Re: [OpenPnP] Issues and solutions problems and questions

[mark maker]

It seems your camera setting are not stable. That's why Issues & Solutions keeps lamenting about them.

I've seen such instability on Linux, are you on Linux?

If yes:

After everything is set up and working for the moment (i.e. inside one OpenPnP session going through Issues & Solutions and/or manually adjusting the camera settings), it may help to enable the Freeze Properties checkbox on the camera:

https://github.com/openpnp/openpnp/wiki/OpenPnpCaptureCamera#freezing-camera-properties

Note: this is just a "band-aid" to alleviate the symptoms of the underlying (Linux) problem. Don't expect a miracle cure.

> Is there a preferred way to go through the I&S procedure?

I&S automatically proposes only those next steps that are ready according to the prerequisites. Any one of the proposed steps is valid.

Having said that: the most logical next step often comes first. Doing it in that order might turn out to be more stable, as it corresponds to the "beaten path" with the most testing/optimizing applied.

> Bottom camera I have moved the bottom camera up so it is closer to the height of my pcb. I think I am going to lower it again

The ideal bottom camera has its focal point at PCB surface Z. That way, the nozzle extension is the same when you align and when you place. So any mechanical imperfection is the same (I'm not saying there is one 😉).

Furthermore it is better to have a longer lens (not very wide angle), as you then get smaller parallax errors (exaggerated illustration):

So both point to lowering the camera quite a lot, assuming your table design can do it (hole?).

As to the view size: you don't need the camera view to be large enough for the largest ICs:

https://youtu.be/P-ZudS7QQeE?feature=shared

> How much distance is needed to eb able to move in x and y for the homing fid?

For the homing fiducial: The minimum is a few millimeters, maybe 10mm. There are good reasons to put it more central, but if you really can't, that's fine too.

And just to make sure there is no confusion: Don't confuse the homing fiducial with the calibration fidcuial(s). The latter really must be quite central, there is no excuse there.

> but need to figure out my whole layout as I remove and replace currently a board with x and y lines for double sided taped down strips.

Sorry, I don't understand what you mean. 

But perhaps this can help reasoning about how to arrange your machine table:

https://makr.zone/planning-the-work-area/117/

The Liteplacer frankly uses a bad layout, with the PCB and bottom camera in the corner like that. That layout doubles the average motion distances.

> Any links to images of users with light placers have modified the machine setup to better accommodate openpnp?

To be clear: OpenPnP can work with the bad Liteplacer layout too. It is just less optimal.

> but then seems like I need to raise everything up somehow so the bottom camera focal length to nozzle tip is the same height as the board.

I believe these are two separate questions. Even the Liteplacer recommends making a hole in the table and sink the camera, AFAIK. That is still an option regardless of central/not central.

Here's a video of my table design. Also note how the nozzle tip changer is sunken, which prevents any collisions (I had a very bad one):

https://youtu.be/poLwQDpfeJ8?feature=shared

The table design itself I no longer recommend. But you see the basic principles of having a substructure, where you can go "below deck".

Other videos that may answer Liteplacer specifics, like contact probing:

https://youtu.be/9uFxV1-vnXw?feature=shared

_Mark

On 27.01.2024 02:36, zzimdog wrote:

I am on my third try of getting my machine back in working condition. I just noticed there was a newer version on 1-8-24 that I installed while being stuck on the advanced bottom camera calibration. I installed it just in case it would help fix something. I have been through all the calibrations on the top camera, offsets etc. Setup visual homing and it was working properly every time I would home. During this process I would keep getting the settings for gamma, white balance etc pop up. I would click accept on all of them and it would say solved and then they would all pop right back up. this kept happening and I noticed when I went to the camera settings and tried to manually set change them the apply button would stay grayed out. I ended up exiting openpnp and then restarted it and then solved them with I&S and then it set them all. Not sure what caused this.

I am now having issues when I exit and then restart the program and click on home i get a cannot not move to position x again because it is below the soft limit. I assume I will need to move my homing fid again as it must be to close to the x limits. I wanted to make sure though as it had been working fine and I had made it through everything but the bottom camera advanced calibration and it gave an error about unit pixel dimensions were off or something. I am unable to get exact error as now I need to correct this before I can go any further

Is there a preferred way to go through the I&S procedure? I try to go through each option as they appear in the list and solve them sometimes though things like camera white balance pops up before all the gamma etc settings and directs that these must be completed first. This might just be because they popped up in the right order but then would come back as I mentioned before. Maybe instead of chronological like I have been trying I should jump to all warnings and errors first and then move to suggestions on my next try through this?

Bottom camera I have moved the bottom camera up so it is closer to the height of my pcb. I think I am going to lower it again some though as even though I get really good detail to my eye I had to adjust the focus way out and when it is doing certain calibrations it drives the nozzle down so low it looses focus and is close to hitting the lens. Where do these low settings come from it is going lower than what I have defined for z movement. Is this setting disabled while it does the small incremental z movements. 

What are current recommendations for cameras for openpnp and lite placer? I saw an old link to ebay I think from one of marks post but they are no longer valid. and do I need to get a smoothie board or something just to get light actuators to turn on and off lights? I am assuming this is beneficial other than just saving power?

How much distance is needed to eb able to move in x and y for the homing fid? I am trying to figure out a way to change the positions to the center as mark suggests in one of the wiki pages but need to figure out my whole layout as I remove and replace currently a board with x and y lines for double sided taped down strips. This worked well in liteplacer software as feeders were assigned to each board I was running but looks like it will be a hassle with all the feeders showing up in openpnp. I am only setup for 2 boards right now and have just named feeders differently so I know which positions to enable for each job.

Any links to images of users with light placers have modified the machine setup to better accommodate openpnp? Specifically I would like to see camera,bottom light placement and nozzle changer. It seems like moving them to the center is recommended but then seems like I need to raise everything up somehow so the bottom camera focal length to nozzle tip is the same height as the board. I am attaching a pic of how I am currently setup if anyone has some suggestions. 

Thanks,

Mike

--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/18867ef9-6dd8-4ba6-994b-2fdb4b88207bn%40googlegroups.com.

[zzimdog]

Thanks Mark,

I am running linux and one of the main reasons I looked into openpnp to begin with. I did not know there were stability issues with it I may try windows and see if that helps if things go south again. I have only built a couple of boards since I finally got back up and running last night but so far everything seems to be working better then the last time I went through this process though I think my speeds are way slower. I need to find what settings to use for the jerk acceleration etc. So far just guessing really. I would like to replace the cameras with what you would currently recommend. The link attached to your blog is bad and putting the generic description in amazon seems to bring up all different models. Can you give me a specific camera to order? Also do I need to order lenses and light rings? How are you controlling your lights on the liteplacer? Did you move over to the smoothie board as your controller instead of the tinyG. I thought I had read that you didn't recommend the smoothie and that the tinyG jerk stuff was better. This stuff is really over my head so trying to keep piece it together going through old info in blogs and wikis and then current posts in the forum gets confusing for the beginner or at least this one ;) 

On Sunday, January 28, 2024 at 11:07:52 AM UTC-6 ma...@makr.zone wrote:

It seems your camera setting are not stable. That's why Issues & Solutions keeps lamenting about them.

I've seen such instability on Linux, are you on Linux?

If yes:

After everything is set up and working for the moment (i.e. inside one OpenPnP session going through Issues & Solutions and/or manually adjusting the camera settings), it may help to enable the Freeze Properties checkbox on the camera:

https://github.com/openpnp/openpnp/wiki/OpenPnpCaptureCamera#freezing-camera-properties

Note: this is just a "band-aid" to alleviate the symptoms of the underlying (Linux) problem. Don't expect a miracle cure.

> Is there a preferred way to go through the I&S procedure?

I&S automatically proposes only those next steps that are ready according to the prerequisites. Any one of the proposed steps is valid.

Having said that: the most logical next step often comes first. Doing it in that order might turn out to be more stable, as it corresponds to the "beaten path" with the most testing/optimizing applied.

> Bottom camera I have moved the bottom camera up so it is closer to the height of my pcb. I think I am going to lower it again

The ideal bottom camera has its focal point at PCB surface Z. That way, the nozzle extension is the same when you align and when you place. So any mechanical imperfection is the same (I'm not saying there is one 😉).

Furthermore it is better to have a longer lens (not very wide angle), as you then get smaller parallax errors (exaggerated illustration):

So both point to lowering the camera quite a lot, assuming your table design can do it (hole?).

As to the view size: you don't need the camera view to be large enough for the largest ICs:

https://youtu.be/P-ZudS7QQeE?feature=shared

My camera was mounted to the bottom of the table which is 45mm thick with an 88mm hole drilled through it with  standard liteplacer setup.  When I first went through I&S before this round everything looked good at first. Then sometime after the advanced calibration it would shrink the image down to a mush smaller size with a tilted jagged outline around the image. Then finally before I went through this iteration of I&S I had rerun the calibration as it quit calibrating the nozzle and would not finish the homing process. When I did that it shifted the whole image over to the right and was so small I could not work with it. 

I created a cup to fit in the hole with the 3 spring holes that light placer used in their mount. The setup now seems to be working better. The lens is currently about 20 mm below the pcb height. I currently have a bigger view of the image on the screen though it is blurry when the nozzle is up. 

> How much distance is needed to eb able to move in x and y for the homing fid?

For the homing fiducial: The minimum is a few millimeters, maybe 10mm. There are good reasons to put it more central, but if you really can't, that's fine too.

And just to make sure there is no confusion: Don't confuse the homing fiducial with the calibration fidcuial(s). The latter really must be quite central, there is no excuse there.

> but need to figure out my whole layout as I remove and replace currently a board with x and y lines for double sided taped down strips.

Sorry, I don't understand what you mean. 

I have a 12"x20" board with the cutout of whichever pcb job I am running. On this board I have grid lines for taping all the part strips for that job down where they need to go. when that job is done I unscrew the board lift it out and place a new board down with all the strips of parts layed out for it. It worked well before with the light placer software but looks like it will become cumbersome in openpnp. I am currently looking at printing some strip holders. I tried the ones from psycogenics posts but has some issues I am going to try the ones from the blog you just linked below and see how that works it is really a brilliant design. 

 

But perhaps this can help reasoning about how to arrange your machine table:

https://makr.zone/planning-the-work-area/117/

The Liteplacer frankly uses a bad layout, with the PCB and bottom camera in the corner like that. That layout doubles the average motion distances.

> Any links to images of users with light placers have modified the machine setup to better accommodate openpnp?

To be clear: OpenPnP can work with the bad Liteplacer layout too. It is just less optimal.

> but then seems like I need to raise everything up somehow so the bottom camera focal length to nozzle tip is the same height as the board.

I believe these are two separate questions. Even the Liteplacer recommends making a hole in the table and sink the camera, AFAIK. That is still an option regardless of central/not central.

Here's a video of my table design. Also note how the nozzle tip changer is sunken, which prevents any collisions (I had a very bad one):

I definitely need to do this as I have had several collisions especially during this last configuration of the nozzle changer. I really need to redo the deck on this. I plan on making a new one but first trying to sort out where everything will end up going and makae sure I have enough experience to be able to take the plunge. I think once I have the cameras sorted out and hopefully can get the bottom vision figured out for placement then I can look more into this. It seemed to work better but the only vision that seems to do anything when I select it is rectlinear the other options give me a subject not found or something like that. The rectlinear worked really nice on the first couple resistors and perfect on the 48 pin micro but then would start placing resistors and caps at worse angles than what they were picked up for some reason and my camera would show the cross hairs at an angle and it would not center the part afterwords. I just have basic pipelines and settings as far as I know as I do not understand the pipeline stuff yet and was thinking that something else might be the culprit before trying to mess with them. Your machine is awesome You need a liteplacer upgrade kit to sell :)

[mark maker]

I hope others have links, I don't.

The recommended model is still the same: ELP Model USB100W03M, 720p USB

You can connect two of those to one USB2.0 root hub.

If you positively know you can use two USB root hub ports (positively root!), you can take the FullHD model. It has twice the FPS (60) at the recommended 720p, which might save some camera settling time.

_Mark

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/31ca38a3-5057-4d36-a3d4-d61542ffce53n%40googlegroups.com.