Electronics > Projects, Designs, and Technical Stuff
Oscilloscope ASIC Kickstarter instead of Open scope.
<< < (30/37) > >>
nctnico:

--- Quote from: Marco on January 14, 2020, 06:56:12 pm ---I don't see why you need field programmable logic, they are just a convenient way of connecting ADCs to memory when you use COTS ICs. For a digitizer ASIC just connect the ADC to the memory with fixed function logic, no FPGA needed ... just a license for a DDR3/4 macro or a couple thousand more expert man hours.

--- End quote ---
And then you find out that much of the functions you like to perform needs lots of processing grunt (*) which is typically where you need more dedicated logic (and development and testing) OR where a GPU comes in very handy. The GPU on an NXP iMX8M SoC for example provides 64Gflops of processing power. This likely allows to take over lots of jobs which are typically done in an FPGA / ASIC.

* This is a problem Rigol seems to be facing currently. Appearantly they left a few things out of their new ASIC which are now slowly done in software.
excitedbox:
You seem to be missing the point. You realize a scope doesn´t DISPLAY 300,000 waveforms a second. You have a 60hz LCD tops.

The ADC captures the signal and sends it to a DSP which processes the signal and puts it into memory. Where is there an ARM CPU in there?

The arm CPU just updates an LCD at 30FPS and renders some pretty touch screen controls. Arm SOCS can render millions of pixels a second so yes it will keep p with drawing a waveform

That is how it is done right now only that the display processing is on the FPGA

https://www.embedded.com/wp-content/uploads/media-1113967-2011-02-15-crh-agilent-infini-2000-3000-2.jpg

Or maybe you like how keysight does it better? ADC ASIC-> Memory Controller and another line form ADC ASIC -> TRIGGER ASIC which is NEXT TO an FPGA (Now another ASIC).

https://keysightoscilloscopes.wordpress.com/author/colinfmattson/
nctnico:
It is funny that you bring up Agilent/Keysight. It is the prime example of why you shouldn't go the ASIC route for a modern day oscilloscope unless you can't move the amount of data around using an FPGA. Keysight's Megazoom ASIC is what is holding them back for half a decade already. There isn't enough memory in the bloody thing and math functions are performed on sub-sampled data.
excitedbox:
They are using ASIC in ALL their scopes now and currently produce the worlds fasted scope at 110GHz and have a ADC ASIC with 80GSPS. light moves 14mm in that time giving you 7 linear mm on a chip to do your sampling.

An ASIC is the same as an FPGA only without the switching. If their megazoom wasnt fast enough either tech wasn't fast enough or they tried to save too much money. memory interfaces like ddr4 move 2 times the data half way across a motherboard.

Also you missed the difference in that picture to my plan. I want to put the DSP on the same chip you say the Megazoom is bad. Now look at that picture again. The DSP is so far away it isn´t on the picture.


Literally anything you can do on an FPGA can be done on an ASIC only faster because no switching. You have high up front costs but after the chip is done. EVERYTHING becomes cheaper, faster, easier, more efficient. That is the point of an ASIC it is a F1 car compared to a Porsche. Yes an FPGA has advantages but those are exactly the problems this project would be solving; the high up front investment cost.

FPGA VS ASIC

https://www.google.com/url?sa=i&url=http%3A%2F%2Fhardwarebee.com%2Ffpga-vs-asic-difference%2F&psig=AOvVaw2kEKdoYB5jPPmiDwweGmXi&ust=1579118057093000&source=images&cd=vfe&ved=0CAIQjRxqFwoTCKjBtanvg-cCFQAAAAAdAAAAABAS

https://www.google.com/search?q=difference+between+asic+and+fpga&client=firefox-b-d&sxsrf=ACYBGNQgmUKzHKCMb3Pw5XFuZh8ZiVgiqg:1579031653426&source=lnms&tbm=isch&sa=X&ved=2ahUKEwjdlaqm74PnAhWSZVAKHay0BcsQ_AUoAXoECAwQAw&biw=1920&bih=1005#imgrc=3ALqRl9BVok-jM
ogden:

--- Quote from: excitedbox on January 14, 2020, 07:38:15 pm ---You seem to be missing the point. You realize a scope doesn´t DISPLAY 300,000 waveforms a second. You have a 60hz LCD tops.

--- End quote ---
:palm: I did say "30000 waveform updates per second". It means 30000 waveform trigger/processing events. The faster update speed - the more chances you will not miss that glitch. It has nothing to do with LCD update speed which OBVIOUSLY is slower.  :palm: Kids...

[edit] Dave measure Rigol DS2202 waveform update rate (46KHz) here: https://youtu.be/_TSr9nFN1GU?t=563
Navigation
Message Index
Next page
Previous page
There was an error while thanking
Thanking...

Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod