Zitat von mbruch im Beitrag #16

---

For the nozzle to center pictures are shown after reset move to camera center or VO was used before?

---

Zitat von mbruch im Beitrag #16

---

Placement deviations could happen because of thickness deviations.
PCB bend or component height not well adjusted.

---

Zitat von mbruch im Beitrag #16

---

Chip parts use full alighnment?

---

Zitat von mbruch im Beitrag #16

---

Mostly placement looks good.

---

Zitat von mbruch im Beitrag #18

---

Please zip all your *.fig files and config.ini and send to me michael@smallsmt.biz we need your settings to trace down problem.

---

Zitat von mbruch im Beitrag #18

---

for example thos one is ok
http://files.homepagemodules.de/b649197/...dae4/b109a4.png

some others too except your SOT23-5.

---

Zitat von mbruch im Beitrag #18

---

Do you really use measured part height -0.1 to -0.2mm?
What's your PCB height setting?

---

Good time.

The boards



were mounted on the VP2800HP-WNE machine using VisionPlacer software versions V2.060 from 01/14/2021 and V2.057 from 12/10/2020.
During the mount, the same configuration files were used, except for the "camera.fig" file which does not support version compatibility.
Mount was carried out under the same conditions, under the same lighting and temperature, with the same nozzles.
The alignment of the components was configured (ajust bridges and contrast) using the version V2.060, which has a display convenient precision function alignment of components.
With the help of each version, 16 boards were mounted (two boards on the panel), with a total of 3264 components in each batch.

The results are as follows:

- spread of positioning deviations when placing large cases such as TQFP64 in both versions is approximately the same and is about 0.2 - 0.3 mm, and these deviations always occur in one direction. Exactly at the given coordinates the component is more often found in version V2.060. To estimate the spread of deviations in angle difficult, but small and the same in both versions.

- components with dimensions 1206, 0805, DPAK are more precisely placed in the version V2.060 (positioning deviation 0.2 - 0.3 mm) than in V2.057, where deviation reached 0.7 mm, but in the V2.060 version there is often deviation of the angle of place of such components by large amount - about several degrees.

- smaller components (SOT23, 0603, SOD80) are much more accurate placed in version V2.057, where the deviation did not exceed 0.2 mm, in the difference from the V2.060 version, where the deviation reached 0.5 mm.

- when working in version V2.060, components are drop much more often due to the impossibility of alignment - photo of dropped components:

Version V2.060:


Version V2.057:


In addition, about 10 components (0603, SOT23) when using V2.060 were placed with a spread 90 degrees, which was never observed in V2.057.

- also in version V2.060, the error "Y position error" often occurs, when trying alignment of the component with camera 2, presumably due to the that the Collection Time for camera 2 was set to 100ms. When this error occurs after clicking "OK" the held components are dumped directly to the camera. This situation is not observed in version V2.057.

- when working with both versions, several components were not placed on their places, but they were found quite far from their place, which suggests that it would be nice to check the vacuum level not only during an optical aligning the component, but also just before placing it on the board.

As a result, when mounting boards with components 0603 and SOT23, software version V2.057 works much better, but no version allows you to do without additional checking and manual alignment of some components.


Also, there are some annoying points in both versions, namely:

- in the list of feeders the feeders of group G are not grouped together, but are located between the corresponding positions of the feeders of group E.

- when moving the camera to the selected component on board, it is desirable to save set the brightness level of the backlight of the upper camera, and not set its each time to the value wich set for the Visual Original alignment.


Sorry for attention.

Dear Michael.

Thanks for your answers, but the accuracy settings were the same for both versions, however the machine worked completely differently.

In addition, while working with version V2.060, it was sight that in some cases when trying to align a component, with the first step, it rotates in one direction at a rather large angle - about 30 - 45 degrees. The next step rotates the component in the opposite direction, but at an angle less than it was rotated by the first step. Then the same attempts continue and when the number of attempts has been exhausted, the component either goes to throw away or is placed at the wrong angle.
With the same alignment accuracy settings with the FutEdge2 algorithm, SOT23 cases in version V2.057 have no problems at all, and in version V2.060 about 10 percent of such components go to throw away, and about 2 percent are placed incorrectly. This does not happen in version V2.057 with the same alignment precision settings.

And yes, since I am a consumer of your product, I want it to really become better suited to work seamlessly with it.

Zitat von mbruch im Beitrag #24

---

Your placement result you posted before looks perfect.

---