High Speed I/O

[Longbow]

I can see that maybe there is a possibility with a SPEED7 in stead of a Siemens S7.
But I still doubt it. With propagation delay of 6 µs for the input, plus unknow propagation delay for the output, plus CPU housekeeping time, plus program scan time, it will soon be more than 20-30 µs.

I would divide high speed tasks into two sections:
predictable and unpredictable.

For this task, it should be possible to determine e.g. 5ms in advance, when to switch on the Outputs for 1-30µs.
With VIPA SLIO ETS modules it is possible to control decentral I/O in steps of 1µs over Profibus, Profinet, CAN, Ethercat, Modbus/TCP, Devicenet.

Example: Activating the Laser for 11µs or 25µs is no Problem with this System.
Setting Multiple Laser Impulses: 10µs on, 20µs break, 5µs on: No Problem
If the Laser is fast enough ;-)


http://www.vipa.de/en/news/news-details/article/vipa-slioR-ios-exactly-to-the-us-via-field-bus/

 

[rmb550]

You might consider CompactRIO ..... but only if you have a ton of $$$.

 

[JesperMP]

Looks interesting.
The prerequisite is that there is a significant timedelay between sensing and acting (at least 250 µs for Profinet), whereafter the acting happens with a resolution of 5 µs.
The "paper cutting" example they show describes it better than I can.

 

[JRW]

Wow
This has to be a partnership with Siemens, no company can
copy (looks and names) without a lawsuit
This is ET200s i/o.

 

[RMA]

I'll add my support to PLC's suggestion of the FM352-5 High speed boolean processor. I used it to get variable delays for triggering Thyristors with the delay in the range 10 ms - 2 seconds being entered by the operator on the panel. I got an accuracy of about 3 micro-secs over the whole range.

One interesting point is that the FM352-5 can be used as a stand-alone CPU, you don't need to use a PLC with it at all, never mind having to use it in a Siemens environment.

 

[Longbow]

Wow
This has to be a partnership with Siemens, no company can
copy (looks and names) without a lawsuit
This is ET200s i/o.

Not at all!

Smaller and faster than ET200s and additional features.
The only relationship with SIEMENS is competition and the OPTION
to use their Engineering Tool.
But you can also connect the system to Ethercat (Beckhoff),
Modbus/TCP, CAN, Devicenet.

If ET200s and SLIO is physically in front of you, you will see
the difference!

 

[JRW]

Did Vipa steal Siemens part numbers on this product
Too??

Check out the FM352-5. Vipa doesn't have anything that fast

 

[Longbow]

Did Vipa steal Siemens part numbers on this product
Too??

Check out the FM352-5. Vipa doesn't have anything that fast

That's right, VIPA has no bool processor.
FM352-5 is unable to communicate with a fieldbus as master,
no online change,
no change in run,
no debug (on target system),
no analog In/OUT,
limited instructions,
no SCL,
no floatingpoint operations

Useful for some things: yes! (extreme reaction times with onboard digital I/Os)
A CPU: No!
For accurate control of timing: SLIO ETS provides same timing resolution (1µs), but with much more flexibility (for predictable tasks)

 

[PLC]

I tried to get some support in the US on a
Vipa product ONCE.
Their support is terrible in North America

 

[RDS]

Take a look at National Instruments. http://www.ni.com/

 

[RocketTester]

jcr,

I have used Beckhoff's EtherCAT for several years, but only tested for our needs (<0.5 ms AI to DO response). Using the simplest "traditional PLC" approach, a 50 usec cycle time is max. However, they have special digital modules that can be synchronized module-module to 1 usec. You might employ their "digital pattern" DO modules. You trigger on a cylce and the modules then outputs a pre-programmed digital pattern using an internal clock (I expect accurate <1 usec). Any of above would be a rugged "industrial approach" with solid I/O wiring.

Another approach is a custom design using discrete logic or dedicated digital chips like CPLD, FGPA, or DSP, as others suggest. A semi "grad school" design might use NI's high-speed digital PC boards. NI also has FPGA boards, which might allow avoiding custom PCB design. If low-volume that might be the best way. NI long promoted CompactRIO without publishing performance figures. When they finally did, they were underwhelming (100 Sps max for >4 chan), much slower than EtherCAT, but claim you can do better if you custom program the FPGA.