Electronics > Projects, Designs, and Technical Stuff
Oscilloscope ASIC Kickstarter instead of Open scope.
<< < (31/37) > >>
nctnico:

--- Quote from: excitedbox on January 14, 2020, 07:45:19 pm ---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.

--- End quote ---
Your analogy is quite good. With a Porsche you can go everywhere and take people and/or luggage with you. An F1 car in contrast does 1 thing great but sucks at everything else. Probably 90% of the oscilloscopes sold nowadays are FPGA based and for a good reason.
Kleinstein:
A DSO has different parts that are hard to integrate into a single ASIC as they need different processes: Input amplifier, ADC, Memory and control / logic. The Rigol MSO5000 has thus 2 types of custom chips.

For the ASIC to FPGA comparison one has to keep in mind that FPGA tend to use state of the art processes very early, as the structure is simple and just repeats a lot of times. An ASIC for the logic part of a scope on the other side is more complicated and usually would be used over a long time. So it would be more like a well matured process (e.g. some 5 years behind in development)  and thus lower performance transistors. So the advantage is not that clear and may only be for a short time. Modern FPGAs may also have optimized memory interfaces and may be ahead of an ASIC there, unless you could compete with the big guys.

There is quite some competition in the scope market. So it would be hard to do a new start here.
ogden:

--- Quote from: Kleinstein on January 14, 2020, 09:54:26 pm ---There is quite some competition in the scope market. So it would be hard to do a new start here.

--- End quote ---
Right. Manufacturing scope ASICs w/o any demand from market is risky business. Better drop idea of scope ASIC. If it is so easy to make scope ASIC then OP can make "just ADC" part of it, something like 1/2 cost HMCAD1520 and put Analog Devices out of business.
excitedbox:
If you read, triggering would also be on the ASIC so you wouldn´t miss anything. EVERYTHING to do with the signal is done on the ASIC. The only thing done on the CPU is display the data and interface same as every Oscilloscope. Why did I bother writing it and including pictures. If you can´t understand a basic sketch I really can´t help you

Why do the 2 of you insist on arguing about something you seem to not understand and have no interest in learning. This does EVERYTHING EXACTLY the same as an FPGA does concerning the signal. The ONLY thing changed is how you display the signal. Once the signal is captured to memory it doesn´t matter how slow you are in retrieving the data. I don´t know how many times I can explain the same thing to someone that doesn´t want to understand. The reason Oscilloscopes are made on FPGAs today is because of the cost of developing the chip. You have to invest money in developing a chip which is expensive and that is the only reason anyone uses them. It makes sense for small volume or if you don´t have the resources to finance developing a chip.

Imagine if in the early days of transistors people said everything is made with vacuum tubes today how can it be better to use a transistor. It is like the People don´t want broadband internet argument.

What kind of reality do you live in? Earlier you claimed digikey doesn´t have big margins on single quantities when you can see the price difference for 1 and 10,000 on many items and on something like a capacitor or usb connector you can see that buying 10k saves you 99% on some items. A resistor costs like 10 cents and 10k resistors it costs 1/10th of a cent. That is a 1000% difference. Of course they make huge margins on the parts because packaging on 1 or 100 costs the same and literally cost 10 times as much as the item.



Kleinstein You are right that some parts are hard to put on the same chip due to interference but that can and has been designed around and they can be moved off chip if need be. Those things that need to be close together for the benefit of saving money can mostly go on 1 chip.

The 130nm process is being used by CERN to make 20GSPS ADCs and is used on particle accelerators. an ASIC is not nearly as complex as an FPGA. An FPGA has to maintain signal quality and low noise while going all kinds of detours due to switching and cope with noise due to all those transistors. keep in mind silicon is not copper and every transistor is a jump from copper to silicon back to copper. Yes it conducts under the right conditions, but not nearly as well. Those FPGAs are not on the newest processes and in most cases an older process doesn´t make a difference like you think it does. It offers lower power consumption which you get much more of a boost from ASIC.

Just go read about this stuff you don´t need me to explain it to you. If you look at the link of the comparison I posted above you would see that your argument is not the case.
ogden:

--- Quote from: excitedbox on January 14, 2020, 10:17:51 pm ---Why do the 2 of you insist on arguing about something you seem to not understand and have no interest in learning.

--- End quote ---
LOL :) Year ago you were just PHP developer, now you **ck off professionals who have decade(s) of experience in electronics. You have much to learn kid.

* howz your PCB printer project going? No posts about it for some time.
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