Fieldbus Standard

R

Thread Starter

Ranjan Acharya

What does the Automation List have to say about the eight-headed hydra that the Fieldbus Standard has evolved into?

What is the exact point of this standard now that anyone with any muscle has their own pet bus written in?

"Fieldbus Standard" or "IEC (6)1131.3 Standard" -- what are the point of these standards if all it takes is a little whining and you are in?

Perhaps I am being too cynical.

RJ
 
R

Ralph Mackiewicz

Yes you are being cynical. But cynicism seems to be appropriate here. Every large manufacturer wants their own standard to be adopted. And because the group will only consider defacto standards, nobody can agree on which defacto
standard because it means adopting someone else's standard. Pride can be ugly. The 8 headed hydra is the perfect bureacratic solution because it avoids having to make a decision.

BUT, there is also something else at work here. One of the goals of the Fieldbus people is a single world-wide standard for all fieldbus applications. The assumption underlying the feasibility of this is that there is only one
problem to solve. Unfortunately this is a flawed assumption. The industrial fieldbus market has many varied requirements. Technology is not sufficiently advanced to eliminate that financial and performance tradeoffs that would have to be made if every single application was going to be solved by one solution. [Example: putting Ethernet into a limit switch is not feasible today. Maybe in the future, but not today] Very few people are willing to sacrifice the economic or technical performance of their systems in order to use just one fieldbus for every application from limit switches to flow controllers to PLCs to DCS to mainframe interfaces. Although I think the 8-headed hydra is
downright silly, the existence of multiple "standard" communications schemes is an economic and technical necessity for now.

To look on the bright side: 8 incompatible standards is better than 80.

Best Regards,
Ralph Mackiewicz
SISCO, Inc.

"Standards are great. Everyone should have one." -- Bob Metcalfe (co-inventor
of Ethernet. NOTE: He was being facetious.)
 
R

Rob Hulsebos

>What does the Automation List have to say about the eight-headed hydra
>that the Fieldbus Standard has evolved into?

I now have the 61158 on my desk and am surprised by the haste in which it has been assembled - the chapter on the application layers has separate paragraphs explaining the drawing-techniques used for the specifications of the different bus-systems! So 'the' standard itself is internally not standardised. Strange but true.

BTW, did you notice that little article in the German magazine IEE about the underlying patents owned by certain companies on the technologies in
one of the eight 'heads' ? According to the article the technology is 'free' for 5 years only. What happens next?

>"Fieldbus Standard" or "IEC (6)1131.3 Standard" -- what are the point

The 61131.3 is something entirely different: standard programing languages for PLC, and not at all related to 61158. Which, while I'm typing it, makes me wonder what ever happened to the 61151.6 (?) which intended to specify a common interface for remote I/O systems?

>Perhaps I am being too cynical.

Cynical is sometimes another word for 'realistic'....


Greetings,
Rob Hulsebos
 
W
>BTW, did you notice that little article in the German magazine IEE about the
>underlying patents owned by certain companies on the technologies in
>one of the eight 'heads' ? According to the article the technology is 'free'
>for 5 years only. What happens next?

It's also been reported in Controls Intelligence and Plant Systems Report. According to the report, Endress and Hauser has sent a letter to IEC and CENELEC stating that whoever uses parts of the standard may be risking infringing on various patents. E+H is asking that descriptions and releases for the involved patents be added to the specification.

Interesting stuff. Maybe the "open" Ethernet Fieldbus standard doesn't exist yet.

Willy Smith
Numatico SA
 
Rob Hulsebos asks :

>What does the Automation List have to say about the eight-headed hydra
>that the Fieldbus Standard has evolved into?
>Perhaps I am being too cynical.
>Cynical is sometimes another word for 'realistic'....

Jim Pinto reminds Rob and the List :

The comical situation can only be described in poetry. Take a look at my Fieldbus Poem :
The 8-part Fieldbus Voting Fiasco at :
http://www.actionio.com/jimpinto/fieldbusvote.html

Cheers:
jim
----------/
Jim Pinto
email: [email protected]
Tel : (858)279-8836 (direct)
Web: www.jimpinto.com
San Diego, CA. USA
----------/
 
W
Actually, the Fieldbus Standard, per se, is dead. It may flop around wildly for several years yet, but it is dead as a dinosaur. Why do you think that very smart man Dick Caro bailed?

Industrial Automation + Business Process Integration = DeadBus

The evolving continuum between Industrial Automation and Business Process Integration has already decreed that Industrial Ethernet, or some successor bus will be the winner, because of economies of scale and ubiquity.

Economies of Scale and Ubiquity Rule

Economies of scale guarantee that it will be nearly costfree to implement, while ubiquity guarantees lots of people besides specialized (read, from a manager's perspective, "expensive") process control engineers to keep it fed and happy.
Economies of scale are already happening. For example, it is possible (not
just in a laboratory setting) to produce a sensor with a TCP/IP stack and a DNS address for an additional $15.00 manufacturing cost using a Microchip PIC microprocessor. There is a grad student at MIT who has been running an 8KB "web server" for over a year now. That means that ANY cost sensor, from a lowly prox switch or flow alarm can be "web-enabled" nearly for free. Or
at least, at a 60% gross margin, for a list price adder of less than $50 USD!
Ubiquity is happening too. I have a friend, a member of this list, as well, who used to be a project manager for a process control engineering firm. Now he is an eBusiness Consultant for a major Business Consulting Firm. What does he do? He is a project manager for process control projects. But, it costs lots more to hire him now.
The ubiquity of ethernet and firewire are going to drive the cost of implementation (which is 75% consulting, as a rule of thumb) way down. If anybody with an MSCNE cert can hook up and run decent networks, the requirement for PE in CSE will go down radically.

Ethernet is a Consumer Product Already

We are already seeing wireless ethernet down to the household level, for Christmas' sake! For less than $1000 I can put an Apple AirPort system in my house and network all my Macs. In another year, somebody will have additional systems to be able to network Macs and PC's wirelessly at
consumer prices.
We aren't going to see Fieldbus prices get that low that fast.
The least expensive way to implement Fieldbus as an add-on to existing equipment that I know of costs (NOT "sells for" but "manufacturer's cost") $300 US for the board and the interface. In quantity.
Who, in her right mind, will Fieldbus a $75 prox sensor? Who will pay $500 for a Fieldbus version of a $75 prox sensor?


Walt Boyes

----------------------------------------------------------
Walt Boyes--MarketingPractice Consultants
21118 SE 278th Place - Maple Valley, WA 98038
425-432-8262 home office - 253-709-5046 cellphone
eFax: 801-749-7142 - email: [email protected]
----------------------------------------------------------
 
A

Andy Piereder

> Industrial Automation + Business Process Integration = DeadBus
>
> The evolving continuum between Industrial Automation and Business
Process
> Integration has already decreed that Industrial Ethernet, or some
successor
> bus will be the winner, because of economies of scale and ubiquity.
>
> Economies of Scale and Ubiquity Rule
>
> Economies of scale guarantee that it will be nearly costfree to implement,
> while ubiquity guarantees lots of people besides specialized (read, from a
> manager's perspective, "expensive") process control engineers to keep it fed and happy.
> Economies of scale are already happening. For example, it is possible (not
> just in a laboratory setting) to produce a sensor with a TCP/IP stack and a
> DNS address for an additional $15.00 manufacturing cost using a Microchip
> PIC microprocessor. There is a grad student at MIT who has been running an
> 8KB "web server" for over a year now. That means that ANY cost sensor, from
> a lowly prox switch or flow alar m can be "web-enabled" nearly for free. Or
> at least, at a 60% gross margin, for a list price adder of less than $50 USD!

An interesting comment in light of a visit I made early this week to a semi-conductor equipment customer who actually researched, with an eye
towards implementation, a web device interface. They too were assured that it was going to be much less expensive than their current DeviceNet
implementation, but after a preliminary design was evaluated, it was soon discovered that rather than $15.00, it was going to cost $100.00 and that
such an interface could not possibly be implemented on a PIC.

Since we are dealing in second-source information, I just thought I would add my own source ;-)

> Ubiquity is happening too. I have a friend, a member of this
list, as
> well, who used to be a project manager for a process control engineering
> firm. Now he is an eBusiness Consultant for a major Business Consulting
> Firm. What does he do? He is a project manager for process control
> projects. But, it costs lots more to hire him now.
> The ubiquity of ethernet and firewire are going to drive the cost
of
> implementation (which is 75% consulting, as a rule of thumb) way down. If
> anybody with an MSCNE cert can hook up and run decent networks, the
> requirement for PE in CSE will go down radically.

I hear this quite often, but it appears to me to be a somewhat superficial analysis of economic cause and effect. The problem I have with this
statement is that it reverses the causality vector--Ethernet is ubitquitous not because it is cheap and everybody uses it, but rather because it
provides a desired quality of service for a general market. Ignoring quality of service is a major marketing error. In my experience, companies and engineers make very considered choices when it comes to choosing a bus for their application. The bus with the most value to them is the one which provides the quality of service at a price that is reasonable within their own cost structures. To really understand this, you only have to look at the market history of LonWorks--it is very cheap to implement, uses twisted pair wiring and the interface cards are in the order of $300-500. It has found wide acceptance in the HVAC market but has an insignificant presence in manufacturing and processing markets. The reason is fairly
obvious--quality of service. Something that doesn't provide quality of service, no matter how cheap, is still too expensive.

> Ethernet is a Consumer Product Already
>
> We are already seeing wireless ethernet down to the household
level, for
> Christmas' sake! For less than $1000 I can put an Apple AirPort system in
> my house and network all my Macs. In another year, somebody will have
> additional systems to be able to network Macs and PC's wirelessly at
> consumer prices.

This comment again ignores the fact that networking home computers and implementing industrial networks are significant different challenges.

> We aren't going to see Fieldbus prices get that low that fast.
> The least expensive way to implement Fieldbus as an add-on to
existing
> equipment that I know of costs (NOT "sells for" but "manufacturer's cost")
> $300 US for the board and the interface. In quantity.
> Who, in her right mind, will Fieldbus a $75 prox sensor? Who will
pay $500
> for a Fieldbus version of a $75 prox sensor?

This begs the question--a $75 prox sensor with a thin server is not a $75 prox sensor either. The price one is willing to pay to monitor the state of a point of data is not fixed. If a company has to monitor its cell tower every 30 minutes or face a $10,000.00 fine by the FCC, then the cost of that point(s) is at minimum $10,000.00. Any technology that could possibly fail to obtain the state of that point even once, is too expensive to use. A reasonable engineer will select a technology that can at minimum guarantee reliable monitoring of that point at the lowest price for which it can be obtained.

In the final analysis, ubitquity and cost are only a killer app when we assume that quality of service issues have been rendered a moot point.
Ethernet has long had a role in industrial automation, but not surprisingly, it has been applied where it was and is best suited technically and I would expect that we will continue to see a diversity of bus technologies for the foreseeable future.

Additionally, one should consider that the 'politics' of the buses is not an insignificant factor in determining their success. Choices to use DeviceNet, Profibus, Foundation Fieldbus or Interbus are often as much dictated by the choices a supplier of industrial equipment has made as they are by a customer. Where is the 'big dog' championing Ethernet I/O?

Andy Piereder
Pinnacle IDC
 
Walt Boyes <[email protected]> said :

>Actually, the Fieldbus Standard, per se, is dead. It may flop around wildly
>for several years yet, but it is dead as a dinosaur. Why do you think that
>very smart man Dick Caro bailed?

Jim Pinto comments :

The subject of Dick Caro's resignation first came up at the ISA/99 in Phiadelphia, at the Great Fieldbus debate. See the complete report with comments from 3 end-users and 3 major suppliers, plus Dick Caro's "final word" at :
http://www.jimpinto.com/writings/debate.html

Dick Caro was Convenor of the IEC committee and he resigned in protest against the 8-standards standard. Dick is still Chairman of the ISA SP-50 Fieldbus committee.

Walt Boyes continues :
>Ethernet is a Consumer Product Already
> Who, in her right mind, will Fieldbus a $75 prox sensor? Who will pay $500
>for a Fieldbus version of a $75 prox sensor?

Jim Pinto suggests :

Walt is right! IMHO, SP-50 Fieldbus is also dead - any committee driven standard has no chance in todays fast-moving environment. I respectfully suggest to Dick Caro that he should resign from that "dead horse" too.

Comments and feedback to this list would be welcome from Dick Caro on why he is still at SP-50....

Cheers:
jim
----------/
Jim Pinto
email: [email protected]
Tel : (858)279-8836 (direct)
Web: www.jimpinto.com
San Diego, CA. USA
----------/
 
Forgive me all for not monitoring this list very often anymore.

Fieldbus (capital F) lives in Foundation Fieldbus H1, and HSE. The ISA/ANSI S50.02 Standard is very much alive and in active use. Type 1 of IEC 61158 parts 3,4,5 and 6 is practically identical to S50.02, and is the IEC standard. Type 2,3,4, 5, 7, and 8 are also very much alive and in broad use.
Here are those Types:
1 Original IEC Fieldbus (ANSI S50.02)
2 ControlNet
3 Profibus
4 P-NET
5 FF HSE
6 SwiftNet
7 WorldFIP
8 Interbus

One can argue about the value of the IEC 61158 standard, but it does raise each of the 7 implemented protocols to International standard level. All of these 7 are in wide use. The problem with this document is that no one in
their right mind would implement ALL of them in a single system or device.

I resigned from this committee for several reasons:
A. I had no interest in running the group responsible for "harmonizing" these into one even more impossible document.
B. No company or consortium has any interest in harmonization.
C. I cannot ask ARC Advisory Group (my day job) to sponsor me for the purpose of harmonization
D. I finished the task for which I agreed to accept the job as Convenor. (so it didn't come out the way I wanted)
E. The next generation of Fieldbus architecture is already included as Type 5. (this came out the way I wanted)
F. The methodology to generate complex standards by volunteer committee is broken and there is no fix in sight

I am still technically Convenor, since the terms of my resignation were that I would remain until replaced. So far, there has been no successor.

I remain as ISA SP50 Chairman for 2 reasons:
1) To make sure that IEC 61158 parts 2, 3,4, 6, 7, and 8 are NOT added to this operational standard
2) To make sure that IEC 61158 part 5 is integrated into this standard.

Walt Boyes, Jim Pinto and others are correct, nobody will pay for a $70 (cost) interface to be added to a $75 (selling price) prox sensor. The best example is right on the Honeywell Micro Switch web site. I "shopped" for a limit switch in standard and SDS format. The plain old YZ-2R-A2 large basic limit switch was $7.97. The lowest cost SDS-C2-GLHT81C limit switch was
$171.60, and this uses the "low cost CAN chip."

However, since the microprocessor was added to early HART process transmitters along with the HART chip, the cost of the basic "Smart Process
Control Transmitter" has been reduced to parity with the old "dumb transmitters." The devices are electrically simpler, less expensive to make, and use less power. Software replaces lots of custom analog linearization and calibration circuitry. Use of the Foundation Fieldbus in
these devices actually reduces cost by replacing the need to generate a precision analog 4-20mA signal. This was the projected market for Fieldbus
and we were right. The Foundation Fieldbus protocol is necessary to realize the final vision of the original committee - to do control entirely in field instrumentation. It is now commercially for sale and is proving to be a
financial success for both users and suppliers.

Dick Caro
==============================================
Richard H. Caro, Vice President
ARC Advisory Group
3 Allied Drive
Dedham, MA 02026 USA
Tel: +1.781.471.1123
Fax: +1.781.471.1023
Web: http://www.arcweb.com
============================================
 
Andy Piereder says:

<regarding snipped WB comment on an existing PIC-enabled webserver>

An interesting comment in light of a visit I made early this week to a semi-conductor equipment customer who actually researched, with an eye
towards implementation, a web device interface. They too were assured that it was going to be much less expensive than their current DeviceNet
implementation, but after a preliminary design was evaluated, it was soon discovered that rather than $15.00, it was going to cost $100.00 and that
such an interface could not possibly be implemented on a PIC.

Since we are dealing in second-source information, I just thought I would add my own source ;-)

WB> Okay, if I can't be done, why is there a PIC-based webserver sitting on the MIT website? It ain't my device. It _is_ however there. And it has been there for about a year. And I know of at least three companies who are planning to introduce thin server enabled devices this year.

> Ubiquity is happening too. I have a friend, a member of this list, as
> well, who used to be a project manager for a process control engineering
> firm. Now he is an eBusiness Consultant for a major Business Consulting
> Firm. What does he do? He is a project manager for process control
> projects. But, it costs lots more to hire him now.
> The ubiquity of ethernet and firewire are going to drive the cost of
> implementation (which is 75% consulting, as a rule of thumb) way down. If
> anybody with an MSCNE cert can hook up and run decent networks, the
> requirement for PE in CSE will go down radically.

AP>I hear this quite often, but it appears to me to be a somewhat superficial analysis of economic cause and effect. The problem I have with this
statement is that it reverses the causality vector--Ethernet is ubitquitous not because it is cheap and everybody uses it, but rather because it
provides a desired quality of service for a general market. Ignoring quality of service is a major marketing error. In my experience, companies and engineers make very considered choices when it comes to choosing a bus for their application. The bus with the most value to them is the one which provides the quality of service at a price that is reasonable within their own cost structures. To really understand this, you only have to look at the market history of LonWorks--it is very cheap to implement, uses twisted pair wiring and the interface cards are in the order of $300-500. It has found wide acceptance in the HVAC market but has an insignificant presence in manufacturing and processing markets. The reason is fairly
obvious--quality of service. Something that doesn't provide quality of service, no matter how cheap, is still too expensive.


WB> Whoa! Hold on there, pardner. Ethernet is ubiquitous because it robustly provides a high level of service. You are mistaking the quality of service issue for the 80/20 rule. I do not believe that 80% of industrial automation data interface connections require the level of robustness (still in some cases to be proven in practice) that the Industrial Fieldbusi deliver. There may continue to be a market for industrial strength fieldbusi, and I truly believe there is, but it is gonna be leetle! AT least in comparison with the market for Ethernet and TCP/IP enabled devices.


> Ethernet is a Consumer Product Already
>
> We are already seeing wireless ethernet down to the household level, for
> Christmas' sake! For less than $1000 I can put an Apple AirPort system in
> my house and network all my Macs. In another year, somebody will have
> additional systems to be able to network Macs and PC's wirelessly at
> consumer prices.

AP>This comment again ignores the fact that networking home computers and implementing industrial networks are significant different challenges.

WB> Oh, no it doesn't. You just chose to ignore the fact that systems integrators went to "home computer system" style products 10 years ago,
because proprietary DCS and SCADA products were far too expensive, both to purchase and to maintain. Remember when OnSpec debuted its Windows HMI? Everybody derided it because it wasn't a "good robust DCS". Tell the investors in WonderWare, USData, and the other MMI companies! With the inexpensive backbone of Ethernet products, and the ability to buy them as
ASICs and VLSI silicon, it is going to be very easy to make it possible to engineer the required robustness into Ethernet.


> We aren't going to see Fieldbus prices get that low that fast.
> The least expensive way to implement Fieldbus as an add-on to
existing
> equipment that I know of costs (NOT "sells for" but "manufacturer's cost")
> $300 US for the board and the interface. In quantity.
> Who, in her right mind, will Fieldbus a $75 prox sensor? Who will
pay $500
> for a Fieldbus version of a $75 prox sensor?

AP>This begs the question--a $75 prox sensor with a thin server is not a $75 prox sensor either. The price one is willing to pay to monitor the state of a point of data is not fixed. If a company has to monitor its cell tower every 30 minutes or face a $10,000.00 fine by the FCC, then the cost of that point(s) is at minimum $10,000.00. Any technology that could possibly fail to obtain the state of that point even once, is too expensive to use. A reasonable engineer will select a technology that can at minimum guarantee reliable monitoring of that point at the lowest price for which it can be obtained.

WB> Agreed, and the relevance of your point to my argument is?


AP>In the final analysis, ubitquity and cost are only a killer app when we assume that quality of service issues have been rendered a moot point.
Ethernet has long had a role in industrial automation, but not surprisingly, it has been applied where it was and is best suited technically and I would expect that we will continue to see a diversity of bus technologies for the foreseeable future.


WB> Nah, the bus wars are over. There will be a couple of factory floor busi, and there will be Ethernet. It will not happen in the "foreseeable
future." It will happen very rapidly. Look at how fast 'firewire' is invading the consumer space. "Firewire" is a very dependable industrial
bus, as well. If you can buy "Firewire" video cameras now, with "Firewire" less than three years old as a standard, how fast do you _think_ the adoption of consumer-grade busi in industrial environments will take?


AP>Additionally, one should consider that the 'politics' of the buses is not an insignificant factor in determining their success. Choices to use DeviceNet, Profibus, Foundation Fieldbus or Interbus are often as much dictated by the choices a supplier of industrial equipment has made as they are by a customer. Where is the 'big dog' championing Ethernet I/O?

WB>Oh, I think you'll see at least two of them by the end of 2000.

Walt Boyes

---------------------------------------------------------------
Walt Boyes -- Director of New Business Development
Branom Instrument Co.-- P. O. Box 80307-- 5500 4th Ave. So.
Seattle, WA 98108-0307
Phone: 1-206-762-6050 ext. 310 -- Fax: 1-206-767-5669
http://www.branom.com -- http://www.branomstore.com
mailto:[email protected]
---------------------------------------------------------------
 
R

Ralph Mackiewicz

Jim Pinto wrote:

> IMHO, SP-50 Fieldbus is also dead - any committee driven standard
> has no chance in todays fast-moving environment.

Can anyone point out one fieldbus that is not comittee driven? It seems to me that all successful (in terms of the market) communications schemes are committee driven including Ethernet, and TCP/IP as well as most of the more popular "fieldbuses" like DeviceNet, Profibus, etc. If you want interoperable communications to occur between diverse groups of equipment it would seem that you need to get involvement from a diverse group of people who understand the equipment. A committee seems inevitable. Even the ubiquitous WWW standards are determined by committees cooperating under the IETF of W3C groups (or others).

Regards,
Ralph Mackiewicz
SISCO, Inc.
 
A

Armin Steinhoff

Hi All,

At 08:49 03.04.00 -0600, Willy Smith wrote:
>>
>>BTW, did you notice that little article in the German magazine IEE about the
>>underlying patents owned by certain companies on the technologies in
>>one of the eight 'heads' ? According to the article the technology is 'free'
>>for 5 years only. What happens next?
>>
>It's also been reported in Controls Intelligence and Plant Systems Report.
>According to the report, Endress and Hauser has sent a letter to IEC and
>CENELEC stating that whoever uses parts of the standard may be risking
>infringing on various patents. E+H is asking that descriptions and releases
>for the involved patents be added to the specification.

That problem is related to the foundation fieldbus ...

>Interesting stuff. Maybe the "open" Ethernet Fieldbus standard doesn't
>exist yet.

Yes ... ETHERNET is very "open":

the 10Mb/s version is a collection of four standards -> 10BASE2, 10BASE5, 10BASE-T and 10BASE-FO
Each standard represents a different media base with different encodings.

The 100Mb/s market is split at least in 3 different standards -> 100BASE-TX, 100BASE-TE, 100BASE-FX
All of these standards are based on different media with very different encodings.

Well ... that's the 'seven-headed hydra' of the ETHERNET standard.

The new IEC fieldbus standard gives me more REAL choices ... in speed, performance and determinisms.

Best Regards

Armin Steinhoff
 
Ralph Mackiewicz <[email protected]> asks :

>Can anyone point out one fieldbus that is not comittee driven?

Jim Pinto responds :

yes - most (if not all) of the others are driven primarily by ONE company. When they get a committee involved, it becomes "politics".

<Ralph>
> It seems to me that all successful (in terms of the market)
>communications schemes are committee driven including Ethernet,
>and TCP/IP as well as most of the more popular "fieldbuses" like
>DeviceNet, Profibus, etc.

<jim>
No, there are no committees for all the ones you mentioned. Ethernet and TCP/IP are user-driven de-facto standards. The standard portion FOLLOWS the de-facto useage.

DeviceNet was developed by Allen-Bradley, and handed over to ODVA (the Open DeviceNet Vendors Assocciation) - note the word "Vendors" - it is driven by a committee of vendors. Profibus was initially driven by Siemens, and now by a committee of vendors.

There are very few end-users who are part of the above committees.

<Ralph>
>A committee seems inevitable. Even the ubiquitous WWW
>standards are determined by committees cooperating under the IETF of
>W3C groups (or others).

<jim>
Yes, these are committees - but NOT composed only of the primary Suppliers - and - to repeat, driven by "de-facto" usage (following, NOT restricting).

This is a peculiar "political" ground, and Fieldbus (and profibus and others) are the center of a LOT of fighting between suppliers. The IEC Fieldbus Vote is indeed a joke : The say that the
Fieldbus standard is comprised of 8 standards ! Surely no company will try to integrate all the standards.

Take a look again at Dick Caro's response to this Automation List regarding why he resigned from the IEC committee. And read again the commentary on "The Great Fieldbus Debate of ISA/99"
at : http://www.jimpinto.com/writings/debate.html

And the poem "The 8-part Fiedlbus Voting Fiasco" at :
http://www.jimpinto.com/writings/fieldbusvote.html

Cheers:
jim
----------/
Jim Pinto
email: [email protected]
Tel : (858)279-8836 (direct)
Web: www.jimpinto.com
San Diego, CA. USA
----------/
 
A

Armin Steinhoff

>From: [email protected]
>
>>What does the Automation List have to say about the eight-headed hydra
>>that the Fieldbus Standard has evolved into?
>
>I now have the 61158 on my desk and am surprised by the haste in which
>it has been assembled - the chapter on the application layers has
>separate paragraphs explaining the drawing-techniques used for the
>specifications of the different bus-systems! So 'the' standard itself
>is internally not standardised. Strange but true.

Well ... have a look to seven-headed hydra of the ETHERNET standard. I don't believe that this 'unique standard' looks better internally ..

>BTW, did you notice that little article in the German magazine IEE about the
>underlying patents owned by certain companies on the technologies in
>one of the eight 'heads' ?

The ironic point is that the companies pushing the Foundation Fieldbus as 'The Single Fieldbus' STANDARD didn't release their patents on parts of the underlaying technology of the FF for free and unlimited usage.

As long as these companies maintain their monopoly on parts of the FF technology... the FF can't be a part of an open standard.

The big question is ... why could it happen that the IEC accepted the FF as part of an open IEC standard ???

Regards

Armin Steinhoff
 
W
I generally stay away from the fieldbus wars. Howerver Dick Caro made a comment about HART that is frequently misunderstood and I would like to correct.

> However, since the microprocessor was added to early HART process
> transmitters along with the HART chip, the cost of the basic "Smart Process
> Control Transmitter" has been reduced to parity with the old "dumb
> transmitters." The devices are electrically simpler, less expensive to
> make, and use less power. Software replaces lots of custom analog
> linearization and calibration circuitry. Use of the Foundation Fieldbus in
> these devices actually reduces cost by replacing the need to generate a
> precision analog 4-20mA signal.

This last statement is simply not true.

The same strategies that dramatically reduced the sensor signal processing have long been used to reduce the cost of supporting the 4-20mA current loop. Looking at the Loop Current interface alone actual costs are very low. In fact a large number of custom HART modem designs include the D-A converter on the same ASIC. As a result the cost to produce the "precision" 4-20mA signal is at worst minimal.

This trend is continuing. Smar Research is a major supplier of HART modem chips. Smar Research recently announced an ASIC that include the Loop Current PWM, HART modem, LCD display controller, and IEEE Floating Point Coprocessor all on one low power chip. They expect the cost to be only slightly more than a HART modem alone. Even complex chips are cheap if you have the volume. HART has the volume

But the cost to manufacture a field device does not stop with the 4-20mA interface. HART is much simpler than Foundation Fieldbus. HART take
about 10 times fewer resources (the size of the microprocessor, the amount of memory, the number of lines of software) then Foundation Fieldbus. Most HART applications can be implemented on a single chip microprocessor as opposed to a larger micro (or more then one micro) with external RAM and ROM. No matter how cheap memory and
microprocessors become, this large of a difference increases manufacturing costs.

I believe fieldbus (some flavor) will eventually displace HART. However, the cost to produce a HART device whether measured in development effort or manufacturing costs will stay lower then
Foundation Fieldbus for many more years.

> This was the projected market for Fieldbus
> and we were right. The Foundation Fieldbus protocol is necessary to realize
> the final vision of the original committee - to do control entirely in field
> instrumentation. It is now commercially for sale and is proving to be a
> financial success for both users and suppliers.


I will not speculate about the financial success for users. However, due to the higher development and manufacturing costs, I think the jury is still out as to the financial success for field device manufacturers.

OBTW. There are a large number of ethernet chips sold. However, it is possible to implement HART using a Stereo Codec chip. In fact, the HCF's new bus analyzer will use a PC's sound card not a HART modem. I think it is safe to say there are more CD players, boom boxes and televisions sold then ethernet adapters.

Best regards

Wally Pratt
[email protected]
HART Communication Foundation
http://www.hartcomm.org
------------------------------------------------------------------------------------
 
R
>BTW, did you notice that little article in the German magazine
>IEE about the underlying patents owned by certain companies on
>the technologies in one of the eight 'heads' ?

>The ironic point is that the companies pushing the Foundation
>Fieldbus as 'The Single Fieldbus' STANDARD didn't release their
>patents on parts of the underlaying technology of the FF for free
>and unlimited usage. As long as these companies maintain their
>monopoly on parts of the FF technology... the FF can't be a part
>of an open standard.

The statement made by Fisher-Rosemount is that they grant licenses to anyone allowing use of their patents, as long as the licensee does the reverse for their own technology which F-M wants to use. But only for members of the Fieldbus
Foundation.

So it can be argued that FF technology is not closed, but in my opinion it can't be called open either.

A similar thing with patents happened to Profibus in 1994 (or so) when it turned out that Siemens owned a patent on a part of Profibus, but Siemens quickly issued a statement allowing the patent to
be used.

Patents can be tough, I recall the case where the inventer of the token-ring (happens to be a Dutchman) got IBM on its knees and got paid an awful amount of money. So it is not for nothing that Endress & Hauser would like to see a judicial clear sky.

>The big question is ... why could it happen that the IEC accepted the
>FF as part of an open IEC standard ???

Well I don't think the IEC has the resources to check on everything offered as part of new standards.

Groeten,
RH
 
M

Matthew da Silva

In apologia, it would seem, is the only option in relation to this thread, since my current employer is a manufacturer of fieldbus devices.
Maybe I should just sit in the corner and shut up. But, as many of you know, that is hardly my style (at least online, heh heh; actually I'm a
dweeb-asaurus).

I don't know the answer (yawn). What is a manufacturer gonna do? We've gotta patent somethun', right? Look at it this way. Say, Monsanto (chorus: Monsanto!). (No, no! Don't SAY it, fools...) Let's just say that Monsanto patents a drug that is marketed with a price tag of 100 dollars per pill, and is prescribed to people who suffer from a rare heart-ailment brought on by alcohol and other common indulgences. No problem, right? But -- what happens when an article is published that
documents an ancillary use of the same drug, and which could help people living in truly desperate circumstances, who cannot afford to buy the
treatment. What then? To me, that's sign of far greater ethical problems, compared with what is going on in this list. But, hey, keep it up. Gives me a chance to practice my debating skills.

Matthew, Tokyo
 
R

Ralph Mackiewicz

> >Can anyone point out one fieldbus that is not comittee driven?
>
> yes - most (if not all) of the others are driven primarily
> by ONE company. When they get a committee involved, it
> becomes "politics".

A list would be helpful. I really don't think that there are that many. I would agree that politics is an issue anytime more than 1 person is involved. But I don't agree that ANY standard developed by a committee is only politics.

> > It seems to me that all successful (in terms of the market)
> >communications schemes are committee driven including Ethernet,
> >and TCP/IP as well as most of the more popular "fieldbuses" like
> >DeviceNet, Profibus, etc.
>
> No, there are no committees for all the ones you mentioned.
> Ethernet and TCP/IP are user-driven de-facto standards.
> The standard portion FOLLOWS the de-facto useage.

The IETF is a committee. It sets standards for the Internet. The IETF and Internet standards are loaded with politics. The core IETF standards, TCP and IP to name two, were developed before being used by a committee under the direction of DARPA. I belong to the RFC mail list. There are many many RFCs that are developed prior to usage on the Internet. The IETF does insist that the protocols be developed and in use prior to their official endorsement but this occurs during the development of the standards, not prior to it. IPV6 is another example of a standard, still touted by the IETF and IAB as the future of the Internet, that is developed by committee. Its usage is growing. Just because its acceptance is not immediate does not mean that all committee standards are a failure. Ethernet was developed by
a committee of vendors (I think it was Intel, DEC, and Xerox but I may be missing one). Most of the other Ethernet standards 10BaseT, 100BaseT, etc. were developed by IEEE committees.

> DeviceNet was developed by Allen-Bradley, and handed over to ODVA
> (the Open DeviceNet Vendors Assocciation) - note the word "Vendors" - it
> is driven by a committee of vendors. Profibus was initially driven by
> Siemens, and now by a committe of vendors.

A very large, and more significant, portion of the DeviceNet system was developed by a committee, not by A-B. I don't have the specific
knowledge about Profibus myself, but it has been asserted here many times that Siemens did not develop Profibus. "driven by" is a euphemism for being active in the standards process (a committee).

> There are very few end-users who are part of the above committees.

Not surprising. Most users want to be focused on making stuff (in the manufacturing industry) not developing comm standards.

> >A committee seems inevitable. Even the ubiquitous WWW
> >standards are determined by committees cooperating under the IETF of W3C
> >groups (or others).
>
> Yes, these are committees - but NOT composed only of the primary
> Suppliers - and - to repeat, driven by "de-facto" usage (following,
> NOT restricting).

XML, a W3C standard, was created by a committee prior to its use. I will grant that I am not on those committees and I do not know the ratio of users to vendors. I don't know why that is so critical anyway. Web users (like an Amazon.com buyer) really don't have much interest in what makes the web work so I don't know what value they
would add to the process.

> This is a peculiar "political" ground, and Fieldbus (and profibus
> and others) are the center of a LOT of fighting between suppliers.
> The IEC Fieldbus Vote is indeed a joke : The say that the
> Fieldbus standard is comprised of 8 standards ! Surely no company will
> try to integrate all the standards.

I agree here completely. It is a joke. The IEC fieldbus standard is completely unnecessary. The standards already exist under other organizations.

> Take a look again at Dick Caro's response to this Automation List
> regarding why he resigned from the IEC committee. And read again the
> commentary on "The Great Fieldbus Debate of ISA/99"
> at : http://www.jimpinto.com/writings/debate.html

I won't speak for Dick but I'm not sure he would agree with your assertion that committee driven standards are useless. The IEC standard may be useless, but not all committee driven standards are useless. I think he feels that the FF standard is very useful.

Regards,
Ralph Mackiewicz
SISCO, Inc.
 
A

Armin Steinhoff

At 15:29 11.04.00 -0700, Jim Pinto wrote:
>Ralph Mackiewicz <[email protected]> asks :
>
>>Can anyone point out one fieldbus that is not comittee driven?
>
>Jim Pinto responds :
>
[ clip ... ]
>DeviceNet was developed by Allen-Bradley, and handed over to ODVA
>(the Open DeviceNet Vendors Assocciation) - note the word "Vendors" -
>it is driven by a committee of vendors.
>Profibus was initially driven by Siemens,

Sorry ... again, that's not correct.
The root of the Profibus standard was initially a research project (1987-1990) with the following members:

ABB
AEG
Bosch
Honeywell
Kloeckner-Moeller
Landis & Gyr
Phoenix Contact
Rheinmetall
RMP
Sauter-Cumulus
Schleicher
Siemens

and five German research institutes

FZI Karlsruhe
LRT
IITB
WZL
LPR

> and now by a committe of vendors.

Regards

Armin Steinhoff
 
M

Mike Tennefoss

Dear Mr. Piereder:

I read with interest your comment about LonWorks and "a lack of quality of service." I'd appreciate knowing what you mean by "quality of
service" - are you using this term in the sense of an IP channel quality of service? Were you aware that Echelon recently released the iLON
Internet Server, a Cisco NetWorks product that is Cisco certified and incorporates Cisco's QOS technology?

With regard to LonWorks in industrial control applications, LonWorks is widely used in industrial controls by a wide variety of manufacturers including Hitachi, Omron, Edwards High Vacuum, Wago, Durr, and others. Our market share of the industrial controls market has been variously estimated by different research firms at between 25-35%. Is it possible that you're
simply not aware of the many industrial applications using LonWorks?

I look forward to receiving your comments.

Michael R. Tennefoss
Vice President, Product Marketing
Echelon Corporation
[email protected]

>
> > > -----Original Message-----
> > > From: The Automation mailing list, managed by Control Technology
> > > Corporation [mailto:[email protected]-CONTROL.COM] On Behalf Of Andy Piereder
> > >
<clip>....(Walt Boyes wrote:)
> > > > Ubiquity is happening too. I have a friend, a
> > > >member of this list, as well, who used to be a
> > > >project manager for a process control engineering
> > > >firm. Now he is an eBusiness Consultant for a major
> > > >Business Consulting Firm. What does he do?
> > > >He is a project manager for process control
> > > >projects. But, it costs lots more to hire him now.
> > > > The ubiquity of ethernet and firewire are going to
> > > >drive the cost of implementation (which is 75%
> > > >consulting, as a rule of thumb) way down. If
> > > >anybody with an MSCNE cert can hook up and run decent
> > > >networks, the requirement for PE in CSE will
> > > >go down radically.
> > >
> > > I hear this quite often, but it appears to me to be a
> > > somewhat superficial
> > > analysis of economic cause and effect. The problem I have with this
> > > statement is that it reverses the causality vector--Ethernet
> > > is ubiquitous
> > > not because it is cheap and everybody uses it, but rather because it
> > > provides a desired quality of service for a general market. Ignoring
> > > quality of service is a major marketing error. In my experience,
> > > companies and engineers make very considered choices
> > > when it comes to choosing a bus for their application.
> > > The bus with the most value to them is the one which
> > > provides the quality of service at a price that is reasonable
> > > within their own cost structures. To really understand this, you
> > > only have to look at the market history of LonWorks--
> > > it is very cheap to implement, uses twisted pair wiring
> > > and the interface cards are in the order of $300-500. It has
> > > found wide acceptance in the HVAC market but has an
> > > insignificant presence in manufacturing and processing
> > > markets. The reason is fairly obvious--quality of service.
> > > Something that doesn't provide quality of
> > > service, no matter how cheap, is still too expensive.
<clip>
> > >
> > > Andy Piereder
> > > Pinnacle IDC
 
Top