Plc suggestion

http://samplecode.rockwellautomation.com/idc/groups/literature/documents/pm/1756-pm018_-en-p.pdf

there are some important items with the IEC standard
Most usual PLC's Claim to be at that level
But they are clearly not.

I had one specified here in Ozz recently - looked like Schneider had partly been in someones ear from some of the specification as it called for 'hot standby' PLCs - impossible to have two separate PLCs in hot standby mode.
Two x PLCs in separate cabinets on separate 24VDC battery sets.
One running and one hot standby.
Changeover to be totally seamless - do not know how you do that with 2 separate PLCs - can do with dual processor and dual power supply types with a central monitoring unit.
Fault detection on all I/O cards.
If an input or output card fails switch to the corresponding card in the other PLC - getting really impossible now. Do not know how you do that when the other PLC is in standby mode.
Even worse, the specification called for all the same nonsense for analogue input and output cards and comms cards!
Yeah - right!!
I asked the consultant why they would not allow a dual processor and power supply PLC with a single set or racks - his reply was to eliminate a single point of failure.
I asked the question how many extra single points of failure is this format going to add to the control system? Question from him - what do you mean? It was for a generator system - answer was - you would have to add a mile of relays to get two signals for everything, dual output isolators for all analogue inputs, how could I duplicate communications from two sources particularly Modbus RTU sources where only one master is allowed, are you going to duplicate the generator controllers on the engines, how do you detect which generator controller is in use to allow communications to the correct controller? Quite a few more questions along the same lines. There was a deadly silence.
I sent the specification back with the comment 'regret unable to quote'. They wondered why as they wanted me to do the job.
It was another one of those silly data centre jobs that are around at the moment.
It was for a large US based computer company. They are absolutely paranoid about single points of failure and you finish up adding thousands of single points of failure to overcome the paranoid belief that it is the PLC that will fail - have not had one fail in over ten years but have seen plenty of other stuff fail!

aec said:

PLC REQUIREMENTS:

i. PLC shall be designed, constructed & tested in accordance with IEC 61131.

ii. PLC shall have self diagnostic feature, to detect failure. When a failure happens it should raise alarm.

iii. Failed modules within the PLC system shall automatically detected, identified and alarmed via output contact(s), diagnostic status words & LEDs without interruption of normal processing functions. On-Line removal/replacement of any PLC system module shall be possible without having reconfigure system software, alter system wiring or cabling, or re-initialize the PLC system.

iv. The PLC shall support PID closed loop control, open loop control, and logic & sequence functions.

v. The PLC shall support Report-By-Exception and Report-All status data.

vi. The PLC shall automatically upload data to the Master station upon receiving a poll request.

vii. The PLC shall support remote configuration, remote down loading, remote uploading of stored data, and remote data base modifications.

viii. The PLC shall be capable of being rebooted via a remote SCADA signal.

ix. The PLC Mean Time between Failures (MTBF) shall not be less than 20,000 hours. This is applicable to the full complement of hardware comprising the PLC.

x. It shall be possible to configure the same PLC to function as: a) Unit PLC Station (slave to a master).

2. CPU:

CPU shall be redundant.


When PLC failure occurs, the discrete outputs shall freeze to its current state or turn off.

3. I/O DETAILS:

Digital ip � 31
Digital op � 7


A failure or fault indication on any I/O module shall initiate a system status alarm.

Discrete outputs that drive solid state inputs, i.e annunciator points, shall be provided with dry relay contacts.

Digital O/P modules shall be 120VAC


4. POWER SUPPLIES:

i. Incoming Power source to the PLC system will be 120VAC

ii. Power fail detection & recovery.

5. COMMUNICATION:

i. Two (2) 10/100 Ethernet communication ports with RJ-45 connector.

ii. Isolation: Ethernet port isolated to 500Vdc

iii. One (1) RS-232 & One(1) RS-485 serial communication port.

iv. The PLC shall be capable of operating in a stand-alone manner, with full functionality, independent of communication with the Master station.

v. In case of a communication failure with the master, the PLC shall be capable of archiving data points. The number of points shall be specified in the PLC Specific Requirement Sheet.

vi. Upon recovery from communication or power failure, the PLC shall be ready to accept automatic download of database changes from the master station and also it shall be ready to respond to a poll request within 30 seconds of re-establishing communication.
6. Environmental Conditions:

Operating Temperature range: 10� to 70�C ,
Relative Humidity:15-95% non-condensing

Please suggest me a PLC .

Click to expand...

check ABB's PLCs. Looks like them. 800 series.
Since there is no requirement for failover time... check also S7-300, siemens (I'm not sure for the rest of specification, but CPU redundancy can be obtained)

CPU redundancy is IMHO nonsense with required configuration (31 DI, 7 DO).
CPU is the most reliable element in redundancy chain. Making it redundant doesn't contribute the system reliability, unless you make I/O cards redundant but then you should make field equipment redundant as well. Makes no sense.
Usually when I see something like this, it is either that someone is aiming to specific vendor or it is copy/paste job (datasheets or previous specifications).

CPU redundancy makes sense when you have a lot of independent I/O cards (distributed as well) and central CPU that controls all of them.

DO card(s) should be failsafe (to have deterministic state in case of failure), which means CPU also should be failsafe.

good luck

AEC,
Does your company have a standard PLC brand already in use? If yes then I would suggest starting there.

At our company we have standardized on Rockwell Control Logix so that would be my starting point. I believe they could meet all of your requirements. But then again Siemens most likely could along with others.

Rich

rkukl said:

AEC,
Does your company have a standard PLC brand already in use? If yes then I would suggest starting there.

At our company we have standardized on Rockwell Control Logix so that would be my starting point. I believe they could meet all of your requirements. But then again Siemens most likely could along with others.

Rich

Click to expand...

Hi Rich
you need to read through the post
it is not his company

Happy new year to all

Really I am much relieved now. I have had all this doubts and started thinking that my knowledge is tooooo dumb and now after reading this post's I have gained some confidents and could understand that it should be a collection of specification from various PLC brands.

I will speak about this to my boss and let him know about this; also I would ask him to contact the manufacturers about these specifications.
Also I would ask him to contact the engineer who gave these specification and know which brannd and PLC he had in his mind.
I will also browse for Siemens s7 400 H , s7 300 and Omron CSID .

I will update my post as I get more result or news . mean while
PLC.net and to all the members in it HATS OFF :site:

Happy new year to you to AEC