Yamaha YV100 conversion to OpenPNP.

[Andrew Frazer]

Well, the Yamaha YV100's that i bought at the auction have been picked up from LabTams old premises and will be delivered to a top secrect location in Melbourne Australia, for a few weeks, until i get there, where i'll be preparing and crating them up for shipping to New Zealand.  I'm the meantime been doing some digging and have discovered that pretty much all of the IO to run the machine is exposed. 168 ins and 190 outs approx!       This sets up the possiblity of the conversion to OpenPNP a much more realsitic proposition.  

These machines are rock solid meachnically, but are running 1990's styled vision..    Swapping them over to a more modern control system and vision will turn these into extremely capable machines..      

While i think i proably got a real bargin,    i have seen similar machines for sale for 10-15k..  However there may well be broken machines that are much cheaper, and it woudl seem we probalby only the arms,legs of the thing, not the brains/eyes.       I'd say it woudl be realistic to get a failed machine in the region of $4-5k..   Clearly this is not a machine for home use, it needs 3phase power and weighs 1500kg..   

Since i have two of them, i'll probably be able to run one in "Yamaha" mode, while i tinker with the other one..     This was'tn the path i was headed down, but its looking like the possiblitys are really good!

[Andrew Frazer]

Quite a lot to digest.

[Wireb]

Oooo that thing has AC server motors in it. If the drivers are integrated into the part your getting rid of they may be fun. Even if the drivers stay and they use a non step/dir or other standard input may still be fun :)

[Wireb]

*servo motors

[Cri S]

Exactly what models you have? Just yv100, there are many different model with different drivers
And from 2 to 8 heads with or without cyberoptics. You to know the subversion. Not only the frame. Cameras are global shutter with special lenses for downlooking, I would not change it only to get more resolution penalizing speed.

[Andrew Frazer]

yv100-2e I think.

[el...@pico-systems.com]

On Friday, September 30, 2016 at 1:28:47 AM UTC-5, Andrew Frazer wrote:

Well, the Yamaha YV100's that i bought at the auction have been picked up from LabTams old premises and will be delivered to a top secrect location in Melbourne Australia, for a few weeks, until i get there, where i'll be preparing and crating them up for shipping to New Zealand.  I'm the meantime been doing some digging and have discovered that pretty much all of the IO to run the machine is exposed. 168 ins and 190 outs approx!       This sets up the possiblity of the conversion to OpenPNP a much more realsitic proposition.  

I will be watching this closely.  I have a Philips CSM84, which was actually made by Yamaha.  I think the equivalent Yamaha model is YM84.  The machine I have has 3 nozzles, no Z axis, no vision.  It runs fine, I have had minimal problems with it.  One reflective sensor on the conveyor went bad some years ago, and occasionally one of the nozzle air/vacuum valves sticks.  That's about it.  But, I do worry that the ancient 286 computer board will fail some day, and I'll be in big trouble.

My big concern with OpenPNP or any other retrofit is to get the same level of error recovery that the existing CSM84 has, which is a LOT!  It automatically retries parts 4 times if it doesn't get good vacuum after closing the centering jaws.  It allows you to resume building the same board after clearing a whole raft of problems.  There are a few things that get by the checks it does before starting the first board and are not restartable.  In fact, and this one amazes me, it can apparently be powered down in the middle of building a board and then finish the board after powering back up!  So, I've done thousands of boards on mine, and only one time had to sweep all the parts off and start over, when I got the machine to exceed the travel limits.

Of course, if I ever did retrofit this machine, I'd add vision to it!

Jon

[Andrew Frazer]

The nice thing is that since i have two machines, theres a good chance i use one and play with one.   

My big fear is that the controller boards will fry and getting parts might be troublesome.

[Cri S]

I think it's a lot of works , for example clamping, PCB pusher, movable toolpin, pressure plate, smema and conveyor interface incl. Sensor.
Programming safety keepout positions, and I think max speed will be 30% from actual speed, where programming with pcbsynergy is probably easier and faster. As max component height is 6.3mm,
do you not need more or what PNP do you have planned for higher component and pitch below 0.5mm? The speed reduction is principally because of taking image of single components sequentially and waiting for that instead of taking picture during movement. Further using polling instead of queuing for motions commands.

Why not buying old vintage win95 PC with is a slots and running some tsr utility to slow down and sleep CPU before these utilitys disappears.
Prices for such PC is low. Further USB floppy disk emulator is nice, but floppy disk on USB is another viable possibility, just hard to find this time at least in Europe.

[Andrew Frazer]

>I think it's a lot of works ,

I think its a big challenge. Just the kind of thign that keeps me awake in the middle of the night!  Love it.

 

for example clamping, PCB pusher, movable toolpin, pressure plate, smema and conveyor interface incl. Sensor.

Yes, identified some 168 IO's so far.. Quite a few.   

do you not need more or what PNP do you have planned for  higher component and pitch below 0.5mm? The speed reduction is principally because of taking image of single components sequentially and waiting for that instead of taking picture during movement. Further using polling instead of queuing for motions commands.

I am quite confident that the mechanics of the machine will deliver good placement at tighter pitch than 0.5mm.  Remember that the vision and processing on these things is 1990' tech.    The mass'es of IO boards will be easily replaced by a couple of FPGA's and micros..   

 

Why not buying old vintage win95 PC with is a slots and running some tsr utility to slow down and sleep CPU before these utilitys  disappears.
Prices for such PC is low. Further USB floppy disk emulator is nice, but floppy disk on USB is another viable possibility, just hard to find this time at least in Europe

Becuase al those things just end up being a painful thing too look after..   This proect is as much about the challenge as the end result.   If I 'fail', i will ahve failed learning a lot.  If i 'suceed' then i'll ahve some thing i can share as well.   There is not a lot to loose. 

If i bought a machine off the shelf ( which i would do if i was settign up a factory ) we would be having a different disussio.

 

.

[el...@pico-systems.com]

On Friday, September 30, 2016 at 10:07:31 PM UTC-5, Andrew Frazer wrote:

The nice thing is that since i have two machines, theres a good chance i use one and play with one.   

My big fear is that the controller boards will fry and getting parts might be troublesome.

Yes, exactly!  On the CSM84, it has what is essentially a PC-AT (80286 CPU) on an 18" square board.  Your machine, if it has the AC servos, may be just a few years newer.
Worse, the service manuals do NOT have schematics of the main CPU board.  So, if it croaks, it would be fairly hard to diagnose and repair.  I have plenty of test gear here, but without the schematics, it would be fairly hard to fix it, unless I got real lucky.  The rest of the machine has detailed schematics.

But (knock on wood) it is still perking along.

Jon

[el...@pico-systems.com]

On Friday, September 30, 2016 at 11:31:52 PM UTC-5, Cri S wrote:

I think it's a lot of works , for example clamping, PCB pusher, movable toolpin, pressure plate, smema and conveyor interface incl. Sensor.
Programming safety keepout positions, and I think max speed will be 30% from actual speed, where programming with pcbsynergy is probably easier and faster. As max component height is 6.3mm,
do you not need more or what PNP do you have planned for  higher component and pitch below 0.5mm? The speed reduction is principally because of taking image of single components sequentially and waiting for that instead of taking picture during movement. Further using polling instead of queuing for motions commands.

Yes, if you run one and see all the functionality that the software provides, you realize how well it was designed and made to make the operator's job easy.
For instance, at any time, you can hit stop, go into manual mode and move things around, change the feeder and placement setup and thn go back and finish assembling the board that was interrupted.  Pretty handy when something doesn't work right on the first article.

Why not buying old vintage win95 PC with is a slots and running some tsr utility to slow down and sleep CPU before these utilitys  disappears.

An interesting concept would be to rig up an ISA PC or maybe even a Beagle Bone and tap into the I/O controller so you could run the original software and hardware with a replaceable CPU.  But, that would be a LOT of work.  Note that while at least my CSM84 has an 80286 CPU that is pretty much a PC-AT on a board, it is not compatible with ISA or any other PC interfaces (other than the serial port.)

Jon

[SMdude]

Probably the most common cause of failure with an old of motherboard is from the battery in the real time clock ic leaking and damaging the board. It would probably pay to get rid of the battery(clock only advances while the machine is on) or make it so it is externally mountable.
The other common fault was ram.

[Andrew Frazer]

Well, first job done. the two machiens have been safetly extracted out of Labtam and are in a safe warehouse so i can get them ready to export to NZ in a couple of weeks.

[Mark Harris]

Nice work. Can I have that mad gantry router in the background?

--
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+unsubscribe@googlegroups.com.
To post to this group, send email to ope...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/1f9d9644-9f6c-46d3-a4fa-138ce5fbe80b%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

[Andrew Frazer]

its not just a router, its also a printer, ( foam, wax, ).. work envelope starting at 2.0 x 4.0m, 2.0m and if you want to make it longer, you can..   It terms of "can you have it".. Thats what our "big brother" company does.. Build these thigns.

On Monday, October 3, 2016 at 6:12:11 PM UTC+13, Mark Harris wrote://

Nice work. Can I have that mad gantry router in the background?

On 2 October 2016 at 22:33, Andrew Frazer <andrew...@stellascapes.com> wrote:

Well, first job done. the two machiens have been safetly extracted out of Labtam and are in a safe warehouse so i can get them ready to export to NZ in a couple of weeks.

On Friday, September 30, 2016 at 7:28:47 PM UTC+13, Andrew Frazer wrote:

Well, the Yamaha YV100's that i bought at the auction have been picked up from LabTams old premises and will be delivered to a top secrect location in Melbourne Australia, for a few weeks, until i get there, where i'll be preparing and crating them up for shipping to New Zealand.  I'm the meantime been doing some digging and have discovered that pretty much all of the IO to run the machine is exposed. 168 ins and 190 outs approx!       This sets up the possiblity of the conversion to OpenPNP a much more realsitic proposition.  

These machines are rock solid meachnically, but are running 1990's styled vision..    Swapping them over to a more modern control system and vision will turn these into extremely capable machines..      

While i think i proably got a real bargin,    i have seen similar machines for sale for 10-15k..  However there may well be broken machines that are much cheaper, and it woudl seem we probalby only the arms,legs of the thing, not the brains/eyes.       I'd say it woudl be realistic to get a failed machine in the region of $4-5k..   Clearly this is not a machine for home use, it needs 3phase power and weighs 1500kg..   

Since i have two of them, i'll probably be able to run one in "Yamaha" mode, while i tinker with the other one..     This was'tn the path i was headed down, but its looking like the possiblitys are really good!

--
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 post to this group, send email to ope...@googlegroups.com.

[el...@pico-systems.com]

On Saturday, October 1, 2016 at 5:38:37 PM UTC-5, SMdude wrote:

Probably the most common cause of failure with an old of motherboard is from the battery in the real time clock ic leaking and damaging the board. It would probably pay to get rid of the battery(clock only advances while the machine is on) or make it so it is externally mountable.
The other common fault was ram.

My CSM84 has a bigger battery, it supports the static RAM that holds the machine settings, the current feeder and placement files.  There is no real time clock, as far as I know.
It does keep "delta time", ie. it knows how long the machine has been on and adds it to the run time clock.  This is a NiCd rechargeable battery.

The DRAM is soldered into the main board, hope I don't have trouble with that.

Jon

[Andrew Frazer]

 

Well, the two Machines are on the floor and have power on them.    A few little jobs needed doing to get them running again.  The Monitor cables had been cut but that was an easy fix and i've replaed the old CRT's with new LCDS.

The machines have VIOS 1.22 on them.      VIOS is Yamahas Operating system, and wow its a dinosaur.  Its complex and confusing.

 

I got it to run its warm up

https://www.youtube.com/watch?v=Gezfs-ZrA_Q 

[Jon Elson]

On Sunday, January 8, 2017 at 12:50:38 PM UTC-6, Andrew Frazer wrote:

 

Well, the two Machines are on the floor and have power on them.    A few little jobs needed doing to get them running again.  The Monitor cables had been cut but that was an easy fix and i've replaed the old CRT's with new LCDS.

The machines have VIOS 1.22 on them.      VIOS is Yamahas Operating system, and wow its a dinosaur.  Its complex and confusing.

 

I think once you find out all the smarts that are in the machine to handle errors, you will take a long second think about conversion.  Yes, i worry about my CSM84 failing in such a way that I can't fix it myself (mostly, major breakdown in the main computer board).  But, I am a member of the google group openPnP, and they have a LOOONG way to go to get anywhere near what capability my old CSM has.

Assuming it is still much the same as my CSM, there are two major files you need to set up for each job.  One is the placement file, it lists the board size, fiducials, and then the components, one per line, with feeder location, head, and X, Y and rotation.  The other is the feeder file, and it tells what size feeder is in each location, pick up orientation, coordinates if a non-standard feeder or more info if a tray.  I saw your machine has vision, so there is another file that specifies the pads and other info for the vision system to properly align the part.  I don't have vision on my system, so I don't know much about that.

I wrote some small C programs to take my CAM system's output and turn it into the placement file.  I still do the feeder file manually, usually on the machine, but then save it on a PC.
You can find these files in the files section of this group.  Another member added a bunch more functionality to these programs, but I've forgotten where he put the source.

Jon

[Andrew Frazer]

Assuming it is still much the same as my CSM, there are two major files you need to set up for each job.  One is the placement file, it lists the board size, fiducials, and then the components, one per line, with feeder location, head, and X, Y and rotation.

Ok, i've been able to create that placmeent file ok, using pcbsynergy.   ( http://members.iinet.net.au/~sarason/ ) which is a pretty handy bit of code.

>The other is the feeder file, and it tells what size feeder is in each location, pick up orientation, coordinates if a non-standard feeder or more info if a >tray.  I saw your machine has vision, so there is another file that specifies the pads and other info for the vision system to properly align the part.  I >don't have vision on my system, so I don't know much about that.

yeah its got three cameras in it.     How do you create this 'feeder' file and import it?

 

I wrote some small C programs to take my CAM system's output and turn it into the placement file.  I still do the feeder file manually, usually on the machine, but then save it on a PC

How do you export any files off the machine..  I cnat' seem to do that.

You can find these files in the files section of this group.  Another member added a bunch more functionality to these programs, but I've forgotten where he put the source.

Thanks. I'll have a look.

[Jon Elson]

On 01/10/2017 02:23 AM, Andrew Frazer wrote:

Assuming it is still much the same as my CSM, there are two major files you need to set up for each job.  One is the placement file, it lists the board size, fiducials, and then the components, one per line, with feeder location, head, and X, Y and rotation.

Ok, i've been able to create that placmeent file ok, using pcbsynergy.   ( http://members.iinet.net.au/~sarason/ ) which is a pretty handy bit of code.

>The other is the feeder file, and it tells what size feeder is in each location, pick up orientation, coordinates if a non-standard feeder or more info if a >tray.  I saw your machine has vision, so there is another file that specifies the pads and other info for the vision system to properly align the part.  I >don't have vision on my system, so I don't know much about that.

yeah its got three cameras in it.     How do you create this 'feeder' file and import it?

Well, at least the Philips CSM can create it ON the machine, by entering the DATA IN / components
menu.  Then, you just scroll down the list of feeder locations and select feeder size, pick up angle, and a bunch of other options.  This file (FDR) can then be saved to a PC for later use.

Then, since your machine has vision, there will be another option to enter the vision menu.  This has a kind of graphical editor where you can describe the component by type, number of pins, pins pitch and row spacing, etc. and then it will present a picture of the part, especially the leads.  At least, that is how it works on the CSM, I'm guessing your machine may have similar utilities built in.  This file can also be saved to a PC.

The placement file can also be saved/reloaded to/from a PC.  The file name is of the form
B<ABCDE>
where ABCDE is the name of that file in the RUNNING menu.

I use a terminal program to load and save the files in plain ASCII.  So, to save the feeder file, the command
is :
@READ FDR

To download the feeder file, it is :
@WRITE FDR

The save the placement file :
@READ B<ABCDE>

To load a placement file :
@WRITE B<ABCDE>

There are similar commands for the vision file.

 

How do you export any files off the machine..  I cnat' seem to do that.

My machine only has an RS-232 serial port.  You set up the comm parameters in the CMU menu.
Not sure if your machine may have something newer than serial.  Anyway, you mostly set up a 3-wire null modem cable between the P&P and a PC serial port, and use a serial comm program that has file save and file send capability.  (I use minicom on a Linux system, there are similar things for Windows OS's.)

It sounds like you do not have the manual for your machine.  That is like flying blind!  I have several generations of the manuals for my machine, the first version is a REALLY bad translation and organization job, it is clear they rewrote it several times.  No idea how good or bad the Yamaha manuals might be, I've had bad luck with a few Japanese manuals that were NOT written by native English speakers well versed in the field.
Hopefully, you can get hold of a copy of a manual for it.  If Philips/Assembleon sold a version of the machine under their label, such a manual might be easier to find.

Jon

[Andrew Frazer]

On Wednesday, January 11, 2017 at 6:03:36 AM UTC+13, Jon Elson wrote

My machine only has an RS-232 serial port.  You set up the comm parameters in the CMU menu.
Not sure if your machine may have something newer than serial. 

Yeah, newer.. 3.5" Floppy Disks!

Anyway, you mostly set up a 3-wire null modem cable between the P&P and a PC serial port, and use a serial comm program that has file save and file send capability.  (I use minicom on a Linux system, there are similar things for Windows OS's.)

 

It sounds like you do not have the manual for your machine.

I've got hte 'operation' manual and the 'maintaince manual', but no reference manual.        Theres big hunks of 'knowledge' missing thats for sure.

[Mark Harris]

Are these running DC servos? If so, feel free to take the schematics and/or board design of what I made for my Siemens machines :)

--

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+unsubscribe@googlegroups.com.

To post to this group, send email to ope...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/0c1bcb3d-78a5-42f8-bf0b-878285fb6ee0%40googlegroups.com.

[Andrew Frazer]

They are DC Servos' but they are high voltage ones; 200V

On Wednesday, January 11, 2017 at 9:19:01 AM UTC+13, Mark Harris wrote:

Are these running DC servos? If so, feel free to take the schematics and/or board design of what I made for my Siemens machines :)

On 10 January 2017 at 11:01, Andrew Frazer <andrew...@stellascapes.com> wrote:

On Wednesday, January 11, 2017 at 6:03:36 AM UTC+13, Jon Elson wrote

My machine only has an RS-232 serial port.  You set up the comm parameters in the CMU menu.
Not sure if your machine may have something newer than serial. 

Yeah, newer.. 3.5" Floppy Disks!

Anyway, you mostly set up a 3-wire null modem cable between the P&P and a PC serial port, and use a serial comm program that has file save and file send capability.  (I use minicom on a Linux system, there are similar things for Windows OS's.)

 

It sounds like you do not have the manual for your machine.

I've got hte 'operation' manual and the 'maintaince manual', but no reference manual.        Theres big hunks of 'knowledge' missing thats for sure.

 

  That is like flying blind!  I have several generations of the manuals for my machine, the first version is a REALLY bad translation and organization job, it is clear they rewrote it several times.  No idea how good or bad the Yamaha manuals might be, I've had bad luck with a few Japanese manuals that were NOT written by native English speakers well versed in the field.
Hopefully, you can get hold of a copy of a manual for it.  If Philips/Assembleon sold a version of the machine under their label, such a manual might be easier to find.

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 post to this group, send email to ope...@googlegroups.com.

[Mark Harris]

The HBridge IC is good for 600V on my one, the FETS are only rated for 150 though, easy enough to swap out for 300v fets though - the big 

caps will need a swap too... the board's isolation is good for about 600v too, if you disregard the mask - well over 1000v with soldermask :)

Mine is essentially a higher voltage/amperage version of Gino's original: https://www.crowdsupply.com/citrus-cnc/tarocco

I figured the fastest conversion process for the pnp would be to get everything on step/direction inputs :)

To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+unsubscribe@googlegroups.com.

To post to this group, send email to ope...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/f72f2c86-42fb-479b-82bc-7b4ae126a48d%40googlegroups.com.

[Glen English]

Actually they are 3 phase Mitsubishi servomotors 1kW or 750W .

the power supply makes 280VDC approx, by a STAR recifier, soft start capacitor fill etc . it can operate from single or three phase.

that is the far right power supply.

this is fed to the driver board that turns it into 3 phase 200V per winding approx

open the bottom two doors at  the front, remove all the cables from their connectors and slide out the control box.

do not yet undo front screws to pull cards out ! they are connected...

remove the four screws that hold the lid on. then , you will see....

EVERYTHING in the machine has an encoder... the main XY encoders are high count like 2500 ppm.....

[Andrew Frazer]

I have discovered that what you are talking about is a much earlier version of the software,    

Interestingly i'm going to have a look at the serial ports and see if i can talk to those.

Have probed all the pnematics now, the feeders will be easy to control.

[Jon Elson]

On 01/11/2017 05:20 PM, Andrew Frazer wrote:
>
> I have discovered that what you are talking about is a
> much earlier version of the software,
>

Yes, for sure, my CSM is about 15 years old, I think. But,
I think some of the major parts of the programming are still
very generally the same.

Jon

[Andrew Frazer]

does your system use VIOS?

[Andrew Frazer]

Vios TXT files jsut are a single file, that contains all the data, you were talking about somethiung that creates two files.

[Jon Elson]

On 01/11/2017 11:04 PM, Andrew Frazer wrote:
>
> does your system use VIOS?
>

No, apparently it might be called UFOS, although the only
place that appears in the manuals is
the last chapter, UFOS file specification, that details what
each field in all the files (placement, feeder, machine
setup, etc.) do.

The fancy video board is called a VICS1000, but that is
likely different. I don't have any video on my machine.

Jon

[Jon Elson]

On 01/11/2017 11:05 PM, Andrew Frazer wrote:
>
> Vios TXT files jsut are a single file, that contains all
> the data, you were talking about somethiung that creates
> two files.

OK, on the CSM there could actually be 4 files. There is a
machine setup file (MCH) a feeder file (FDR) and a placement
file (named as you wish), and if you have vision then there
is also a vision setup file that defines you to measure
centering and orientation of the parts.

The CSM is the only machine I know. It was sold by Philips,
but made by Yamaha.

Jon

[Andrew Frazer]

have found out that my YV100's at some stage would have been running code that you are talking about. The conversion ultity i have found on the disks that were with this machine, creates those files. after talkign to the local Yamaha Engineers about this,  i got filled in on the story.

Hoping that shortly we can find or locate a copy of cad2cad.

THis will be stepping stone on the wya to doing a OpenPNP Conversion..   Theres so many of these yamaha machines still lurking aroudn and they are really solid!.     Swapping to OpenPNP will let me run image chips and packages that the system currnetly doe'snt know. but doing all teh checking will be a big project.       Things like having the machine automaticaly compenstate for twist/bow by determinign the fiducals.

For a machine of its vintage, its Image recognistion and processing was really leading edge.

I'm told that the going price for these machines in similar state to the ones we have is between 20 and 25k USD.   So i got a bargin for what i pay for them.    And Glen English got an even better bargin. And now i regret i did'nt just buy all three machines.

[Andrew Frazer]

So the machine is up and running now, thats a great starting point.

https://www.youtube.com/watch?v=96LbwYYobcE

[Glen English]

fark, it's fast with the vision.....

wow

nice one andrew

[Jon Elson]

Great! Now, you need to compare what it does to OpenPnP.
Especially, see how well it handles mis-picks, sticky
feeders and the like. If you have boards with 100+ parts on
them, you don't want to have to restart the whole board due
to one mispick from a feeder.
(Not sure how well OpenPnP does with this, but I know it is
not quite as capable.)
I've run thousands of boards through my old Philips CSM84,
and there is only ONE time I was not able to complete a
board after a problem. (Crazy error message that wasn't
even in the book, I had to GUESS what the message meant.)

Jon