What's new
What's new

Look ahead, HSM, NURBS old controllers

Hardplates

Stainless
Joined
May 8, 2019
Location
What once was a free country
Curious on others opinions on what controllers can run HSM tool paths. I have a Mitsu Meldas that runs fairly well for an 1989 if I play around with smoothing and whatnot. I'm looking at a machine with a Siemens 2100 that is supposed to be 660 blocks a second and 180 lines of look ahead. How about older Yasnacs? Are the Meldas 500s better than the 300s? 600s? I have never even tried on an 32 bit OM but I assume that is out of the question. Mazatrol? Whats the oldest Fanuc that may work. Ideally I would like to be able to rough at 300-400 ipm and finish with a ball around 100 ipm. This isn't a high dollar part I will be running on this machine so I am looking for something old and cheap.

I'm interested in hearing peoples stories on being able to get old iron to work with new tool paths. Some day I'll get a Okum Genos M470V but today is not some day......
 

mmurray70

Hot Rolled
Joined
Jan 11, 2003
I used to work on a mid 90s Fanuc with 0M control. I would certainly avoid this control if you want that kind of performance. It would slow down to a crawl doing anything 3D, acceleration was slow, following error was terrible. We used to have to finish tight tolerances in single digit feedrates lol.
 

13engines

Hot Rolled
Joined
Jun 30, 2015
Location
Saint Paul
Not sure exactly when it came out, (90's?) but the Fanuc 16/160 with HPCC (High Precision Contour Control) using a RISC (Reduced Instruction Set Control) board was a stab at what you're after. The 18 control is typically associated with the 16 but is not configurable with the Risc Processor. Only the 16 is. Both can have Look Ahead and other options like advanced acceleration /deceleration control. Mmurray is right about the OM. I've run the same 3d tool path on an OM and an 18 with Look Ahead. The Om starts shaking when you try to push it too hard. My 18 with Look Ahead handles the same code easily at five times the speed or more.

I believe the Fanuc 15 is smarter then both the 16 or 18. I know almost nothing about it. Not sure about it's HPCC or RISC configuration options.

Of course a hot rod control needs the machine to match. Adding racing slicks to an old dog is likely not going to get you anywhere.

Pretty sure Fanuc has moved well past what the 16 or 18 controls can do, but you did say you were looking for old stuff.
 

DouglasJRizzo

Titanium
Joined
Jun 7, 2011
Location
Ramsey, NJ.
Not sure exactly when it came out, (90's?) but the Fanuc 16/160 with HPCC (High Precision Contour Control) using a RISC (Reduced Instruction Set Control) board was a stab at what you're after. The 18 control is typically associated with the 16 but is not configurable with the Risc Processor. Only the 16 is. Both can have Look Ahead and other options like advanced acceleration /deceleration control. Mmurray is right about the OM. I've run the same 3d tool path on an OM and an 18 with Look Ahead. The Om starts shaking when you try to push it too hard. My 18 with Look Ahead handles the same code easily at five times the speed or more.

I believe the Fanuc 15 is smarter then both the 16 or 18. I know almost nothing about it. Not sure about it's HPCC or RISC configuration options.

Of course a hot rod control needs the machine to match. Adding racing slicks to an old dog is likely not going to get you anywhere.

Pretty sure Fanuc has moved well past what the 16 or 18 controls can do, but you did say you were looking for old stuff.

The Fanuc 16 was a real HOT control when introduced, but that was around 30 years ago. You can make old controls do some cool stuff, but nowhere near as fast as a modern CNC will. In my shop I have Fanuc 11s, old Mitsus and an Okuma OSP 5000. All are O-L-D, but still make good parts. Just not that fast. A new Okuma is in my future, just not quite yet.
 

gustafson

Diamond
Joined
Sep 4, 2002
Location
People's Republic
I think acceleration is the issue on older controls. Not speed, acceleration. When you watch one of those trochodial toolpaths it is the rapid back that makes it fast
 

Hardplates

Stainless
Joined
May 8, 2019
Location
What once was a free country
So would an 11 work better than an OM? Is an 11 comparable to a Meldas 300 as the 11 was available on my OKK but it seems most came through with a Meldas 320.

I've yet to come across a resource that compares all the controllers and their capabilities. That was one of my hopes in starting this thread.
 

Hardplates

Stainless
Joined
May 8, 2019
Location
What once was a free country
I think acceleration is the issue on older controls. Not speed, acceleration. When you watch one of those trochodial toolpaths it is the rapid back that makes it fast

My limitations are in the controller being able to process code fast enough and being limited by 9600 baud drip feeding. I have no trouble at all hauling ass back when not in the cut.
 

316head

Aluminum
Joined
Jul 27, 2008
Location
finland
I program two HMC's with a 18i-mB control and they have massive lookahead buffer as it's easy to get hypnotized by looking at the running CAM-generated code, but it SUCKS at following a path. AICC or HPCC would be more than welcome (boss says no - not yet). I cannot program too small arcs in HSM toolpaths to avoid premature endmill breakage; on the other hand, high feed toolpaths work great (for example, area roughing). The machines were apparently used for aluminium only before.
 

cameraman

Diamond
Joined
Nov 24, 2014
So would an 11 work better than an OM? Is an 11 comparable to a Meldas 300 as the 11 was available on my OKK but it seems most came through with a Meldas 320.

I've yet to come across a resource that compares all the controllers and their capabilities. That was one of my hopes in starting this thread.

I've been looking at new/ newer Mits based controls and NURBS (haven't dug very deep yet).

Wondering what vintages of controls claim NURBS capabilities ?

Also wondering about vector descriptions vis a vis Fanuc versus regular G -code / efficiency for older lower bandwidth controls ? Is that just more of a 5 axis thing.

Not 100% how the sampling of a NURBS description is interpreted by a control in a more direct way ? In terms of conversion to direct motion control bypassing G-code based descriptors/ commands.
 

Vancbiker

Diamond
Joined
Jan 5, 2014
Location
Vancouver, WA. USA
So would an 11 work better than an OM? Is an 11 comparable to a Meldas 300 as the 11 was available on my OKK but it seems most came through with a Meldas 320.......

Pretty comparable performance, but keep in mind that there were many options from both manufacturers that could significantly sway the choice. Most of the time the Mitsu controls were offered with better standard specs than the Fanuc. Things like larger standard memory, macro B, helical interpolation, which some builders made optional with Fanuc were standard with Mitsu. It was their attempt to improve market share and interest in their product.

My limitations are in the controller being able to process code fast enough and being limited by 9600 baud drip feeding. I have no trouble at all hauling ass back when not in the cut.

Whenever you have to drip feed a program with lots of small moves you will hit data transfer limits at pretty low feeds. I've done a bit of playing around with this on my M50 and found that a program that won't run smoothly above ~25IPM while drip feeding, will run smoothly at ~80IPM if a clip of it is stored in memory and run.
 

Hardplates

Stainless
Joined
May 8, 2019
Location
What once was a free country
Pretty comparable performance, but keep in mind that there were many options from both manufacturers that could significantly sway the choice. Most of the time the Mitsu controls were offered with better standard specs than the Fanuc. Things like larger standard memory, macro B, helical interpolation, which some builders made optional with Fanuc were standard with Mitsu. It was their attempt to improve market share and interest in their product.



Whenever you have to drip feed a program with lots of small moves you will hit data transfer limits at pretty low feeds. I've done a bit of playing around with this on my M50 and found that a program that won't run smoothly above ~25IPM while drip feeding, will run smoothly at ~80IPM if a clip of it is stored in memory and run.

Although I am getting higher feeds (likely from playing around with smoothing and arcs) this is why a while back I asked about memory upgrade which is not an option for the 300 but is for the 500.
 

Jashley73

Titanium
Joined
Jan 24, 2013
Location
Louisville, KY
I've been looking at new/ newer Mits based controls and NURBS (haven't dug very deep yet).

Wondering what vintages of controls claim NURBS capabilities ?

Also wondering about vector descriptions vis a vis Fanuc versus regular G -code / efficiency for older lower bandwidth controls ? Is that just more of a 5 axis thing.

Not 100% how the sampling of a NURBS description is interpreted by a control in a more direct way ? In terms of conversion to direct motion control bypassing G-code based descriptors/ commands.

My understanding is that the code gets posted to the CNC in normal G-code. The 'nurbs' simply translated into XYZ. It's then up to the control to interpret all of the incoming moves, and smooth/tweak the actual commands to the servos in order to end up with a smoother path. I may be dead wrong - and someone correct me if I am - but that's the best I can make of it...







Hardplate what's your budget on this? Occasional I see some older S/V series Makino's on eBay pop up for 30K or thereabouts. There's an V33 with a Pro-3 control for $27,500 now... Hopefully Vancbiker can chime in, but I'm guessing that Makino's Pro-3 is built on a 21i control?
Used Makino V33 High Speed CNC mill | eBay

Edit: We had one of these Bridgeport branded (Hardinge?) machines when I was in tech school. Came with a 21i control. Probably not the greatest machine in the world, but it would move, no problem. I'd double check on the spindle speed though. The stated 5k seems suspiciously slow...
2000 Bridgeport VMC3020XV CNC Vertical Mill | eBay
 

Vancbiker

Diamond
Joined
Jan 5, 2014
Location
Vancouver, WA. USA
........ Hopefully Vancbiker can chime in, but I'm guessing that Makino's Pro-3 is built on a 21i control?......

All the PRO3 controls I have worked on were on Makino HMCs and were all 16 or 16i based. PRO3 is just a U/I mainly on the PMC side of the control so could run on other models but I've not seen that.
 

Hardplates

Stainless
Joined
May 8, 2019
Location
What once was a free country
My understanding is that the code gets posted to the CNC in normal G-code. The 'nurbs' simply translated into XYZ. It's then up to the control to interpret all of the incoming moves, and smooth/tweak the actual commands to the servos in order to end up with a smoother path. I may be dead wrong - and someone correct me if I am - but that's the best I can make of it...







Hardplate what's your budget on this? Occasional I see some older S/V series Makino's on eBay pop up for 30K or thereabouts. There's an V33 with a Pro-3 control for $27,500 now... Hopefully Vancbiker can chime in, but I'm guessing that Makino's Pro-3 is built on a 21i control?
Used Makino V33 High Speed CNC mill | eBay

Edit: We had one of these Bridgeport branded (Hardinge?) machines when I was in tech school. Came with a 21i control. Probably not the greatest machine in the world, but it would move, no problem. I'd double check on the spindle speed though. The stated 5k seems suspiciously slow...
2000 Bridgeport VMC3020XV CNC Vertical Mill | eBay

To be honest I'm all over the map. One day I was looking at a machine for 2G (I have machines I've paid half that and hold tenths all day) and the next day I'm looking at something listed for $35,000. I want another machine but don't NEED one and the way I see it right now is a buyers market so I'm not going to jump on anything that doesn't make me feel like I am getting a great deal. Also I have the ability to haul up to around 47K pounds but I don't want to go thousands of miles either. I understand these factors are limiting but in the past I just sit and wait cash in hand till someone needs to get rid of something I want.

BTW that V33 looks pretty sweet. Way more advanced than I am used to but definitely in line with the direction I want to head.

PS I much rather an older high end machine than a newer commodity grade machine.
 

barbter

Diamond
Joined
Oct 27, 2007
Location
On Tour...
BTW that V33 looks pretty sweet. Way more advanced than I am used to but definitely in line with the direction I want to head.

PS I much rather an older high end machine than a newer commodity grade machine.
That Makino looks a lot of bang for your buck. My concern would be cost of spindle rebuild.
Ref old machines... what always worried me (and I stayed away from) was old turcite machines.
So i'd always stick with linear rails.

16 control is very capable. Has loookahead but usually not configured properly, if at all. But that's only playing with parameters.
21 control is basic - can be available with a pcmcia card option if you're lucky to find one, in which case you can output your HS roughing part of the program (ie lots of code and big program) as an M198 external sub prog, and run that part off the card.
BUT, it would need servo tuning to get anywhere near accurate toolpath at a feedrate of 2metres...

Ref commodity machines - depends upon what you're doing and your budget.
I bought 4x new Chevalier machines (Fanuc 0iMC, and 0iMD) and they had 10kRPM #40 spindles. Good for HSM once set (talking the C, the D was far better out the box) and would cut steel pretty well (12mm trocoidal paths in Stainless).
They had 35mm linear rails and the machines were great value for money.

If money hadn't been an option, I'd have gone for Quaser machines. 45mm wide rails and another level up.
 








 
Top