Page 1 of 4 1234 LastLast
Results 1 to 13 of 42

Hybrid View

  1. #1
    Join Date
    Feb 2007
    Location
    Helena, Montana
    Posts
    2,155

    JACE or Appication/PLC

    Reading another thread where this started to be a heated topic and became curious. Realistically, what is the difference between utilizing a JACE as your "critical equipment" controller as compared to another device? Now let me be specific, I mean a dedicated JACE with I/O modules for this purpose.

    Both are suseptible to the same dangers. Both basically have the same weaknesses and similar strengths.

    I would lean toward the JACE for more flexibilty, but that probably comes from my lack of experience with some of the other "better" controllers on the market.

    Thoughts?

    Don't worry zombies are looking for brains, you're safe...

  2. #2
    Join Date
    Dec 2001
    Posts
    214
    Exucution times are more critical in the PLC world vs the BAS environment.

    Its important that modified code; implementation/updates are performed "on the fly" without a re boot etc.

    Initial thoughts....

  3. #3
    Join Date
    Feb 2005
    Location
    Where it's dark & damp
    Posts
    609
    We've used a Jace as a field controller before and will again. I will say that it doesn't seem as robust as most of the field controllers we use. The one thing I hate is the fact that it reboots with the loss of comuniction with the NDIO. I have seen this happen a few times and it also takes more time to start executing it programs on reboot than all our field controllers as well.

    We do have plant controllers with almost twice the point count so there are benifits in both applications.

    The one thing the Jace has to it's advantage is the multiple protocol capability. We have used it as a slave modbus device, a BACnet device and a Lon device to comunicate to other supervisory controllers. I see that as its advantage hands down.
    It took actually learning something to realize how much I had to learn.

  4. #4
    Join Date
    Dec 2001
    Posts
    214
    The Jace does not look to appealing for this application.

    The JACE fastest is 500ms (note they really mean minimum 800ms).

    You may consider this
    Loytec Linx products can be configured to 10ms PLC friendly execution rates, this allows you to run faster inputs when you have better program cycle rates.

    Some more information:
    LINX 150
    30,000 total points
    24 I/O modules (over 500 I/O)
    20 simultaneous web clients

    LON (FT/IP), BACnet(MSTP/IP), Modbus (master and/or slave) and OPC are standard with no licensing.

    This would be a good comparison for a platform that crosses both spectrum's of HVAC/BAS and PLC environments.

  5. #5
    Join Date
    Jul 2008
    Posts
    1,475
    Quote Originally Posted by acddc View Post
    The JACE fastest is 500ms (note they really mean minimum 800ms)
    .
    ...are you sure about this?
    I can resolve a timer in milliseconds in a jace so I guess I'm wondering how this is possible with an execution time ofwhat...800ms
    I guess one of us is just misinformed.
    A bit weak claiming an authoritative position when yr facts are plain wrong...could backfire and make you look very stupid

    ...not saying you are of course. I'd never do that!

    ...sky still falling?

  6. #6
    Join Date
    Jan 2003
    Location
    USA
    Posts
    1,689
    Quote Originally Posted by MatrixTransform View Post
    I can resolve a timer in milliseconds in a jace so I guess I'm wondering how this is possible with an execution time ofwhat...800ms
    Not saying I have any clue how the Jace handles timers or the exact cycle rate, but....

    In a RTOS, timers can just hold a static count based on "system time” when an event was triggered (sys time is copied into them). System time may be from a RTC, system clock cycles, etc. Each cycle of the program these are updated by comparing their static value with the current system time.

    Soo. They may well have resolution that is in ms or less, but they may not be updated anywhere near that rate.

    I would be interested in what the real Jace cycle time is. I fail to see any need in the HVAC world for 10ms cycle times. Lighting maybe, but even that could done at 5x that rate (blink warnings).
    Propagating the formula. http://www.noagendashow.com/

  7. #7
    Join Date
    Jul 2008
    Posts
    1,475
    Quote Originally Posted by orion242 View Post
    ...In a RTOS, timers can just hold a static count based on "system time” when an event was triggered (sys time is copied into them).
    public static Clock.Ticket Clock.schedule ( BComponent target, BRelTime time, Action action, BValue arg)

    Schedule an single action at the specified relative time from now. This action is scheduled independently of changes to the system clock.
    Parameters:
    BComponent target - BComponent to invoke the action upon.
    BRelTime time - relative time from now to invoke the action.
    Action action - the action to invoke on this schedule's component.
    BValue arg - argument to pass to the action or null if no argument is required.
    Returns:
    Ticket instance which may be used to cancel the scheduled action.
    Exceptions:
    IllegalArgumentException - if time less than or equal to zero.
    1 + 1 = 3 ( *** for very large values of 1)

    ...everybody wants a box of chocolates and long stemmed rose

  8. #8
    Join Date
    Jul 2008
    Posts
    1,475
    Quote Originally Posted by orion242 View Post
    I would be interested in what the real Jace cycle time is.
    Event based ... ECHO...ECHo...ECho...Echo...echo...ech...ec...e... .
    1 + 1 = 3 ( *** for very large values of 1)

    ...everybody wants a box of chocolates and long stemmed rose

  9. #9
    Join Date
    Jan 2003
    Location
    USA
    Posts
    1,689
    Quote Originally Posted by MatrixTransform View Post
    Event based ... ECHO...ECHo...ECho...Echo...echo...ech...ec...e... .
    Yea I read the manual.

    Event based if the event is slow enough maybe.

    Pretty sure I could hook my freq generator to output 100hz square wave to an input. Tie that input to an output on a wire sheet, then measure the output.

    Want to bet the output is far less? How about 1,000Hz or 1MHz, its ON/OFF "events" right??

    Load the jace up with lon, bacnet, modbus trunks with 30+ modules each, think it will remain the same rate?

    Its slicing the processor time between all the tasks. Load it up and the effective cycle time is bound to drop. Your testing of the timers already proves this.

    Again I fail to see any reason any HVAC controller needs to have shear speed, so the disscuion is pointless IMHO.
    Propagating the formula. http://www.noagendashow.com/

  10. #10
    Join Date
    Jul 2008
    Posts
    1,475
    Quote Originally Posted by orion242 View Post
    Load the jace up with lon, bacnet, modbus trunks with 30+ modules each, think it will remain the same rate?
    Its slicing the processor time between all the tasks. Load it up and the effective cycle time is bound to drop. Your testing of the timers already proves this.
    Again I fail to see any reason any HVAC controller needs to have shear speed, so the discussion is pointless
    of course the testing of the timers proves this is true ... and thats why I posted it.
    It was a quick and dirty test ... more than 10ms on a partially loaded JACE(6) ... possibly in the order of 40ms....certainly not 800ms!
    Like I said could always try the same test on a faster platform I suppose ... pointless though it is.
    and yeah anyways, who is going to use a loaded up 'anyController' for critical processes?

    Yes, this is a hvac controls forum and yes the discussion is pointless....but it was framed to invite comment
    So Im happy to draw the fool out and let him keep making wild statements for as long as it takes for him to quit nagging about it.

    Who gives a jot about how crap he thinks a Jace is ... I dont have a very high regard for anything JCI ... occasionally I let something snide out about it but generally I keep my mouth shut. I certainly dont get into pissing competition with the JCI-junkies here....

    that would be very silly.
    1 + 1 = 3 ( *** for very large values of 1)

    ...everybody wants a box of chocolates and long stemmed rose

  11. #11
    Greetings to ControlMT. I would like to respond to this topic will confine my response to you. As such, my comments are directly solely to you out of a necessity.

    I would realise a boiler and cooling plant on different levels of consideration in a BEMS.

    For a program change I can see that the benefits of instant reloading in a boiler plant may not be necessary. Yet, in a cooling plant I can visualise many issues with the I/O cycling off and PID loops resetting when this is not advisable or desired. Or, that all the logic is in one place unless it is very carefully planned distributed. The downfall of an event based system could be this very point of where the control is executed. Even a device with high cycle rate or response timing could be lost due to the nature of the I/O as user Orion deftly points out. If the I/O is Modbus communication, then fast cycle times are not a consideration. It would be impossible for a fieldbus or even Modbus TCP to respond on network in 10ms or less. Such a timing comparison is reckless. But, if the I/O is directed connected via a higher speed, regulated internal bus then faster cycling rates can be advantageous for some operations such as lighting and sunblind integration and operation, or control of a generator. The ability to check twice for an error once is good practise.

    From management and documentation level if a user is maintaining a standard PLC environment much of the programming is across vendors platforms so code can be shared. This ultimately gives options to an integration specialist to move to multiple vendors with largely the same code library. This is 100% in theory and 60% in reality. But, 60% is better than 0%. 0% means you have a single vendor offer. Scalable control offering is also a consideration. Can you obtain both small and large I/O counts with the same programming and features? Save engineer time.

    Some PLC manufacturers products are very inexpensive but may lack for features found in building automation such as scheduling. What is expensive? And what is your labour time? Your service time? One thing I most certainly disdain is coming upon a controller which only has blinking lights. Blinking lights are better suited on a tree. I prefer a display interface locally. I place value for this, whereas another frustratingly not.

    What I suggest is bring in some documentation and make some comparisons. Lastly, compare pricing and any costs associated with the product name. One popular PLC vendor I have used charges thousands for software. They do this because their name is big where others have no charge. Compare. You can assign any cost premiums, warranted or not.

  12. #12
    Join Date
    Jan 2003
    Location
    USA
    Posts
    1,689
    So let me ask this about what your shilling here....

    Quote Originally Posted by AtticusFinch View Post
    For a program change I can see that the benefits of instant reloading in a boiler plant may not be necessary. Yet, in a cooling plant I can visualise many issues with the I/O cycling off and PID loops resetting when this is not advisable or desired.
    So I'm going to load in a new program that's "instantly" reloaded.

    Can you explain to me how none of the outputs are going to change state after this? All the PIDs are going to be at their same state?? What if they are removed, added, connected to something else??

    Even IF the controller has some magic inside that saves states of everything, how would this be applied to modified logic?

    This 10ms cycle time would be more of a curse than a benefit 99.9% of the time.

    Gee <insert tech name here>, that program snafu managed to cycle the output on/off every 10ms and blew out the control relays, contactors, and pump couplings in record time..... If only I could only find a tech that can react in 500ms and kill the controller before the $hit hits the fan....
    Propagating the formula. http://www.noagendashow.com/

  13. #13
    Join Date
    Jul 2008
    Posts
    1,475
    Quote Originally Posted by AtticusFinch View Post
    ... the I/O cycling off and PID loops resetting...
    Look, every control system may behave differently in this situation.
    All that is important, is that IO and LOOPS and stuff go to a known state when rebooted or reprogrammed or whatever....AND that the clown that installed it has appropriately engineered the system to account for it.

    Yr all playing semantics with the word 'critical' here ...
    does critical mean ... continue to run without interruption?...or
    does it mean...not doing something destructive in a reboot?
    or does it mean ... engineering the control system appropriately???

    ...i started my working life in a steel mill ... worked in an engine casting and milling plant ... pointed Solar Concentrators to track the sun ... and of engineered more HVAC control systems and more PLCs than some people here have had hot-dinners.

    I can say one thing ... for many of those applications I would never consider using a JACE ... and I certainly wouldn't consider using a Loytec anything either. Some of them, we built our own dedicated embedded controllers (LON based actually)

    ... get the question right first before applying half-arsed solutions.
    1 + 1 = 3 ( *** for very large values of 1)

    ...everybody wants a box of chocolates and long stemmed rose

Page 1 of 4 1234 LastLast

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  
Comfortech Show Promo Image

Related Forums

Plumbing Talks | Contractor Magazine
Forums | Electrical Construction & Maintenance (EC&M) Magazine
Comfortech365 Virtual Event