RF causing USB to malfunction

[djacobow]

I've got a new ICOM rig with control and audio over USB. It's pretty cool and very convenient.

However when transmitting, many USB connections to my PC (including the rig itself) disconnect and reconnect. I believe that RF is interfering with the connection. With minimal RF measurement gear, what's the best way of debugging this? Plan now is to just trial and error putting ferrites here and there to see if I can figure out the issue.

Antenna is G5RV type (I know, not great, but it is my first HF antenna) connected to shack through < 30 ft of RG213. Is it worth the investment ($30) for an "RF isolator balun" on the feedline? I believe these are just a length of coax strung with a bunch of beads.

For USB I use a hub and have a lot of devices connected to it because I use this machine for HW development, so JTAG, serial adapters, programmers, etc. I can yank all that stuff while doing ham stuff, but it does not seem to make a difference.

[Fank1]

I won't get into the G5RV.
What kind of SWR do you have?
A balun may possibly help but first make sure you have a good solid ground on the tranceiver.

[djacobow]

I won't get into the G5RV.
What kind of SWR do you have?
A balun may possibly help but first make sure you have a good solid ground on the tranceiver.

I'll have to check what the antenna is getting by itself. The radio sees 1.1 or something from the tuner.

A ground is there, but not great. Using AC wiring ground right now, which I know is not good for RF. But in this case, it's quite short, probably 20 feet of #10 to a copper to the breaker box, then a few more feet of #8 to copper rod.

[xrunner]

Plan now is to just trial and error putting ferrites here and there to see if I can figure out the issue.

Yep, I've had the same issue before using a laptop in my living room in the evenings with a small rig running it to work JT65. Some issue can be solved by placing ferrites at the right places on cords. Others can't. For example, if I use 40 m, it locks up my mousepad. I've found there is nothing practical I can do about it. The antenna I'm using is right above me outside, and the RF in that band is coming into the computer directly. I can't shield it.

[djacobow]

Yep, I've had the same issue before using a laptop in my living room in the evenings with a small rig running it to work JT65. Some issue can be solved by placing ferrites at the right places on cords. Others can't. For example, if I use 40 m, it locks up my mousepad. I've found there is nothing practical I can do about it. The antenna I'm using is right above me outside, and the RF in that band is coming into the computer directly. I can't shield it.

Interesting. I was doing JT65 (forum members will remember I trash talked this mode recently) at the time I noticed this problem. I did not notice it when I did PSK31 or phone earlier, but it didn't even occur to me that the mode might be a factor.

[T3sl4co1l]

At what frequency/band?

It's an awful antenna design.  It's made, I can only assume intentionally, to deliver a near maximum of common mode current.  Why?  The feedline acts as an antenna too, perhaps giving more isotropic response, or wider bandwidth.  It can give the illusion of having a larger antenna, but carries the downside of being different for every user, plus it blasts crap all over your equipment.  It's piss poor design from the kind of people who don't test their shit! 

Your ground will most likely need to be a copper sheet down to a ground rod, as near to the antenna as possible.  There's no "ehh close enough" in RF. 

So yeah, add a balun and use a conventional dipole.  Doesn't need to be fancy.  At least a stack of ferrite beads (in conjunction with a good actual ground) for starters, but preferably a ferrite or powdered iron core with proper matching balun windings on it.

As for your USB devices, some of them are hopeless.  HID (keyboard and mouse?) almost always use unshielded cables.  They're low bandwidth too, but the ports can't be filtered that low (no one makes HID-dedicated USB ports), so they're still susceptible to RF.

Other devices should be responsibly shielded.  If they aren't... it sucks, but they're crap.  Basically you wasted your money on something that doesn't meet basic standards.  You'll have to buy new ones that actually work.  If you can't find one (at any price) that's not susceptible, you're screwed.

Properly shielded USB should withstand tens, even hundreds of volts of ambient RF.  It's nothing special, that's just the nature of good shielding.  For some reason, almost no one understands how to accomplish this, though.

Touchpads are the first to go, even with responsibly balanced RF.  You simply shouldn't even be trying to use them in an RF field.  They're not intended, by any means whatsoever, to be a reliable, dedicated input device.  (Which seems kind of strange, when you consider how few choices of input device there are on something like a smart phone, but alas...)

Tim

[djacobow]

At what frequency/band?

40m. (I don't seem to get problems on 20m.) I spent some more time on it today and got very frustrated, trying all sorts of things. But it looks like making sure the antenna coax and USB coming out of the radio are as far from each other as possible helps. I also added a "air coil" choke (ie, a few loops of coax). Not sure if that helps at all.


I bought the G5RV last year and only got a transeiver recently. (Before that, had SDR receive). Anyway, shortly after I learned how little people like this antenna. I'm not sure why it's suggested to beginners so often.

I'll have to take it down and replace it with a simple dipole cut for 20 or 40. I have a very small lot, and a single tree. So, it'll have to be inverter vee and folded as well. It was hard enough to get that antenna up in winter when there wa no foliage. Should be a trick to get a new one up when the tree is leafy.

  It's piss poor design from the kind of people who don't test their shit! 

Well, it's been broken for a long time, and yet it's out there. I'm only out like $60, at least.

Your ground will most likely need to be a copper sheet down to a ground rod, as near to the antenna as possible.  There's no "ehh close enough" in RF. 

I need to much better understand this. Are there antenna arrangements that do not need ground at all? I mean, I know there are antennas recommended for apartment dwellers, etc.

Tim

Thanks for all the advice, most of it snipped here for brevity. Yeah, I have to go through these USB devices one by one. At least they're easy enough to plug, unplug.

-- dave j

[T3sl4co1l]

Yes- any balanced antenna design gives (ideally) no common mode.

Obviously(?), anything symmetrical, like a loop or dipole, has equal and opposite voltages (and equal currents) at the feedpoint.  All you need to provide is equal and opposite voltage, usually from a transformer.  Then you can use unbalanced (coax) anywhere you please.

Nonideal situations, like the antenna elements being of different lengths, or being nearer to some branches or structures, or the ground, than others, mean you can still get feedline current.  Ferrite beads are good for this (just looping the transmission line isn't, unless it's very specifically a 1/4 wave helical resonator).

Don't count on 20/40 action, unless you're using a short dipole and a powerful tuning box (two tuning settings), or a 40m resonant antenna with 20m traps (so it looks like a 20m resonant antenna at 20m).

That is,
1. If you want best performance at 20m, cut the antenna to be resonant at 20m.  You can still use it at 40m, but it will be electrically "short", meaning you need to switch in a tuner box to fix the length problem.
2. If you want best performance at 40m, cut the antenna to be resonant at 40m*.  You can still use it at 20m, but you have to insert 20m traps, so the extra antenna length becomes electrically inactive.  (Otherwise, 20m is antiresonant, and you can't really get any power out of the thing at all, even with a rather aggressive antenna tuner.)

*Includes loading coils.  Note that resonant traps look like inductors below f_o, and capacitors above.  The added inductance means the physical length (for 40m tuning) will be shorter.  Which is probably an advantage, really.

Ed: visuals: http://degood.org/coaxtrap/

Tim

[Kilrah]

For USB I use a hub and have a lot of devices connected to it

I've had several bad experiences with hubs that were very susceptible to RF (things like disconnecting when a mobile phone would receive a message if it's less than 10cm away from the hub).

Also as there are tons of cable going to it that's as many antennas for it to pick up nasty stuff. Simplifying the wiring might be a good idea...

[djacobow]

Obviously(?), anything symmetrical, like a loop or dipole, has equal and opposite voltages (and equal currents) at the feedpoint.  All you need to provide is equal and opposite voltage, usually from a transformer.  Then you can use unbalanced (coax) anywhere you please.

OK, that is consistent with my limited intuition. It's been >20 years since I took one class in E&M Fields and I distinctly remember that there was no practical content whatsoever.

Nonideal situations, like the antenna elements being of different lengths, or being nearer to some branches or structures, or the ground, than others, mean you can still get feedline current.  Ferrite beads are good for this (just looping the transmission line isn't, unless it's very specifically a 1/4 wave helical resonator).

What do you think of this mighty-fine junk?
http://www.mfjenterprises.com/Product.php?productid=mfj-915

Don't count on 20/40 action, unless you're using a short dipole and a powerful tuning box (two tuning settings), or a 40m resonant antenna with 20m traps (so it looks like a 20m resonant antenna at 20m).

My tuner has, in addition to the unbalanced input, a balanced input with a built-in 4:1 transformer. Could I just take the antenna I have, cut the coax connector off after the ladder line and extend with more ladder line all the way into the shack, essentially making it a dipole?

I believe this antenna (G5RV jr) is cut for 25.5 ft on each side. Lengthening seems awkward, though I guess I could solder on more wire or maybe fisherman's knot it on. Cutting it down to 20m is feasible. But I'm curious about how badly it would perform (obviously, with a tuner) at the current length.


I had a little more success last night. My antenna feed line was about 10 feet too long so I just had it coiled behind the radio and tuner. When I moved that extra coax away from the radio, it seems that it stopped the USB hiccup, and I was able to make some contacts using JT-65, mostly Japan and a few China, on both 20 and 40. (I'm SF Bay Area). I also had a PSK31 contact with Sakhalin Island two nights ago on 20.

[borjam]

Interesting. I was doing JT65 (forum members will remember I trash talked this mode recently) at the time I noticed this problem. I did not notice it when I did PSK31 or phone earlier, but it didn't even occur to me that the mode might be a factor.

I'm doing JT9 and JT65 with a loop antenna quite close (MFJ loop by the window), 20 - 30 W, and I've had no problems. IC-7200 and a Mac Pro. I added ferrites to the USB cable just in case. My USB cable is connected to an Apple keyboard.

JT9 and JT65 are quite similar to a constant carrier,

[T3sl4co1l]

OK, that is consistent with my limited intuition. It's been >20 years since I took one class in E&M Fields and I distinctly remember that there was no practical content whatsoever.

Bleh... I hate classes like that.  I think the worst one I had was Control Theory.  's' everywhere.  They could've put in an RC here and an LC there, but nope.

Wasn't a problem for me, because I already understood all those things (systems and their equations in terms of components), but I don't think there was another person in the class who got it...

What do you think of this mighty-fine junk?
http://www.mfjenterprises.com/Product.php?productid=mfj-915

"Isolator" isn't a good term to use for it (that's a microwave component that absorbs reflections), but the description sounds like the thing.

Or you can make your own, but I suppose a sack of 50 ferrite beads would cost about the same, really.

My tuner has, in addition to the unbalanced input, a balanced input with a built-in 4:1 transformer. Could I just take the antenna I have, cut the coax connector off after the ladder line and extend with more ladder line all the way into the shack, essentially making it a dipole?

I believe this antenna (G5RV jr) is cut for 25.5 ft on each side. Lengthening seems awkward, though I guess I could solder on more wire or maybe fisherman's knot it on. Cutting it down to 20m is feasible. But I'm curious about how badly it would perform (obviously, with a tuner) at the current length.

Ah, great!  Then no worries.  Tune that sucker on the balanced terminals.  If it's 'short' for both bands, it'll look capacitive (give or take feedline action), just needing a varying amount of tuning.  As long as the tuning range is enough, and the balance is good, you should be fine.  May be a help to ground the antenna box too, just in case.

Tim

[zl2wrw]

Hi Djacobow,

I strongly suspect that during TX, common mode RF current is flowing in the USB cable between the Icom and the USB hub.
You could try installing a common mode choke on the USB cable (wrap several turns of it through a ferrite toroidal core), and if that does not work, you could try using a USB isolator (one that provides complete galvanic isolation between its input and output).

If your Icom has an ethernet LAN port, I strongly suggest that you use that instead of USB - unlike USB, ethernet is both galvanically isolated and the data signal is completely differential.

Good luck

[djacobow]

Ah, great!  Then no worries.  Tune that sucker on the balanced terminals.  If it's 'short' for both bands, it'll look capacitive (give or take feedline action), just needing a varying amount of tuning.  As long as the tuning range is enough, and the balance is good, you should be fine.  May be a help to ground the antenna box too, just in case.

I sort of resolved my USB problem by separating cables as widely as I could and also by working at reduced power, but just yesterday I got around to changing my G5RVjr to a balanced input all the way to the tuner, and it seems to work pretty well. On 20m, it tunes 1:1, no major symptoms of RF in the shack, and it seems to be hearing better. On 40m I can only get it to 1:1.5. This is actually worse than before. Maybe I should try lengthening or shortening the feedline. I haven't determined if it is actually working better or not.

The main barrier for me to making the conversion to ladder line feedline was putting a new hole in the wall of my shack. Ended up using a covered outdoor electrical cover box around a couple of pieces of PVC that penetrate the wall. Used an outlet box on the inside, too, to position a blank cover that I cut holes in. Stuffed some fiberglass in the PVC to block critters and provide some insulation. Looks ok, works ok.

[tatus1969]

that isn't necessarily RF interference. USB is quite picky with ground potential shifts. Yes, it is differential transmission, and no, it is not really. If you have more than some 250mV (don't have the exact value, its in the USB spec) of ground offset then communication will be disturbed. So if you have some USB connected devices like this that have their own power supply, and computers like laptops, and not everything goes to the same protective earth, or your supplies are isolated, and, and, and, then you can run into similar problems as your audiophile neighbor.

[cncjerry]

I had these same problems whenever I used my amp.  Interesting thing though, was I was using a tuner with an off center fed dipole.  I had a balun that took the coax from the tuner to open wire about 50ft from the tuner.  I also have a good ground with multiple bars, good soil, tinned copper braid, etc.  Without the tuner, on bands where the OCFD had resonance and low SWR, no problems with USB; with the tuner - major problems.  I found later that the tuner had two tuning settings for the band, 80M, where I had problems.  The low inductance setting had the least impact on the USB.  I cleared up the problem by proper tuning along with ferrite cores wound on everything. 

I had RF sparks everywhere when using that tuner and 1500W.  I was getting little burn spots in the corners of the amp and power supply even though the SWR was tuned to ~1:1.  The SWR without the tuner was about 1.5:1 with no sparks and no USB issues.  I never thought thru it all.