Hi,

I wanted to talk about this new project, and show some photos. It is a USB test switch, or switcher, aimed for test automation. Basically, it allows you to connect or disconnect the bus power or the data lines separately, for enumeration testing.

For now, you can see a photo of the board, and the schematic. I'll post more details soon. I'm still designing the circuit, since this project is still a prototype. Many things were already green wired onto the board, in order to fix the current measurement glitch. Also, forgot to add a pull-up resistor to an open collector output.

​

Anyway, this board employs a TS3USB221A to connect and disconnect the data lines. For interrupting Vbus, I've used a LM3525-H. Originally, a LM3525-L was used, but then I've verified that Vbus would connect when the USB on the control side was disconnected. The solution was to use the "H" version instead, with its "EN" input negated via a transistor.

I'm yet to characterize the board, and also create software and drivers. Anyway, I just wanted to share this with you.

Kind regards, Samuel Lourenço

 

Sorry, it seems that the original pdf containing the schematic was not properly formatted, so I had to re-upload. However, I can't upload it on the OP, due to the action being interpreted as spam, which is strange. So I have to post it here.

Kind regards, Samuel Lourenço

 

Did some mods to the board in the last few days. I still have to test the current measurement (none of my USB safety testers agree with any of my multimeters). I suspect that the current measuring resistor is a bit off, despite its 1% precision. When measuring the voltage across it, I get a reading that corresponds to the value being read by the on-board ADC.

Anyway, here are some photos of the corrections made to the board. I think we should start a "Patched boards that now work collection" tread with pictures of green wired PCBs and tented components.

Forgot to add a pull-up resistor (soldered to two of the IC5 legs). Also, the LM3525M-L was replaced by a LM3525-H.​

Added a MOSFET with a pull-up resistor in order to negate the signal going to the "EN" input of the LM3525-H. Also, an RC filter was needed between the current sensing amplifier and the ADC. To finish it, some epoxy glue.​

Enjoy!

Kind regards, Samuel Lourenço

 

Hi,

It has been some time now. Meanwhile, I've finished the layout and ordered the boards. Now I'm in the process of building the software and testing. Already tested the current measurement feature, and it seems to be precise.

I've included the layout and some images. For your viewing pleasure, enjoy!

Kind regards, Samuel Lourenço

 

Here are the photos and pictures of some tests. The control commands are still in a preliminary stage. Once this project is complete, I'll probably post it in "The Completed Projects Collection" forum, just like in the old times.

Kind regards, Samuel Lourenço

 

Hi,

I've done some current measurements on two devices, a SanDisk Cruzer Slice and a SanDisk Ultra. The results are interesting. We can see the effect of the LED blinking on both devices. On the SanDisk Cruzer Extreme, we can see a very smooth ramp when the pen drive is idle. Notice the current bursts that are a consequence of file reads. On the Ultra, the LED is clearly controlled in a PWM fashion (wavy lines with lots of "noise").

I've attached the charts to this message. The way I did the measurements is very crude. The current is expressed in mA, but the time scale does not represent time in an accurate way (just measurements taken, roughly 100 every second).

Kind regards, Samuel Lourenço