Sveriges mest populära poddar

The Industrial Wi-Fi Shop Podcast

The Industrial Wi-Fi Shop Podcast – Ep. 9 Wi-Fi Fight Club!

70 min • 3 september 2024
  • CWISA Guide Giveaway details
    • There are two copies of the Certified Wireless IoT Solutions Administrator (CWISA) study/reference guides
    • Winners will be drawn randomly on episode 10 (episode title TBD)
    • ONE winner will be from IT
    • ONE winner will be from OT
    • Click the link below to enter the drawing! (One entry per person)
    • —>  https://forms.gle/eE4rYtixhbppyMti8  ←- (Link now inactive)
  • WiFi6 and Mobile Robots:  AS/RS, Conveyors, AGVs, AMRs, Automotive Skillet Lines, Electrified Monorails…. All have one thing in common:  THEY ROAM….  A lot!
    • Challenges going to WiFi6
      • The OT traffic is different than IT.
      • Instead of pushing big-huge files through the air, mobile robot applications need to move lots of small files – rapid fire, fast!
      • PLC-to-PLC communication.
      • The reliable, uninterrupted, consistent nature of the OT traffic is different with mobile bots. 
      • The OT networks are different.
        • Think 20ms RPI, so the roam time has to be a fraction of that.
  • The generation level of the WiFi doesn’t necessarily matter.   … except for 6E (that “E” is really nice). How will we handle the common need for “purpose built” antennas/ signal reciprocity. For example a linear stacker crane usually only needs directional/ bi-directional antennas. No external antennas means you waste energy. The power differential is also a concern.
    • The changes in generation levels focus on more volume of data thru the air.
    • The changes are not necessarily focused on reliability.
    • The fast-roaming standard hasn’t changed.
    • 95% of the chips end up in enterprise IT, and the current roaming standard satisfies that IT market.
    • Adding video, bandwidth, machine learning, etc. is important to enterprise IT, usually not to the bots.  
  • The ProSoft radios do not support multiple SSIDs, on purpose.  The idea is to have the SSID dedicated to the robot “fleet” network.  Any additional WiFi required nearby will be handled by a completely different AP.
    • Automotive skillet lines and AS/RS roaming bots don’t need that.  They need really, really reliable connections and ultra-fast roams.
    • Think Safety I/O
    • Think CIP Safety, ProfiSafe.
    • Think 20ms RPI and 40ms timeouts.
  • ProSoft buys the RF modules like everyone else.
    • Ex:  We buy Qualcomm, package it, and sell it with our software running on it. I don’t know exactly how we do it, but somehow we hijack the chip and take control of many of the low-level decisions.
    • WiFi4 (802.11n) gave us access to the chip’s low-level functions.  
    • We could monitor data from the RF chip so our software makes the roaming decisions, not the chip.  The chip’s roaming standard is too slow.
    • We run our own calculations.  We had a lot of control, resulting in consistent roams under 10ms – often down to 2ms if the client and AP were both the ProSoft RLX2-IHNF-A.
    • We can roam on the same channel.
    • These are ProSoft’s killer features for mobile applications.
  • What is the most common installation case for ProSoft? Were you always geared towards single-band applications?  The radios have always offered 2.4 and 5.0.  We had another line of 900MHz frequency hoppers, but discontinued them in 2020.
  • Most common installations are AMRs in automotive, warehousing, oil/gas, cranes.  
  • With WiFi6, the downside is we’ve been locked out by the chip makers, and don’t have access to those low-level functions.  We can only do high-level computations.
    • So the WiFi6 chip handles the roam. Is this part of the standard? Does this lock you out from using the chipset and modified frames?  There is a roaming standard, but from a ProSoft point of view, it’s too slow.  That’s why we take control of the roaming decision.   For whatever reason, we were not locked out of accessing the low-level functions required to take over the roaming decision.  Now, with WiFi6, the manufacturers of the chipsets have locked that down.
    • Unfortunately, the 802.11r standard is way too slow.
    • Challenge:  Can we figure a way to still get to that low-level data?  ProSoft is working on it.
    • Fluid Mesh puts 2 radios in each bot, and this might be the only way to fast-roam with WiFi6. Literally 2-4 radios that link logically. Newer hardware has two internal radios. Similar function. Splitting fleets is also a use case here.
      • Radio A talks to AP1
      • Radio B talks to AP2
      • When radio A is moving out of range of AP1, it starts roaming to AP3… in the meantime Radio 2 is still linked with AP2, 
    • Siemens says IPCF does not currently work with WiFi6 to accomplish fast roaming. They also lost the second radio model from the W788 so those deployments are getting forced into WiFi6 one way or the other. Give and take with RPI and timeouts, some of these applications will have to relax constraints to get it to function.  I see that give and take a lot, especially before I start talking to them.  Slowing down RPIs and timeouts to allow for the longer roams.  
  • BENEFITS of WiFi6:
    • If you need more bandwidth,
    • If you have pushing gobs of data,
      • Machine vision.
      • video
    • If you have a huge number of bots (clients) tied to 1 AP.
      • As the bot count goes up – 700, 800, 1,000 bots all in one warehouse – WiFi6 manages the data and traffic more efficiently.
  • WiFi6E is nice, opening up that 6GHz spectrum, you go from 8-9 channels to almost 60 more channels (at 20MHz)
    • 2.4 is almost never used by ProSoft.  It’s there.  You can use it.  But it’s too crowded with bluetooth, cordless things, microwave ovens, everyone’s cell phone which makes for an unreliable connection handling 20ms RPIs. Unreliable when you’ve got 500+ bots in a warehouse, and there are people there too with cell phones, and airpods, and apple watches. I learned the hard way that airpods really are a no-no when surveying 2.4 :).
  • We use 5.0, including the DFS channels.
  • More channels, more likely to get customer’s IT to STAY OFF a few of them.
  • Avoids all that saturated 2.4 traffic.
  • Siemens iPCF will not work on DFS channels. 
  • WiFi4 is going away, one day. RIP
    • We need to get WiFi6 functioning in the OT space similar to how we currently have it with WiFi4, i.e. Ultra Fast Roaming <10ms in the RLX2-IHNF-A
    • For mobile applications, it has to support Safety I/O, CIP Safety, ProfiSafe, etc., etc.
    • For mobile applications, it has to support high client density (bot swarms), 6E is the answer, if we could only now just get it to fast roam.

If you would like to know more about our guests, check them out on LinkedIn:

Vic Paraenglo – https://www.linkedin.com/in/victorparangelo/

Prosoft Technology wireless products – https://www.prosoft-technology.com/Products/Industrial-Wireless

If you would like to connect with me or learn more about my employer, Global Process Automation (GPA), then check the following:

Scott McNeil – https://www.linkedin.com/in/americanmcneil/

GPA – https://www.global-business.net/

If you would like to connect with Jeremy or learn more about his employer, Prism Systems Inc, then check the following:

Jeremy Baker – https://www.linkedin.com/in/jeremyabaker/

Prism Systems Inc – https://www.prismsystems.com/ 

Förekommer på
00:00 -00:00