| Electronics > Beginners |
| Shottky diodes for USB ESD |
| << < (3/3) |
| EEEnthusiast:
Stay clear of adding through hole components on the USB lines. They act like stubs, can pick up noise and cause multiple problems due to the vias added. Even if you plan to make a kit, you could use 0603 components which can be easily hand soldered. |
| Blitzschnitzel:
--- Quote from: pigrew on October 16, 2019, 04:04:44 pm ---I'd normally suggest diodes sold as TVS diodes (as they will have a somewhat different die layout), but the ones listed on DigiKey all have way too high of a capacitance (1 to 10 nF) for an IO signal. For low-speed and full-speed USB, application notes suggest 5V stand-off, even though the logic shouldn't go above 3.3V. Probably a "standard" zener would be the way to go (like BZX79-C4V7,143 which has ~300 pF capacitance). However, the datasheet of the NXP zeners do not mention ESD ratings. Another note is that FTDI suggests adding a small capacitor (~47 pF) to the data lines which also helps with ESD. Double-diode arrangements would work, but require twice as many diodes. :) If this were my design, I'd go with a SMD component, as appropriate THT diodes don't seem to exist. Something like ST USB6B1 in an SOIC package should be manageable. --- End quote --- Thanks I'll try the BZX79-C4V7 And if that causes interference I'll change the design to a USBLC6-2SC6. I had to buy a 50 strip for the last project anyways. --- Quote from: EEEnthusiast on October 16, 2019, 04:34:25 pm ---Stay clear of adding through hole components on the USB lines. They act like stubs, can pick up noise and cause multiple problems due to the vias added. Even if you plan to make a kit, you could use 0603 components which can be easily hand soldered. --- End quote --- hmmm... I've looked at many open harware keyboards which use none of the best practices of USB 2.0 design. So far my impression was that HID-device speeds are so low that you can send the signal through a rusty spoon. |
| jhpadjustable:
--- Quote from: pigrew on October 16, 2019, 04:04:44 pm ---If this were my design, I'd go with a SMD component, as appropriate THT diodes don't seem to exist. Something like ST USB6B1 in an SOIC package should be manageable. --- End quote --- Lead inductance defeats the entire purpose of providing a path to ground that has very low inductance, relative to the trace inductance to the protected devices. ON Semi application note AND8232 is an excellent explanation of how to use TVS for ESD protection. --- Quote from: Blitzschnitzel on October 16, 2019, 05:12:43 pm ---hmmm... I've looked at many open harware keyboards which use none of the best practices of USB 2.0 design. So far my impression was that HID-device speeds are so low that you can send the signal through a rusty spoon. --- End quote --- 1.5Mbps USB could be sent through a rusty fork, but it's still not a good engineering practice, and will still contribute to the smog in the HF and MF radio bands. (Still waiting for Dave to take some EMC probes to that 4-layer Gigatron he was soldering a while back...) |
| Bassman59:
--- Quote from: Blitzschnitzel on October 16, 2019, 05:12:43 pm ---hmmm... I've looked at many open harware keyboards which use none of the best practices of USB 2.0 design. --- End quote --- Don't confuse the spec revision with the wire speed. USB 2.0 added the 480 Mbps High Speed transfers, but devices did not have to run that fast. USB 2.0 supports the 1.5 Mbps Low Speed transfers commonly used by keyboards. |
| langwadt:
--- Quote from: wraper on October 16, 2019, 02:15:36 pm ---You don't need 3.6V TVS. ESD protection devices made specially for USB are usually rated for 6V minimum breakdown voltage. Also you shouldn't just slap any TVS or varistor onto USB data lines, they need to have low capacitance. --- End quote --- It looks like the hack used for bitbanging low speed USB with a 5V mcu (V-USB) I've seen some use blue leds |
| Navigation |
| Message Index |
| Previous page |