Circuit analyzer (computer based)

[burningman]

Does anyone know of freeware/shareware that can analyze circuits?

[isildur100]

Hi,

I'm not sure if you mean a circuit simulator, but you can try LT Spice, which is free and works great.

http://www.linear.com/designtools/software

[CynicalMan]

Hi,

I'm not sure if you mean a circuit simulator, but you can try LT Spice, which is free and works great.

http://www.linear.com/designtools/software

+1 for LTSpice

Or if you mean software to analyze physical circuits:
http://www.sillanumsoft.org/prod01.htm

[burningman]

I'm not even sure if what I am talking about exists. Essentially, what I am looking for is a simulator/analyzer that will take a circuit diagram and allow you to verify current flow, check voltage values at points etc. I'm sure it could all be done in the real world with breadboarding, but I'd like to be able to do this faster on the computer.
Thanks.

[CynicalMan]

I'm not even sure if what I am talking about exists. Essentially, what I am looking for is a simulator/analyzer that will take a circuit diagram and allow you to verify current flow, check voltage values at points etc. I'm sure it could all be done in the real world with breadboarding, but I'd like to be able to do this faster on the computer.
Thanks.

You want a circuit simulator then, like LTSpice.

[burningman]

excellent thanks!

[PRR]

> I'm not even sure if what I am talking about exists. Essentially, what I am looking for is a simulator/analyzer that will take a circuit diagram and allow you to verify current flow, check voltage values at points etc. I'm sure it could all be done in the real world with breadboarding, but I'd like to be able to do this faster on the computer.

SPICE will take a net-list and compute stuff about it. A number of commercial packages allow you to (tediously) draw-out a schematic, compile it to a netlist, SPICE it, and run PROBE on the output.

I used PSpice since days of DOS and 8088 CPUs (I had to get an 8087 math-chip; even then a one-transistor plan could take 20 minutes for a basic bias-check.) The oldest versions required you to draw (on paper) your circuit and hand-type the netlist (what connects to what). The 1997 version has an OK graphical schematic editor. Here's a handy example:



It can display voltages (or currents) on the schematic, as shown here (with stupid "precision").

Newer packages have a few more frills. A full-blown commercial package is $2K-$8K per seat, there are several good packages around $300, but this LTspice works for free and seems to be popular among audio hobbyists.

BUT...

Computers are stupid. They don't know how things are usually connected. They don't know what you want to do. They happily give you the right answer to the wrong question. And circuit analysis is multiple-guess: SPICE is very prone to get jammed-up on some small detail and bomb.

I believe you should be able to analyze most audio circuits on the back of an envelope. It's mostly about knowing V/I=R and some notion how splits work. Knowing how tubes/transistors work is a bit harder, but part of the craft. In the plan above, I'd already guessed each triode would act like a resistor much higher than the data-sheet's 7K, but maybe under 50K. Assuming 30K, then a happy plate load might be 30K to 60K, and the plate voltage would be like 1/2 to 1/3rd of the supply voltage, 3V to 2V. I put 47K on the plan, SPICE chewed a while and declared "2.499V". However in a slightly different plan, SPICE can get some utterly wrong answers. So I always hand-figure a 1-digit precision answer first, to know when SPICE has come up with a very good looking wrong answer.

Another point: the above plan runs a tube on 6V. SPICE does not know any actual devices, only "models", a list of numbers that someone collected about one particular device. Tube data is collected for 50V to 500V, cuz' we usually have to run them at 100V-300V to get useful power output. Can we trust the model at "2.499V"? Experience counts. I know from lots of mucking-around that tubes conduct at low voltage, but poorly. And I suspect more poorly than this particular 12AU7 model predicts. I'd bet the plate is really closer to 3V. And that the sine output won't reach 1.2V peak. But to know if either SPICE or I are even close, we MUST build it!

SPICE (most packages) will never object to a teeny TO92 transistor carrying 400V at 1 Ampere. In real life, the smoke gets out. SPICE -can- calculate dissipation, and in some non-trivial systems it does a much better job than hand analysis. But you have to "ask", and the question must be phrased very carefully. So you still need some seat-of-pants understanding about which parts can't get hot, and which parts need to be checked.

And like email or online poker, it is very easy to get lost for hours in SPICE and come out with little to show. Pencil and envelope sharpens your mind and builds intuition. Breadboarding gives you actual real-world results.

> do this faster on the computer.

SPICE was invented to predict MOS chips. When it takes weeks to have your idea fabricated, and there is strong reason to get the fastest thriftiest design on the least silicon, you really need something like SPICE. Can you shave 13 micros off your Gate and still meet the clock-rate requirement? Will it still work in arctic cold and deep-well heat? When the supply wanders from 5.9V to 4.3V? Most audio design does NOT have to be so sharp-pencil.

I think using SPICE too soon cripples your design skills. However I am an old grump, out of touch with current trends.

[burningman]

Thank for your reply PRR. I am looking for help from a simulator more to clarify questions that I have. As I am learning more I'll probably just use it as a reference. Cheers.

[T1bbles]

Hey guys, just drawn up a schem on LTspice, and it runs a simulation, but how do I tell LTspice where to take readings from? Is there a component for that?

[JKowalski]

When you have a transient analysis up and running, you should be able to click on all the different nodes on your circuit (it will show a little scope probe symbol when you can) and that will display the transient analysis (referenced to groiund) up on your scope window. If you click on one node and drag the scope symbol to another node, it will come up with a transient analysis of the voltage difference between the two nodes you chose.

[puretube]

> I'm not even sure if what I am talking about exists. Essentially, what I am looking for is a simulator/analyzer that will take a circuit diagram and allow you to verify current flow, check voltage values at points etc. I'm sure it could all be done in the real world with breadboarding, but I'd like to be able to do this faster on the computer.

SPICE will take a net-list and compute stuff about it. A number of commercial packages allow you to (tediously) draw-out a schematic, compile it to a netlist, SPICE it, and run PROBE on the output.

I used PSpice since days of DOS and 8088 CPUs (I had to get an 8087 math-chip; even then a one-transistor plan could take 20 minutes for a basic bias-check.) The oldest versions required you to draw (on paper) your circuit and hand-type the netlist (what connects to what). The 1997 version has an OK graphical schematic editor. Here's a handy example:



It can display voltages (or currents) on the schematic, as shown here (with stupid "precision").

Newer packages have a few more frills. A full-blown commercial package is $2K-$8K per seat, there are several good packages around $300, but this LTspice works for free and seems to be popular among audio hobbyists.

BUT...

Computers are stupid. They don't know how things are usually connected. They don't know what you want to do. They happily give you the right answer to the wrong question. And circuit analysis is multiple-guess: SPICE is very prone to get jammed-up on some small detail and bomb.

I believe you should be able to analyze most audio circuits on the back of an envelope. It's mostly about knowing V/I=R and some notion how splits work. Knowing how tubes/transistors work is a bit harder, but part of the craft. In the plan above, I'd already guessed each triode would act like a resistor much higher than the data-sheet's 7K, but maybe under 50K. Assuming 30K, then a happy plate load might be 30K to 60K, and the plate voltage would be like 1/2 to 1/3rd of the supply voltage, 3V to 2V. I put 47K on the plan, SPICE chewed a while and declared "2.499V". However in a slightly different plan, SPICE can get some utterly wrong answers. So I always hand-figure a 1-digit precision answer first, to know when SPICE has come up with a very good looking wrong answer.

Another point: the above plan runs a tube on 6V. SPICE does not know any actual devices, only "models", a list of numbers that someone collected about one particular device. Tube data is collected for 50V to 500V, cuz' we usually have to run them at 100V-300V to get useful power output. Can we trust the model at "2.499V"? Experience counts. I know from lots of mucking-around that tubes conduct at low voltage, but poorly. And I suspect more poorly than this particular 12AU7 model predicts. I'd bet the plate is really closer to 3V. And that the sine output won't reach 1.2V peak. But to know if either SPICE or I are even close, we MUST build it!

SPICE (most packages) will never object to a teeny TO92 transistor carrying 400V at 1 Ampere. In real life, the smoke gets out. SPICE -can- calculate dissipation, and in some non-trivial systems it does a much better job than hand analysis. But you have to "ask", and the question must be phrased very carefully. So you still need some seat-of-pants understanding about which parts can't get hot, and which parts need to be checked.

And like email or online poker, it is very easy to get lost for hours in SPICE and come out with little to show. Pencil and envelope sharpens your mind and builds intuition. Breadboarding gives you actual real-world results.

> do this faster on the computer.

SPICE was invented to predict MOS chips. When it takes weeks to have your idea fabricated, and there is strong reason to get the fastest thriftiest design on the least silicon, you really need something like SPICE. Can you shave 13 micros off your Gate and still meet the clock-rate requirement? Will it still work in arctic cold and deep-well heat? When the supply wanders from 5.9V to 4.3V? Most audio design does NOT have to be so sharp-pencil.

I think using SPICE too soon cripples your design skills. However I am an old grump, out of touch with current trends.

hehe: for fun`s sake, try two 2N3819`s in that circuit instead of the triodes...   

[edit]: and UP the battery to 300V...

[igor12]

Use MULTISIM, you will never go back to anything else.  Real time Oscilloscope and pots you can turn in real time. That is powerfull. PM me if you want a "copy".  You can download a free demo. Google Multisim.