Thread: DRO option "switching inputs/outputs" ??

Heidenhain PT880 DRO offers option of "switching I/0's (available via separate AMI I/0 unit)". I hope this is what I think it is. Look on page 105 of the PT855 manual here

(PT880 manual link doesn't work for some reason but 855 is the same except monochrome)

http://filebase.heidenhain.de/doku/o.../en/index.html

See where it says under features "allows motor cutoff" ?

Does this mean the DRO could be configured to send a signal at predetermined datum settings such that if one had a servo for table feeds, one could create a sort of "active" DRO for simple one axis at a time moves ?

I suspect the answer is "yes, but" with the "but" being the inability to ramp up and ramp down the feed automatically, and therefore actual accuracy of the setup (on a mill anyway) might not wash.

Thoughts ?

Also, what is "seperate ~AMI~ I/O unit" ? Any idea what AMI stands for ? AMI semiconductor, by chance ??

I can't believe none of you guys know about this...don't make me go to that "other" forum now ! [img]tongue.gif[/img]

I think I know what you want to do, and there is a very simple and cheap solution.....

No designing or inventing.. Just get These parts...

Why don't you tell in more detail what you want to do?

Pete

Pete, I would like to be able to duplicate what a Heidenhain TNC 113 control can do, but in a modern new control. The only modern control that duplicates the TNC 113 functions "officially" is the Heidenhain TNC 124. Fine, but it's overkill in many ways and cost $6,000.

And what the TNC 113 is, is basically 'point to point' CNC in it's simpliest form...sort of the "Pong" video game of CNC [img]smile.gif[/img]

Basically I'm talking automatic programable electronic feed "trip dogs" that take the place of mechanical feed trip dogs.

You have a machine with one DC drive for all 3 axis that has manual lever engagements for each (typical for late model Deckel mills) You have a simple pot for adjusting feed speed.

Typical example, you want to move X axis +3.222", so you push the incremental key, and type in 3.222". You push the "X" key, an LED comes on. You then engage the appropriate feed lever for X axis positive direction...the table moves 3.222" at which point the control shuts down the power to the DC drive and the table stops dead on 3.222".

But to be accurate, and stop dead on +3.222" no matter what the feed rate selected via the pot, I suspect there may need to be automatic ramp down in feed rate that is triggered a few thou before it reaches +3.222" and that may be the fly in the ointment.

AMI stands for Auxiliary Machine Interface.
I'm not sure what the AMI's capabilities are but I will say this the Positip 880 Interface is exactly the same as the acu-rite millpwr control.
Heidenhain owns Acu-rite if you didn't know that already. But having said this I'm sure that the hiedenhain 880 can be set up to run in cnc once the AMI is purchased.

OK I looked in the manual and here is what it says;

AMI (Auxiliary Machine Interface)(X51)
(Optional)
If AMI hardware is connected to the POSITIP 880, the SWITCHING
OUTPUTS form will be available from the installation setup list box.
The AMI hardware may be configured with 8 output relays. See Fig.
II.20.
To configure an individual relay output, move the cursor to the relay
number and press ENT. To remove a relay’s setting, highlight the relay
and press CE.
�� The AXIS field is used to specify which axis position controls the
relay. The axis is selected by pressing the NEXT AXIS soft key.
The DIRECTIONAL soft key is used to select the directional signal
mode. When selected, the relay’s operation is based on the position
value’s sign. The relay is activated when the value is non-negative.
The relay is de-activated when the value is negative. If
DIRECTIONAL is selected, the axis selection may be set to ALL
AXES. In this configuration, the relay’s output is set when an
operation sets any of the distance-to-go mode display values.
�� The CONDITION field is used to specify the required relationship
between the current position display and the switching point. It is
set by pressing the soft keys.
�� The PRESET field is used to define the switching points and to
select whether the point is an actual value or distance-to-go
position. The position source is selected by pressing one of the soft
keys.
�� The POINT/RANGE field is used to define whether the condition
refers to a point on the axis or refers to range about zero. If a range
is specified, the sign of the value is ignored.
�� To fill in the RELAY ON/OFF field press the OFF or ON soft keys.
When the condition has been satisfied the relay will turn OFF or ON.
�� The RELAY OPERATION field is used to further define how the relay
will operate when the switching condition is met. The options
include CONTINUOUS, PULSED or SINGLE. SINGLE is only an
option when the distance-to-go position is selected in the PRESET
field.
In CONTINUOUS mode, the relay remains activated (ON or OFF) as
long as the switching condition is met.
In PULSED mode, the relay is activated only for a predefined, period
of time. After the period has timed out, the relay is de-activated. The
time is entered in the RELAY OPERATION field.
If the PULSED soft key was pressed, a delay value must be entered
using the numeric keypad.
In the SINGLE mode, the relay will transition to the activated state only
one time after the axis value is set. It only applies to distance-to-go
mode position values.
The AMI settings are retained on a power cycle.

Well, what I had in mind, but not knowing what you wanted to do...
Was to use the output of your scale and drive the inputs to a Geko drive for the motor.
Then use any one of the cheap step and direction controller software packages that run on on a PC to tell the Geko drive how far and how fast.
The Geko drive would drive the motor, and it would track via the scale outputs, the software would provide the ramp up, ramp down and stop and the feedrate speed.

Now, That I find out more of what you want...
What you want is to have three sets of scales, and only one motor, with hand clutches to drive each axis.
Now that is a puzzle for what I was envisioning.
Electrically, and cost wise with stuff that I know about, it might have been cheaper to have three motors, one for each axis. Then you can drive them from a very cheap step and direction software package, and use scales and DC servo motors.

If you have a clutch between the motor and the axis, then you need a scale on the driven axis. And a contraption to drive the motor while it is connected to the axis, and then stop worrying about the motor once it reaches position, and hopefully it stays put...

I know it could be done cheaply enough with individual motors and drivers, otherwise it sounds like a specialized piece like the $6,000 heidenhain. Unless someone else has another idea.

Maybe Mr. HammerHead can whip something up, but again, you need inividual feed back from each axis and a counter at the PC to track each axis.
Mr. Head's proposed solutions all involve using a Geko drive or similar to track the scale feedback for an indiviual motor/axis.

Don is right; acceleration/deceleration is required - even with servos (otherwise you overshoot your target position; and if the move is in a cut, you gouge the part ). Although I could indeed whip something up, it doesn't make sense to - for the parts cost to go to a switched axis drive, you might as well go for 2 or 3 axis functionality (a G320 is $114, and add an encoder for $50). I'm using a trackball for jogging on DiamondBack, but it could be setup with a conventional MPG as well...

And about using linear scales to feedback to a Gecko - there can be mechanical latencies and non-linearitys in the system, which can cause resonance; which is why it's usually best to run the servo loop off of a close-coupled encoder (ideally, one that's integral to the motor's shaft)...

[img]smile.gif[/img]

And about using linear scales to feedback to a Gecko - there can be mechanical latencies and non-linearitys in the system, which can cause resonance; which is why it's usually best to run the servo loop off of a close-coupled encoder (ideally, one that's integral to the motor's shaft)...

However, in this case, we are not talking about ballscrews, or a high performance contouring application. Additionally, Dan wants to see .001" accuracy I believe, and was talking about having scale encoders anyway.

So... What I implied but don't take the time to type is.... The scale feedback could be fed back to the Gecko drive, but it would have to be de-tuned to avoid hunting at any great rate.
Here is an example:
I once looked at a Large gantry machine that would overshot the target in X axis since it was about 80,000 lbs of gantry, then slowly back up. that is how the servos were tuned for that application. If the servo gain were cranked up, that servo motor would beat the ballscrew to death trying to rattle that 80,000 lb gantry into position.
Clearly it needed more tuning, but it did have a fair amount of hysterysis from the ball screw to the gantry to the (40')linear encoder.

So for Don's aplication, we can reduce the tuning, and let the deccel coast the slide into position.
Sometimes one must compromise their (servo)ideals to match reality.

And again, this discussion revolves around using seperate servo motors. For the application that might not be a practicality.
What we end up with is a full (hobbiest grade) three axis control, without ballscrews and a low performcne axis control, probably not suitable for any sort of countouring, just point to point.
Too bad it doesn't seem a practical matter to switch the encoder feed backfom one axis to another depending on which clutch is engaged, so the motor drive tracks one axis at a time. and threfore meet the cost and (mechanical)trouble goals desired.
Pete

I'm still wondering if it might work even without decell as there seems to be a way to adjust the "reaction time" so to speak, via the "measured value ouput" parameters.

On the TNC 113 control, there is a noticable accelleration upon starting the feed, but ironically, the decell is not noticable at all unless you're running at near full feed rate. At most normal feel rates one would actually use when milling, it seems to just stop dead when it reaches the programmed datum point. And manages to do so +- .0002" every time.

I called Heidenhain about this yesterday, and damned if even *they* don't know the nitty gritty on the switching I/0 option....still waiting on a callback from the resident guru.

Even Deckel Doctor knew nothing about it...I must be plowing new ground here... :rolleyes: