Page 5 of 20

Re: S3 stops mid-job

Posted: Thu Aug 27, 2015 10:11 pm
by apspurgeon
It shouldn't, ground loops (especially audio) are often a result from different earthing points in large structures (building). e.g the Earth potential in one socket in one end of the house is different to the other as they are earthed in different locations.

In this instance, I have all grounding connected together and to a single power socket.
Any difference will be negligible. IMHO

Re: S3 stops mid-job

Posted: Thu Aug 27, 2015 10:18 pm
by veng1
There can be some safety concerns about grounding logic voltages to equipment ground.

Adafruit and others make a USB isolator:

This uses an Analog Devices part similar to one I've used successfully in the past to connect a tablet to a machine with a large compressor switching on and off.

An advantage to this design is it provides 100 mA to power something on the USB port.

Re: S3 stops mid-job

Posted: Thu Aug 27, 2015 10:20 pm
by apspurgeon
it's already grounded through the SO3 Power supply. That may in fact be the source of the problem, in how it's grounded (e.g indirectly).
The case of a PC running of the mains is grounded, so what I am suggesting is essentially doing the same for the laptop.

The point being, that the shielding of the USB Cable has a earth/ground. In any case, it does seem to make a big difference (if not fix the problem)

Re: S3 stops mid-job

Posted: Thu Aug 27, 2015 11:51 pm
by veng1
Have you actually measured that?

Most supplies I'm familiar with do not tie the negative side of the power to earth ground but I agree if it is so, it may be a major part of the problem. Even the ones I've seen typically have to do it with a removable jumper.

I guess I need to measure mine also although there may be some differences due to localities.

Re: S3 stops mid-job

Posted: Fri Aug 28, 2015 12:17 am
by apspurgeon
Probably what I'm not communicating well is the shielding of the USB cable, the metal of the SO3 and the metal of the PC (which is required by law) and the laptop should be earthed to 'drain' the high freq noise. Ideally all to the same earth point (to stop the above mentioned ground loops)

I'm not suggesting tying the 0v of 24v SO3 power supply to earth. Perhaps I've confused the ground term as 0v is often called GND.
That said, I'll meter the SO when I get home and check if the electronics 0v is indeed tied to the power supply ground pins (not the 0v) and USB shield on the board connector.

A quick google search on USB shielding ... not.58811/

Re: S3 stops mid-job

Posted: Fri Aug 28, 2015 1:46 am
by apspurgeon
Just got home and double checked, SO3 0v (e.g negative of the 24v) is at electrical earth. As is my home PC.

What I believe the issue is, is that 0v and USB shield on the board while electrically connected, and earthed by the power supply (Neg/0v and Wall outlet earth connected) the way and routing on the board what connect 0v/Neg to Earth of the power supply is causing the issue. Earthing the USB shielding at the PC/Laptop end (already is by default on PC's - at least mine is, but not my laptop) means the electrical noise is taken to earth there rather than through the SO3 board. Anyway, my 2c, after the change I have had zero failures after repeated testing (touch wood)

Re: S3 stops mid-job

Posted: Fri Aug 28, 2015 2:20 am
by veng1
I agree there can be a problem with terminology. I use the convention that the "earth ground" is also the "safety ground". In the States, it is the round pin on the 3 pin outlet, connected typically by a green wire. By regulation, it is only allowed to connect to AC neutral for each circuit at the fuse/breaker box at a terminal bus bar that would go directly to a copper ground rod driven into the earth, hence the name.

Cross connecting earth ground and/or AC neutral between circuits is almost guaranteed to create ground loops and safety issues due to differential currents. Often one can measure quite a few volts difference between the supposed ground or neutral on two different circuits in the same room of the house.

I think the USB shield is connecting things in a manner that was not contemplated by the safety testing.

If I remember from your video. you used two alligator cables to reach from the laptop to the SO3. You might try to clip a small value capacitor between the two ends and see if it works just as well.

Re: S3 stops mid-job

Posted: Fri Aug 28, 2015 4:38 am
I have the first generation of DC power supply. The 24VDC ground is isolated from earth ground, this should be referenced as a floating ground or 24VDC negative.

When testing ground points or negative DC points on the CM board, it is isolated from the mounting holes, which should keep the board isolated from chassis ground in theory.

if you don't have a 3 prong plug for your power, you do not have a direct path to earth ground. If you ohm or continuity test the neutral or return to the ground plug, you will be ohming back to the bonding point at the service entrance and measuring the resistance thru the shortest path to ground in the overall circuit. In operation it is possible you may experience noise on the neutral line from other items sharing the path. In the U.S. Only one neutral bond to ground is allowed at the service entrance.

Re: S3 stops mid-job

Posted: Fri Aug 28, 2015 10:16 am
by cvoinescu
Even a floating supply has a low-impedance path, at high frequency, from the line conductors to the DC output, even if it's just parasitic capacitance in the transformer.

I have just received an ADUM4160 USB isolation board, so I'll be testing it over the next few days. Now I just need to install a garage door opener... :D

Re: S3 stops mid-job

Posted: Fri Aug 28, 2015 4:41 pm
by veng1
Your testing could go faster if you add a bigger garage with multiple door openers.

Better yet, a multi car drive through garage with door openers at both ends.