EL4004 analogue output setup with EL6692 I/O Mapping KRC4

beckhoff modules have their own PLC CPU - CX8090. mapping of all IO is done in TwinCat project for that CPU.

so first get your analog IO working in TwinCat

on the end of the unit are two EL6692 bridges.it is likely that one of the two is used to connect PLC to robot.

that connection will have large IO blocks. you need to find free space in there to transfer values from robot, so that PLC can write them to analog output card.

WorkVisual configuration of course need to match this.

but this need to be done correctly. size is everything. digital IOs use 1 bit per signal. single analog output is likely to be 16bit. so to map four analog outputs you would need some 64 bits. this is way more data than 5bit for the five outputs you have in mind.

Quote from akaaschool

Another question I have is how is it possible for the Digital Outputs 49-63 to be converted from digital signals to analogue values as seen below?

Scaling. You already showed how each $IN effects the Analog value. Each bit represents a certain power of 2.

For an 8-bit analog signal mapped to $IN[1] through $IN[8], $IN[8] represents 2^0, $IN[2] represents 2^1, and so on. All "group" I/O signals work like this.

"Signed" vs "unsigned" has an effect. While for non-analog signals, KRL only treats 32-bit numbers as inherently signed, in an analog signal configuration, you can have smaller bit ranges that are signed or unsigned -- it depends on the analog device you're connecting to. So 10000001 would represent 129 on an unsinged signal, but -127 on a signed signal. You can use the Calculator in Windows -- select the Programmer mode, then select Word, Byte, or DWord to represent the limit of the bit range.

Unfortunately, some industrial analog units come in unusual bit ranges, like 12bit or 14, but the principle remains the same -- you just can't use Windows Calculator to predict their values the same as for Byte, Word, or DWord.

That sums up the digital side of the equation. The physical side is hardware-dependent. But you can treat the binary signal pattern as a % of maximum range. So, for an 8bit unsigned analog module, 11111111 (255) would be 100%, and each increment from 0-255 would represent 1/256, or about 0.4% of the maximum range.

On the physical input/output side, it depends on the hardware. You can buy modules whose max range is 5V, 10V, 24V, or current-based modules whose analog signals range from 4-20mA. Assuming we have an 8-bit unsigned 10V module, then each increment from 0 to 255 would represent 0.04V, or 40mV:

00000001 = 40mV

00000010 = 80mV

00000011 = 120mV

...

10000000 = 5.12V (almost exactly 1/2 of the total 10V range, as one would expect, though there is a bit of rounding error)

...

11111111= 10V

Quote

From now understanding the TwinCAT System Manager a bit better, I can now see that indeed one EL6692 handles the I/O for the KR30 and the other handles it for the KR60. There are 4 banks of Inputs and Outputs for each EL6692. I can see that the first 2 Input banks handle the digital output signals [1]-[32]. I can also see that in WorkVisual that indeed that $OUT[1]#G maps to the first Input bank on the EL6692. What decides the order of the mapping between WorkVisual and TwinCAT? Is it just the order in which the bits are received?

yes, size of IO block transmitted and order of signals need to be the same on both ends (PLC and WoV).

Quote

Is it possible to mix digital and analogue outputs (size permitting) in the same input bank on the EL6692?

yes... you can do anything you like as long as size and order match.

Quote

And finally, the physical I/O terminal with 5 digital outputs and 5 digital inputs that I mentioned in my original post... what is the likely order of physical connections to allow this to happen? Is it Beckhoff Terminals --> KRC4 SYS-X44 --> Physical I/O terminal?

you can see that in WoV project in IO mapping. those outputs are mapped to EL6692 therefore they are controlled by PLC.

robot outputs 1-5 -> robot io mapping (created in WoV) -> X44 -> EL6692 -> PLC backplane -> PLC CPU (io mapping done in TwinCat)

Quote

Does additional mapping need to occur in WorkVisual or somewhere to allow for the physical I/O terminals near the robot to be used?

yes...

those five outputs are controlled by robot to send signals to PLC. PLC may do with them anything needed, for example it can send them to physical outputs of the PLC. in this example physical outputs are controlled by PLC.

but you do not need PLC for that. robot can also control the IO directly if there is IO block (a slave) that belong to robot (robot is master of that IO block). then this new IO block would need to be also connected to KEB (X44 bus) and mapped in WoV. You can connect the new IO block either between X44 and EL6692 or after EL6692.

new IO block would require bus coupler EK1100 along with any IO modules you wish to add. KUKA robots that have been ordered with IOs would usually have

EK1100+EL1809+EL2809

Quote

Another question I have is how is it possible for the Digital Outputs 49-63 to be converted from digital signals to analogue values as seen below? E.g. If $DO[63] is HIGH then the aiSpindleSpeed value is 16384, if $DO[62] + $DO[63] are both HIGH then the aiSpindleSpeed value is 24576. Is this done in the PLC configuration somewhere?

note that signal direction matters...

OUTPUTS are signals sent by some node, INPUTS are signals arriving to some node.

OUTPUTs of one node can be INPUTS of other node.

so:

Robot outputs are signals sent from robot to PLC (on PLC they would be inputs)

Robot inputs are signals received by robot from PLC (on PLC they would be outputs)

in other words - turning robot outputs into robot inputs makes no sense.

name aiSpindleSpeed suggests that this is an analog INPUT. topic title mentions analog OUTPUT.

the data can be grouped so you can turn bunch of digital outputs into one analog output. (all are outputs...)

also you can turn one analog input into bunch of digital inputs. (all are inputs...)

but you cannot magically alter signal direction (convert input to output or the other way around).

you can have bunch of robot digital outputs wired to bunch of robot digital inputs, then map those digital inputs as an analog input. but doing this makes no sense either, using hardware workaround makes no sense when information is already in the memory. therefore unless you REALLY want an ANALOG INPUT, it is enough to create suitable signal in software, such as:

SIGNAL myValue $OUT[49] TO $OUT[63]

if you add that to $CONFIG.DAT for example, you can turn individual outputs and observe changes in myValue - or the other way around.

btw. your "TwinCat project" attachment is not a TwinCat project.

signal mapping is explained rather well in the WoV help and documentation.

here showing same thing on EL6695 rather than EL6692 (don't feel like importing ESI files)

signals can be edited in both panes but... they need to match in order to map them

and of course signal direction must match too (arrows point in the same direction)

With signal editor you can specify exactly what bit range corresponds to particular signal etc.

Here is first DWORD split into bunch of bits for digital outputs and two different size and polarity analog signals. When splitting signals it is a good idea to name them to avoid mixup.

First split ranges - click on first bit and drag mouse down then release, then rename each range.

If you want to split range into bits, it is better to first rename the range and then split.

This way new name is going to be used in each of the bit signals. during splitting, WoV will automatically increment numeric suffix. Here the top group that was 8-bits i called "Digital Output" and WoV added underscore and number.

After clicking OK, new structure is shown. Note that DWORD0 is now replaced by parts we created.

Note that signal direction remains the same (arrow on the left side of the pane)

Now we can map Analog1 to $ANOUT[1]

and do with the rest as you please, for example map those 8 digital outputs to $OUT[1] to $OUT[8].

you can add more IO that are not part of the PLC (IO directly controlled by robot). Here i added two more blocks and named them accordingly:

pasted-from-clipboard.png

Each can have variety of IO modules. And yes one or more of those could be analog IO too.

And of course, all of those would need to be mapped too.

Important thing is that field wiring of the EtherCat bus (vertical lines) and module order must match topology exactly:

In this case ECAT cable would need to go from X44, to EL669x in the PLC, then to the IO at the robot foot, then to the IO at the robot flange.