200Mbps link between FPGAs

Check out Wikipedia, "Cables USB 3.1 cables are considered full-featured USB-C cables. They are electronically marked cables that contain a chip with an ID function"

So no buffering of the high speed data, but the chip is in there. Did you not read my other post?

Rick C wrote in news: snipped-for-privacy@googlegroups.com:

Perhaps, but what you said was "driver chips", which kind of implies buffering.

ID chips

Would be different.

Hey I know... Handshake chips with ID feature.

I was referring to the "slot" between the cabinet door and cabinet frame.

Two hinges on one side of the door and a single lock on the opposite aide does net sound as a proper RF tight space. At least have a few flexible jumper cables between the frame and door on the hinge side and three locks on the opposite side of he door helps a lot.

Microwave oven doors form a 1/4 wavelength trap at 2450 MHz, strongly attenuating any leakage. Unfortunately this is a narrow band solution only.

just

imulation package could probably give better information.

one close to it, he isn't being all that careful about his packaging.

'd expect the boards involved would all be plugged into a single rack, and that rack would be packaged up in a metal box which would look quite like a Faraday cage from an electrical point of view.

gh (if your boards run hot, as most of them do).

Of course you were. You even mentioned "Fingerstock between the frame and frame door" which kills that problem, when it actually is a problem.

I was talking about a single rack, not a cabinet full of them, but I'm happ y to admit that slots can be a problem, and note that there are solutions i f you are prepared to pay for them

I was talking about a single rack, not a stack of them.

Too true. But cheap.

snipped-for-privacy@downunder.com wrote in news: snipped-for-privacy@4ax.com:

EMI/RFI shielded racks are way more expensive than the standard gear mot get, and even those are not cheap.

But there is braided shield "wire rope" That can be put in as 'gaskets' if the contact surfaces are prepperd right, to seal up the biggest leaks in the cabinet.

Remember Class A and Class B personal computers?

And now you can go buy cases with plastic windows on plastic cabinets. Unreal how much less modern circuitry emits.

Good thing we got better, and tighter (and faster)than the old 5 volt TTL days.

I can still remember when EISA came out and they decided to keep it handcuffed at 4MHz, the Old ISA bus speed. FAIL.

The biggest change was going from two side printed circuit boards to four a nd six layer boards with buried ground planes - a whole lot of boards that had needed a grounded aluminium plate bolted on the track side to minimise board-to-board interference didn't need them any more if you buried a groun d a power plane inside the baord.

Even back then you could always use balanced ECL signals where you needed d igital logic that didn't upset nearby analog circuit elements. The fact tha t ECL was current steering logic rather than current switching logic made t he digital power rails a whole lot quieter, which also helped.

You can buy a lot of different c0nnectors labeled as XLR, but they aren't even similar to the real thing. Just more greedy idiots that hope you don't know the difference.

Most everybody calls them DB9 and everybody knows what they mean.

e close to it, he isn't being all that careful about his packaging.

expect the boards involved would all be plugged into a single rack, and th at rack would be packaged up in a metal box which would look quite like a F araday cage from an electrical point of view. You might need lots of small holes for cooling air to circulate through (if your boards run hot, as most of them do).

at

USB-C is very nice and gets you four high-speed twisted pairs plus a lower speed twisted pair and other conductors including 3A to 5A power. However, if you don't feel that it is robust enough then another option is passive copper SFP cable. The cables do not need to be expensive and they contain two twisted pairs, each rated at up to 10Gbit/s. They are incredibly robust, well shielded and lock in place. There is an eeprom at each end, but you can ignore it.

If you get the passive type there is a copper connection all the way through whereas the active ones have buffers in the connectors. There may be dc blocking capacitors - I would need to check the spec or measure one. Optical SFPs typically have 10nF DC blocks.

John

snipped-for-privacy@gmail.com wrote in news: snipped-for-privacy@googlegroups.com:

Now *THAT* is a solution.

My next post corrected that.

If your purchaser is trying to find it in the catalog,

then calling it DB9 is a dirty trick.

Our purchaser buys exactly the part numbers that Engineering specifies.