Hi,

Answering some of the questions:-
*I guess what would be interesting, would be what address is going over the
wire for "30001" for example.*
The address that gets sent over the wire is the address starting from 0 i.e
300001 would be address 0. I didn't know that.

*Also as a register is always a 16 bit value, the increments by two sort of
puzzle me.*
The Modbus registers are numbered 300001 thru to 365536 (or whatever the
highest is for the device), they are 16 bit registers they don't increment
by 2. If you were mapping 32-bit data types to the registers then of course
you would increment by two but I don't know of any other situation or
device that increments the addresses by 2.


*Looooong time ago, back when the Modbus came to life it was simply a
memory area*
I do vaguely remember working on a device that didn't differentiate between
the memory areas and would just use the address when reading and writing,
the leading digit would just specify data type. 000001 would be a bit in
400001.


*so would you suggest we leave things the way they are or should we change
something?*
I would, it is the most commonly used. I think I have also seen the format
0x00001, 4x00001 being used but is less common.

*I guess we might think about adding different address parser flavors in
the future. So we could use the current one per default, but we could add a
Schneider Field parser that helps convert %MW1 to our
holding-register:4000001 ... or whatever the scematics are.*
You would want to be careful doing this because a lot of the times the
mapping between Modbus Registers and the internal addresses is
configurable. Newer Schnieder PLCs I think have an option to start the
numbering at %MW0 , Rockwell Micrologix PLCs have a configurable mapping
table so you can select the data table the Modbus addresses get mapped to.

Some other things that might not have come up,
There is also an extended memory (Leading digit 6) area on some devices.
Schneider Quantum controllers I know use this.

There are a lot of devices that support Modbus, but the quality of some of
the implementations can be very poor. I wouldn't be relying on some of the
non-essential fields (Transaction Ids, Lengths) as they can sometimes not
be populated.

I also wouldn't assume that every device supports all the function codes.
Some devices (Honeywell C200/C300) allow you to select which function codes
you want to use for each device to try and work around this.

For Modbus TCP, some servers won't allow more than a certain number of
connections at any one time. Sometimes this number is 1. It can be very
annoying to troubleshoot.

Kind Regards

Ben





On Wed, Jul 15, 2020 at 9:53 AM Niclas Hedhman <nic...@hedhman.org> wrote:

> Let me try again; There is no 3xxxx addresses in the products. It is a
> notation written with text, by people. the first written digit convey the
> type. That's it.
>
> Looooong time ago, back when the Modbus came to life it was simply a memory
> area that one could read and possibly write. And then documentation
> specified what was in each bit/byte/word. And with such approach, it is
> probably possible to implement Modbus in less than 100 bytes of very
> expensive (E)PROM back then. Not until later was there address validity
> checks and what not.
>
> FTR; It is completely beyond me how Modbus became as popular as it has, and
> that this happened decades after it was introduced on the market. When I
> was part of a startup and we were developing a serial protocol in
> 1984/1985, we knew of Modbus and thought it was a really poor protocol
> (couldn't even do floating point numbers). I don't like it, but just about
> every electric thing can be purchased with Modbus option...
>
> Cheers
> Niclas
>
> On Wed, Jul 15, 2020 at 9:39 PM Christofer Dutz <christofer.d...@c-ware.de
> >
> wrote:
>
> > Hi Niclas,
> >
> > Protocol-wise the numbers could be up to 65535 as they use an unsigned 16
> > bit integer as an address.
> >
> > I guess what would be interesting, would be what address is going over
> the
> > wire for "30001" for example.
> > Also as a register is always a 16 bit value, the increments by two sort
> of
> > puzzle me.
> >
> > I think Modbus is an extremely simple protocol but we need to get a
> > feeling for it's usages.
> >
> > It does seem as if the industry tends to wrap some aspects in software
> > bubble-wrap ... we gotta find out what's below.
> >
> > Chris
> >
> >
> > ´╗┐Am 15.07.20, 15:23 schrieb "Niclas Hedhman" <nic...@hedhman.org>:
> >
> >     It is a common format in equipment documentation. I am currently
> > working
> >     with an electric meter. See screenshot from its manual;
> >
> >
> https://ipfs.subutai.io/ipfs/QmPsra6ExrkSX9GCYWgHa6CTfJbBnddLVXJ3NmJjmAYrAk
> >
> >     At protocol level, there are no such "high numbers", just that the
> > industry
> >     got used to write "4" instead "holding register". It is a "people
> > thing",
> >     not a technical one.
> >
> >     HTH
> >     Niclas
> >
> >     On Wed, Jul 15, 2020 at 6:11 PM Christofer Dutz <
> > christofer.d...@c-ware.de>
> >     wrote:
> >
> >     > Hmmm ...
> >     >
> >     > Regarding: 10012 being a coil and 4xxxx a register.
> >     >
> >     > Well the coils and registers are completely different things ...
> > They are
> >     > accessed via completely different requests.
> >     > That might be some sort of convenience convention, but I wouldn't
> > call
> >     > that a standard (I've actually never seen it before)
> >     > Would you then just subtract 40000 from every register or would you
> > just
> >     > configure registers starting with numbers 40000?
> >     >
> >     > What do the others think? The address format does reference
> > registers ...
> >     > should we be doing the translation?
> >     >
> >     > I guess that the general use-case would be that someone has a look
> > at his
> >     > Modbus config and says that he wants "Register X"
> >     > and not "The Register Y references" (which seems to be one less)
> ...
> > I
> >     > might even swing my vote in favor to using register numbers instead
> > of
> >     > addresses.
> >     >
> >     > But then I ask myself: How can I address register 0 in ModbusPal
> ...
> > that
> >     > would require an address that's impossible to send.
> >     >
> >     > Chris
> >     >
> >     >
> >     > Am 15.07.20, 11:51 schrieb "Niclas Hedhman" <nic...@hedhman.org>:
> >     >
> >     >     Personall, I prefer that Register Numbers are used in APIs and
> > that the
> >     >     address is only seen if you analyze the over-the-wire format,
> > but maybe
> >     >     that is just me. A big reason for this is that any interface
> > presented
> >     > to
> >     >     an operator would need to use RegNumbers, as most (possibly
> all)
> >     >     documentation of equipment has RegNumbers and some are lacking
> > the
> >     >     addresses. It is also common that the register type is part of
> > the
> >     > Register
> >     >     Number, such as 10012, where the first 1 indicates "coil"
> > (IIRC). and
> >     >     holding registers are in 4xxxx. Ideally(!), this is also
> handled
> > by the
> >     >     register/address parser.
> >     >
> >     >     Niclas
> >     >
> >     >     On Wed, Jul 15, 2020 at 5:43 PM Christofer Dutz <
> >     > christofer.d...@c-ware.de>
> >     >     wrote:
> >     >
> >     >     > Hmm .. so are we doing it correctly?
> >     >     >
> >     >     > I mean Wireshark isn't the ideal reference here as I have
> > several
> >     > valid
> >     >     > packets in the KNX space, where WireShark just says "corrupt
> >     > package".
> >     >     >
> >     >     > So if you enter an "holding-register:42" address it tries to
> > read the
> >     >     > register number 43 in ModbusPAL.
> >     >     >
> >     >     > I guess as we are reading an "address" we are correct and
> > probably as
> >     >     > ModusPAL says register number, that might be correct too?
> >     >     >
> >     >     > Chris
> >     >     >
> >     >     >
> >     >     > Am 15.07.20, 10:54 schrieb "Niclas Hedhman" <
> > nic...@hedhman.org>:
> >     >     >
> >     >     >     The issue with 1-offset is that the "Register Number"
> > found in
> >     >     >     documentation is at an "Address" one position less. So
> Reg
> > 41 has
> >     >     > Address
> >     >     >     40. Som libraries expect Register Numbers and some expect
> >     > Register
> >     >     > Address.
> >     >     >     And every so often, one mixes that up.
> >     >     >
> >     >     >     On Wed, Jul 15, 2020 at 2:53 PM Christofer Dutz <
> >     >     > christofer.d...@c-ware.de>
> >     >     >     wrote:
> >     >     >
> >     >     >     > ....
> >     >     >     >
> >     >     >     > While looking for more Infos I found out I once wrote a
> >     > tutorial
> >     >     > page for
> >     >     >     > PLC4X __
> >     >     >     >
> >     >     >     >
> https://plc4x.apache.org/users/plc4j/virtual-modbus.html
> >     >     >     >
> >     >     >     > I so totally hate searching the web and getting my
> > answers
> >     > answered
> >     >     > by
> >     >     >     > myself.
> >     >     >     >
> >     >     >     > Chris
> >     >     >     >
> >     >     >     > Am 15.07.20, 08:31 schrieb "Christofer Dutz" <
> >     >     > christofer.d...@c-ware.de>:
> >     >     >     >
> >     >     >     >     Hi Niclas,
> >     >     >     >
> >     >     >     >     when preparing a workshop for the Building IoT we
> > were
> >     > looking
> >     >     > for
> >     >     >     > something where you could simulate a Modbus Slave.
> >     >     >     >     Most of these were Windows only solutions, so
> > ModbusPAL
> >     > was one
> >     >     > of the
> >     >     >     > very few solutions that were pure-java.
> >     >     >     >
> >     >     >     >     Don't quite understand what the thing with the
> > offset is
> >     > however
> >     >     > ...
> >     >     >     > should be easy to fix ... I mean if I say address 42 in
> > PLC4X,
> >     >     > Wireshark
> >     >     >     > says 42, but in Modbus Pal I think it has to be 43 ...
> > don't
> >     > quite
> >     >     > get it.
> >     >     >     >
> >     >     >     >     Chris
> >     >     >     >
> >     >     >     >
> >     >     >     >
> >     >     >     >     Am 15.07.20, 03:54 schrieb "Niclas Hedhman" <
> >     > nic...@hedhman.org
> >     >     > >:
> >     >     >     >
> >     >     >     >         What is ModbusPAL?
> >     >     >     >
> >     >     >     >         The 1-offset in Modbus has caused a lot of
> > confusion
> >     > over the
> >     >     >     > years.
> >     >     >     >
> >     >     >     >         On Tue, Jul 14, 2020 at 11:03 PM Christofer
> Dutz
> > <
> >     >     >     > christofer.d...@c-ware.de>
> >     >     >     >         wrote:
> >     >     >     >
> >     >     >     >         > Sorry for the noise ...
> >     >     >     >         >
> >     >     >     >         > this was a false positive ... the driver was
> > doing
> >     > things
> >     >     >     > correctly.
> >     >     >     >         > It seems the ModbusPAL was just off by one
> ;-)
> >     >     >     >         >
> >     >     >     >         > Chris
> >     >     >     >         >
> >     >     >     >         >
> >     >     >     >         >
> >     >     >     >         > Am 14.07.20, 15:51 schrieb "Christofer Dutz
> > (Jira)" <
> >     >     >     > j...@apache.org>:
> >     >     >     >         >
> >     >     >     >         >     Christofer Dutz created PLC4X-214:
> >     >     >     >         >     -------------------------------------
> >     >     >     >         >
> >     >     >     >         >                  Summary: [Modbus] Holding
> > register
> >     >     > addresses
> >     >     >     > have an
> >     >     >     >         > offset of 1 (Not reading the correct address)
> >     >     >     >         >                      Key: PLC4X-214
> >     >     >     >         >                      URL:
> >     >     >     > https://issues.apache.org/jira/browse/PLC4X-214
> >     >     >     >         >                  Project: Apache PLC4X
> >     >     >     >         >               Issue Type: Bug
> >     >     >     >         >               Components: Driver-Modbus
> >     >     >     >         >         Affects Versions: 0.7.0
> >     >     >     >         >                 Reporter: Christofer Dutz
> >     >     >     >         >                 Assignee: Christofer Dutz
> >     >     >     >         >                  Fix For: 0.8.0
> >     >     >     >         >
> >     >     >     >         >
> >     >     >     >         >     If I set holding the following holding
> > registers
> >     > to:
> >     >     >     >         >
> >     >     >     >         >     1000: 41
> >     >     >     >         >     1001: 42
> >     >     >     >         >     1002: 43
> >     >     >     >         >
> >     >     >     >         >     And then read: holding-register:1001 ...
> I
> > get
> >     > the
> >     >     > value 43
> >     >     >     > returned.
> >     >     >     >         >
> >     >     >     >         >
> >     >     >     >         >
> >     >     >     >         >     --
> >     >     >     >         >     This message was sent by Atlassian Jira
> >     >     >     >         >     (v8.3.4#803005)
> >     >     >     >         >
> >     >     >     >         >
> >     >     >     >
> >     >     >     >
> >     >     >     >
> >     >     >
> >     >     >
> >     >
> >     >
> >
> >
>

Reply via email to