Re: TWGS vs. ZipChip

[foegelle@ccwf.cc.utexas.edu (Michael Foegelle)]

In article <1do0vfINN47t@gap.caltech.edu> toddpw@cco.caltech.edu (Todd P. Whitesel) writes:
>foegelle@ccwf.cc.utexas.edu (Michael Foegelle) writes:
>
>>I would be interested to see a GS tested for class B with a Zip running at 
>>about 15 MHz or so..  The fact that the low power Zips and RamFASTs flake
>>out sometimes could very well be due to them failing class A certification 
>
>The GS case is one large faraday cage, so I'm not sure how much this really
>matters. In the case of the Zip there is very strong evidence that the original
>two-layer board design simply could not filter out noise in the power rails --
>which low power CMOS is very sensitive to. My best info on the RAMfast so far
>is that the low power drivers simply can't compete with noise on the data bus
>generated by some peripherals and by the //e motherboard.

Well, we've already held our discussion in E-mail on most of this, but just
in case anyone else is interested, I'll re-iterate a few of the finer points.

As soon as you poke a hole in a case, whether to hang a cable out, or for
ventilation, or whatever, you no longer have a faraday cage.  And with a 
cable, you've got one whopping big antenna instead!  (Power cords, printer
cables, modem cables, monitor cables, joystick port, etc.)  Sheilded or not,
these things can radiate if there are alternate ground paths available.  And 
then there's MAGNETIC fields, which a simple conductive shielding doesn't
handle.  

There's also the consideration of these low power cards being susceptible to
external fields produced by the card next to them, etc.  Even though there 
may be no problem OUTSIDE the case, when you've got a card with wires running
every which way past another that is an inch away and might not be too clean, 
(noise-wise) you're likely to run into trouble.

>
>Your point about 15 mhz is a good one, though; I wouldn't be surprised if the
>old Zip design simply doesn't run once you get much higher than that. Hopefully
>the new four-layer board will be sufficiently quiet. Drew says his RAMfast D
>board has Zero noise on the power rails, it's the busses to the rest of the
>system that are the problem.
>
>Todd Whitesel
>toddpw @ cco.caltech.edu

Anyways, I think that covers most of the points we tossed back and forth well
enough for anyone else's edification...

Michael Foegelle

-- 
-------------------------------------------------------------------------------
   Michael Foegelle  |                    |   foegelle@ccwf.cc.utexas.edu
     ____________    |     You want it    |   foegelle@utaphy.ph.utexas.edu
                     |                    |   GEnie:   M.FOEGELLE2
    University of    |        WHEN?       |   Wunderland BBS  (512) 472-0544
   Texas at Austin   |                    |   14.4kbaud, v.32/bis:  Sysop