Hi all,
We're using the TPS65950 USB Transciever for our design (with an OMAP35x). Basically we're identical to the Beagle board layout in regards to how the USB is connected. We're running Windows CE 6.0 and have the OTG configured for client mode.
There's been a constant problem where it seems that USB is quite picky about what cable it uses. Deciding to look at what's going on, I've put DM, DP and ID on a scope to see what the problem is. Basically it looks like both DM and DP are looking VERY unhappy : they measure at a peak or around 760mV (I'd expect well above 3v...) and are quite noisy. First I thought it could be our USB driver but if I enable peripheral boot mode, the USB looks like this as the OMAP tries to peripheral boot over usb via the microcode.
I'm perplexed. This looks like it's a hardware issue but there's nothing in our layout that could account for that level of attenuation. The only thing I could think of is that maybe the transciever's pull-downs are on but according to OTG_CTL they are disabled. I checked to see if the TERMSELECT could affect it but according to the truth table this should not affect the pull-downs.
Following this line of thinking I changed OTG_CTRL_DM_PULLDOWN and OTG_CTRL_DP_PULLDOWN to enable pulldowns to see if this would further reduce the voltage level and prove these bits were actually affecting the pulldowns. After doing this I can see that the bits are set and the voltage observed on DP and DM has not grown or dimished. Could this indicate that for some reason the pull-downs are fixed on?
Beyond this has anyone else observed similar or have any good suggestions for other things to investigate that would cause this sort of behaviour?
Cheers,
~Pev
