Machsupport Forum

Mach Discussion => General Mach Discussion => Topic started by: mikep_95133 on January 11, 2012, 10:26:00 PM

Title: What's your fastest feed rate for your mill/parallel port.
Post by: mikep_95133 on January 11, 2012, 10:26:00 PM
This past week I 'upgraded' my pentium III 800mhz machine with a built in parallel port, to a core 2 duo 2.4ghz machine with a built in parallel port. The pentium could hit 135ipm with accel at 1 and have no pulses dropped. The new to me core 2 duo, and only hit 135ipm with accel at 1.5 and no dropped pulses.

Why did I not get a least a double in speed with the computer update.

All of this cutting air in a straight line.

Mike
Title: Re: What's your fastest feed rate for your mill/parallel port.
Post by: ger21 on January 11, 2012, 10:55:21 PM
What was limiting your speed to 135ipm with the old PC?

Generally, the only reason your PC will limit your speed is if it can't run at a high enough kernel speed. So if you were''t running at the maximum velocity for whatever kernel speed you were using, a new PC won't go any faster.

Say you were running as fast as you could at 25Khz kernel speed, but the PC wasn't capable of running at 35Khz kernel speed. IN that case, if the new PC could run at 35Khz kernel, then it might be able to go faster.

Generally, it's the motors torque that limits your speed. Not usually the PC.
Title: Re: What's your fastest feed rate for your mill/parallel port.
Post by: mikep_95133 on January 11, 2012, 11:29:27 PM
Sorry I left those details out. The pentium III could only go to 45khz. The new pc can go to 100khz but it skips pulses until I lower it to 45khz. Anything higher than 45khz just misses too many pulses.

Mike
Title: Re: What's your fastest feed rate for your mill/parallel port.
Post by: mikep_95133 on January 11, 2012, 11:32:20 PM
The motors are dc servo. I have run them at 75v for 12 years. The power supply is variable from 0-200vdc. I did raise them from 75v to 90v. That made them stiffer, but not faster.

Mike
Title: Re: What's your fastest feed rate for your mill/parallel port.
Post by: mikep_95133 on January 12, 2012, 12:16:01 AM
Here is a video of the mill at 135ipm cutting a pocket in the air.

http://www.rotordesign.com/servolite/VID_cnc_135.3gp

I found this guys's cnc site. His ideas for debug of the parallel port are just what I needed. I will bring my scope downstairs and get started. http://cockrum.net/cnc.html

Mike
Title: Re: What's your fastest feed rate for your mill/parallel port.
Post by: RICH on January 12, 2012, 06:42:17 AM
One needs to understand the complete system. By that i am meaning the pulse capability must be adequately supplied by each electronic  component.
PC ( sharing time with windows), PP, BOB, drive, and motor.
Then you have the other factors which relate, like the eletro mechinical of the motor and the mechanical.

The max stream of pulse delivered and used is only as good as the weakest link in the path.

In practical applications one defines an acceleration and velocity and finds what is required to achieve it …or … accepts the values based on
given parameters.

RICH
 




Title: Re: What's your fastest feed rate for your mill/parallel port.
Post by: mikep_95133 on January 12, 2012, 08:00:58 AM
Thanks Rich. Understood. Has anyone ever hit 200ipm for example??

I'd just like to know what others have achieved as a max feed rate while cutting air. I don't care if their hardware is different from mine.

After finding Mr. Cockrum's site, I'll take my isolated scope to the shop and start finding the weakest link as I posted earlier. My BOB uses NEC 2501 optos.  The data sheet looks slow to me. My amplifier boards have a 6n137 opto. Much faster.

Mike
Title: Re: What's your fastest feed rate for your mill/parallel port.
Post by: ger21 on January 12, 2012, 08:20:00 AM
 Is your velocity slider maxed out in motor tuning? If not, then kernel speed isn't your problem. What is your steps/unit set at? That's the only setting that may require a higher kernel speed to go faster.

Also, many PC's are not capable of using kernel speeds faster than 45Khz, and it's usually not recommended to use higher kernel speeds, as the PC can become unstable.

It doesn't sound like the kernel speed is the issue, but if it is, A smoothsteper or Kflop would help/

I can run my router with a 250oz stepper and 24V at 200ipm using 25Khz.
Title: Re: What's your fastest feed rate for your mill/parallel port.
Post by: mikep_95133 on January 12, 2012, 09:15:35 AM
No, the velocity slider is not even close to being maxed. I don't use it. I have always adjusted my max feed rates manually by changing the acceleration and feed rate individually. I could experiment with it. The acceleration has to stay under 2.0 to not miss pulses. I have gone as high as 10 for accel. Just looses pulses.

Do you think a 2.4ghz core 2 duo processor would really become unstable at higher than 45khz? My last pc was just a lowly PIII at 800mhz and it handled 45khz for years on end. These 2 comparisons alone show me it really isn't the pc.

200ipm on 25khz is great. I have zero experience with steppers. Mine is a servo.





Title: Re: What's your fastest feed rate for your mill/parallel port.
Post by: ger21 on January 12, 2012, 09:17:47 AM
If the velocity slider is not maxed, then you don't need a higher kernel speed.

Title: Re: What's your fastest feed rate for your mill/parallel port.
Post by: HimyKabibble on January 12, 2012, 11:22:45 AM
And if you can't set accel over 2 without losing steps, you have some pretty seriously under-powered motors.  I run accel of 25, which is not particularly high, but going higher just makes for more shaking when the big machine changes direction.  Steppers lose torque rapidly as speed goes up.  If you can't handle high accel, even at low speeds, then you simply don't have a lot of torque to work with, and you won't be able to reach high rapid speeds.  LOTs of people have hit 200IPM, and well beyond, with Mach3.  Your problem is your hardware.

Regards,
Ray L.
Title: Re: What's your fastest feed rate for your mill/parallel port.
Post by: ger21 on January 12, 2012, 11:28:02 AM
Ray, he's using servos
Title: Re: What's your fastest feed rate for your mill/parallel port.
Post by: stirling on January 12, 2012, 11:38:46 AM
As I think has been more or less said, your kernel frequency only needs to be fast enough for your machine. You can't make a machine that doesn't need a faster kernel go any faster by upping the kernel speed.

200ipm on 25khz is great. I have zero experience with steppers. Mine is a servo.

How does a 15m/min @ 2500mm/sec/sec grab you? (that's 590 ipm @ 98 in/s/s) ;D and that's steppers on a 25KHz kernel. It'll actually do quite a lot more both in vel and accel but it gets a bit scary. It's a router but it'll snap cast iron G clamps if they get in its way.

Ian
Title: Re: What's your fastest feed rate for your mill/parallel port.
Post by: HillBilly on January 12, 2012, 11:48:20 AM
If you are losing steps with servoes just by changing acceleration you have a problem. Servoes should either fault or catch up.

Something like Wrong active setting of step line????
Title: Re: What's your fastest feed rate for your mill/parallel port.
Post by: HimyKabibble on January 12, 2012, 11:57:58 AM
Ray, he's using servos

Gerry,

Well, then, what I said goes double.  He should be able to set accel MUCH higher with servos.  Either they're not tuned right, or they're gorssly under-
sized for the machine.

Regards,
Ray L.
Title: Re: What's your fastest feed rate for your mill/parallel port.
Post by: mikep_95133 on January 12, 2012, 01:37:24 PM

How does a 15m/min @ 2500mm/sec/sec grab you? (that's 590 ipm @ 98 in/s/s) ;D and that's steppers on a 25KHz kernel. It'll actually do quite a lot more both in vel and accel but it gets a bit scary. It's a router but it'll snap cast iron G clamps if they get in its way.

Ian

Ian, you set the bar very high!! I'm thrilled! What size is your table/gantry? My table is a 42" wide cast iron Bridgeport mill table. So it's heavy.

Have any of you looked at the video of mine in action? Does it look sluggish? It's cutting air using code for a spiral cut pocket. It's set to 135ipm. I don't remember what the acceleration is set to.

 http://www.rotordesign.com/servolite/VID_cnc_135.3gp
Title: Re: What's your fastest feed rate for your mill/parallel port.
Post by: mikep_95133 on January 12, 2012, 02:08:13 PM
My x and y axis are set to 20,000 pulses per inch. This is why I needed a faster computer that can handle a faster kernel.

1. Kernel=25khz, vel=135ipm, accel=1 the velocity slider is at 100%.

2. Kernel=45khz, vel=135ipm, accel=1 the velocity slider is at 100%.

3. Kernel=60khz, vel=135ipm, accel=1 the velocity slider is at 75%.

4. Kernel=65khz, vel=135ipm, accel=1 the velocity slider is at 66%

5. Kernel=75khz, vel=135ipm, accel=1 the velocity slider is at 55%.

6. Kernel=100khz, vel=135ipm, accel=1 the velocity slider is at 45%.

So I have to apologize for being wrong about my slider position earlier in this thread!!

Right now my accel is set around 1-2 as i recall. It's 3 stories down stairs. So I'm replicating my system settings on it's twin computer upstairs.

20,000 pulses per inch uses up a lot of cycles. So I was pretty lucky to get 135ipm out of the old pentium III 800mhz machine. But the machine won't go any faster at 2.4ghz. In fact, it actually skips slightly more steps if the kernel is set higher than 45khz, if I leave all of the motor settings the same. This is why I think the scope needs to be applied to the hardware to see if there is a mismatch in timing between the step and direction pulses. Or maybe my PS2501 opto board on the parallel port is the bottle neck.

Mike



Title: Re: What's your fastest feed rate for your mill/parallel port.
Post by: mikep_95133 on January 12, 2012, 02:13:53 PM
I think the reason having a kernel over 45khz causes more missed pulses, is that the velocity slider is already at 100%. So going to a faster kernel actually increases the pulse rate and makes a bad situation worse. All because the slider is already at 100%.

Then on the other hand, once the slider comes down off of 100%, I still get missing pulses. Dang. Sorry, just thinking out loud.

Mike
Title: Re: What's your fastest feed rate for your mill/parallel port.
Post by: RICH on January 12, 2012, 06:05:40 PM
Mike,
You are missing basic concepts of motion (the motion profile in tuning) which is used by the controlling software to provide pulses to dowstream devices. Changing the Kernel Speed allows Mach to assume some pulsing is possible in Motor Tuning. The driver test confrims if the selected Kernel speed is usable.
Thus, higher is of no value if the axis "system" is not capable of using the pulses. Note that an axis is made up of both electronic and mechanical items and all of them are relative to each other.

Velocity is calculated or tested by trial to find the max velocity. It is a user defined input value in motor tuning. If calculated then one assumes a safety factor. If tested one again provides a safety factor. The value can be restricted if the steps per unit are high, but if the axis can provide for a great velocity
and the computer train output is proper, then usable velocity can be varied over some range. Since velocity is constant speed in time, the tuning will show you the IPM avialable based on input velocity and steps per unit. Just because the controller indicates a high IPM dosen't mean you can use it.
So velocity is related to other inputs. What velocity can be acquired by a axis system is based on the motor characteristics to provide power / rpm / speed to overcome the forces acting on the motion of the axis. The mechanical aspects can be used to change the rotational to linear motion to provide for a different end use velocity. So how fast you can constantly go is dependant on a number of  related items.

Acceleration is different than velocity since it is defined as a time rate of change of velocity. It also can be calculated, tested to find, and is a user defined input value. What determines the acceleration is again defined by available power to overcome the forces acting on the axis. The forces are different in that greater power / force is required to accelerate to some velocity. You need to overcome the inertia ....resistance to a change in motion.
Same goes for decelerating.

What motion is available by others is totally irrelevant to YOUR machine unless you are comparing "apples to apples".

Just because you have changed the computer and increased only one part of a number of interelated factors may not necessarily give you a great
feed rate. It's all related and usually will require a compromise.

RICH
Title: Re: What's your fastest feed rate for your mill/parallel port.
Post by: RICH on January 12, 2012, 06:11:50 PM
I forgot to ask a very important question.

What are you trying to use / accomplish with your cnc machine?

RICH
Title: Re: What's your fastest feed rate for your mill/parallel port.
Post by: mikep_95133 on January 12, 2012, 08:52:14 PM
I'm trying to maximize my feed rates so I can make parts faster.

Be patient Rich. Over time you'll figure out that I have a clue.

Mike
Title: Re: What's your fastest feed rate for your mill/parallel port.
Post by: RICH on January 13, 2012, 06:33:46 AM
Quote
I'm trying to maximize my feed rates so I can make parts faster.

Then you must determine  the required parameters which will optimise the machine and provide
the appropriate components to accomplish the task reliably and within tolerance. The required or desired acceleration, velocity, power are determined FIRST.

Have fun,

RICH
Title: Re: What's your fastest feed rate for your mill/parallel port.
Post by: mikep_95133 on January 13, 2012, 07:13:11 AM
With some experimentation yesterday I achieved 175ipm up from 135ipm with and acceleration value of 10. Here I come Ian ;) Didn't miss any pulses. Over about 4" of travel is was off by .005". It's usually off by .000".

I better find the rpm limit of the ball screw.

Kernel went from 45khz to 60khz. So at 200ipm, it missed just a few pulses. The 60khz kernel was limited to 200ipm max with my 20,000 pulses per inch. So at 195ipm it was better, just not zero lost pulses. Testing yesterday and in the past showed that I need to get the kernel higher in order to go faster since the velocity slider is now almost pegged at 100%. I hope to see more of what is happening at over 65khz+ with the scope. I'm hoping today is scope day.

Both the step and dir pulses go through the PS2501 opto's. I still have to figure out where the 6n137 opto is in the circuit. There are no schematics.

Mike
Title: Re: What's your fastest feed rate for your mill/parallel port.
Post by: cjmerlin on January 13, 2012, 09:09:48 AM
Hi, What servo driver are you using, I just read the thread through and as said previously servo drives dont lose steps or pulses. If the motor cant drive the axis to keep up with the sent pulses it will fault the drive. That said, if the drive does not fault but the position if off a small amount then the drive's PID or tuning needs tweaking to provide a stiffer positional response.

I also found that if I ran a high kernal on the 1.2ghz computer I am using the pulses output on the printer port became less square and erratic when viewed on an oscilloscope which would cause the drives to fault intermittantly.

If you cannot tune the drive to get the accuracy then you need more powerful motors.
Title: Re: What's your fastest feed rate for your mill/parallel port.
Post by: HimyKabibble on January 13, 2012, 09:20:22 AM
And high speed, without high acceleration, is kinda pointless....
Title: Re: What's your fastest feed rate for your mill/parallel port.
Post by: mikep_95133 on January 13, 2012, 09:24:39 AM
Servolite is the name of the hardware. This hardware can't fault. It's simpler than that. It just slips. But it's done very well for me for 10+ years.

I'm anticipating the same thing you found with a mal formed pulse train. If that's the case I wonder if there is a parallel port card that actually works, but also has a higher rate of throughput. Or maybe I can build a high rate buffer that can clean up the pulse train.

The motors are Ametek Servo 600in-oz on 38v IIRC. At double that voltage the horsepower is quadrupled. So I don't worry about the motors. They keep the repeatability and accuracy up nicely.

Mike



Title: Re: What's your fastest feed rate for your mill/parallel port.
Post by: mikep_95133 on January 13, 2012, 09:27:30 AM
And high speed, without high acceleration, is kinda pointless....

My experience is different. I've done fine making parts and robots with accel at 1 for 10 years. So far I'm up to 10 and that's more helpful.

Mike
Title: Re: What's your fastest feed rate for your mill/parallel port.
Post by: cjmerlin on January 13, 2012, 10:16:09 AM
Ah, Is that Mr Bill's Servolight? I just looked it up on the net, The web page states that tops was 135ipm for the board.

When you get the scope, check the printer port and then do some checking on the driver board. Check part numbers for chips and specs to check the upper limit for pulse through put. It might be that the board just cant run those high freq pulses.

A drive that doesn't fault, not sure I'd be happy with that, I've had a few mishaps knocking into fixtures and have been thankful the drive faulted instead of carrying on regardless.

Title: Re: What's your fastest feed rate for your mill/parallel port.
Post by: mikep_95133 on January 13, 2012, 10:20:44 AM
Ah, Is that Mr Bill's Servolight? I just looked it up on the net, The web page states that tops was 135ipm for the board.

When you get the scope, check the printer port and then do some checking on the driver board. Check part numbers for chips and specs to check the upper limit for pulse through put. It might be that the board just cant run those high freq pulses.

A drive that doesn't fault, not sure I'd be happy with that, I've had a few mishaps knocking into fixtures and have been thankful the drive faulted instead of carrying on regardless.



The website is mine. I reconstructed it once Bill closed shop. I'm updating it as I gain progress to let the remaining users out there know the hardware is still viable.

I'll scope the entire pulse train path and compare the timing of the step and dir pulses.

A hard enough fault just pops a fuse.

Mike
Title: Re: What's your fastest feed rate for your mill/parallel port.
Post by: HillBilly on January 13, 2012, 12:33:39 PM
The problem with that drive is you have to avoid the 127 count error limit and where you are wanting to run is right there.

Darek
Title: Re: What's your fastest feed rate for your mill/parallel port.
Post by: mikep_95133 on January 13, 2012, 01:38:43 PM
The problem with that drive is you have to avoid the 127 count error limit and where you are wanting to run is right there.

Darek

Have you owned one? Tell me more.

Mike
Title: Re: What's your fastest feed rate for your mill/parallel port.
Post by: HillBilly on January 13, 2012, 02:01:22 PM
"One  area that my controller falls short of the more expensive control boards is it only has an 8 bit counter, compared to 24 or 32 bits of an expensive servo  board. This has the effect of only being able to control position error up to +/- 127 encoder or computer pulses. When pushed past these limits the motor  slips like a stepper motor does when under load except instead of 1.8 deg. slip there will be a slip in multiples of 127 encoder pulses. This means that the amount the motor slips, or cogs, is a bit bigger step than a stepper depending  on the resolution of the encoder mounted on the motor. This motor slipping I  speak of is no problem under normal operating conditions, this is because the  error should never get close to being that big. If it does it means you are asking to much from the system and need to get bigger motors and amps or slow  the thing down. Although my servo controller is not as sophisticated as a full blown motion control card, it is more than the average hobbyist needs, and cost only as much as a stepper system that has no position feedback, if you take into account the fact that stepper motors will cost more."

This is from the website
Title: Re: What's your fastest feed rate for your mill/parallel port.
Post by: mikep_95133 on January 13, 2012, 02:48:29 PM
So when the system hits the 127 count limit, it just slips instead of flagging an error. OK.

Mike
Title: Re: What's your fastest feed rate for your mill/parallel port.
Post by: stevehuck on January 14, 2012, 07:03:17 AM
Thanks Rich. Understood. Has anyone ever hit 200ipm for example??
Hello Mike!

I have a 1.8G machine with a BOB and Gecko. When I first setup I wrote a piece of Gcode to move around for a dozen or so moves and then return to a spot where I had 2 indicators setup. I would keep increasing the speed to see where I would start to miss steps. I got to 350ipm and the table zipped around in the pattern and returned to the indicators and zeroed them right out never losing a tenth. I finally backed the setting down to 200 and that's where it is today. I would have to check the setup but I think I'm at 35Khz kernel and 10,000 Steps per inch. Can't remember

Home built mill
12" X
7" Y
570oz steppers
Gecko @ 51 Volt
$5 breakout board
Title: Re: What's your fastest feed rate for your mill/parallel port.
Post by: mikep_95133 on January 15, 2012, 10:07:37 PM
Hi SteveHuck,

That is an amazing feed rate! Thanks for letting me know!

Mike