turret encoder

[Karl_T]

I've attached two files:

1) a picture of the turret encoder on my Hardinge CHNC
2) the Electrical print of the encoder.

I've not seen or dealt with a device that looks like this before.

I need to get turret position into Camsoft. Looking at this print, is it as simple as connecting A B C D to four inputs on the Opto 22 board? Then decode position as four bits of base 2? (I'm curious; there's eight turret locations - 3 bits would be enough) Assuming this encoder is still working, that is.

Or should I replace this device with something else? What? FWIW, I'm using a 1840 Galil card and have an extra axis available for the turret. I'm out of inputs and would prefer to use an encoder. But the only was I see this working would require homing the turret on every startup.

Karl

[Al_The_Man]

It looks like eight hall effect ic's , but the picture is blurry, normally four bits would give you up to 16 inputs, so you should be able to decode a binary pattern from them for 8 positions, they may already have pull up resistors, but you may have to reverse engineer the electronics to find out what type of output the sensors have, for interfacing to the opto's.
You would not need to home with this arrangement, with an encoder you would have to.
BTW have you used the Galil general inputs? these can be used on 5v or 24v isolated inputs.
Al.

[Carlo Rechi]

Mr. Karl,

To me I dont understand this encoder so much. I think this is not a normal encoder from the wiring. The encoder is ABCD bits like an old time optical eye counter. Al said this would be 16 bits. This is right. You need all 16 bits not just use a few cause the binary pattern would not be the same.

I have seen binary bit macros and turret routines in the Camsoft system that would work but I dont know if I can post these from their books cause it said copyright. I have asked Camsoft for many routines and help before and they have been very nice to me. You should ask if they have something like this.

If you are out of inputs you can find a cheap encoder for about $100 I think.

Carlo

[Al_The_Man]

Originally Posted by Carlo Rechi Mr. Karl, To me I dont understand this encoder so much. I think this is not a normal encoder from the wiring. The encoder is ABCD bits like an old time optical eye counter. Al said this would be 16 bits. This is right. You need all 16 bits not just use a few cause the binary pattern would not be the same. Carlo

Actually 3 bits would give you 8 positions, but they probabally made it 4 bits to avoid using all zero's as a position, because this could mis-interpreted when zero occurs between sensors.
So you would only use 0001 to 1000 (8), there is probabally a binary decoder ic that you may have to hook into.
Al.

[HillBilly]

Alot of these had two sensors for each tool location. The first was the decel switch and the second was the stop sensor. They would only index in one direction if this is the case.

Darek

[Karl_T]

Originally Posted by HillBilly Alot of these had two sensors for each tool location. The first was the decel switch and the second was the stop sensor. They would only index in one direction if this is the case. Darek

This makes sence. Its an air motor and I can see two solenoids for fast and slow. Now it would be REAL NICE TO index both ways. Is this possible with an air motor system?

Originally Posted by Al_The_Man It looks like eight hall effect ic's , but the picture is blurry, normally four bits would give you up to 16 inputs, so you should be able to decode a binary pattern from them for 8 positions, they may already have pull up resistors, but you may have to reverse engineer the electronics to find out what type of output the sensors have, for interfacing to the opto's. Al.

I gave you all the info I have. What sort of reverse engineering should I do? Just put a VOA meter on on an output?

Encoders are a cheap item, should I just jerk this thing?

Karl

[Evodyne]

Karl,

I think you are right on the money. With a cheap logic proble from Radio Shack you can determine the coding.

Without knowing any better, I'm thinking about Hillbilly's comments about a slowdown and stop: the board might still use all four bits as follows:

3 2 1 0
2 2 2 2
1 Slowdown 0 0 0 0
1 Stop 0 0 0 1
2 Slowdown 0 0 1 0
2 Stop 0 0 1 1
3 Slowdown 0 1 0 0
3 Stop 0 1 0 1
4 Slowdown 0 1 1 0
4 Stop 0 1 1 1
5 Slowdown 1 0 0 0
5 Stop 1 0 0 1
6 Slowdown 1 0 1 0
6 Stop 1 0 1 1
7 Slowdown 1 1 0 0
7 Stop 1 1 0 1
8 Slowdown 1 1 1 0
8 Stop 1 1 1 1

Good luck. Let us know what you find out.

Evodyne

Crud: my formatting went away-but from the sound of it you understand the truth table anyway.

[Al_The_Man]

The turrets with slow down that I have seen like Mazak etc, use 2 sets of reed switches per location, in the photo it looks like there is 8 hall effect ic's which for an 8 position turret may not offer slow down, unless the method use was to detect one-position-away in order to activate slow down.
If you can trace the four ABCD code lines back to whatever IC is driving them, It should be possible to tell what interface is neccessary to hook up to the Galil.
If it was working well before, it may be worth resurecting it.
But if you go with an encoder, get a 360 pulse/rev which Galil reads as 1440 so you can divide by an even number of pulses.
Al.

[camsoft]

Hello Karl,

If you can find some spare inputs the current version has a binary decoder routine that we can give you. The front office should give us permission to send it to you today.

To cut cost there are also 2 new devices and 1 not so new low cost device listed in our price list, downloadable from the home page for reading a spare digital encoder or adding more optically isolated I/O if needed. These items are not meant for high speed but would fit the application perfectly in your case since you're low on inputs. If you change to an encoder and don't have a spare connection we could make the Hand Wheel Device read quad pulses for positioning like a digital encoder.

Standalone Digital Input/Output units - Provides 14 Inputs or Outputs. Uses simple RS232 commands. $159

Standalone Hand Wheel Pulse Counter unit- Reads digital quad handwheel pulses through RS232 box. $159

Optically Isolated Spare Relay Racks. Provides Isolation and Noise Guard (Qty 4 $35) - (Qty 16 $120)

Al is correct the binary bit pattern only need 4 bits. Another point Al mentions is that on this type of device the bits may or may not change to zero or some odd value in-between turret rotations. Also what Evodyne suggest would be great if it was designed for slow and stop bit patterns.

You could purchase a probe or simply view and record to a file the bit patterns on the Digital I/O panel in Diagnostics in real-time to see what device really does. If this was the case our binary turret logic routine could easily have a table to accommodate this. There are so many options, besides what has been mentioned already. The most common variety we see is the type that the binary bits would only be used for position confirmation and not monitored while the turret is moving. Instead there are prox sensors or digital I/O counters that count the turret positions as they go. This may then require more I/O points. If this is the case we would recommend a modern digital encoder be used instead of I/O. A real encoder would be monitored in real-time quickly and precisely and be less money than buying our Standalone Digital Input/Output unit.

Since you can see the two solenoids on the air motor and if you have 2 spare outputs then we can show you how to modify the turret logic to toggle between fast and slow mode as well as logic to calculate turret positions for clockwise and counter clockwise tool changing if the air solenoid device will permit us to reverse directions physically.

Tech Support
CamSoft Corp.
951-674-8100
support@camsoftcorp.com

[Karl_T]

Thanks for all the help. You guys are the best.

Karl

[Karl_T]

I accidently discovered something that makes NO SENCE to me.

All the I/O on this machine is 12 VDC. So that's what I tried on this encoder, no joy. Doing something else, I switched the hall effect limit switches to 5 VDC and now the encoder works. WTF????

I guess I'll take my bit of luck. But are there any theories as to why this might happen? The wires run in the same cable, but as far as I know, are electrically separate.

Karl

[Al_The_Man]

If the IC's on the board have a number it should be possible to tell if they are TTL (5v) or CMOS (12v).
Al.