PLC performance on network with makeshift SCADA

My Question is:
Is there an impact on individual PLC scan rate as an plant Ethernet network size increases, where a single HMI is requesting data via modbusTCP from the multiple individual PLCs (i.e. in place of a 'real' SCADA package). i.e. does the communications 'noise' from the HMI requesting increasing levels of data from so many sources confuse/slow a PLC trying to listen for its own IP?

Follow on question:
What benefits would come from operating a SCADA platform on a server on the network, over using a HMI with remote view / website

**I have posted this same question on another forum as well, so apologies to subscribers of both, however I know the readers of this forum and that forum (reddit/PLC) often have very different views of the world, so I am interested to see the conversation develop.**

Full Background / Scenario:

I work for a small manufacturer. My role is part Management, part Engineer. Until very recently all the machines we operated were stand alone units with a local control panel only. As the number of machines has grown, I decided it was time that a supervisor could see the running data from a number of machines in one place.

As the number of machines is not large (10-15 different units), the applications not critical (stand alone, no cascading failures, not dangerous goods), and the traditional operating method (walking from machine to machine and looking at each control panel) was currently still effective (just), this is not the sort of project with a budget to implement something like Ignition (yet..).

The solution I have implemented is to connect all the machines that have some data we would like to overview to a local network (factory only) via ethernet and unmanaged switches. One device on the network is an AutomationDirect 'Headless' HMI- of the same family as the C-More EA9, but no local display.

The headless HMI had a few features which were useful for the project:

  1. The firmware has inbuilt webserver, which will show each HMI page as a webpage available to any PC on the network. Not real time, but refreshed with live data each time the web browser is refreshed.
  2. Application / screen design was easy even with no prior HMI programming experience.
  3. Ability to connect to multiple PLCs simultaneously was easy to set up.
  4. Hardware was affordable (<$1000)
  5. Remote access of the HMI can be done with a tablet/smartphone (HMI program is only viewing data and has no outputs, so no major 'risk' here- although I understand the industrial vandalism/blackmail risks still exist in the broader network security discussion. For now, the local area network is not connected to the internet, and will not be without suitable IT support (VPN etc).
I have installed a cheap PC-on-a-stick with a monitor in the supervisors office. The PC has a web browser open with a 1 minute auto refresh. The supervisor can now see the running status of a bunch of machines at all times (up to 1 minute delayed), without having to constantly walk back into the production area. It has been well received and I can see it adding real efficiencies and increasing machine utilisation.

I am now working to add additional data to the system- storage tank levels, and few other machines to begin with.

Which brings me to question 1 in full detail:

As the network gets larger (planned to have ~15 PLCs and more in the future), and all connected using unmanaged switches (in some cases via multiple switches to minimise cabling requirements), is there an effect on the devices on the network. I am not so concerned if the data I am trying to view is slightly laggy, or misses a refresh to the HMI, but the existing machines' PLCs should still be able to do their thing without being slowed down due to being bombarded with requests from the HMI to an address they need to determine is not theirs.

Question 2

While I consider the system I have implemented to be a rudimentary SCADA system, what would be the advantages of moving to a 'proper' server + software environment, rather than the hardware + firmware + webserver setup at present. Remembering the Headless HMI are only $400USD. A proper historian would be very useful, but maybe there is better way of using the cost effective "viewing" attributes of the existing set up, along with a database for trending, rather than starting the whole system again.

Looking forward to peoples views. Thanks
I may be missing something but a couple of things...

1. You can change the poll rate of the headless display. This might cut down on the network conjestion. Since you're refreshing the data only once per minute from the computer anyways, this probably wouldn't have any practical effect on slowing down the data collection. It's in the same part of the programmjing software where you add the PLCs the display is reading from.

2. You can set the viewing computer to show the data live. There's a little"widget" you can download. Just connect to the display via a web browser and there should be the option to download the widget.

You have to go into the network configuration using the programming software and add a user (and if desired password). The widget will simply have you log on and display the screen as if it were a regular display. You can also set it for view only or with full control.

Note. On a regular display the widget shows whatever the screen showed and vice versa. This is some came up to the display and switched the screen the widget would also change.

I hope this information was helpful.