Open hardware PAC/PLC/Edge controller

[FalconeG]

Hi,
I’d like to inform the community about YABA, a new Open Hardware  Project for a PAC/PLC/IoT Gateway/Edge Computing systems.
YABA is a modular backplane/slot controller, based on multiple platform from Raspberry module to Nvidia Jetson, Google Coral, standard SMARC module and so on.
The expansion boards will provide digital and analog input/output as well as motors control, field bus integration, video input, HMI etc.
The triple bus (USB/I2C/LVDS-Ethercat) on the same connector allows high performances but also easy interfacing for hobbyists and low budget producers.
Every maker can develop hardware and software solutions to works on YABA platform.
We just launched our Indiegogo campaign and we would love like to receive not only backing, but also suggestions, criticism and development ideas.
Visit campaign page at:
 
https://www.indiegogo.com/projects/yaba-yet-another-backplane-architecture

[wizard69]

Interesting offering.   A couple of thoughts:

• If the platform doesn't support ladder logic it is pretty much dead in the water.
• The use of USB is fantastic but in an ideal world the backplane would have 24 VDC DC bus's for Inputs and Output cards.   I know that this is uncommon these days but it sure does make field maintenance a lot easier.   Also in some organizations it is common to have a separate input supply and an output supply.
• The compact box offerings need I/O built in.   Adding an expansion card to get a couple of inputs and outputs doesn't cut it in practice.   Frankly this means at least 16 inputs and 8 outputs.   Special purpose I/O may be included also.
• USB sucks for ruggedness.   You need to pay attention to the strain put on the connector.

Just some things that popped into my head after giving the site a quick skim.

[senso]

How resilient is it?
I can power cycle an S7-1200 PLC hundreds of times a day and not have a problem, will the RPi handle that with grace or it will end up corrupted at the third power loss?

What is the IDE/software used to program it?

What software to program the touch display? Having a touch display, and it being an HMI is not the same thing, AT ALL..

There is already plenty of cheapo PLC's that you can add IoT gateways, what is your value added proposition.

At least running Codesys should be an objective that you want to fulfil.

Also, IO-Link is all the rage currently(at least in EU), having at least one IO-Link master port might be in your best interest, because IP68/69K field distributors that talk IO-Link are pretty cheap and simplify the wiring a lot(3 wires for power/data) and you have 16 I/O's in a nice robust field device with standard M12 5 pin connectors.

[excitedbox]

I don´t like USB because of the length restriction. Ethernet can go much longer distances and in a factory environment you will quickly go over that USB limit. In addition USB is not meant for multiple devices on 1 port. USB always has 1 active device at a time. That is another reason Ethernet is much better because you can have many many devices running simultaneously which is a must for a industrial setting for synchronization. This is dead on arrival in my opinion. Use POE like everyone else or you won´t get far. If you want to be different write a new protocol that is geared towards industrial applications.

You could make it even better by making a whole new cable standard. Custom connectors cost pennies each and besides needing a new crimping tool won´t cost more than POE cable. If you combined Ethernet with thicker wires for a High powered POE with an industrial protocol for powering things in addition to networking. Call it CAT7i and CAT5i for the unshielded core version.

[jackthomson43]

A person working on PLC, can't leave Ladder Logic, so you can't replace Ladder Logic with any other Language.

Replace USB with Ethernet or if you could, the best option would be Wifi, you create a static page for configuration and then communicate over Wifi, as in wifi routers. I have worked on Arduino YUN, it has the similar Wifi connectivity. You can also check ESP8266 Datasheet, it will help.

[wizard69]

A person working on PLC, can't leave Ladder Logic, so you can't replace Ladder Logic with any other Language.

Replace USB with Ethernet or if you could, the best option would be Wifi, you create a static page for configuration and then communicate over Wifi, as in wifi routers. I have worked on Arduino YUN, it has the similar Wifi connectivity.

My understanding that USB was to be used as the interconnect method for a back plane and the module that plug into it.    Ethernet would be i deal for any thing supporting remote I/O though.   

With all of these attempts at DIY PLC's the designer seems to forget the importance of a well implemented ladder logic facility.   I suspect that the designers are coming from an embedded programming back ground and not an industrial controls background.    It isn't uncommon for somebody working in a plant to be given 5 minutes to fix or update code and make sure it works as intended.    It is hard to beat ladder logic for that.

[Doctorandus_P]

Indiegogo has very little information at the moment. Just enough to annoy me for wasting time...

A quick search however:
https://www.yaba.tech/

If this bus is USB, then a standard USB connector would be a big plus. It would allow to use these modules standalone.
Starting with a single data acquisition unit on USB would lower the threshold to try this Yaba thing, without commit directly to the whole ecosystem.
For this I would deem a standard USB connector mandatory. Euhm, it seems to use an "USB C internal  connector (also named U31G2)" with also EtherCAT &I2C
https://www.yaba.tech/the-usb-choice/

There is also a mention of a "passive adapter", but that would still add an extra obstacle for connecting one of these boards to a PC. But still, there is very little real information. The whole project still seems to be in a concept stage.

Long ago at school I was forced to learn ladder logic after multiple years of experience with C and C++, and I still have nightmares from it. The idea to use a microcontroller (in the PLC) which executes instructions sequentially, then interprets the ladder where everything is supposed to happen "at the same time" (You can't rely on order) and then implement some sequential conveyor belt control. The whole thing is so convoluted that it twists my mind.

There probably are a multitude of ladder logic interpreters for linux, and as this thing has a pc, it should not be difficult to add it for the people who want this though.

One of the possibly uses for me would be to run LinuxCNC on it. Some years ago I tried Machinekit on a Beaglebone, ran into some troubles and then settled on a "Blue Pill" with GRBL, which works OK for now, but GRBL is quite limited. For "professional" use of LinuxCNC one of the Mesa FPGA cards is one of the most used options, but together with a (u-ITX?) PC it gets bulky fast.

Another potential problem is the USB / EtherCAT thing. These are powerful busses but also need quite powerful hardware, which drives up the cost. I'm also a bit apprehensive for I2C over a backplane, but it may work reliably. Conforming to more standards would be a plus. ModBUS RTU seems a logical choice.