keyence laser gauging system

[jboetto]

My company is developing a system to measure the underside of our parts. We are employing a Keyence LKG032 Displacement sensor on an X/Y Stage. We will be measuring "z" @ an assigned x, y position @ a very high speed. The data will then be "stitched" together to create a 3D visualization of the measured geometry. Question 1. Which stages play well with the NI MID 7652 Servo Drive? Looking for a linear motor arrangement with a nominal 50 x 50 mm travel. I already tried the motor and stage advisor. Less than helpful. Question 2. Please illuminate any circumstances we should consider in determing hardware requirements. Thanks.

[Brian_Beal]

What x-y resolution are you looking for?  I believe that in some scanning applications like this that a stepper system is the better way to go.  You don't have to worry about tuning and any possible dithtering.

I have used that Keyence system before, and measurement results greatly depend on having the many configuration parameters set correctly.

[jboetto]

Brian

Thanks for your response. Can I email directly to you greater description of our device for your reference? I'm not sure if steppers are a consideration. We need to scan a length of 150 mm in 300 msec. Can you elaborate on which configuration parameters you set to achieve your results?

Thanks

Joe

[Brian_Beal]

Joe-

Yes, you can email directly if you like to brian@bjbeng.com.  The LG unit I used was a demo unit, and I don't remember all the parameters.  You need to be careful with averaging and the frequency of the actual laser measurements.  There are other parameters, but I don't recall them right now.  I also wrote a Labview program to get the measured values.  This was all done at a previous job, so I don't have any more information.

I am a little confused about your requirements.  You said you need 50mm travel, but then you state that you need to scan a length of 150 mm in 300ms.  Are you saying that each scan is 300ms long, and the laser measured distance is 150mm?

Brian

[ThomasGoodey]

Sirs: I see you are working with a LK-G 032. My question is: are you using the analog voltage output from the controller? My problem is that I am using a LK-G82 and a LK-G152, and the sensors are capable of providing much greater accuracy than is represented on the analog output. Accordingly I want to use the digital output, or the buffered output of the 3001 controller. (Obviously the RS-232C output is no use; it is too slow.) I am a bit stymied, due to the inadequacy of the Keyence software. Can you help?

[Brian_Beal]

For accuracy, the RS-232 is the best way to go.  How fast do you need measurement updates?

[ThomasGoodey]

Thank you so much for being interested in my problem!

What I want is to obtain, together, a combination of high speed (number of measurement values per second) and high accuracy. The two Keyence LK-G sensors and controller, for which I paid about $10,000, are capable of providing such measurement values; this is why I paid so much for them. But there appears to be no way of getting the data out automatically.

I have purchased the pathetic Keyence control software ($800), and, using that software manually, I can get the accurate data I want at the speed I want it, buffered and then sent via USB. But I have to operate the Keyence software manually with the mouse. I can't operate it from another computer program.

This is because the software is not furnished as an API, which in 2007 I frankly consider to be a disgrace. I need .NET Framework, or at least, a LabView driver.

In other words, the analog voltage output is fast but not accurate enough, and the RS232-C is accurate but not fast enough (to cope with the full flood of data from the sensors).

The controller has a so-called binary output, which consists of 23 binary lines and a trigger, and this might work, but it involves substantial programming effort, and moreover some interface to get those binary values into the computer. Since all this stuff is already inside the controller, it is very annoying not to be able to use what I paid for. It is most irritating.

[Richard_Jennings]

Hi Thomas,

We are currently using a Keyence controller in a measurement application. I definitely recommend using the binary output port for high-speed acquisition. Place a scope on the strobe output on pin 34 and you should see the binary output continuously updating at 5 kHz. An NI 6534 DIO board can interface with this binary stream fairly easily. Just make sure to connect the strobe output to PCLK input (PFI2) on the 6534 and use an external clock for your DAQmx task.

However, this won't help you switch between heads or trigger the measurement. For high-speed applications with complex triggering and timing I think LV-FPGA and NI RIO cards are the only way to go.

Hope this helps,

Richard

[ThomasGoodey]

Thanks Richard. Yes, if I connect to pin 34, the strobe output, I should be able to see the strobe. And I should be able to switch between the outputs using pin 35. There are also RESET inputs for both OUTs. And I understand you saying that a NI 6534 board could be interfaced to the binary output lines 13 through 33. Then I could, I suppose, program the computer for reception of the data in "twos' complement 21 bits" format. What a pain.

But surely all this is a function already existing in the controller in a more sophisticated mode? The nature of my experiment means that the data comes in bursts. One burst can easily be accommodated in the buffer memory of the 3001 controller, and I could download it later during an interval between bursts.

According to the scenario you advocate, I would only be using the controller as an expensive power supply for the heads, and its data processing functions would be ignored.

Here in England the 6534 board costs about $2000. I have already purchased the controller, which does the required job, and it cost $2000. I am rather resentful that, because of the inadequate Keyence software, I can't use the circuitry in the controller which I have bought. Surely there is some solution that involves communicating with the controller via USB? That's all that's required: the USB interface is perfectly capable of handling the data volume and speed.

[ThomasGoodey]

And, jboetto, if you are scanning your LK-G 032 at high speed, how do you propose to get the data out? Is the analog voltage output sufficiently accurate for your requirements?