KAYO A8L conversion

[Trevor White]

Hi All

There is a Kayo A8L on UK eBay but I have heard the software isn't very good on them. I wondered if anyone knows if the machine shares the same electronics with the Kayo A4 that has been successfully converted? 

Wondering how much work it would be to get it working with OpenPNP is the general question. 

Thanks

Trev

[Jonathan Oxer]

I don't know for sure because I don't have one, but I heard that the original Kayo software can support multiple machine models so the protocol is most likely to be common between them.

I wanted to allow for support of 6 and 8-nozzle machines in OpenKayo so I made the gcode parser easy to extend. For example, it maps axis letters to nozzles by assigning A to nozzle 1 rotation, B to nozzle 2, C = 3, D = 4. It would be a trivial code change to add axes for nozzles 5-8.

https://github.com/superhouse/openkayo

In my code you can see for example that the command for rotation of nozzle 1 is 0x11, nozzle 2 is 0x12, etc. I have a strong suspicion that if you extended the code to use command 0x15 for nozzle 5, 0x16 for nozzle 6, etc it would just work:

https://github.com/SuperHouse/OpenKayo/blob/15181e1d61ce416175e057e14db9fdf4a5951c1c/Firmware/OpenKayo/machine_actions.h#L322

OpenKayo isn't perfect and there is a bug in the conveyor width setting that I need to track down one day, but it's been at the "good enough for daily use" stage for a while now. I've run many thousands of boards through my Kayo A4 with OpenPnP+OpenKayo.

Cheers

Jon

--
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/1b914560-0c1e-4006-b145-c79ab8ee7870n%40googlegroups.com.

[Trevor White]

Thanks for the info, Jon. I think the machine advertised is too expensive for a project and also a bit too big for my space. Ideally something like the A4 is more my requirements. Do you find the hardware works well with openpnp? Placement is accurate and repeatable? 

Trev

[Jonathan Oxer]

I've found my A4s to be quite repeatable. I've just finished running a panel a few minutes ago, and I was watching the bottom camera view in OpenPnP and wondered why the image wasn't updating with each pick. I looked more closely and noticed that the part on the nozzle was changing each time, but the rest of the image encompassing the nozzle and its surround were pixel-identical every time.

Before you jump in though, be warned that I couldn't get the original bottom cameras working. From the factory it has a single "precision" camera and an array of 4 "high-speed" cameras to allow it to image all nozzles simultaneously. What I ended up doing was removing all the bottom cameras and mounting a single ELP camera in the location of the original precision camera using a 3D-printed mounting adapter. I'm very happy to provide all the details, design files, etc of changes I made.

Cheers

Jon

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/e3b60d63-e3bc-40cf-ae6b-466ddfbad58an%40googlegroups.com.

[Dave McGuire]

I'd love to get my hands on one of those A4s. Do these ever come up
in the US?

-Dave

On 8/29/24 06:31, Jonathan Oxer wrote:
> I've found my A4s to be quite repeatable. I've just finished running a
> panel a few minutes ago, and I was watching the bottom camera view in
> OpenPnP and wondered why the image wasn't updating with each pick. I
> looked more closely and noticed that the part on the nozzle was changing
> each time, but the rest of the image encompassing the nozzle and its
> surround were pixel-identical every time.
>
> Before you jump in though, be warned that I couldn't get the original
> bottom cameras working. From the factory it has a single "precision"
> camera and an array of 4 "high-speed" cameras to allow it to image all
> nozzles simultaneously. What I ended up doing was removing all the
> bottom cameras and mounting a single ELP camera in the location of the
> original precision camera using a 3D-printed mounting adapter. I'm very
> happy to provide all the details, design files, etc of changes I made.
>
> Cheers
>
> Jon
>
> On Thu, 29 Aug 2024 at 19:26, Trevor White <tswelec...@gmail.com
> <mailto:tswelec...@gmail.com>> wrote:
>
> Thanks for the info, Jon. I think the machine advertised is too
> expensive for a project and also a bit too big for my space. Ideally
> something like the A4 is more my requirements. Do you find the
> hardware works well with openpnp? Placement is accurate and repeatable?
>
> Trev
>
>
> On Wednesday 28 August 2024 at 12:22:29 UTC+1 j...@oxer.com.au

> <mailto:j...@oxer.com.au> wrote:
>
> I don't know for sure because I don't have one, but I heard that
> the original Kayo software can support multiple machine models
> so the protocol is most likely to be common between them.
>
> I wanted to allow for support of 6 and 8-nozzle machines in
> OpenKayo so I made the gcode parser easy to extend. For example,
> it maps axis letters to nozzles by assigning A to nozzle 1
> rotation, B to nozzle 2, C = 3, D = 4. It would be a trivial
> code change to add axes for nozzles 5-8.
>
> https://github.com/superhouse/openkayo
> <https://github.com/superhouse/openkayo>
>
> In my code you can see for example that the command for rotation
> of nozzle 1 is 0x11, nozzle 2 is 0x12, etc. I have a strong
> suspicion that if you extended the code to use command 0x15 for
> nozzle 5, 0x16 for nozzle 6, etc it would just work:
>

> https://github.com/SuperHouse/OpenKayo/blob/15181e1d61ce416175e057e14db9fdf4a5951c1c/Firmware/OpenKayo/machine_actions.h#L322 <https://github.com/SuperHouse/OpenKayo/blob/15181e1d61ce416175e057e14db9fdf4a5951c1c/Firmware/OpenKayo/machine_actions.h#L322>

>
> OpenKayo isn't perfect and there is a bug in the conveyor width
> setting that I need to track down one day, but it's been at the
> "good enough for daily use" stage for a while now. I've run many
> thousands of boards through my Kayo A4 with OpenPnP+OpenKayo.
>
> Cheers
>
> Jon
>
> On Wed, 28 Aug 2024 at 20:59, Trevor White
> <tswelec...@gmail.com> wrote:
>
> Hi All
>
> There is a Kayo A8L on UK eBay but I have heard the software
> isn't very good on them. I wondered if anyone knows if the
> machine shares the same electronics with the Kayo A4 that
> has been successfully converted?
>
> Wondering how much work it would be to get it working with
> OpenPNP is the general question.
>
> Thanks
>
> Trev
>
> --
> 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/1b914560-0c1e-4006-b145-c79ab8ee7870n%40googlegroups.com <https://groups.google.com/d/msgid/openpnp/1b914560-0c1e-4006-b145-c79ab8ee7870n%40googlegroups.com?utm_medium=email&utm_source=footer>.

>
> --
> 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

> <mailto:openpnp+u...@googlegroups.com>.

> To view this discussion on the web visit

> https://groups.google.com/d/msgid/openpnp/e3b60d63-e3bc-40cf-ae6b-466ddfbad58an%40googlegroups.com <https://groups.google.com/d/msgid/openpnp/e3b60d63-e3bc-40cf-ae6b-466ddfbad58an%40googlegroups.com?utm_medium=email&utm_source=footer>.

>
> --
> 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

> <mailto:openpnp+u...@googlegroups.com>.

> To view this discussion on the web visit

> https://groups.google.com/d/msgid/openpnp/CADBU0ecCDKKuaaPkgHJQ3aYdzbfNQZxBpaA4DSTu%2BkkdAA83Gg%40mail.gmail.com <https://groups.google.com/d/msgid/openpnp/CADBU0ecCDKKuaaPkgHJQ3aYdzbfNQZxBpaA4DSTu%2BkkdAA83Gg%40mail.gmail.com?utm_medium=email&utm_source=footer>.


--
Dave McGuire
McGuire Scientific Services, LLC
New Kensington, PA

[Trevor White]

Thanks again, Jon. Very generous of you to share all your hard work in getting the machine up and running. 

Repeatability seems very very good. I will get a quote from Kayo but it feels a bit wrong to buy a new machine and butcher it so much and to not use some of the features. Maybe if a second hand one comes up in the UK it's a good idea. If I had the space the 8AL would be a consideration for sure even though it would be reduced speed due to the cameras not being supported. 

Trev

[Trevor White]

Jon, what kind of cph do you get from your Kayo running OpenPNP? I am wondering what speed impact you see compared to if you were running the original Kayo software. Did you ever use the original software provided with the machine? Some people say it's terrible, others say it's okay. There isn't a lot out there on it. 

Trev

[Mike Menci]

Hi - I was looking for the way KAYO is handling the Juki nozzles but I could not find anything - are there any Manuals or maintenance instructions anywhere to download?  
I came across these two pdf pages only- attached - might be of some use to somebody out there....

Cheers and good luck getting KAYO rolling! 

[Mike Menci]

And a video here; https://www.youtube.com/watch?v=bd11fD_XaFA&t=12s

[Trevor White]

Thanks Mike. 

I got the following manual from them. 

[Mike Menci]

Thanks Dave 
I see for nozzle rotation are used.... 

- what about Z - up/down  --- is one stepper used for two nozzles? or each nozzle is driven by ??? 

Do you know ?   
Jon - can you explain or you can take a photo ? 

Are Juki nozzles connected with adapter like RobotDigg has it = 
https://www.robotdigg.com/product/517/Adapter-for-Juki-Nozzles-to-5mm-Hollow-Shaft-Stepper

Thanks

Mike

[bing luo]

It looks similar to this     

[Mike Menci]

Thank you Bing,

Mike

[bing luo]

This design is very interesting, similar to the drive wheels of a steam locomotive. Can make the mouter  very compact. And the movement of the nozzle is completely independent. It's much better than belt drive and seesaw.    I have never seen an openpnp designed like this before. 

[Jason von Nieda]

That's really clever! I have not seen a head design like that before.

Looks easy to 3D print, and then if you want high precision you could add a linear magnetic scale.

I like it!

Jason

--

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/19857817-be87-4f5e-8acc-7645f7244322n%40googlegroups.com.

[Jonathan Oxer]

== Speeed: I'm currently only getting about 1200 CPH, for a few reasons.

1. The machines are not mounted on solid frames. Even worse, I put them on bases with castor wheels originally to make it easier to move them around, and I never took them off. If I run them at more than about 50% speed they shake themselves around.

2. The ELP cameras take a while to settle. This is one of the biggest speed killers. It has to pause at the camera for every placement.

3. I removed the quad camera array and put in place a single camera, because I couldn't figure out a way to ingest the video from the original cameras. This means it has to move, settle, and take an image for every nozzle individually instead of imaging them simultaneously.

4. I mostly use just one nozzle. Most of my placements are 0603 and I have different sized tips on the other nozzles. I could optimise this by putting the same tip on multiple nozzles. Most of the time the machine is acting as if it's single-nozzle, because the oddball size parts get loaded in parallel with the first 0603 parts and then the rest of the job is just placing 0603 using nozzle #1.

If I cared enough to make it faster, I'm sure it'd be possible to get closer to 5k CPH by running at 100% speed, tuning or replacing the camera, tuning my nozzle loadout, etc. However, my use-case isn't volume manufacturing, it's small runs of typically 8 - 10 boards or panels at a time. The time the boards spend in the machine is tiny compared to the time taken to paste them up, load manual parts, go through reflow, etc. Even at just 1200 CPH, the machine is rarely the bottleneck.

== Z-axis: the 4-nozzle machines use a rocker cam mechanism.

From what I understand all Kayos from 6 nozzles up use individual steppers on each Z-axis. However, the A4 uses a double-sided cam system with one stepper per 2 nozzles. This has several implications:

1. The Z-axis is non-linear, and the Kayo protocol expects you to tell it the number of steps and NOT the absolute position. In my driver code I do some compensation for this. It's not perfect, but I've found it to be Good Enough(TM) for actual use.

2. OpenPnP needs to be configured as you would with a Peter's Head setup, with paired Z-axes that track one going up while the other goes down. This works just fine.

With a 6-nozzle machine you may still have to care about point #1 if it uses a cam system, but you shouldn't have to worry about point #2.

Cheers

Jon

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/7849f681-403f-4c04-8561-758a2d421dc6%40app.fastmail.com.

[Mike Menci]

Jon - you could consider swapping your wheels with such:  https://a.aliexpress.com/_Ev0rP6V

than your KaYo"s will dance only when you Want them to dance....
Mike

[Mike Menci]

6 x Z is easier with the belts... 

[mark maker]

> The ELP cameras take a while to settle. This is one of the biggest speed killers. It has to pause at the camera for every placement.

Can you post details? If you really talk about "settling" per se, then if the ELP is setup correctly (all "auto" settings off),  it is not the camera that slows this down, but the mechanics, i.e. the nozzles swinging. Replacing the camera will then not help.

https://github.com/openpnp/openpnp/wiki/Camera-Settling

https://youtu.be/Pxg6g3KI5_E

> The Z-axis is non-linear, and the Kayo protocol expects you to tell it the number of steps and NOT the absolute position. In my driver code I do some compensation for this. It's not perfect, but I've found it to be Good Enough(TM) for actual use.

Use Issues & Solutions to select the right nozzle configuration. It does support all the usual cam configurations, i.e. it will transform the cam to linear for OpenPnP. Also note the Number of Nozzle Units spin control at the top to make two pairs for your machine. Enable Include solved and then Reopen the solution

https://github.com/openpnp/openpnp/wiki/Issues-and-Solutions#welcome-milestone

_Mark

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/CADBU0eceQjyJeqWimsxesU3tSAz%2BUhCgg_FSu5AuuGqHQ-0CMQ%40mail.gmail.com.

[Trevor White]

Thanks for the detailed information, Jon. That was extremely useful. 

I decided to buy a HWGC HW-T4-50F machine in the end. From what I have read I think the Kayo was an older design from HWGC. I think the Kayo machine has better hardware spec with Panasonic servos and ground screw than the current HWGC T4 machine. 

What I am now wondering, more as a curiosity at this point is if the electronics in the two makes of machine are the same. Do you have any photos of the circuit boards inside the A4? It might be that the OpenKayo also works with HWGC machines or at least is very similar. 

There is a 2 nozzle version ( HW-T2-40F ) which is also sold as a SMT220. This looks to be a very good spec hardware wise but it does not have tool changer options. This machine might be a nice option for openpnp builds as it supports CL feeders, ball screw and servo drives.  I think its placement accuracy will be fantastic.

This is the machine running the native HWGC software. It's pretty fast. 

https://www.youtube.com/watch?v=jJ1C7fJ4rYw