SoftPac Questions


#1

Mary, since the architecture of a PC is entirely different than the Pac controllers, is there still an unpredictable aspect to the “up to 100ms delay time” for the stopchart, startchart and callchart commands?

Is the time required for these commands a fairly predictable amount and approx how much time?

Is it similar to running inside an LCM4?


#2

Hi Barrett,

Yes. No. No.

Yes, there is an unpredictable time for those chart commands to action.
The PC is using the same core code as the controller.

No. The time is not predictable, thus we cant say how much time.

No. Its not similar to a LCM4. That controller does not support all the cool Ethernet commands that the SNAP range does.

That said… with 64 charts, do you really need to run more than that by stopping some and running others?
64 is a lot of charts… a real lot…

Ben.


#3

It’s more similar to PAC Sim, which you could try on a variety of PCs, if you’d like to time those numbers.

However, if you’re worried about startup time of charts, I’d recommend using a different method. For those charts that must start immediately, add a loop at the top of the chart that waits for a flag to exit the loop (don’t forget your short delay, of course). Then, in your powerup chart you’ll start the chart right away. That chart will hang out in the loop until you throw the “Go” flag.

BTW, you implied in another post about how a loop without a delay will suck up ALL the machine’s time – not true, it’ll just squander that chart’s turn when you’re better off adding a delay to give the other pieces running a bigger slice of the overall pie. Page 4 of form 1776 gets into this concept.

-OptoMary


#4

Ok, guys I did not start programming Opto yesterday…hehehe. Yes I understand, I just brought this up becuase a lot of people used to (LCM4) and unfortunately still do use the charts like subroutines. For instance, the Callchart command is actually a great command, but because of the architure issue, it is too slow and unpredictable. The point is this command provides a simple means of sequentially processing logic and still putting logic into separate charts. I wanted to make it a point here, about the slow and unpredictable aspect, because newbies will try to use charts this way. When in FF I used the call chart command all the time, but since PAC I rarely start and stop charts, there really isn’t need to do so for the most part. It’s fine if one likes to use run once charts who’s start time is not crutial.
No I did not mean it would suck up all cpu time, but what I was refering to is that if carefule attention is not paid to how long that loop is going to run versus other items in the chart that might need to be processed or “missed” while that loop is running, odd things may “appear” to happen in your program…