0

I am trying to load a firmware up-rev to a device with a USB stick. It doesn't see the stick. My USB D+ and D- signals have what looks like a clock signal embedded onto them. The chip is an ISP1362 USB OTG controller see www.farnell.com/datasheets/1668574.pdf

I have not found any components that seem to be bad.
To see a picture of the D+ incoming signal that looks like it has the clock signal mixed into it, copy and paste the following into a separate window.

enter image description here

The scope is set at .5 usec/div.

Schematic Pin 48 is wrong on the schematic, it is pulled low. Here are the readings I measured: on ISP1362.

Pin
20 3v 21 3v 22 3v 28 3v 29 3v 23 0v 59 0v 60 3v 48 0v 45 3v 30 3v 41 5v
46 signal shown when USB first inserted, otherwise 0v
47 inverted signal shown when USB first inserted, otherwise 3v
33 0v 34 0v 61 3v 62 3v 56 5v
4, 14, 26, 40, 52, 56 3v
1, 9, 19, 27, 57, 51 0v

Any guess as to what the problem might be?

Dmitry Grigoryev
  • 25,576
  • 5
  • 45
  • 106
MAT
  • 11
  • 1
  • It looks noisy, that's it. – Eugene Sh. Apr 19 '17 at 21:57
  • The image is at http://imgur.com/a/U0sil – MAT Apr 19 '17 at 21:57
  • More info to come later – MAT Apr 19 '17 at 22:00
  • Shown is one complete cycle, if this was just noise it would be random but this is constant and it matches the output of the clock at 3rd and 6th division – MAT Apr 19 '17 at 22:02
  • 3
    Noise is not necessarily random. It can be a cross-talk with some other signal (like clock). And the D signal is differential, so you should look at the difference between D+ and D-, not on one of them. – Eugene Sh. Apr 19 '17 at 22:04
  • 6
    "More info to come later" is not a way to post questions that you actually want answers to. And I do hope you mean _more_ information, because we know hardly anything here. – Asmyldof Apr 19 '17 at 22:16
  • Show us your schematic and layout. – uint128_t Apr 19 '17 at 23:20
  • 1
    What's the scope bandwidth? – Matt Young Apr 20 '17 at 02:01
  • 1
    Apart from the low scope bandwidth it looks normal to me. Does your scope have dual trace? If so then show us both D+ and D- signals together. Use 10x probes and make sure they are properly adjusted for flat square wave response. Here's an example of what you might see:- http://kevinpt.github.io/ripyl/_images/usb_data_hs_example.png – Bruce Abbott Apr 20 '17 at 05:12

2 Answers2

1

Could be these things:
1) The scope does not have enough bandwidth to capture the signal, get a better scope. (or turn up the time resolution if possible). Make sure the probes can handle the bandwith also
2) The scope impedance is too high. Make sure the probe is clipped to the shortest possible distance between the signal that your monitoring and the nearest ground on the IC.

Voltage Spike
  • 75,799
  • 36
  • 80
  • 208
1

If it is 0.5 us/division, I think it's just a bandwidth problem.

A full speed device will have 12Mbps. In your image, in one division I see up to 6 periods, which means exactly 12 MHz (83 ns per bit). That signal is just normal.

next-hack
  • 5,297
  • 15
  • 33