On 5 Apr 2007 at 1:09, Jay West wrote:
Alison wrote....
You have to be kidding me. I have that
arrangement and it's solid. The
RX02
bottom is well shielded as it the top of the RL02 logic.
Not only am I not kidding, but I can PROVE that RFI/EMI is the problem.
I wonder if it's more EMI than RFI, because it reminds me of a nasty
problem we had when developing the Durango F85.
If you look at the machine, there's a 9-inch CRT placed above the
dual floppy drive. Before we actually had a production model, we
worked from a lash-up with a keyboard connected to a card cage
connected to the PSU, monitor and diskette drives--all laid out on a
benchtop in traditional develpment fashion. It all worked fine.
Then came the first production prototypes. Naturally, the initial
ones were given to the marketing people to take pretty pictures of
and show around at shows and to prospective customers. The software
folk were the last to get a real one. Except for one problem...
The diskette drives refused to work on the production model.
Grounding braid was added, power supply connections were re-routed
and everything else was tried. Nothing worked--but the laid-out-on-
the bench lash-up worked perfectly, even when the production cards
and drives were substituted. But the pretty production models
refused to even boot.
The engineers were baffled--until someone took the monitor off of the
top of the diskette drive area to attach a scope probe--and the
problem went away. It turns out that the FBT was kicking up noise
something fierce in the read channels of the diskette drives. An
shield made of sheet aluminum didn't get rid of the problem, but one
made of steel worked just fine.
We also added a shield around the paper feed stepper--it seems that
when a line- or form-feed was performed, the CRT image shuddered
something awful.
Cheers,
Chuck