Your right I haven't, which 1 did you do? Has anyone ever done it with "super spice" ?
This thread has nothing to do with chip design.
No reason why to cant simulate a curcuit after you have it working if you needed that information but the components with the biggest variation are semiconductors and iv'e not seem a Monty Carlo test that included their variables but it's been a while so maybe.
Again this thread has nothing to do with chip design.
I'm not sure what the needs of the original poster were, but tools that might be applicable include: Gnucap, ngspice, tclspice, gnetlist, WinSpice, XSpice and MacSpice. I've got a page on my website where you can access details:
Neither breadboarding nor simulation is a safe substitute for design. I rarely simulate, and rarely breadboard, and then just a few parts tacked together. We think, design, lay out multilayer boards, let Production build a couple, and test them. Most are sellable first try.
We don't measure a product's performance so much as we predict it and sell it, then verify.
"Getting anything usefull from a simulator is usually much harder than building the circuit and mearuring its performance" ranks right up there with most stupid statement of the year (even after ignoring the illiterate spellings).
I didn't know there WAS an argument.
...Jim Thompson
--
| James E.Thompson, P.E. | mens |
| Analog Innovations, Inc. | et |
The facts are truly that 10,000s of analogue ics engineers simulate as the main bread and butter tool of design, for 40 hours a week and produce billions of chips. Its just the way it truly is. No amount of banter is going to change the facts that is done every f*&^%ing day of the week at companies such as linear technology, analogue devices, Texas Instruments, Intersil, On-semi conductor. The list is endless. Unfortunately, many just don't have the knowledge as to what is actually done at these companies, no matter how many times I try to point out the real, verifiable facts.
You are probably not designing complex transistor level circuits. Anything more than 2 transistors is not truly solvable by hand. I would like anyone do an accurate pole-zero calculation of even a 4 transistor circuit, and this don't even account for non-linear storage, or non-linear capacitor effects.
The reality is that despite having a good handle on the overal operation of complex circuits, without simulation, one knows nothing about how the circuit will really work in detail.
People need to get out of this, pen ana paper is the only way "real" engineers design. Like, consider solving the equations for colliding black holes. Like, you think people can do this by hand?
Most modern products now are too complex to do without extensive simulation.
Ahmm...
Kevin Aylward B.Sc. snipped-for-privacy@anasoft.co.uk
formatting link
SuperSpice, a very affordable Mixed-Mode Windows Simulator with Schematic Capture, Waveform Display, FFT's and Filter Design.
"There are none more ignorant and useless,than they that seek answers on their knees, with their eyes closed"
I am working on one right now. Quite impossible without simulation.
To be fair, SS is not really designed for circuits that large. It will simulate them, but its not really up to speed for that size of a job. However, it has been used extensively for ic building blocks.
For the above larger job I am using Cadence/Spectre.
Its to do with analogue design. IC design is being illustrated to show just how daft the idea that:
"getting anything usefull from a simulator is usually much harder han building the circuit and mearuring its performance."
is.
With all due respect to the writer, the writer of that statement has very limited knowledge of professional analogue design.
Again, with all due respect, all this shows is that you are quite ignorant of 10,000s professional analogue designers and the tools such designers use on a daily basis. Monty carlo and worst case simulations are de-facto standard to produce a reliable product in short time scales. Many tools have specific support for these features.
This debate is pretty much futile as its unlikely you will be able to take your blinkers off and accept what is actually known and done by
10,000s of professional analogue designers. Its 40 hours a week running sims. That's just the way it is, mate.
Its about analogue design, and as such, for anything other than a micky mouse 2 transister circuit, simulation tools are, in todays market place, indespensible.
Kevin Aylward B.Sc. snipped-for-privacy@anasoft.co.uk
formatting link
SuperSpice, a very affordable Mixed-Mode Windows Simulator with Schematic Capture, Waveform Display, FFT's and Filter Design.
"There are none more ignorant and useless,than they that seek answers on their knees, with their eyes closed"
I could show you a thousand. The last project I did, a temperature controller/pulsed gradient driver, had no sim or breadboarding. This week's gadget, a sub-ns singlemode laser driver, will have none either.
About the only time I ever breadboard is when a part is poorly characterized, and then I only test a single part or a very simple circuit involving the part, like when I TDRd a trimpot last week to see just how high a frequency signal can be trimmed with a pot. Ditto simulation, just little snippets, usually for nonlinear control loop tweaking, based on idealized component models. I never breadboard or simulate an entire product. I design maybe 10 products a year and fire up a simulator two or three times a year at most.
I do appreciate that it makes sense to simulate an IC design, but I don't do IC design. I do FPGA design, but don't simulate that either.
Really, designing is a good investment in time. That especially applies to software, where most programmers spend four times as long debugging ("breadboarding") as they do coding, and have the same problem as hardware breadboarders have, namely that ad-hoc debugging never finds all the bugs. My ratio is more like 4:1 in the other direction, not because I'm so smart, but because I *know* I'm not smart enough to speed-type reliable code without looking back.
I think that the worst part of design by simulation/breadboarding is the bad-habits-training loop that results.
I do note that you admit to regular failures. My rate of need for debug is under 1%... usually due to the customer's failure to spec what he really wanted.
...Jim Thompson
--
| James E.Thompson, P.E. | mens |
| Analog Innovations, Inc. | et |
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.