[sdiy] How to design out "usb noise"?

[brianw]

On Feb 4, 2024, at 2:30 PM, mskala at northcoastsynthesis.com wrote:
> On Sun, 4 Feb 2024, brianw wrote:
>> You cannot optoisolate a (DC) power source. You can only optoisolate
>> (AC) signals.
> 
> When I said "optoisolate it" I thought it was clear that "it" meant the
> data.

The noise is most likely coming in over the USB shield, ground, or power - not the data. It might even be more problematic because of the ground, not the VBUS, per se, since it's easy to keep the VBUS separated.


> The point is to not have a connection between the USB power and the
> synth power.  Whatever receiver electronics need to be powered by USB
> power can be, but then that stuff transmits only data and *not* power
> across the optoisolator to the rest of the synth, which is on its own
> power with some effort to keep that power clean.  The synth's power is not
> coming from the USB cable.  The original question was pretty clearly
> describing a synth with its own power system that needs a USB cable
> plugged in to receive data, but then when that cable gets plugged in, oh
> no!, everything gets noisy because of the ground loop created.  It's a
> losing battle trying to clean that situation up unless the ground loop can
> be broken and the noise voltage on the USB +5V prevented from influencing
> other power lines.

I didn't quite make the assumption that the synth would be a USB Device, because it's becoming more common for products to offer USB Host features as well, allowing USB-MIDI controllers to be attached easily. Most consumers still don't understand that USB products are divided into Hosts and Devices, because quite often there will be Customer Support questions like, "Why can't I connect a USB-MIDI control surface to my synth?" The answer is that you can't connect one USB Device to another USB Device.

Granted, if the synth is a USB Device, then self-powered is a good option, but there is also a good argument from the users' perspective to design for Bus Powered. At the very least, you have one less wall wart to keep track of when packing for a gig.


>> even filtering for (DIN) MIDI, and if you follow those instructions
>> you'll have solved a lot of the potential problems. But you still need
>> to keep the CPU from introducing noise on the ground and power rails.
> 
> Ideally, any CPU involved would as you say be on the USB side of the
> isolation.  However, even just the USB transceiver being on the USB-power
> side, with most of the digital stuff being on synth power, would probably
> be enough (requiring two-way optoisolation because USB needs to transmit
> in order to receive).

I do not think that it is possible to optically isolate in two directions. USB uses differential signaling, so both the D-/D+ are driven by the Device at some times, and by the Host at other times. In fact, that's one of the reasons why a Self-Powered Device still has to wait for VBUS, because it can't step on the Host driving those Data lines.

If there is a chip that optically isolates the bidirectional differential USB Data lines, I'd be fascinated to look at a data sheet to see how it works.

> The very common complaint of synths becoming noisy
> after connection to USB isn't caused by a USB-receiving CPU built into
> the device.  It's caused by garbage coming through the USB power lines
> from the much larger computer on the host end of the USB cable.  Keeping
> those noisy power lines separate from the synth's own power lines is a big
> step forward, even if some digital electronics built into the synth still
> end up being powered from the synth's own power.

Exactly. Isolating the data lines won't keep the noise on the power lines out. But shielding properly - without direct grounding - and then filtering the power as well as following proper grounding techniques to separate digital and analog will all help.

> -- 
> Matthew Skala
> North Coast Synthesis Ltd.