Eyebot is a revolutionary inspection device.  It is a trainable machine vision system that installs fast and allows extreme flexibility.

Eyebot requires no PC, no operating systems, no frame grabber, and no software.  It is inexpensive to install and support, and you can teach it several products.  You can use Eyebot wherever automated optical inspection was previously too complicated or too costly.  So throw away your PC, strobe and frame grabber, and enjoy the power of Neuro-RAM! 

Eyebot's Easy Operation

 Four buttons and one knob are all you need to use to make this machine vision system
 work. The intuitive knob takes you step by step through all the key steps. The buttons
 help you customize the Eyebot to your needs. Here is a quick overview of how it works: 

• ERASE: Permanently erases Eyebot memory.
• VIEW: See what Eyebot sees. This shows you the field of view and helps you identify if you need to improve your lighting.
• LEARN: Once in this position, Eyebot starts learning whatever you show it, even if it is moving or rotating.
• TEST: After you train Eyebot to learn an object, this mode will alert you whenever it detects any deviation from the learned object.
• IGNORE: This is rarely used, but is useful when you want to perform object recognition. Eyebot will learn to ignore whatever you show it while in this mode.
• RUN: All systems go! This is the same as the Test mode, but now it will activate the RS-232 and optically isolated outputs.

Eyebot can function in two ways: 

1. Object Recognition: Eyebot alerts you when it recognizes the object it learned.
2. Objection Inspection: Eyebot alerts you when it does not recognize the learned object, thereby detecting a defect on the learned object.

INSPECTION METHODS

Eyebot has four methods inside the box: Shape, Spectrum, Color 1 and Color 2.  You choose which is best for your application needs. Select which method you want to use by going to OPTIONS.

Shape Method

Spectrum Method

Color 1 Method

Color 2 Method

Applications

• Inspecting to see if a filter is correctly placed in a pipette tip (it cannot be damaged, misplaced, or missing).

• Verify that a robot has put a crimp on a nut.

• Inspect Ball Grid Arrays (BGAs) for surface defects (e.g., improper placement in tray, missing or double balls)

• Verify that there are threads on a nut.

• Verify that a robot placed adhesive on the inside of a car door.

• Verify that a car label is placed in the correct orientation and is not askew.

• Verify that a bearing is lying on the correct side.

• Verify that the correct torque converter (out of 7 models) is being manufactured. 

• Inspect fruit and vegetables for color differences.
• Verify that french fries are not burned or moldy.
• Verify that the correct color label is applied.
• Inspect the color of grass to determine if it needs fertilizer or water.
• Inspect for corn in chick peas.

Can Shape Method Solve My Problem?

However, there are two simple questions that you can answer: 

1. How controlled is your process? This includes the control of part presentation, orientation, and lighting. Is there substantial acceptable variation within good parts or do they all look exactly alike?
2. Are the defects that you want Eyebot to detect obvious or hard to see? In other words, are you trying to catch gross or subtle defects? Imagine looking at a monitor screen: could you easily see the defect? If so, it is probably a gross defect. If it would be difficult to see, then it is probably a subtle defect.

The more control you have in your process, the further down the X axis you are. The more obvious the defects you are trying to see, the higher up the Y axis you are. 

Gross Defects

Medium Applications for Eyebot 

Easy Applications for Eyebot

• Seeing a missing ball on a ball grid array. 
• Identifying when an apple, tumbling down the assembly line, has a significant piece missing or is obviously molding.

• Detecting the presence of a label on parts that are being indexed. 
• Identifying when a part has a notch on the right edge, with controlled lighting. 

Subtle Defects

• Seeing scratches on a metal nut that has to be learned in all rotations. 
• Identifying a missing letter on a lot code after learning it in a variety of lighting situations and positions.

• Seeing a missing ball on a ball grid array, which is being indexed on an X/Y tray. 
• Detecting subtle deviations in a stationary image that was learned. 

Little Control

As you can see, the upper right quadrant features applications that are ideal because they are highly controlled and the user is inspecting for gross defects. On the other hand, the lower left quadrant features applications that are extremely difficult because the part presentation is not controlled and the user hopes to find small and subtle defects. 

In reality, most applications do not fall into either one of these extremes. Most are either in the lower right quadrant or the upper left quadrant. The applications are possible for Eyebot to do, but may be somewhat tricky; fortunately, the manual and SIGHTech’s support line will help you overcome these hurdles. 

Can Spectrum Method Solve My Problem?

When in doubt, contact Bender Associates, Inc.

However, there are two simple questions that you can answer: 

1. How controlled is your process? This includes the control of part presentation, orientation, and lighting. Is there substantial acceptable variation within good parts or do they all look exactly alike?

2. Are the defects that you want Eyebot to detect obvious or hard to see? In other words, are you trying to catch gross or subtle defects? Imagine looking at a monitor screen: could you easily see the defect? If so, it is probably a gross defect. If it would be difficult to see, then it is probably a subtle defect.

Now, using the table below, locate in which quadrant your application lies. 

The more control you have in your process, the further down the X axis you are. The more obvious the defects you are trying to see, the higher up the Y axis you are. 

Gross Defects

Medium Applications for Eyebot 

Easy Applications for Eyebot

• Seeing a missing ball on a ball grid array. 
• Identifying when an apple, tumbling down the assembly line, has a significant piece missing or is obviously molding.

• Detecting the presence of a label on parts that are being indexed. 
• Identifying when a part has a notch on the right edge, with controlled lighting. 

Hard Applications for Eyebot 

Subtle Defects

• Seeing scratches on a metal nut that has to be learned in all rotations. 
• Identifying a missing letter on a lot code after learning it in a variety of lighting situations and positions.

• Seeing a missing ball on a ball grid array, which is being indexed on an X/Y tray. 
• Detecting subtle deviations in a stationary image that was learned. 

Little Control

As you can see, the upper right quadrant features applications that are ideal because they are highly controlled and the user is inspecting for gross defects. On the other hand, the lower left quadrant features applications that are extremely difficult because the part presentation is not controlled and the user hopes to find small and subtle defects. 

In reality, most applications do not fall into either one of these extremes. Most are either in the lower right quadrant or the upper left quadrant. The applications are possible for Eyebot to do, but may be somewhat tricky; fortunately, the manual and SIGHTech’s support line will help you overcome these hurdles. 

If you have multiple parts and quick product change overs, check out the Multi-Session Eyebot. 

For a lower end application, learn about Mini-Eyebot.