Author Topic: broken USB line  (Read 1084 times)

0 Members and 1 Guest are viewing this topic.

Offline andrewlaphamTopic starter

  • Contributor
  • Posts: 48
broken USB line
« on: September 10, 2019, 05:33:18 am »
Hey, I have this USB wifi module that I broke. it seems it runs directly off USB but it seems the VCC requires 3.3v. I can easily solder in a 3.3v regulator to drop the voltage but what's confusing me is the DATA+ line, would this need to be regulated to 3.3v as well?

I've attached pictures of both the device and the regulator board that I damaged.
 

Offline Nominal Animal

  • Super Contributor
  • ***
  • Posts: 7198
  • Country: fi
    • My home page and email address
Re: broken USB line
« Reply #1 on: September 10, 2019, 09:20:40 am »
USB D+ and D- data lines use 3.3V signal levels for low and full speed (max. 12 Mbit/s), and 0.4V for high speed (max. 480 Mbit/s), so no, you do not need to do anything to them.  You only need to regulate the USB +5VDC down to +3.3VDC for the module.

You can use either a linear regulator, or a DC-DC converter, but make sure you size it large enough to handle the current needs of the module.

(I have a roughly similar situation, with a Huawei ME909s-120 4G/LTE module.  It takes the same inputs, essentially, and requires up to 800mA at 3.3VDC.  That is a bit of an issue, as technically you shouldn't try to draw more than 500mA from USB.  If there was a lossless DC-DC converter, then 530mA from USB +5V would suffice.  A linear regulator will convert up to 1.4W to heat.  A 85% efficient converter draws max. 620mA from USB +5V, and generates less than 0.4W of heat.)
 

Offline andrewlaphamTopic starter

  • Contributor
  • Posts: 48
Re: broken USB line
« Reply #2 on: September 10, 2019, 11:31:12 pm »
Sweet, I've got a heap of ams1117 3.3v regulators that will do the trick. Yeah my guess was the WiFi module would be pulling max about 200ma if i ran it at 1W transmitting power.
 

Offline Nominal Animal

  • Super Contributor
  • ***
  • Posts: 7198
  • Country: fi
    • My home page and email address
Re: broken USB line
« Reply #3 on: September 11, 2019, 01:23:50 am »
No, the radios aren't 100% efficient, so you can't assume that.  If this is a Pegatron WL-266N22-Duband, it might draw considerably more (based on the hardware it contains; I don't have any specific data though).

I recommend you try with an AMS1117 linear regulator, but use an USB power meter (cheapie off fleabay) to verify the actual USB power consumption.  The point is, you shouldn't assume you can draw more than 500mA off USB +5V, and you can use a cheapie USB power meter to check that.

There have been WiFi USB sticks that draw more than 500mA, and won't work on all machines.  Desktop and laptop PC's often can supply quite a bit more current, so they do work on most of those, but when you have a machine that actually follows the USB standard (say, a low-powered single-board computer), they don't work anymore.  Or even crash the machine, by drawing too much current off the +5V line, sagging it too much.  It is horrible to debug, too, unless you know to suspect the power consumption.
 

Offline andrewlaphamTopic starter

  • Contributor
  • Posts: 48
Re: broken USB line
« Reply #4 on: September 11, 2019, 05:10:05 am »
yep, you're dead on bud.  this thing as getting super hot. might have to rethink my design.   
 

Offline Nominal Animal

  • Super Contributor
  • ***
  • Posts: 7198
  • Country: fi
    • My home page and email address
Re: broken USB line
« Reply #5 on: September 12, 2019, 01:50:20 am »
Texas Instruments WEBENCH Power Designer provides several interesting solutions, like 4-6V in, 3.3V 1A out, using a TPS5432.  It uses a Würth Elektronic 74436411000 inductor, ten capacitors, and five resistors, and is over 96% efficient at output currents above 300mA.  It is basically the same as Figure 19 in the datasheet, just with slightly different component values.

With this circuit, 800mA at 3.3V only requires 550mA at 5V, and only 0.11W of heat is generated. My problem is very similar, as the ME909s-120 module consumes up to 800mA at 3.3V; that's why I've looked at this.

Unfortunately, while I can draw and route that circuit in e.g. EasyEda, using the Figure 31 in the datasheet as a layout guide, and probably even solder it (using hot air, the TPS5432 has a ground pad underneath), I am just an uncle bumblefuck hobbyist, and have no idea whether that would work in practice.  I don't even have the equipment to check for EMI.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf