On Tuesday 16 May 2017 17:29:34 Bertho Stultiens wrote: > On 05/16/2017 09:56 PM, Gene Heskett wrote: > > Is this driver your work, Bertho? > > No, not at all. But, I have dealt with high frequency stuff before. > The biggest problem is that some parts are in the realm of magic ;-) > Not really, but cost is as always, the deciding factor.
> The RPI hardware is not made for such high frequencies. Pushing that > out of a china-model 100mil header is, generally, crazy and you have > impedance mismatches everywhere in the path. Yup. Try to show some bean counter just how bad they are on a decent TDR, is the biggest waste of a good engineers time there is. > > > My point is adding that 7 pf+whatever the tip clipon is, 3 or so > > more, to the pi src side of the termination resistor of 82 ohms, of > > this clock signal, makes it work, everytime. > > There are probably several effects at work here. You are changing the > impedance which probably changes the transmissionline properties as to > feed back into other parts of the circuit (including into the Broadcom > chip). That it works is amazing all by itself. I've had lots of > problems with 20MHz on that header already. Yeah it goes back to the first ide hard drives, whose cables weren't solid beyond a foot long, but they made them 16" long anyway. Or the perfectly good design of a scsi controller that somewhere between the engineers drafting board and the production line, had a power schotky diode switched out for a si type that was only 10% of the cost of the schotky(sp) by a bean counter who had no clue why the designer specced such a high priced isolation diode, who also could never understand why the switchboard lite up like a Christmas tree from people trying to make it work a full hour at a time in a busy graphics production. That bankrupted several back in the heyday of the Amiga. I could go on, but we've been observing things long enough to understand that even if the engineer drew it right, thats no guarantee it will still be right when shipped. The ideal situation in this instance would be to mirror the design of the 7i90 so as to rotate the 26 pin header 180 degrees, which would allow the cable to be only an inch long. Or mount the pi upside down, which would have the same effect. Humm, why not, I have some looong standoffs? While I'm waiting on 7i42TA's, why not try it? More than one way to skin this scrawny cat. Why is it I think best while writing about a problem? Can't be the age of the wet ram. Or is it... :( > > BTW, I think that not only the probe tip has influence, but also where > your probe's ground-lead is placed (and not that fluffy long > aligator-clip lead). > Chuckle, guilty yur honor, but its only a 4" lead. :) In this case clipped onto a corner of a usb sockets frame. > > I think I will solder a tightly > > twisted pair there for a loading cap, but that seems like such a > > kludge that I'd really like to be able to test and adjust for > > maximum data integrity. At some point I'll kill it again, but I > > might be able to arrive at a "this size of cap is best" by using > > half of what it takes to kill it again. > > That, of course, will work for this specific setup. However, changing > the RPI or changing any cable may throw you off again. Cheers, Gene Heskett -- "There are four boxes to be used in defense of liberty: soap, ballot, jury, and ammo. Please use in that order." -Ed Howdershelt (Author) Genes Web page <http://geneslinuxbox.net:6309/gene> ------------------------------------------------------------------------------ Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot _______________________________________________ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
