Matsuura pics

Dang - that's a complicated machine. What's the Mean Time Between Failures - an hour or two? You know, with n parts, each of which has a probability P of failure in the next hour, the probability that at least one of the will fail in the next hour is 1-(1-P)^n. E.g., 1000 parts & .01% (10,000 hours MTBF) each is 10% overall.

You know, you don't have to go to a lot of trouble to rotate a picture - Windows Picture Viewer ("Preview") will do it. My neck is sore from twisting to see them.

Bob

I'll look at them in more detail in the next day or two. I see that your ATC is not the basic carousel style used on the somewhat smaller machines. Looks like you'll be needing at least four outputs and about eight inputs to control it, and a bit more scripting for the tool change routine. Nothing overly complicated, just a bit more than the basic carousel type. On that relief valve I believe that the hex nut seen just past the black knob is the lock nut which prevents the setting from changing due to vibration. You'll likely need to back that off a good half to one turn before the knob will turn.

...

Existing ATC has WAY more I/O than that. About 18 outs and 24 - 30 ins. I could reduce ins considerably with two encoders. But, that's next year's problem.

I'll try again on that relief valve. With two combination wrenches, I couldn't get these to separate. The whole vave end will easily screw out though. You can't see from the pic but that 8x8 hole in front of the relief valve opens up to a cavern that is the whole inside of the machine and coolant sump. That's where the spring and ball would go if I take it apart.

Karl

Good grief, are you sure? Looking at it I can't see where they could be using that many I/Os.

You might need to hold the round part with vise grips while loosening that lock nut if the locknut is tighter than the round part. If you do take it apart, duct tape a gal zip lock bag around the area and remove the part inside the bag.

I've only spent a few minutes with the ins going to the ATC. Just enough to know there's no need to wire now. But I did count 30 and have to leave that much room on the input panel for later. The outs are connected and the solenoids all click, Here's a note listing:

relay green tape power panel - coolant

110- SOL 3R Key lock (Right) 110- SOL 7 Shift H 110- SOL 8 Shift L 110- SOL 11R Drum Revolution CW(Right) 110 SOL 12R Drum Revolution CCW (Right) 110 SOL 13R Drum Revolution Rapid (Right) 110 SOL 14R Drum Pin Out (Right) 110 SOL 9 Oil Mist 110 SOL 10 4 axis clamp 110 SOL 1R Arm Forward (Right) 110 SOL 2R Arm Down (Right) 110 SOL 4R Arm Swing (Right) 110 SOL 5R Unclamp (R ) Keylock (L) SOL 3 (L) 110 Drum Revolution CW ( L) SOL 11L 110 Drum Revolution CCW ( L) SOL 12L 110 Drum Revolution Rapid (L) SOL 13L 110* Arm Forward (L) SOL 1L 110 Arm Down (L Sol 2L 110 Arm Swing (L) Sol 4L 110 Unclamp (L) SOL 5L 110 hydraulic pump Spindle Brake 110 volt Or Z axis clamp

Look to the bottom of the IO panel. Its the Opto with all 24 red wires connected. I'm OUT OF OUTPUTS! The 72 inputs look to all be spoken for also.

I'm working on a trick to get 16 more outs from the two galil boards at the very bottom of the IO panel. Don Foreman did the design work on my last CNC mill when I ran short on that one. Generation 2 will be a slight improvement. Right now, I'm only a couple short that I know of

- oil cooler and air blast, but there will be more.

Karl

Ok, some of those connections look like they have nothing to do with the ATC, such as the 4th axis clamp output, shift H/L outputs and the hydraulic pump output (since hydraulics needed for spindle counterbalance also).

I see this machine actually has two separate ATCs. The obvious answer there is to just use one initially and ignore the other, cutting your I/O needs in half. You're not doing production stuff, right, so I can't see where you would ever really have a use for the second spindle and ATC.

Presumably "drum" is referring to the tool carousel. Since you're not doing production, you can probably dispense with both the CCW rotation and rapid rotation outputs, and simply use one CW rotation output. This of course would be slower to select the correct tool than bi-directional rapid searches for the correct pocket, but for non-production use, does the extra 45 seconds to index fully around vs. reverse one pocket matter?

Looks like that gets you down to around 8 outputs or so:

1 ATC rotate 2 ATC arm swing 3 ATC arm up/down 4 ATC arm forward/back 5 Spindle unclamp 6 Oil mist (spindle blow I think, this could share an output with spindle unclamp)

7 Drum pin out (ATC carousel lock?)

8 Keylock (possibly for spindle orient?)

I would expect around 8 inputs as well:

1 ATC carousel home 2 ATC carousel index 3 ATC arm at ATC 4 ATC arm at spindle 5 ATC arm up 6 ATC arm down 7 ATC arm forward 8 ATC arm back

I plan to go the other way, actually more I/O. This machine is a gift and I promised better than new. On the computer end each I/O point costs about $5 so there's no reason to drop stuff that makes the machine better. I will be at a limit of 40 outs and 90 ins (plus 15 more for limits and homes) The one indisputable strength of Camsoft is how well it integrates I/O with the control.

Now, this tool change logic will be a stone bitch. "The Kid" has asked for ability to load tools every other pocket and then always put the tool in the empty next to the one you need for very rapid tool change. This could get REAL COMPLEX and may not happen.

But, lets not worry about next year now.

Sorry, both EMC/EMC2 and Mach3 also handle as much I/O as you need quite happily :)

That would be pretty insane since you'd have to constantly be changing the tool table. The big machines do fast tool changes by using a double ended tool changer arm, pre-fetching the next tool on one side and then doing the quick swap at the spindle letting it get back to work while the old tool is being put away. Looking at the pictures, you don't have clearance to change to a double ended ATC arm.

The most practical way to modify the machine for a super fast tool change that I can think of would be to add an intermediate stop position to the arm swing and an extra tool pocket above that position to allow you to pre-position the ATC carousel to the next tool, remove the current tool from the spindle, park it up in the intermediate tool pocket, grab the next tool from the ATC carousel, install it in the spindle, and then when the machine is back to work, retrieve the old tool from the intermediate pocket and return it to it's ATC carousel position.

Maybe what Karl meant (or, what he could do) is something like that -- eg, pre-position ATC to empty pocket left of next tool y; remove current tool x from spindle; park x in empty pocket; shift ATC left; pick y; install y in spindle and start mill ops; shift ATC right; pick x; shift ATC to home pocket of x; park x; repeat ad lib.

...

I can see agreeing on which control is best is just like agreeing whose religion is best.

My control is better, but its too expensive. If EMC were this well developed when I started doing serious refits, I would have went that route. I would, and have, recommended EMC to other hobbyists. I personally still wouldn't suggest EMC to a shop in business for $ unless they had a refit guru on site.

Now the other control you mention has a concept design problem for serious machines. No amount of good engineering can overcome a poor concept.

Like I said, its like religion. If you aren't a Lutheran, you're just not right . My two cents. I'm not changing your mind, you won't change mine.

Yea, I'm not modifying the machine. I may play with a dynamic tool table. WAY down the road.

Karl

Yea, that could work, however having to leave every other ATC pocket empty /= "better than new".

Personally, if it were my machine I'd simply treat one side of the machine as spare parts for the other side since a dual spindle machine isn't very useful for non production work, or even production work these days.

This route would work and probably be safer. I really haven't serously thought it all out. The other concept is just always put the tool in the closest empty spot to the next tool you need. My control has look ahead so this isn't a problem. But you would always be changing your tool table as to what tool is where. Screw up, even once, and break some serious shit.

Karl

Nope, it's not about agreeing on which control is best, the answer there is clearly the big guys like Fanuc. The issue is about not making untrue claims that one retrofit control handles lots of I/O better than the others.

Better than EMC/EMC2 or Mach3 how?

Mach3 was that well developed when you started, but you didn't notice that it could handle as much I/O and as much scripting as any of the competitors.

Which is why I also would not recommend EMC to a commercial shop for a retrofit, even if they had a guru since gurus can leave. I'd recommend Mach3 which is very user friendly, or possible EMC2 which I believe is pretty close to Mach3 in terms of user friendlieness now.

What is this mysterious concept design problem you think Mach3 has? People are using Mach3 for serious machines and many commercial machines are using it as well, not just hobbiest ones. You just seem to want to dismiss Mach3 without taking a good look at it.

You haven't provided any details on why you think Mach3 is inferior.

I'm not sure why there is this need for ATC speed, are you going to be producing some product? Heck, I can't see the need for the second spindle at all, I'd just keep one side of the machine as spare parts for the other. I don't recall seeing any dual spindle machines at the IMTS show, they are a barnacle left from the days when controls were really expensive so you wanted to double the output of the machine.

You could use a couple of the outputs and run them through a binary to decimal converter. Two outputs would change to four outputs, four outputs would control sixteen outputs.

John

Many machines use random tool selection. You load the tool in an empty pocket and tell the machine an assigned tool number. The machine remembers what pocket it puts the tool and will go to that pocket to load it when you call up the tool number. The machine usually puts the unloaded tool in the pocket of the tool that just got loaded. It will also compute the fastest way to the new called up tool and rotate the carousel in the shortest direction. Some machines you can also call up the tool by the pocket number rather than the tool number at your option.

John

A binary to decimal converter would change 4 ouputs to 10 outputs.

...

This is what "the Kid" has seen and wants me to do. He looks at the machine mechanically and says, "easy peasy, just make it do this"

If nothing never ever went wrong, I can see how to do it. But, I worked as an engineer making machines do my bidding for far too long to know this will work.

Karl

I don't go BCD, but I use this concept on inputs all the time. A rorary switch with three inputs has eight detents or positions. A rotary switch with two inputs has four positions. Or four inputs can have 16 positions. Useful in operator panels for things like speed, feed, jog increment, mode etc. saves a lot of inputs.

I'm not seeing this concept on outputs. I'm still going to need one opto device per out to go from 5 volt TTL to real world 110 VAC or 24 VDC. The optos are the limiting thing, there's lots of computer resource.

Karl

You would want to latch the data on the outputs if you multiplex it. Old TTL ICs were made for 16 to 1 and 1 to 16 mutiplexer and demutiplexer applications. The 74LS150 comes to mind. It looks like the major manufaturers have dropped all the older logic ICs and only supply some CMOS versions. Jameco might still have some wide de/mutiplexers in stock. I prefer LS around electrical noise.