Personally I'd use the XMOS xCORE
devices, with code written in xC.
?- rock solid *hard* realtime guarantees: the IDE specifies the
*exact*
?? number of clock cycles the code will take. None of this
"measure and
?? hope we've found the worst case" rubbish :)
?- FPGA-like I/O structures: 250mb/s per pin, SERDES, strobed, ,
etc 1-32
?? bit, I/O happens on specified clock cycle
?- up to 32 cores, 4000MIPS MCUs
?- *fun*, in a way that conventional MCUs aren't
?- easy: I had my first program executing correctly within half a
day of
?? downloading the IDE
?- no errata that I've seen
?- mercifully brief and explicit documentation (because the
abstractions
?? are good and well implemented)
?- buy them at Digikey
?- been available for 13 years, actively being developed and
enhanced?
And most notably, the hardware, language and software tools were
designed together as a whole. There are dozens of multicore
designs, but the software is an afterthought - with little thought
given to the hard real time and parallel processing requirements.
The XMOS team uses parallel processing concepts that have around
for decades: Hoare's Communicating Sequential Processes (CSP), the
Transputer, and xC is an updated Occam with additions for hard
realtime I/O.
An example of what one medium-sized MCU can do in software:
?- input two 62.5Mb/s data streams and count the 0->1
transitions,
?- front panel i/o and control
?- communicate control and results over USB to a PC
guaranteed not to miss any edges.
Other example is generating/consuming 100Mb/s serial ethernet
packets in software (but personally I think it better to use an
internal ethernet peripheral)
A four page overview is at
Overview of the language and how to use it:
That's longer, but is beautifully written and pleasingly concise.
On 20/04/20 03:10, Kuba Ober wrote:
toggle quoted message
Show quoted text
I imagine that pretty much all legacy HP displays
can be emulated using a single chip and a few screenfulls of
code, outputting straight to VGA or HDMI :) Parallax Propeller 2
is the magic sauce that makes it possible. It¡¯s quite an amazing
design, done in the open (yeah, open source silicon with full
commercial backing) and it¡¯s actually fun to program it. It was
meant to ¡°mess with¡± - it¡¯s not like the usual ARM chips with a
boatload of erratas and data sheets thousand pages long. You can
do very meaningful things without leaving assembler, if you like
that sort of a thing. Want to sample XY signals at 100Ms/s? You
can. On each pair of I/O pins. Simultaneously :)
They are currently offering 2nd silicon revision
engineering samples, and you can actually talk to the designer
of the chip on the forums. Those samples work very well, and the
¡°errata¡± fits in a paragraph last I checked. I imagine that the
production silicon won¡¯t have an errata. There¡¯s a whole bunch
of people that used that design before it was available in
silicon, so lots of kinks got straightened out long ago.
Also, it comes with a FORTH interpreter in ROM,
with runtime library that actually supports the peripherals.
What¡¯s not to love :)
Cheers, Kuba Ober
? Stay OFF the Danaher 57x series. My replacement
for these is almost ready.
As for the 53310A: not sure if it pays. You would need to use
an FPGA to harvest the digital data being farted out. I wanted
to do it once, but never had the time. However, afaik, the
video board on these units is a separate board (as per the
late Jzon Geller), and some instructions in the service manual
of the 53310A even give some data about the pinout.
Tam
With best regards
Tam HANNA (emailing on a BlackBerry PRIV)
Enjoy electronics? Join 14k other followers by visiting the
Crazy Electronics Lab at
Am 16. April 2020 19:56:40 MESZ
schrieb Bill E
<solartron@...>:
That's something I've been thinking about for a while now.
I have several instruments where I'd like to replace the
CRT. I know there have been a few people in the past that
have done some for specific instruments, but I don't know
of a general solution. One would think that for many
instruments, should be fairly easy. They usually used OEM
CRT displays that had simple analog or pseudo-digital
intensity signals, and standard or close-to-standard scan
rates. But, for the older style oscopes where it was a
classic deflection CRT with the text additions being done
by vector graphics, whole different game.My 54540 really
needs a new CRT, modern replacement would rock. I'll add
it to my list. :)
_._
