Hello my question is this, i wuld like to put more computer controlled featurs on my cnc, but the problem is that i dont have much input output pins as an option.
If i wanna use 3 axes with encoders it wuld take 6 outputs and 6 inputs and that is too much for 1 printer port, couse it has only 4 inputs. I know i can use another printer port card for just a few $ but it too have verry litle inputs if i wanna conect the spindel with encoders so that i culd make screews, and a coolant pump... is there any alternative (USB...)?
i wuld also like to make a costum key board, when my hands are filthy, than a tuch screen is out of the question, and ordenary keyboard is not to handy to use in the shop.

For more IO, you could look at using a PLC and running it over Modbus. I think it is Peter Homan that alos has a ModIO board you can use. For up to 54 more inputs, no outputs, you could use an Ipac with keygrabber but it is the least robust and IPac is not supported by the developers any longer. It works pretty good, but you are on your on to work it out except from other members helping.

Also might want to look at this thred about Cubloc.
http://www.artsoftcontrols.com/forum/index.php?topic=1105.10

Brett

If you are familiar with embedded processors, Microchip's PIC 24FJ********* series of processors has a hardware peripheral for CRC. Very nifty.
Microchip's link http://www.microchip.com

Whacko

I thought someone was in the stages of developing an input output board specifically for Mach 3. We were asked what features - specifically the number of input and output pins we would like.

Can't be too far down the posts. it was only a couple of weeks ago (or so).

The i/o is expandable as said previously via the modbus interface. It could also be implemented with the primitive serial facility. But I think the poll was maybe to determine the needs of users, maybe to implement a USB i/o interface. Then there is also the TCP modbus now available. I know that the i/o on the parallel ports is propably the easiest and cheapest, but this is an interesting subject I have been thinking about, and maybe Art and Brian can think about it, if the parallel port is used with a parallel interface controller that can turn the 8 bit databus into a bidirectional communication as per ECP and printers, you could get 255 inputs and 255 outputs on a second parallel port, or even do the pulser the same way. There are many embedded controllers that can easily handle a 8bit databus protocol and do the decoding. Might I be bold enough to say that I could design such an interface easily. ( I'm not blowing my own horn here  ) It would be as simple as interfacing with SDRAM, bidirectional and with a 70 ns strobe.
Think about it, let me know how many people are interested, we can create an open source and hardware project.

Whacko for Real

i am interested, but i am not sure i understand exactly how the parallel port would work in both ways

The same way as a printer interface. Look at the signals and their terminology. There are some well documented information on the three parallel port interfaces at this link under "Parallel port interfacing" on this page. The link is http://www.beyondlogic.org/. It could work that I know, the question is if Art would be willing to code for this facility. With other words, the 8 bits on the databus is a byte which could be resolved as 255 bits and can be implemented bidirectional.

Lets do an example:
The 8 bits out reads 0000 0001 and could mean 0000000000000000000............00000000000000001 (255 bits wide decoded) that is 1 of 255 or 01 of hex FF
or 1000 0000 would be 100000000000000000......000000000000000000000000 being bit 128 0f 255 is set and the hex number 80

If this don't make sense, I can explain more

Whacko

Very interesting idea Whacko.

Bett

Whacko:

  This comes up about every year or so. , its a common misconception. While your right, a 1 woudl be 254 o's followed by a 1, and a 255 woudl be a 1 followed by 254 0's , image the need to turn on 3,4,6,3,1,23,45 bits at the same time.
Cant be done. What your describing woudl take a strobed input changeing every n microseconds with MAch3 reformatting the data. Trouble is, the printer port is strobed for data only every 40us so 255 x 40us woudl be 10ms, which means the highest speed any input coudl be is 10ms, AND thats only if you can find a way to sync the pattern change to the input strobe internally. Of course this could be done with MAch2 putting out a bit to say "change state", then accepting the resultant bit combination as an input byte. BUT then youd still have to sync what number is expected, so I guess in that case MAch3 could put out a number from 0-254 to indicate what pattern to accept next, BUT that still makes all inputs only 10ms
in granularity, vs the 40us they are now. 40ms is bad. Its a long time, Youd be much better of simply doing TCP modbus input where you could get much lower than 1ms per input with over 16000 inputs done each 10ms.
  While many have thjought at first glance that doing a binary decode of an 8 bit input byte is possible, its actually a very slow way to get data in a polled input system. Each input is limited to 1 bit of data in reality, so the interface is 255 times slower than youd think..

'Art


 

SSSHHHHH (48-96 channel PCI I/O Card) I hear tale that MACH USED to have A PCI I/O card that would interface, what happened to it along the way??

(;-) TP