Your post and others in this discussion seem to indicate that production hi-pot testing is done with the AC/DC power supply connected to the mains. Is this the norm?
The hi-pot test with the P/S connected was done on the initial CE qualification tests (many years ago, as we have been using the same P/S for a long time), but the lab engineer/inspector indicated that we could disconnect the P/S prior to production hipot testing. This prety much turned the test into a chassis wiring safety verification as there are no mains EMI filters in this unit. There is a possible misinterpretation of the statement. Perhaps he was saying to disconnect the DC side of the P/S, versus the AC input.
This was indeed a DOS-based system. We were using a PC-104 embedded CPU running custom C code. PC-104 when it first came out was the answer to our prayers. It made for a robust system in a small common form-factor with some real processing power.
Then it came time to add USB to our system to enable the use of a flash disk (thumb drive) for data extraction and program upload. We tried the Linux path, but could not find the right person / integrator to pare down the multi-media centric kernel to boot fast enough for our embedded control system. We did not want to go down the Windows path for a critical control system.
We finally ended up doing a dedicated control system based on a PIC microprocessor that has worked very well. No OS royalties or misbehavior.
Sorry, I missed your question until today. The test involved was a hipot on a medical device. Every circuit added its own leakage, and in some cases the "medical grade" power supplies used up all of the leakage we were allowed.
AC hipot leakage is usually negligible unless EMC capacitors are large & then leakage can become so big that the hipot shows a fail & you have to use DC hipot source. Chassis leakage through ground lead at mains voltage is also usually very small, but EMC caps can cause it to be large. Most ordinary eauipment is allowed up to 3.5 mA leakage through 1500 ohms (representing human body resistance). Medical and certain other types of equipment have much lower allowed leakage currents, and often lower body resistance. Safety standards for ordinary use typically fail any one component (medical typically fails two components) and leakage must remain in limits.
That was a standard joke of the BIOS on all DOS era PC's - if you disconnected the keyboard, at startup you would get the error message - "No keyboard found - press F1 to continue". So now you know where the embedded software came from.
It is very easy to make mistakes in respect of safety 'grounds' and 'signal' grounds. Such mistakes are common and reinforced by poor advice in EMC and safety texts, which each ignore the problems of the other.
In the UK the electrical supply industry now expects to provide a safety conductor (CPC - circuit protective conductor) and treats 'ground' as a source of danger (faults attack you from there!). Fault currents used to be directed toward 'ground', now they (one's own faults) are directed at the CPC to be terurned safely to the source, and faults from elsewhere may come from 'gound' (e.g. hidden faults in a wall space).
Likewise keeping signals clean requires careful thought about return paths, and about safety under fault conditions. Bill Whitlock (Jensen Transformers, and commentator here) has some great articles on his web site (http://www.jensen-transformers.com/apps_wp.html AN003, AN004), and there is a good video of his UK lecture to the IET http://tv.theiet.org/technology/electronics/1048.cfm on the same issue. His techniques (aka Physics rules) work at all frequencies, as I comment in the video. For those working at higher frequencies Howard Johnson's articles are a good source.
Yes, all the DC and signal grounds inside the cabinets of the cameras are all tied to the chassis, and the AC power entry ground tied to chassis at the entrance point. Other pieces of equipment (typically not under our control or made by us) in a system were pretty much on their own, but the commercial stuff all was bolted into metal rack and run from a common AC power strip. Medical cameras are much the same inside, but generally reside as independent items in a surgical equipment cart. Though the cameras have what's called an "equipotential post" on the rear of the chassis, it's not always used if the camera isn't involved in open-heart surgery or the like. However, electrical leakage of all sorts is VERY carefully regulated and tested, and must allow for single ground faults. For everyday use, the AC power ground is what ties the different units together.
Video signals were and are always the most critical in the camera systems. AC noise from the power input has never been even a tertiary issue. Of course we have to pass conducted emissions and susceptibility tests, but I don't recall that ever being a problem with commercial power supplies and power-entry filters.
Now we enter the murky world of EMI and EMC. But first, the definitions have to be clarified. Are we again speaking of separate AC grounds, separate DC commons, or separation between AC ground and DC common? I think there is agreement that DC commons are best tied together, and that AC grounds should be tied to together. Whether the two share common earth ground is a gray area.
In order to control EMI and provide the AC (earth) ground, we tie all of the chassis grounds together with dedicated physical wires and terminals. Some people refer to this as bonding, which in a fixed equipmemnt installation actually refers to strapping the equipment enclosure to the AC ground through a ground strap or bus in addition to the ground provided through the power cabling. This is of primary importance on ships to pass EMC criteria and prevent EW (Electronic Warfare) snooping.
I wonder if in your TV camera experience, there were more high frequency (video / RF) signals and their issues with which to deal than simple AC power noise and spikes. The military version of our equipment is housed in a metal case to provide the best chance of immunity to external RF radiation and to contain any internal radiation from the higher frequency processor clocks. Think Faraday cage. I do not know if tying DC common into this system would actually improve or degrade the immunity and radiated emissions levels. This is one for the EMC guys.
I am curious as to what type of leakage test you are referring. Is this a chassis Hi-pot test for AC leakage to geround or something performed at a lower voltage? We do a Hi-pot from the AC mains connection to the metal chassis, but disconnect the AC/DC power supply to prevent damage to the power supply from the 2200VAC test voltage. The test is performed to meet CE Low Voltage Directive safety requirements. We trust the CE mark by the manufacturer on the power supply is sufficient to exclude the power supply itself from the high-voltage stress.
Festo's BionicKangaroo combines pneumatic and electrical drive technology, plus very precise controls and condition monitoring. Like a real kangaroo, the BionicKangaroo robot harvests the kinetic energy of each takeoff and immediately uses it to power the next jump.
Design News and Digi-Key presents: Creating & Testing Your First RTOS Application Using MQX, a crash course that will look at defining a project, selecting a target processor, blocking code, defining tasks, completing code, and debugging.
Focus on Fundamentals consists of 45-minute on-line classes that cover a host of technologies. You learn without leaving the comfort of your desk. All classes are taught by subject-matter experts and all are archived. So if you can't attend live, attend at your convenience.