Hi everyone,

Hoping someone here has dealt with high-static industrial environments and HP workstation grounding quirks. At Core Blast Solutions, we manufacture automated pressure blasting rigs. I’m currently running a custom CLI diagnostic and logging script on an HP ZBook Fury 15 G8 (Win 10 Pro) connected to our cabinet PLC over Modbus RTU (RS-485, 2-wire, 115200 baud, 8N1).

Sensors being polled via CLI (50ms intervals):

• Hopper load cells (tracking real-time steel grit / aluminium oxide consumption rate)
• Air-line moisture separator dew-point sensor
• Nozzle throat pressure transducer (monitoring 80–100 PSI blast pressure)
• Deadman handle safety interlock status

The Glitch: When I run the HP ZBook purely on battery power, my CLI serial monitor polls the PLC all day with 0% packet loss, even while blasting at full pressure.

The exact second I plug the HP 150W Smart AC Adapter (3-prong barrel jack) into the shop floor receptacle near the control panel, my CLI terminal starts throwing Modbus CRC checksum mismatches and garbage hex framing errors (0xFF, 0x00)—but ONLY when dry abrasive media is actively flowing through the blast hose and boron carbide nozzle. If we run "air only" (no abrasive media flowing), the serial comms remain stable even with the HP charger plugged in.

My Hypothesis: High-velocity dry abrasive media (especially garnet and aluminium oxide) generates massive triboelectric static voltage along the blast hose. Even though our blast pot and tungsten carbide/boron carbide nozzle holders are bonded to earth ground with antistatic whip cables, I suspect the HP Smart AC brick’s internal EMI Y-capacitor is coupling the building's 230V AC protective earth directly to the ZBook's USB/Serial 0V reference. That turns my RS-485 cable shield and signal ground into a return path for ESD spikes.

Questions for the forum:

1. Has anyone measured the chassis-to-earth leakage current on HP ZBook Smart AC power bricks in high-ESD environments?
2. Will placing a DIN-rail 2.5kV galvanically isolated RS-485 repeater (like a Moxa TCC-100I or Waveshare industrial isolator) between the PLC and the HP laptop completely break this ground loop?
3. Should I also lift the RS-485 cable shield at the laptop end and terminate the cabinet end through a 100nF capacitor to block 50Hz ground currents while still shunting high-frequency static?

Thanks in Advance for reply

 

This sounds like a "lack of a common RS-485 0V reference" issue.

When the laptop is on battery power, its 0V reference is floating and will naturally sit appropriately halfway between the + and - signals. When the power adapter is plugged in, your laptop's 0V reference is now (likely) tied to local earth ground via the power adapter. When the abrasive media is actively flowing, this is potentially causing noise on the local grounds of the other RS-485 devices, making it different from the local grounds of the laptop.

One potential fix, assuming the USB to RS-485 adapter used with the laptop has a 3rd terminal for 0V reference, is to simply run a 3rd wire along with the + and - wires and connect to all RS-485 devices' 0V reference. Note that if the other devices reference earth ground, you may need to put a small resistor (such as 100 ohms) in series between the laptop's RS-485 adapter and the 0V reference wire.

The other solution is to employ isolation. You can replace the existing USB to RS-485 adapter with a galvanically isolated version, or as you stated, use an isolated RS-485 repeater.

 

As for the cable shield, standard practice is to earth it at one end only and daisy-chain it at each device so that it is continuous.

 

I think you're on the right track with the grounding hypothesis. What stands out to me is that you need both conditions, the laptop connected to AC power and abrasive media flowing, before the communication errors appear. That certainly suggests a common-mode interference problem associated with the blasting operation.

Before modifying the shield connections or adding a signal-common conductor, I'd be inclined to try a galvanically isolated USB-to-RS-485 adapter. If the errors disappear, that would be a pretty strong indication that the laptop's ground connection is contributing to the problem, although it wouldn't completely establish the source of the interference.

I'd also be interested in measuring the ground-potential difference between the laptop-side interface and the PLC, ideally while the blasting equipment is operating, using suitable isolated instrumentation. RS-485 is differential, but it still has common-mode voltage limits.

One other thing worth checking is the continuity of the bonding throughout the blast hose and nozzle assembly. You may be generating significant static charge that isn't being adequately dissipated.

I wouldn't change the shield termination or lift any grounds until you've established where the unwanted current is flowing. Depending on the installation, that could make the problem worse rather than better.

One final consideration: since you're monitoring a deadman safety interlock, I'd verify that any Modbus communication failure or electrical interference cannot compromise the actual safety function. Ideally, the deadman circuit should operate independently of the PLC diagnostics and laptop connection.