Originally Posted By: Ross Wellington
I also noticed that the schematic he referred to has very little decoupling too. It would benefit from a 10uF capacitor on the USB power side of the inductor.


The system I'm having difficulty with, already has a (10nF) cap on the USB power side of the inductor.

The clock crystal/circuit is built into the FT232R chip, and again, this entire USB chip/board is optically isolated from all of the motor electronics. There is no shared ground or signal or voltage lines with the motor driving stuff (also just a single chip).

The motor driver has no clock or frequency to speak of. It's either on, or off, and there's no clock or switching of any kind within that .. just some darlingtons wired as H-bridge, with diodes -- all inside the single chip.

The USB tends to fail when the motor is stalled (not turning, and therefore not generating pulses). Or perhaps the USB fails just before (?) the motor is about to stall, when there may be a rather largish EMI field being output (motor draws about 450mA at 4.8V when stalled, and maybe only 1/10 that when free running).

So the mystery for me really is, how does a stalled motor, on independent circuitry and independent PSU, somehow affect my USB gadget? Yes, it's gotta be EMI, I suppose. Big magnetic field there from the motor at that point.

The FT232R board does *not* have the usb cable shield connected to anything other than the terminating USB socket itself. The cable *is* grounded on the host end, though. It has a 10nF cap from USB power to ground, and some other similar bits aimed at transient suppression or something. smile

I did once try connecting the cable shield ground to USB logic ground on the FT232R board, and was surprised when doing so made things *much* worse. But that was before I decoupled the motor driver from it with the optical isolators.

Here's the stock FT232 board schematic -- very simple, and probably in dire need of some mechanical shielding.

I take the CBUS0/CBUS1 lines off to a pair of optical isolators, and from there back through 10K to GND. Those pins are powered from the VCCIO line, which I have jumpered to VCC.

End of circuit. The rest is on the other side of the isolators, with it's own 5V PSU, a single H-bridge chip, two LEDs, and a motor.


Attachments
1.gif

Description: Description: Schematic for FT232R breakout board.