Has anyone tried adding these to their power strips or chargers?
https://www.amazon.com/eBoot-Pieces-Ferrite-Suppressor-Diameter/dp/B01E5E5IY4/ref=sr_1_3?ie=UTF8&qid=1468533064&sr=8-3 https://en.wikipedia.org/wiki/Ferrite_bead |
Banned User
|
Ferrites typically provide about -3dB or less per single turn (clamp on type) ferrite. Depending on the material, they can go up to about 50Mhz, and you wouldn't expect much if any effect below 500KHz.
I've found that most people with ES are able to just barely perceive a -30dB change in a problematic, symptom inducing symptom. 60dB is felt as a substantial improvement (though still bothersome over time). So clamp on ferrites are disappointing, usually. I use them where practical as a matter of principle, even though I can't feel the difference. We all wish that simple commercial level shielding and filtering fixes would work, but my experience is that much, much higher levels are needed when it's a problematic frequency. Cherry picking equipment that is less bothersome, then shielding and filtering (not the silly motor run capacitors) the heck out of it to avoid sensitizing later does work, but it is a big project. Books on the electrical engineering field of Electro-Magnetic Compatibility are a good place to start for the straight scoop. |
Do you mean -3dB measured by a meter held near the offending device, or -3dB on the line?
I've had so-so luck with home-built inductors/chokes, so I wasn't expecting much. |
Banned User
|
-3dB at best on the conducted emissions, radiated emissions maybe half that, depending on where the ferrite is placed on the cord. If the source is a switching supply with a frequency of 100KHz, the old engineering (EMC) rule of thumb is 10x the fundamental frequency is very strong, 100x significant. The switching transisitor transition (off to on, on to off) speeds have increased dramatically to a couple nanoseconds, allowing even smaller or no heatsinks, this old rule of thumb needs adjustment upwards to more than 1000x. Ferrites will be (slightly) effective in that range, but compared to a well designed commercial level (-40dB) multistage LC filter (see FilterConcepts SW series) they are a bit of a farce.
Another problem with filters on AC is that you really should be using them on a balanced AC line; so you should use an toroidal isolation transformer first, at a distance from the user. Then you can provide an isolated, high frequency, earth ground for the filter and not cause very high ELF magnetic fields. Mostly, I just wanted to say clamp on ferrites do work, and can be effective but unless the source was very, very close to your tolerance level, you would are very unlikely to be able to feel the difference. Trying a different switching supply (they vary wildly in emissions) is more likely to be fruitful. |
In reply to this post by BruceM
<quote author="BruceM">Cherry picking equipment that is less bothersome, then shielding and filtering (not the silly motor run capacitors) the heck out of it to avoid sensitizing later does work, but it is a big project.
That's what I've tried to do. I'm actually writing this on a home-brew panel-PC based on a BeagleBone Black SBC and a 4D Systems 7" LCD, with an Andice Labs power supply to give it enough of a sense of humor about input voltage to run from a 12V lead-acid battery or a 100' DC extension cord. The nice thing about that setup is that there aren't a lot of low-frequency emitters - just a 32K oscillator (I think for the RTC) and a PWM dimmer for the LCD. I run the LCD at 100%, and I could probably also solder a capacitor between the two sides of the backlight circuit to soften the pulses, since there's only one channel. I may be getting close. There seemed to be some improvement when I put chokes in a few important places, so here's hoping :) There's at least one other SBC/LCD supplier that I've looked into, and they were super helpful about providing schematics and datasheets or some of the board level components that I couldn't find docs for (one PMIC in particular). That would let me move up to a quad core, 1.4 GHz ARM9 SoC with gigabit ethernet and a gigabye of RAM, and a 10" display with ~50% higher resolution. The BBB is a nice little board, but it has to work a bit to run modern desktop software. |
Banned User
|
Good project, Karl.
I've gone to shielded (welded steel 21" cube) pocket (LED) projector with rear projection screen. That allows me to put the computer in a separate building. Very elaborate, and a big project, but it does allow me to use the computer. I'm not crazy about the DLP display visually, but it's less bothersome than 60 or 72Hz refreshed LCDs for me. LCD's actually do have 1-2% modulation, a fade to white between refreshes. To me that is just as disturbing as a 60Hz CRT with 99% modulation. The keyboard and trackball were actually much more difficult for me than anything else, I really don't tolerate active matrix (strobed) keyboards or microcontrollers right in my lap. There's not too many well enough and with the technical background to do the disability engineering work of applying EMC to ES disability. Instead what we get are fraud products like pendants, stick on labels, motor run capacitors, etc. |
That's on my project list, but I've never taken a mini-projector apart, so I didn't know how much shielding I'd need. Are you using solid steel or something gapped? It would be really nice to have mill-produced sheet with alternating layers of carbon steel and an austenitic steel like 303 or Mangalloy. That's farther than I was thinking of going, but worth looking at if I can come up with the space. I was just thinking about the cables, but it's been so long since I've pulled apart a keyboard that I have no memory of the electronics inside. Can you clue me in? No kidding. It's stuff that even most EEs don't deal with. I grew up around privately a held R&D/instrumentation company, so it's possible that I could get things built. Right now I'm swamped figuring this stuff out and trying to find suitable off-the-shelf parts. Feel free to shoot me an e-mail through the site if you're interested in collaborating at all. |
Banned User
|
All keyboards and keypads use the "active matrix" method, since it supports N-key rollover. For example, rows pulled high, then columns individually driven low, one at a time. Diodes added to allow for N-key. Search on google for active matrix and you can find some articles. This has been used exclusively since microprocessors were available to do it, starting in the late 70s. If you hold an AM radio near the keyboard and touch a key, you'll hear the screaming of the active strobing of the columns. They are typically 2k pull ups so not a lot of current, but the current loop (as big as the key matrix) is very large and right there in your lap. This is a show stopper for me, EMI wise. Even shielded well, the current pulse and large loop area was too much for me. I tried reducing the pulses down to millivolt level and filtering them but was still unable to use it.
I finally developed my own method I call "passive matrix". Rows pulled high 1M, columns pulled low 1M, key sensed by micropower analog comparitors or other means. Shift, control, alt must be brought out separately. I then created a shielded and filtered, opto-isolated interface to an IR keyboard/mouse processor that I modified to drive plastic optical fiber all the way to my computer in separate building. I've been using this for 20 years. I've helped a couple other guys make their own versions of the same basic design, though both use fiber USB instead of a modified IR keyboard processor. I'd love to do a new version designed from scratch with a micropower processor but my MS has advanced enough that I'm not up for the software anymore. My short term memory is not adequate for much more than a single page of code, and learning new tools is hell. |
Thanks a lot for all the technical information, it is already interesting and useful as far as i can understand it, which is limited.
What about laser keyboards ? : https://en.wikipedia.org/wiki/Projection_keyboard Probably emf-wise also problematic, and possibly hard to get with an usb-exit these days to begin with, but the question entered my head and here it is. |
Free forum by Nabble | Edit this page |