Topic: USB Weatherlink very sensitive to static?

[pfletch101]

With the colder temperatures, and despite a reasonably effective humidifier on my HVAC system, the humidity is dropping enough at home to allow some static build-up. For some reason, this seems to occur particularly when I am working at my computer. The main light in my office is incorporated in a ceiling fan and is turned on and off by the usual metal chain hanging from the fan - the whole assembly being grounded. If I stand up and reach to turn off the light without remembering to ground myself more gently first, a small static spark jumps to the chain and my USB Weatherlink disconnects from the VB.Net program that monitors it (and stays disconnected)  This is absolutely repeatable, and is a particular nuisance because the only reliable way of getting everything going again is to reboot my computer (which is the main server for my home network and takes about 15 minutes to be usable after a reboot).

The light is ultimately on the same circuit as my computer, but the computer (which doesn't otherwise glitch under these circumstances) and the VP2 console are both plugged in to a functioning APC UPS, which should also be protecting both of them from mains-transmitted interference. I wonder, therefore, if the long vertical USB cable, and/or the console's power cable, is acting as an antenna and receiving the static-generated interference pulse over the air.

Has anyone encountered anything like this, and/or can anyone suggest how I might approach fixing the problem?

[dalecoy]

Workaround:

Shorten the chain by a foot or so, and attach a non-conductive string with a convenient non-conductive knob.

[pfletch101]

Workaround:

Shorten the chain by a foot or so, and attach a non-conductive string with a convenient non-conductive knob.

The chain is only about a foot long, now, but I could certainly replace it with a non-conductive substitute. I may have to do that, but it may be challenging to find one that doesn't look cheap & nasty. I would prefer to fix the static-sensitivity at the console end, if possible.

[dalecoy]

Understood, but this problem has a multitude of potential causes.

Please describe "long vertical USB cable".

Have you checked to determine if the "chain" and the physical ground of your computer case (and the outside shell of the USB cable) are actually connected?

What else is plugged into the computer?  What kind of computer? 

Does the VP2 console power supply have the ferrite isolator? (newer model)

Have you tried not using the AC power supply for the VP2 console, for a while?

Why are you "collecting" a static charge when you stand up?  [covering of the chair; clothing; nylon underwear!)

Etc.

[pfletch101]

Understood, but this problem has a multitude of potential causes.

Please describe "long vertical USB cable".

The console is mounted almost 6' off the ground on the side of a bookcase. A USB cable runs directly down from the WeatherLink's short connector and then swoops across to the back of my computer tower, which is sitting on the floor a couple of feet (laterally) from the bookcase.

Have you checked to determine if the "chain" and the physical ground of your computer case (and the outside shell of the USB cable) are actually connected?

As far as I can determine (with a MegOhmmeter), there is no DC connection between the chain and the physical ground of the computer case, but nor is the computer case apparently directly connected to the utility ground.

What else is plugged into the computer?  What kind of computer? 

The computer is a homebrew machine with a new, reasonably high end, Asus motherboard and Intel I7 processor, running Windows 7 64. There are a lot of (principally) USB peripherals plugged in to it, including a Canon printer, a Dymo label printer, a Logitech camera, and powered speakers (directly connected; not USB). None of the other USB peripherals are disturbed under the circumstances when the Weatherlink disconnects.

Does the VP2 console power supply have the ferrite isolator? (newer model)

Yes.

Have you tried not using the AC power supply for the VP2 console, for a while?

No, but I agree that I probably should.

Why are you "collecting" a static charge when you stand up?  [covering of the chair; clothing; nylon underwear!)

Etc.

Probably friction of the man-made fiber of the clothing on my arms against the (plastic covered) arms of my desk chair as I keyboard.

[dalecoy]

Have you checked to determine if the "chain" and the physical ground of your computer case (and the outside shell of the USB cable) are actually connected?

As far as I can determine (with a MegOhmmeter), there is no DC connection between the chain and the physical ground of the computer case, but nor is the computer case apparently directly connected to the utility ground.

Hmmmm.  Thanks for thinking about that "next question" - the grounding of your computer.  There are several threads in the wxforum detailing "grounding" problems with various computers, etc. etc.  And related to Davis USB loggers.  Try searching for those.

The "solution" seems to be to fix the grounding situation. 

[pfletch101]

The "solution" seems to be to fix the grounding situation.

For various reasons, I had to put this issue aside for a few days. On taking it up again I investigated the grounding of my system more carefully, and was able to determine that it was, in fact, properly grounded, with a continuous low resistance path to local utility ground from all shielding and earth planes, and a similarly low-resistance path to 'real' ground from local utility ground. My belief to the contrary was apparently based on a previous bad measurement - probably in trying to do things too quickly. There is not, however, a low resistance path to ground from the distal part of the light switch chain (measured with a meggohmmeter). I imagine that corrosion within the chain offers a high resistance to very low voltage DC but doesn't stop high-voltage static, since static sparks certainly jump to it!

I did look at some of the grounding related problem reports you mentioned (now probably shown to be irrelevant, anyway) and they all seemed to be referring to major problems getting USB WeatherLinks to work at all. Mine works fine except when (indirectly) zapped!

[dalecoy]

A USB cable with ferrite chokes (beads) might help. 

[miraculon]

ESD (Electro-Static Discharge) is pretty tricky stuff. You might have a "ground upset" in the PC's USB circuit injecting the ESD pulse into the motherboard circuits.

You might try a high quality USB cable with 100% shielding. A lot of USB cables are made at low cost. A good one will tout the shielding design. (foil/braid)

As dalecoy mentions a ferrite might help with RF generated by the ESD event.

Another thought that I had was whether a USB hub between the console/logger and the PC would help. The ESD might get absorbed by the hub and somewhat diverted from the PC. If the ESD spike is on the shield it might not help, but it might we a worthwhile experiment.

I had a USB isolator a few years back (acromag) with my old "Green" Bliztortung lightning detector when I was still using a PC "tracker". There are some on the market that I see on Google usb-isolator. This might also be worth a try. Hopefully your ESD event is smaller than 3KV.

Greg H.

[pfletch101]

ESD (Electro-Static Discharge) is pretty tricky stuff. You might have a "ground upset" in the PC's USB circuit injecting the ESD pulse into the motherboard circuits.

If the problem were at the computer system end, I would expect some of my other USB devices to be affected, at least some of the time. In practice, when I have been working for some time at the computer, unless I am careful to ground myself before I touch the light chain, a palpable spark jumps and the Weatherlink reliably (!) disconnects and stays disconnected, but all my other USB devices continue to function normally.

You might try a high quality USB cable with 100% shielding. A lot of USB cables are made at low cost. A good one will tout the shielding design. (foil/braid)

As dalecoy mentions a ferrite might help with RF generated by the ESD event.

I will look for both of these (hopefully combined in one cable)

Another thought that I had was whether a USB hub between the console/logger and the PC would help. The ESD might get absorbed by the hub and somewhat diverted from the PC. If the ESD spike is on the shield it might not help, but it might we a worthwhile experiment.

It might. I have a strong feeling, though, that the ESD/RFI is acting at the Console/logger end, so the hub would probably have to be interposed there, which would present some practical problems with my layout.

I had a USB isolator a few years back (acromag) with my old "Green" Bliztortung lightning detector when I was still using a PC "tracker". There are some on the market that I see on Google usb-isolator. This might also be worth a try. Hopefully your ESD event is smaller than 3KV.

Greg H.

I am not keen on spending $50.00 for a solution that I suspect won't work. These devices are primarily intended to protect the computer's input circuits from being fried by ESD coming in on the USB cable. One of the reasons why I think that this is an RF-transmitted issue is that it appears to be non-destructive - reboot the computer to reinitialize the connection and you are back up and running.

Thanks for all your thoughts. I will report back if I confirm/come up with a solution.

[pfletch101]

Adding clip-on ferrite chokes to both ends of the existing cable didn't help - I had the impression that it actually made the connection more sensitive to the results of static discharge.

Replacing the existing extension cable with a somewhat longer (may or may not be relevant) one advertised as having high quality shielding and with a built-in ferrite choke at the console end does seem to have helped, but, at the same time, our outdoor temperatures have moderated (around freezing rather than low single digits F) and my indoor humidity has gone from the mid to high 20s to the mid 30s, so it is difficult to be sure that the new cable is the reason for the improvement. I have also discovered that disconnecting and reconnecting the cable after the USB connection has dropped out is enough to reset it, so I don't need to reboot the system on the (currently rare) occasions when a static discharge causes it to drop out.

[dalecoy]

So, it's perhaps not an RFI problem as you initially assumed, and may not be a problem at the console end.

What sort of diagnostics can you run when the problem occurs?  That is, can you determine if:

A.  The USB port can't be detected (until you unplug and replug); or

B.  The USB port can be detected, but there's "nothing connected to it"?

[pfletch101]

So, it's perhaps not an RFI problem as you initially assumed, and may not be a problem at the console end.

I see why you might say that this made it less likely to be a RFI-transmitted problem, but I am not sure that I agree. Ferrite chokes may be expected to attenuate (albeit substantially) rather than abolish RFI, and the clip-on ones might be expected to be less effective than ones built in to the cable and thus closer to the RFI-susceptible conductor. Cable length may be relevant to RFI, too.

What sort of diagnostics can you run when the problem occurs?  That is, can you determine if:

A.  The USB port can't be detected (until you unplug and replug); or

B.  The USB port can be detected, but there's "nothing connected to it"?

The error I am getting does not allow me to distinguish between these two possibilities, and I don't currently have the time to devote to doing so. If it gets very cold again, I may do so while I am debugging some other recent modifications to the code that need a low outside temperature to test properly.

[dalecoy]

OK, I understand you won't be able to do that.

A.  The USB port can't be detected (until you unplug and replug); or

B.  The USB port can be detected, but there's "nothing connected to it"?

The information would reveal whether the problem is at the Davis Console end - or at the computer end. 

Although that would not solve the problem, it might suggest a solution.

But since we're reduced to speculation, I'll speculate that miraculon's suggestions (USB hub or isolator) might prevent the problem.  [Or, the string that I suggested].

You did, originally, ask "can anyone suggest how I might approach fixing the problem? "

[Fivestringer]

So, it's perhaps not an RFI problem as you initially assumed, and may not be a problem at the console end.

What sort of diagnostics can you run when the problem occurs?  That is, can you determine if:

A.  The USB port can't be detected (until you unplug and replug); or

B.  The USB port can be detected, but there's "nothing connected to it"?

From my experience with the USB datalogger, I believe you can tell the USB status connection under the following conditions:

1. Datalogger is connected to USB, but not connected to the console: Go to the Communication menu and click Test. If the test takes 10-15 seconds to time out, then the datalogger is connected and communicating with your PC USB port, but it is not seated in the console.

2. Datalogger not connected: Test times out immediately.

Hope this helps.

[pfletch101]

From my experience with the USB datalogger, I believe you can tell the USB status connection under the following conditions:

1. Datalogger is connected to USB, but not connected to the console: Go to the Communication menu and click Test. If the test takes 10-15 seconds to time out, then the datalogger is connected and communicating with your PC USB port, but it is not seated in the console.

2. Datalogger not connected: Test times out immediately.

Hope this helps.

This may well be true, but I don't think that it is relevant. There is no reason to think that the datalogger is losing its connection with the console. The question that Dalecoy is appropriately addressing is whether the computer has lost its connection with the USB port (a problem at the level of the USB hardware or driver software) or the USB connection between the computer's USB port and the logger's USB 'transceiver' is disrupted (a problem at one or other end of the USB connection). Either of those scenarios would presumably result in your second response.

Some further complexity is added by the fact that I am not using the Weatherlink software (or, indeed, any other commercial software package) to communicate with the logger, so troubleshooting requires me to modify my own software to determine the nature of the error, using the rather limited error reporting facilities of the Davis DLL.

[Fivestringer]

You are correct. The immediate time out will happen if there is a problem with the PC USB port or if the datalogger is not connected.
I only mention the data logger seating thing because it happened to me. My logger became unseated from the console and I nearly went crazy trying to figure out the problem.