Robot Integration - Looking for Experience and advice.

Kp61dude! should be able to help.

Whilst I've not developed such a system myself I imagine it well within my capabilities.

I had first hand experience with such a system (developed by someone else) in a UK based capacity (it was an Italian guy who created it) and I was Hex's VB guy at the time so they wanted me to be familiar with it in case there were issues.


The premise was there was a 'Supervisor' piece of software which managed the whole operation.

Parts were palletized and the pallet was transferred from the machine tool to the CMM. There was a RFID chip on each pallet which held the part specific data (such as Op number, measurements/offsets) - I liked this idea as there was no chance of incorrect data being used for the wrong part.

The RFID chip was offered up to readers on the CMM & machine tool before and after being loaded.

I've developed systems using a PCI card before for I/O, but I imagine there are USB options now which are probably easier.

Did all the research, had product demos performed for us, didn't pull the trigger.

A fixture plate on the CMM, good training, and detailed setup pictures allow our inspectors to use the CMM's programs that inspect full "pallets" of parts without having to buy an expensive robot or its associated hardware/software maintenance costs.

We have a robot that "works" in our lathe dept, it only gets fired up when we're walking customers around and they go "ooo that's cool".

We have a cell with 2 CNC lathes that run the exact same part. Can 1 cmm with an automation controller be set up to offset both lathes? I suppose it just needs to know which machine the parts are coming from? Lathe A or Lath B, Or is that something to avoid?

We have a cell with 2 CNC lathes that run the exact same part. Can 1 cmm with an automation controller be set up to offset both lathes? I suppose it just needs to know which machine the parts are coming from? Lathe A or Lath B, Or is that something to avoid?

Shooting from the hip....
Think that one of the main CMM issues would be how to know which part came from which lathe. Maybe use a pallet with a distinguishing feature that could be quickly measured? Some kind of block with different heights that could have a vector point touch with a generous prehit/ retract. 1 mm high block = Lathe A, 2 mm high block = Lathe B

Kp61dude! should be able to help.

I never got notified about this or I must've missed it.

Yeah where do I start.....

I had success at what you're about to embark on. Thanks to some very, VERY important circumstances and they are:
1. many mistakes from the integrator
2. dude in charge of our project internally is a jack-a-s-s
3. the software/hardware used

this is such a deeeeeep discussion and I will only go into some detail here because ya know I got a job to pretend to do.

"BLUF, we are going to set up a robotic cell without a professional
integrator. Not going to pay $$$. Going to learn as we go. We'll buy
the hardware/software but its going to be google and youtube for the
most part."

You'll save a ton of money. When I started giving this a go I
worked off a PO list from the integrator that was 3 times my
salary and I was able to do in about a month. Wow! I should start
my own business lol! The google/youtube part is going to be very
dry but there are a few must see. All I can say is good luck and
hats off to you guys for giving this a try. You don't hear of many
shops doing this for one they hang on to it as secrets (totally
understandable) and two most engineers have zero time. Oh and
there will be some money involved a little less or a little more
depending on who you end up doing business with.

Has anyone here ever implemented a production cell with a CNC, CMM,
and a robot? Either professionally or own their own? I know there are
professional integrators out there but we are more like do it
ourselves kind of company. I don't like the idea of reinventing the
wheel but its not up to me. Even more so due to the state of the
economy and industry.

You're lucky your boss is backing this... you will have many
sleepless nights and weekends where they just won't end quick
enough to get in the office to implement something you thought of
during your hunting/fishing trip. You will have a ton of fun in
the process and loose a bunch of hair too.

I don't have experience integrating from the ground up but I have
been involved with a poorly integrated cell that I've figured out
how to improve and sometimes that required some complete tear-downs
and buildups.

Our setup consists of:
* 2 CNC machines, 5 axis, probing
* 1 robot, 6 axis, 8 detachable end-effector tools
* 2 inventory shelf towers
* 1 parts washer
* 1 printer for part marking
* 1 CMM machine with all the bells and whistles
* many part fixtures for multiple machining operations

Cell handles high mix, low volumes, with 'tight' tolerances.

He wants us to start a small scale project specifically using a UR. I
have a work station that makes only 1 part. Its small, about 2 inches
round, total of 5 dimensions, and high volume. CNC is a lathe and part
goes in as a bar and is completely machined when done. Going to try
and implement a robot for machine and cmm tending and automatic
machine offsets from CMM to CNC. Eventually lights out, email or text
notification for issues, full process control, etc...

Offsets from CMM to CNC are achievable but take a ton of time
after setup and still a lot of checks and balances have to be made
it's less of a headache to just utilize the probe in the CNC
more. It's not impossible to set it's just a long process to have
to do for each part you run though your cell, not good for high
mix, low volumes.

Talking to my sales rep, I need to have an automation controller. Some
of my cmms can be upgraded here onsite but my 454s would have to be
sent back to the factory.

This is true. We had ours upgraded onsite (not a 454).

How do you send measurement feedback to a CNC? I currently send
measurement data to a .csv file on our network. Could I set up my CNC
to read the .csv for tool offsets? Would that be considered
communication? Do I absolutely need to have a specific software.

Sending offsets greatly depends on your CNC controller but most
likely will be over OPC-UA. I wish our machines ran their controls
software off a real PC like some machines do, if you come across
this decision consider it.

I suppose I would need a master controller or system monitor to queue
when its safe to move the robot or operate the CMM. Is there a
specific software out there to use? I think I've read on the forum
where some users created their own font end application/user
interface. So something needs to control when to start and stop the
CNC, start and stop the robot, and start the CMM.

We call ours a PLC. There are many brands in the game. We
Americans are somewhat weird in this field IMHO. Never mind this
discussion but the game has been changing the last few years and
Twincat 3 is gaining popularity and for obvious reasons. Beckhoff
Twincat 3 development environment is free and you only pay for
what is needed in the end for the production version. Oh and they
have an API similar to what Pc-Dmis has for automation/scripting
and that's free also. Popular US brands will cost you thousands of
doll-hairs in development software alone never mind the high costs
of their hardware.

Any ideas, good/bad experiences, mistakes, pitfalls, horror stories
are welcome. If this is accomplished, I will gladly share the steps I
did to achieve this.

These are complicated set-ups. Running lights out won't happen
overnight. There is a bunch to gain and lose first but like
anything else... with patience, and all that other shtuff needed
to succeed will be needed along with a shed load of coffee.

This is a great idea to do in-house but will require a strong piggy bank
but not an outrageously big one.

Kp61dude! : I'm perfectly integrated at work, don't cost too much and have a lot of great ideas... But I'm not a robot !
Also, I have some doubts about my set up !

funny how your AI is learning humor.... robot!

Whilst I've not developed such a system myself I imagine it well within my capabilities.

I had first hand experience with such a system (developed by someone else) in a UK based capacity (it was an Italian guy who created it) and I was Hex's VB guy at the time so they wanted me to be familiar with it in case there were issues.


The premise was there was a 'Supervisor' piece of software which managed the whole operation.

Parts were palletized and the pallet was transferred from the machine tool to the CMM. There was a RFID chip on each pallet which held the part specific data (such as Op number, measurements/offsets) - I liked this idea as there was no chance of incorrect data being used for the wrong part.

The RFID chip was offered up to readers on the CMM & machine tool before and after being loaded.

I've developed systems using a PCI card before for I/O, but I imagine there are USB options now which are probably easier.

The "Supervisor" software is what we are using to "talk" to the CMM from the cell PLC. Parts are kept track of in a SQL database that is running in the PLC, it knows all the information about the part. The cell processes 12 parts at the same time with each part having 3 operations, robot flips parts to other pallets depending on its operation. There are 3 different types of pallets and 12 total with 14 stations with active clamping. It's a monster!