Now looking at it, it is the same page and version.

Unfortunately it's just badly worded, the HMI application is only used for legacy projects. "Web Site Definitions" is a specific feature which has been completely remade, and as such are not necessary in the new integrated HMI.

If you have a look at the tutorial I have linked above it has a simple step-by-step guide for creating an HMI from scratch in the latest version.

I will get in touch with our documentation team to see if they can help you further.

You will have to create a Software Update (.swu) file containing your desired changes and/or containing the partition you want to write to the new device. There is a tutorial for doing so here, though do be careful not to overwrite the entire drive (as opposed to just the main partition) - doing so may render the device inoperable, since the boot partition is also on the drive.

The update archive can the be loaded using the latest Axon OS image. which you can download in the "Downloads -> Software -> Images" section of our knowledgebase.

I do need to stress that while we may be able to repair the device for a charge if things go wrong, performing the steps above is not supported or covered by our warranty in the case of user error.

This looks to be a driver issue with the latest Open Source OS image and L203. It should not affect L203 Mervis or Codesys. We will fix this asap.

Ahh I was assuming you are using Neuron; with the UniPi 1.1 you need to rename the inputs and outputs in the "evok.conf" configuration file; using aliases with the UniPi 1.1 is not supported as the various subdevices are fundamentally different devices. Whereas for Neuron they are, well, subdevices, which are communicated with via Modbus handles. And it is these handles which are renamed with the alias.

I can have a look if we can implement the aliases for UniPi 1.1 as well, but as of right now it's not possible. We should remove the alias-setting option for UniPi 1.1 from the web inteface.

As for the instructions, they will indeed get you the latest Master version.

Hi @roberts_lando!

It should be in /etc/network/if-eth0.conf, which is referenced from /etc/network/interfaces.

@Boy-Lenssen

I believe that @dgajic was asking about updating UniPian as opposed to Raspbian in this topic.

Anyhow - here's an answer to you. All of EVOK installations done since yesterday are broken as of right now, as a recent update of the pymodbus https://pypi.org/project/pymodbus/1.5.2/#history (~3 days ago) library broke the installation. UniPian does not have this issue.

It does not matter which version of EVOK or Raspbian you use, this is an issue with the library breaking the compatibility. We will get it fixed within today, but you will have to perform a new install afterwards.

I have only found this out now, through the issues you are having, and so thank you for the bug report.

We will have a look at figuring out a way to package the libraries with EVOK, but it's not a simple process since python libraries are distributed as source code under various licenses.

I have now released version 2.0.5c (hotfix) which should have the pymodbus library version forced to 1.4. The changes are big enough that it would not be trivial to port EVOK to the newer version, though we'll have a look at it in the future.

The specification is ModBus RTU, with a custom header. For details see: http://www.simplymodbus.ca/FAQ.htm

In practice all the Neuron TCP server does is to repack the messages from ModBus TCP to a customised ModBus RTU. The overhead associated is therefore minuscule, albeit it does exist.

It should also be noted that if you do wish to use the device directly regardless, you may have an easier time using our OSS image from downloads.unipi.technology, which uses a custom driver to get around certain problems associated with spidev. That way you will have a SYSFS interface along with a documented message structure. spidev is only used in EVOK for legacy reasons and ease of installation, both Mervis and CODESYS use the kernel driver instead.

The IO is not asynchronous in any case. The primary reason is that there is only a single communication channel back towards the primary RPi board, and so there is no convenient way to inform it of changes as they happen. The number of interrupt lines which would be required is considerably more than the number of GPIOs provided by the RPi itself. At any rate, the maximum supported polling speed of 12MHz is more than sufficient for a fast response time.

@majorcsa
Unfortunately not, the updates have broken our driver three times within the past year, and so even if it is technically easy to do with a simple build script, it's not feasible for the process to be automated in such a way.

The kernel updates by their nature change the entire API around which our driver is built, often changing things like includes, data structures, macros, function signatures etc. The process is somehow formalised by how the linking process for Linux drivers works, namely the use of a symbol table, but again the table itself changes quite a lot. Not all changes will break the module, but many have.

This is the same reason why Raspbian itself only updates to one version in ~20, and is many updates behind the latest, as it always needs to be tested (and updated) for compatibility.

Hi @mich-boxman!

You should be able to install it via our debian package repository used on Axon; instructions are available on the Evok GitHub repository: https://github.com/UniPiTechnology/evok

Hope this helps

Hi @matt

I believe the latest version of Mervis should be able to handle it better now; you can give it a go and see if it works properly with multi-character strings. For further information please get in touch with support@unipi.technology and/or Martin Kudláček instead, as I will be leaving unipi shortly.

It should be fixed in the latest Master development version on GitHub, but it may take a little while before the next release is done. You can test it out yourself by replacing the file as detailed above.

V případě potřeby lze také nastavit bytové pořadí, v případě jinak-endianového kódování které používají některé vzácnější zařízení.

Hi @Joerg!

This looks like a bug in EVOK. The /rest/wd interface should return it correctly, I'll fix this in the github repository shortly, but it may take a little while before it gets into a debian package.

Once the update appears on GitHub you may be able to update EVOK manually by overwriting your /opt/evok/lib/python2.7/site-packages/evok/evok.py with the new version, or you can wait until the next EVOK release.

Hello @PR-Reger,

Generally speaking that is always the safer approach, if possible.

Regards,
Tomas

Hello @PR-Reger,

The protection slightly differs between the first I/O group and the second, but broadly speaking it is protected against voltage spikes.

Here's the protection circuitry for the first I/O group (taken from a generalised description):

The protection circuit has 2 pairs of Schottky diodes, as you can see from the circuit schematic above. (The B-1000-2 is a newer version which is used in Axon and uses a transile instead)

The first protection pair opens at positive voltage over 18.2V and shorts with negative voltage over 0.2V. The maximum current which the circuit can withstand is 70mA, but the protection characteristic is strongly time-dependent – it can withstand significant currents (>1A) for under 1uS, but it will fail if the current is sustained longer.

The second protection pair then further limits the voltage to 10.5V, which is the maximum voltage that the measurement circuit is capable of measuring, but this is only intended to cover the difference between the 10.5V and 18V voltages, not to act as an external current protection if the first pair fails.

The protection is designed primarily against EMC interference; if the device is connected to a source of strong and persistent interference we recommend the use of other external protection, based on varistors and transiles.

For the second I/O group the protection is similar, but the measuring circuitry is much more complex (to provide greater accuracy) which does make it vulnerable to certain specific faults, e.g. if switched to a wrong measuring mode. I will confirm with my coworker, who is unfortunately not available right now, whether there may be any problems with the second group I/Os and provide a description of the protection circuitry later today.

I hope this helps, if you require more details don't hesitate to ask.

Hi @PhilesG!

That's good to hear, though a bit concerning that it doesn't work with all flash drives. I'm not sure what we can do about that, but we'll have a look at it again.

The size limit in the web interface is unfortunately inserted quite deep into the system; essentially it's the amount of memory-mapped drive space that is reserved for holding the file sent over the network. Which is also why the flash drive method works, as the file is instead stored on the drive.

At any rate - glad you managed to get it to work!

Hi @Martin-Trojan,

This is just one possible solution, but you can use a custom ST FunctionBlock with this code (use the name "DigitalMemoryMultiplexer" if you do use it):

FUNCTION_BLOCK DigitalMemoryMultiplexer
(*
EXTENDS //base type
IMPLEMENTS //interface type list
*)
	VAR
		previous_inp_1: BOOL;
		previous_inp_2: BOOL;
		previous_outp: BOOL;
	END_VAR
	VAR_INPUT
		inp1: BOOL;
		inp2: BOOL;
	END_VAR
	VAR_OUTPUT
		outp: BOOL;
	END_VAR
	INIT
		previous_inp_1 := inp1;
		previous_inp_2 := inp2;
		previous_outp := inp1;
	END_INIT


	IF inp1 <> previous_inp_1 THEN
		outp := inp1;
		previous_inp_1 := inp1;
	ELSIF inp2 <> previous_inp_2 THEN
		outp := inp2;
		previous_inp_2 := inp2;
	ELSE
		outp := previous_outp;
	END_IF;
	
	previous_outp := outp;
END_FUNCTION_BLOCK

And connect it like this:

The DO_2_01 is set as DirectSwitch Trigger and the WEB_LIGHT_BUTTON is tied to a HMI Digital Setter. Whenever you are using DirectSwitch remember to set the Debounce value on the DI sufficiently high for your input device.

Again, there are many other options for doing this, including using two HMI elements on top of each other (Digital Indicator and Digital Setter).

Here is the project file, for reference: https://www.mediafire.com/file/1o8kq9iici2qc8g/ForumProject.7z/file

Hi @bubuche!

Could you please try to switch the relay by writing '1' to the file /sys/devices/platform/unipi_plc/io_group2/ro_2_01/ro_value? You can do this via the commands:

sudo su
systemctl stop evok
systemctl stop mervisconfigtool
systemctl stop mervisrt
echo "1" > /sys/devices/platform/unipi_plc/io_group2/ro_2_01/ro_value

If this does not work then unfortunately it most likely is due to a hardware fault.

I had a look at the testing logs for your unit, and it has passed tests without any problems, so, if the above does not turn the relay on, you may have to email support to arrange further steps.

Hi @bubuche,

I'll reply to both your questions. Apologies for taking a while, I have not been available over the Christmas period.

1. The upper board RS485/RS232 ports are accessible under /dev/ttyS(0-3) as opposed to /dev/extcomm/0/0. This is unfortunately a hardware limitation which we cannot work around easily.

2. If the LED light turns on then the relay should have continuity across it; the LEDs are controlled by the same circuitry as the output itself. If there is no change in continuity across the output it may be due to some sort of a hardware fault. I should note that digital (as opposed to relay) outputs are transistor-based, and will not make any noise. The relays should issue an audible click however. Could you please post a quick schema of your installation as well? LED lights in particular can have a very high inrush current, which a slow-blow fuse may not neccesarily catch.

I hope this helps