HV dielectric solution

manufacturers)

(1Mbits/s)

so under half of what digikey wants for something similar

we normally use a bulkhead version so we can get it bolted to the front plate

I'm a bit reluctant to buy cheap sma stuff, I tried it once, it was very cheap but it was crap, center wasn't properly centered, isolation wasn't teflon so it tended to melt when soldered

-Lasse

Maybe they don't trust the Russkys :-) ... Because AFAICT they are the only ones making new ballast tubes, if they still do:

A HV rectifier, 1B3 types, make great regulators and amplifiers and controlled rectifiers, if you put the control signal into the filament voltage. Bandwidth is a tad low.

You're an idiot. I rigid one.

In showing us how dumb you can be.

OPTO-isolator.

Hard wired forms would not be easy.

"Rigidity"????? What is that?

Incwww.highlandtechnology.com jlarkin at highlandtechnology dot com

The earliest Theremins were controlled that way (using a contemporary filament triode). Clara Rockmore complained that it was still too slow, so electronic gain control was developed.

The 6BK4 is probably a lot faster than the photodiode stack -- I'm willing to bet it has far lower capacitance, and having more current, you can certainly slew it faster. It's not going to have that nasty tempco either -- apparently, above 100C the diode doesn't turn on any more than it leaks. Unfortunately they didn't specify 6BK4 and similar tubes for cutoff current; typical data only gives regulator service, which suggests plate current goes down to at least 45uA (out of a 1-1.5mA max), probably much lower in cutoff.

Tim

Interesting. I often need connectors without teflon, because teflon is awful when exposed to radioactivity. But I always have a world of trouble finding any, and they often end up being more expensive than the teflon stuff, despite the tendency to melt.

Jeroen Belleman

Why would anyone want to bring kilovolts to a PCB ?

My guess is that the simplest would be using multimode optical fiber between boards.

After all, due to the widespread use of Ethernet, such multimode fiber transceivers are quite cheap.

Be careful, teflon/conductor contact can exhibit triboelectric effects and make the noise floor of your cable higher than your electronics noise floor!

ability to resist something?, it doesn't take much imagination to guess that dielectric rigidity is probably what an english speaker would call dielectric strength

-Lasse

Le 18/05/2012 17:42, John Larkin a écrit :

Hello John !

Apart Joerg it's strange that the word "rigidity" is something not well understood in this group in fact of dielectrics, i apologies on my English syntax.

Not talking about the other "contributors" who think they have license to insult me.

Yes, but it is very important to distiguish between the occasional exposure and continuous exposure to a high voltage. You can have an isolation path that is designed to withstand several kilovolt but when testing it on a Hipot tester for too long of a time period the performance of that barrier can begin to deteriorate.

The syntax (arrangement of words) is not that hard to parse, it's the term "rigidity", which would normally be "dielectric strength" or "breakdown voltage".

Rigidity is a synonym for stiffness, so I thought maybe you're asking about the field dependency of the permittivity matrix, not just where it "breaks"!

Best regards, Spehro Pefhany

Yeah, good luck with that having zero experience.

Usenet attracts abusive people. In their real lives, they have probably discovered that people avoid them when they rant insults.

Do you have a PC board that has regions that are 10KV apart? Or do you want to send a signal between separated boards or assemblies that are

A light pipe with a LED on one end and a photodiode on the other would probably be the cheapest way to do this. Data could be async/uart, NRZ, or Manchester, any of the single-wire protocols.

There may be some 10KV optoisolators, maybe even ones that work at 1 MBPS, but as Joerg says, the insulation ratings are often short-term.

Fiber optics would take less mechanical design, but would be expensive in production. Maybe $10 or so for an LED-based link with cheap plastic fiber.

Hey, you could use one of the commercial right-angle light pipes that we discussed in another thread recently. The LED would be surface mount shooting up into the pipe; the other end could dump the light into a TO92 photodiode. Vaguely like this:

Or just a plastic rod with TO92 optos on both ends.

IOW, just like the local saloon. Generally a nice place but once in a while there will be a guy who wants to play bully, let off a major rant or is looking to pick a saloon fight :-)

What ever happened to TOSLink? I just took a look. Their parts are prohibitively expensive and, no surprise there, very low to zero stock levels. That would probably have been a good system for Habib to use.

In a saloon, the physical proximity inhibits most people who aren't really loaded.

HP/Agilent used to make cheap plastic fiber links, and had sample kits, TTL in and our. Avago probably still does them.

The optos+plastic rod could get close to $1 in production.