How Many Devices on H1 - Archive

[Stephen Mitschke]

The questions you ask cannot be answered without knowing more about how youplan to install them. IF you wish, it is possible, though not veryadvisable, to put up to 240 devices on a single segment. A more general ruleof thumb is for an Intrinsic Safe installation to use 4-6 devices/segmentwhile if the segment is N.I.S. you can go as high as 12 devices.Having said this you need to do the engineering calculations for bandwidthand power per segment based on cable characteristics AND device powerrequirements.2. Sorry I don't understand this question. Perhaps you meant to use adifferent word than rope?3. I understand you can transmit about 120 'blocks' per cycle and stillmaintain a reasonable cycle time. National Instruments (www.ni.com) has somebackground about this on their web site.4. Fieldbus control in the FIELD must be on the same segment.5. Has anyone tried connecting one PID output to multiple AO's?Theoretically it should work but I haven't tried it.Thanks for your comments folks.Ian-----Original Message-----From: ECG, Freddy del Toro Leyva [SMTP:FdelToro@moa.minbas.cu]hello: 1. - do I need to know how many instruments or devices to FieldbusFoundation like maximum can be connected in a channel H1??? 2. - how many do I rope of regulation PID like maximum I can connect in busor channel H1 to 31,25 Kbps??? 3. - that quantity of blocks of functions can have a channel H1 to 31,25Kbps?? 4. - is it advisable to have a regulation knot in buses(H1) or differentchannels? 5. - in F.F it is possible to connect at the same time the exit of aregulator PID to several valve at the same time??? greetings, Freddy D.T PCE.- - - - -I have tried to connect the output of a PID block to multiple AO, and Ididn't succeed. I had to use a "splitter" block, which to my knowledge isn'ta common offering on FF devices. I was attempting to do this in a DeltaVsystem.One problem is . . . where to tie the BK_CAL parameter from the 2nd AOblock?John RezabekBP Amoco BDO Manufacturing- - - - - - - John,BKCAL_OUT of an AO block is used by the PID block for two purposes:(1) To calculate OUT parameter (how much difference from the previousoutput is needed) with BKCAL_OUT.value, and(2) To know from BKCAL_OUT.status if the tracking action is necessarywhen the final control element is not under PID control.In theory, you can connect PID.OUT to multiple AO.CAS_INs and leaveAO's BKCAL_OUTs unconnected except one. The PID block is interestedonly in the AO block with BKCAL_OUT connected to the PID's BKCAL_IN. Other AO blocks (and other final control elements also) may be expected to behave in the completely same manner.This assumption fails when one of control elements come to anunusual state (loss of air, friction, manual operation, etc.).Since it could endanger your plant, you should be very carefulin applying this scheme. I am not sure your design software warnsyou.If you want to use PID block in a control valve to control AOblock in the same valve as well as other AO blocks in other devices,it would cause another difficulty. You need TWO Link Objects forthe PID block in the valve, one for internal connection and the otherfor publishing. I don't think all devices can do this.Chuji Akiyama- - - - Dear Mr. Rezabek,I'm on leave now so I cannot test it out, but here is what I belive thesolution could be. I'm giving a longer explanation that I believe you do notneed, just for the benefit of others on the list.For the FF status Quality there is Good, Bad and Uncertain. The Good comesin two forms: Good(Cascade) and Good(non-cascade). Blocks that have aBKCAL_IN (such as the PID) use the Good(Cascade) for OUT when they are OK,blocks that do not have a BKCAL_IN (such as AI and Arithmetic) useGood(Non-cascade) on OUT when they are OK. The Good(Cascade) has substatusto enable a bumpless mode transfer in the lower block using a handshakingmechanism on the "cascade structure" made up of the OUT-CAS_IN in theforward direction and BKCAL_OUT-BKCAL_IN in the backward direction.Blocks that have a CAS_IN input (like the AO) expect to make aninitilization handshake using the substatus if the Quality is Good(Cascade).If the BKCAL_OUT of the block (in this case the second AO) is not connectedto the BKCAL_IN of the higher block (in this case the shared PID) thebumpless handshake will not occurr. Therefore the second Ao will not go intoCas mode. However, if the CAS_IN quality is Good (non-Cascade) the AO blockwill accept the value "as-is" without the handshake.Although using the splitter block is best for a split range solution,another alternative is to use a block that has a Good(non-Cascade) output,i.e. any block without a BKCAL_IN and put it between the PID and the secondAO. For example, you may try an Arithmetic block where the PID goes to INand its OUT goes to the second AO. Of corse this means that for the secondAO there is no bumpless transfer when you take it out of Cas mode, but mostlikely you will not be doing that anyway.The fact that CAS_IN accepts a Good(non-cascade) value without the feedback"cascade structure" BKCAL is used e.g. for ratio control where thearithmetic block goes direct to CAS_IN of PID, or an AI direct to AO.BTW: Field devices from Smar have dynamically instantiable blocks and allbut one of them support the splitter block as does the DFI302 linking device(aka controller) in SYSTEM302.Best regards,Jonas Berge- - - - - You can connect an output to multiple inputs but not the other way around.The input would not know what to do with conflicting incoming data. The Output Splitter is described in the FF Function Block Specifications,part 3. The Interoperability tester does not test this block yet, so it isstill considered a "custom" block. The best location for this block is inthe final control element. Smar has it in their positioner. Fisher DVC willhave it in the valve positioner soon. I am not aware of otherimplementations.The one available in Delta V follows the FF specifications exactly. You havetwo outputs for two final control elements: OUT1 and OUT2. The Backcalculations are respectively Backcal1 and Backcal2. There is a feature that can be included in the standard splitter that allowthe master controller to bypass a faulty slave in cascade control. I amproposing to include it in the standard. You may remember that we discussedthis during the last ISA conference in New Orleans.Regards,Marcos- - - - -- Chuji,I tried the scheme as you propose . . . only connect one BKCAL_OUT back tothe PID. The second AO (in this case a legacy 4-20 mA AO) wouldn't track theoutput of the PID. I don't recall if the "primary" AO continued to function.I vaguely remember the status of the CAS_IN of the second AO to be"Constant", that is, both of the least-significant bits of the status wereset.Configuring a "Splitter", which accommodates BKCAL_OUT from two blocks,solved the problem.John Rezabekrezabejd@bp.com- - - - Hi Ian....For FUN list:We had a question on whether a backup LAS is allowed in the field withDeltaV's redundant H1 card. See 'yes' answer below. -----Original Message-----Yes , you can have backup LAS in the field, together with redundant H1cards. We didn't show that in the video as it would have been too long.So the redundant card has a copy of the LAS , and then field device can havea backup LAS. The card has a little more than backup LAS as it needs tosubscribe to view lists etc.The LAS in the redundant card will alwats take over before the backup LAS inthe field device, due to the addressing schema we employ.RegardsGary Law Product Manager DeltaV- - - - If you want both AO blocks sending back calculation (backward path) to thePID block, you will have to use a Splitter(splitter or sequencing algorithm)or any other FB with two outputs and two back_cal_in's, as the OSDL blockdeveloped by Smar (output selector and double cross limit algorithms).Regards.Sergio Hideki Tateishi