CTRIO2

[johnofeasttex]

Is there any status on the upcoming CTRIO2? It would be great to improve the performance I and O of the current CTRIO. This would open up several posibilities for upgrades for our customers.

[BobO]

Excellent question.

As has been mentioned on this site on several occasions, Host is in the process of developing a new line of controllers. That effort is huge, but is coming nicely. One aspect of that controller is the ability to do a much better job of integration with intelligent modules. With that in mind, we backed off a bit in the CTRIO2, not really sure how we'd want the CTRIO2 to play with the new CPU.

After some discussion, we decided that rather than try to make the CTRIO2 play well with the new controller, it might make better sense to split high speed counting/timing functions and motion functions into 2 separate modules. That would enable us to purify the approach to each function, rather than end up with hybrid that doesn't do either very well. So, the long term plan is to do that. The down side is that those new modules will only work on the new controllers...but you won't mind...trust me. 

Now, we are rapidly approaching a point where the existing CTRIO may start facing parts obsolescence issues. Many of the chip vendors have accelerated the life cycles of some chips to streamline their operations due to the hard economic times. So in the last few days we have started discussing the possibility releasing a CTRIO-1 (rather than CTRIO2) that would be a drop in replacement for the CTRIO. It would use newer technology though, and performance would be significantly better.

Module scan times would be ~250us regardless of what the module is doing. Inputs would go to 1MHz and would be completely hardware based, meaning it would have almost none of the input limitations that the current module has. Outputs are a little harder, though. We could go very fast (perhaps even 1Mhz, although with a dither)...but...in an effort to be compatible with the existing module, we might have to limit the max output rate to 65535Hz per channel, due to the frequency variable being a WORD. I have kicked around the possibility of adding a 1X, 2X, and 4X output configuration option that would allow higher freqs with the existing data, or maybe using a new profile type to drive the higher speed. Not real sure, but we're willing to entertain any option that allows the module to be a replacement for existing CTRIO.

I would welcome thoughts on this because this is something that we may crank up shortly. Ya'll tell me what would give you the most bang for the buck.

[Controls Guy]

After some discussion, we decided that rather than try to make the CTRIO2 play well with the new controller, it might make better sense to split high speed counting/timing functions and motion functions into 2 separate modules.

I think that's a better way to go.

[ATU]

Is the motion module and counter module going to be on similar or the same hardware platform or will there be major differences?

[BobO]

Remains to be seen. I expect them to be very similar, at least for the pulse output version of motion. I'll probably get something running with the CTRIO-1 and then do the next-gen counter/timer on that hardware. The motion module would follow when we've had more time to evaluate what we really want it to be. The CTRIO's input side was rarely the weakness, the output side was a pain. We want the motion module to fix that.

Big functional differences between the CTRIO-1 and the new counter module will be flexible I/O assignment...use any input anywhere, multiple times if you like...and perhaps 8 outputs instead of 4.

The most relevant part of this plan is that we want both the counter/timer and motion functions to be tightly coupled to DirectSoft and the new controller. Which means that those modules become the standardized approach to those functions. Obviously we want to do it right.

[b_carlton]

I want a counter with the capabilities of the D4-HSC (not the H4-CTRIO). Primarily it's the automatic positioning control by means of digital outputs. It is perfect for our application, moving toward a positioning using High speed - low speed - stop type function. Our axes of movement tend to stay in position by themselves so they don't need active servo type operation. The D4-HSC did exactly that. Give me that in a 205 format.

[BobO]

Can that be done with a simple PLS type function?

[b_carlton]

I guess you'd have to check up on the functionality of the D4-HSC unit. It has homing routines. For positioning the CPU merely loads a target and strobes a 'go' bit. The unit determines direction, turns on the appropriate direction output while monitoring a quadrature encoder and limit inputs. At a given number of encoder pulses from the target it turns on a 'low speed' output. It turns off the low speed and direction outputs as it hits the target. Sure there's some minor overshoot but in our situation it's trivial. It retains the position (count-wise, not by active positioning) for the next move. This is with dumb drives. It's an incredible poor man's positioning that works 'well enough' for a lot of situations.

[BobO]

With our plan for the next-gen counter, it would be pretty easy to add that mode to a function block. We wouldn't plan on attempting to emulate the module, however, but would be happy to add whatever behaviors that would fit reasonably well within the design.

[ATU]

The reason I ask about the hardware is that if you had a "standard"  hardware platform  you could download a specific firmware to make the module a direct replacement for the legacy CNTRIO. Another would make the module an enhanced HS Counter module  or motion controller.  You order the module you want, but the differences are only in firmware and faceplates. To me that would be attractive having 1 card to keep on hand and 3 possible functions. Of course if you had the program space, make it software or jumper configurable. Its probably not practical and I have never seen it done that way, but just a thought.

[BobO]

It really isn't a bad idea, but there are definitely other considerations. It could probably be done, but I suspect it would result in a fight with ADC's tech team, not to mention confusion. "Hmmmm. It says it is a motion module...but is it really? I wonder..." Now depending on how I end up implementing the counter/timer and motion function, it might be possible to put both functions into the same module. That would really be quite cool if it could be managed.

No, no. I'm not talking about doing it as the CTRIO...that's the existing problem. But if the module could do four functions, and each function could be a timer/counter variant or a motion axis, and you simply assigned the I/O to each that you wanted, it could just about work...but...

The whole reason I wanted to split them, however, was to free me from the inevitable compromises that would have to happen.

Funny though, that's the way product design happens. We talk ourselves into stuff, and then back out of stuff, and into stuff...etc., etc...