linear technology spice cad

Status
Not open for further replies.
Jump to Latest
#2
.
 

Attachments

  • npnamp.jpg
    npnamp.jpg
    62.3 KB · Views: 622
#5
Start with a simple transistor stage.

Get SPICE models for all parts if you not have them already

Try to get a frequency response

Try to get a step response and sinus response

(Try to get a FFT chart)

When you know this, you can draw bigger circuits but you must test something small first. Note also that big simulations demands a powerful PC. I have a P4 1.8 GHz and a bit faster wouldn't hurt.
 
#6
I also am a beginner with LT Spice and found some strange behavior with a BD139 Model from the OnSemi website.
There you can download Spice2 Model, Spice3 Model, PSpice Model and Saber Model. I assumed that for LT Spice, the Pspice Model would work. But in my circuit, the B-E voltage drop is around 0.3 - 0.35 Volts😕 . So I don’t trust the simulation.
Do someone have any advise?

Thanks, Zinsula
 
#7
hjelm said:
Isn't one of the q5 transistors backwards.

What is the CF there for and why is it only on one of the q5's.
Click to expand...


It is not backwards. It is a quasi-complementary output
stage.

It looks like one of the designs amp_man has posted on
the forum, so the design has been discussed quite extensively
somewhere on the forum.
 
#8
zinsula said:
I also am a beginner with LT Spice and found some strange behavior with a BD139 Model from the OnSemi website.
There you can download Spice2 Model, Spice3 Model, PSpice Model and Saber Model. I assumed that for LT Spice, the Pspice Model would work. But in my circuit, the B-E voltage drop is around 0.3 - 0.35 Volts😕 . So I don’t trust the simulation.
Do someone have any advise?

Thanks, Zinsula
Click to expand...

Try the Spice3 model perhaps? It is a universal problem that
it is difficult to find good transistor models for Spice. The best
is to plot they usual graphs and compare to the datasheet.
Also try different models if something seems wrong.
 
#9
Are you sure that the circuit is OK? That low Vbe may be the result of a circuit mistake, not the model's fault. These are the basics, and I seriously doubt that the model would be that far off. In the circuit, is there any Ie? What are the node voltages for the transistor?

Jan Didden
 
#10
the .op ( operating point ) command is really the first analysis to option to use, if you don't like reading the output text file just close the text window that pops up after the sim and cruise the schematic with the cursor, the values of node voltages and device currents appear in the lower left corner

then all of your "debugging" should be done with test sine and pulse inputs in .tran ( transient ) analysis mode

DON'T use .AC until the circuit is really working or you will get thourghly confused .AC doesn't simulate your circuit, it assumes it works and linearizes the model around the dc operating point - if Qs are reverse biased, stages clipped, ect you will have a hard time finding it with .AC analysis

many basic spice intro books exist and give you an idea where to look in Lt SwCad's really excellent help files
 
#12
Originally posted by jcx
the .op ( operating point ) command is really the first analysis to option to use, if you don't like reading the output text file just close the text window that pops up after the sim and cruise the schematic with the cursor, the values of node voltages and device currents appear in the lower left corner
Click to expand...

Thank you for your help. This is exactly what I have done first, performed the dc operating point analysis.


Originally posted by janneman
Are you sure that the circuit is OK? That low Vbe may be the result of a circuit mistake, not the model's fault. These are the basics, and I seriously doubt that the model would be that far off. In the circuit, is there any Ie? What are the node voltages for the transistor?
Click to expand...

The circuit was OK with default transistors from LT Spice. Ie is around 30mA, as far as i recall. But I have to check that tomorrow evening since I am not at home currently.

I will check the transistor model with standard graphs as suggested by Christer (hopefully I would be able to do that, but seems the only way to learn something).

Thanks to all of you for your help, Zinsula
 
#14
wonder_b said:
how do I put new transistors into lt spice it doesnt have the mj15024's and the amp (from amp man) sim looks to be working on defult transistors but I have been told the amp wont work and I dont realy know what I'm doing can anyone help
Click to expand...

For components where you can use standard models you do
it in the /lib/cmp/ directory. There is one file for each component
type. For instance BJTs are in standard.bjt and diodes are in
standard.dio Use a text editor, Notepad if you don't have any
better favourite, and just add the model for the new component
as one single line. It doesn't matter where in the file you put it.
For instance, I have reorderd all components so the ones I use
most appear first in the file, and thus appear first in the list of
choices.

For more complex models, where subcircuits are necessary, it
is a bit more complicated. Search the forum and you will find
it somewhere.
 
#15
Question to you high-power sim users:

I try to use a chip in floating mode. The model is fine, works in the 'normal' way, but when I float it there is a problem, because the internal .subckt file has a node 0, which is universally used for ground. So, the circuit didnt work with the chip floating but internally connected to gnd.'

I modified the internal node 0 to 200 (which was not used) and brought the pin outside in the model, and connected it to the reference in the floating circuit. Still no joy. What am I doing wrong?

Jan Didden
 
#16
Not a high power sim user but i found this on converging simulations in the Winspice manual
"Check for a DC path to ground from every node"
and
"Have ground (node 0) somewhere in the circuit. Be careful when using floating grounds; a large valued resistor connected from the floating point to ground may be needed"

Do not know if it helps.
 
#18
I am a beginner with Spice and am having trouble simulating even a basic voltage source.

I cant seem to get a voltage source to simulate and get a graph. I then tried editing an existing schematic of the audio amp .asc circuit and only left the voltage sine generator but cant seem to get a simulation and waveform of that either.

Can anyone help me understand my error?

Craig.
 
#19
Craig;

spice circuits must have a ground node and must be a "circuit" - ie have at least 1 loop which current flows through

to see a vsource waveform you need to ground one end of the vsource and connect a resistor (with a value: 1 Ohm is ok) in parallel with the voltage source

Jan;

look for 0 node in poly source arg lists too

i've used the ti/bb opa227m model (be sure to get the "m" sufix for the multiple pole macromodel, not the Boyle model)

i've also "built" my own floating "op amp" with dependent sources and R-C networks
 
Status
Not open for further replies.
Top Bottom