:

1- http://www.ni.com/gettingstarted/

2- LabVIEW Training: Learn LabVIEW in Three or Six Hours:
http://zone.ni.com/devzone/cda/tut/p/id/5247
http://www.ni.com/academic/labview_training/

3- Schedule Your Personal Online LabVIEW and NI Hardware Demo with an NI Engineer

http://zone.ni.com/devzone/cda/tut/p/id/10864

Hope this get you started

Thanks and have a great day.

[Speedworx]

Hi esmaham,

Thanks for the reply.

I have read the getting started section and watched all the movies 🙂

The NI Engineer gave me a 30 day trial and I've built several VI's from uni sites. They are very helpful for any other newbie's.

I've had an NI Engineer out to inspect my situation and see what equipment I need. He showed us the labview software and also showed us the range of hardware available to us. I'm in the process of getting some sensors and once I have them I can tell the NI staff what voltages the sensors will be so we can select the right equipment.

I'm just worried about the correct speeds I need to sample at to have accurate data. The NI engineer didn't speak much on it and I know it plays a big part of accurate testing. Most people I have come across seem to be using the USB hardware. One of the articles said you need to sample 3 times the ideal speed??

I've brought some books from Amazon on labview which should be here in a couple of days. Which will hopefully help with the labview side.

Thanks,

Josh

[GovBob]

Another good resource for how to measure from sensors and at what rates is here. If you root around in there you will be suprised what you will find.

Hope this helps.

[johnsold]

Josh,

Your question about sampling rates has not yet been addressed.  You indicated a maximum speed of 12000 RPM or 200 revolutions per second and inferred a sample rate of 600 Hz based on that.

1. The Nyquist rate for sampling a signal so that it can be accurately reproduced from the samples is greater than twice the highest frequency component in the signal.

2. The pressures in your system vary significantly during a revolution so the highest frequency in the pressure signals may be well above 200 Hz.

3. The temperatures probably vary much more slowly and could likely be sampled much more slowly.

4. The air/fuel sensors are described as "wideband."  This may indicate that they need to be sampled faster also.

5. I do not understand a "720 degree" encoder.  Does it go around twice and stop? or is it 720 pulses per revolution? Are you connecting it to a counter or using a digital or analog input to measure its output?

How fast you need to sample depends strongly on the nature (and frequency content) of your signals and what you will be doing with them.  Are you looking for a peak value or an average value or do you need the waveform data for each degree of rotation?  Do you need to synchronize with the motor speed?

I suspect that 600 Hz is not fast enough for some of your signals while it is excessive for others.  You can learn this stuff from a few books, but you would probably get good results much faster if you had someone knowledgeable and experienced to help you.

Lynn

[Speedworx]

Hi All,

Thanks for all your comments.

In reply to Lynn:

The engine would fire up to 12000 RPM but its a 4 stroke so it would turn over 24000 times a minute. I don't know if that changes anything?? I plan on using a pressure sensor in the inlet manifold and exhaust manifold. Also a combustion sensor in the combustion chamber and a pressure sensor for measuring the torque from the load cells.

The combustion, inlet and exhaust sensors would have to measure everytime the engine fires. The data from this is extremely important and needs to be accurate. The torque on the other hand would not need to be measure so fast.

The temperature and the wideband AFR sensors probably don't need to be measured as fast either. But remembering the temp and AFR sensors save engines from failures. I'd still like it to be accurate. I'll do more homework on this 🙂

The 720 deg encoder would be used to tell where the engine is when the combustion, inlet and exhaust pressures are taking place. Its 720 deg because the engine must revolve twice to complete one cycle. The encoder would be in 1 deg increments.

Also on the note of the encoder could I use this as a RPM counter? Or would it be easier to use a separate counter for just picking up the RPM?

Silly question but can a sensor limit how fast you can sample??

Thanks,

Josh

[johnsold]

Josh,

I am not sure I completely understand, but let me try to clarify.

1. The output shaft of the motor turns at 24000 revolutions per minute and a cylinder fires 12000 times per minute.

2. The encoder produces one pulse per degree of shaft rotation.

3. Some of the sensors change their outputs significantly in a few degrees of rotation (the combustion sensor)

Based on these assumptions:

1. The encoder frequency can be up to (24000/60)*360 = 144 kHz

2. Based on (3) above the sampling rate for the pressure sensors should also be about 144 kHz or faster.

You indicate that the temperature and AFR sensors are at least in part safety systems.  How fast you need to sample them depends on how fast they can change and how fast the engine can be damaged.  I doubt the temperatures can change significantly in less than a few seconds, so the thermocouples can probably be sampled quite slowly.  Do the AFR sensors measure averages or do they respond to the instantaneous flow when the valves operate for each cylinder?  If the latter, they will need fast sampling also.

Typically the encoder would be used to measure RPM.  No need to have a separate sensor just for that.  If the encoder has a Z-axis output (one pulse per revolution), that can be used directly.  When the encoder is installed, align the Z-axis pulse with cylinder 1 TDC and you can tell exactly where the engine is in its rotation.

The sensor does not limit how fast you can sample as much as it limits how slowly you can sample.  For example you could sample the thermocouples at 144 kHz and then average 144000 samples to get one temperature reading per second or you could just sample once per second.  In a noiseless system the two results would be the same.  However, if you sample the combustion pressure sensor once per second, the data you get will be essentially meaningless random numbers.

Lynn

[Speedworx]

Hey Lynn,

Thats excatly what i mean!! Your done this before 😉

The AFR's are sampled slow. I am currently using a hand held system and they change about every second. I think that should be sufficent for the thermocouples too.

I spoke with a encoder manufacture today and they have a standard encoder to do what i'm after. And it has the "Z axis" output for the one pulse per revolution.

I was looking at using a optrand combustion sensor. Its price is 1/10 of a kistler and 1/5 of a PCB sensor. Its comes with a signal conditioner built-in and its in my price range...  Still looking for an absolute sensor for the inlet and exhaust pressurs. I already had the torque sensor so now i'm just looking for some thermocouples.

I have some AFR sensors but i'm trying to figrue out the voltages they use for the output and input. I'm sure its all 5 volt..

To sample that fast i'll need to use a USB data system??

Thanks again,

Josh

[johnsold]

Josh,

I have not worked on that specific type of equipment so I cannot advise you on transducers or DAQ devices.  We generally use Omega for thermocouples.

My primary concern was that it appeared that you had significantly mis-estimated you sampling requirements.  If you want more help with selecting transducers and DAQ devices, I suggest that you carefully evaluate your actual requirements for each type of signal and post a summary of that information along with specific requests.

Lynn