Designer’s Guide Community

hzfeiyun

Junior Member

Offline

Posts: 16

can i use .LOAD to set all operating points in ac analysis?
Dec 24^th, 2008, 10:42pm

to do the ac analysis of a open-loop circuit, usually i need use ac/dc voltage source to set some open nodes and the ac input. the dc values are the results from close-loop circuit. but, for all corner simulation, it will become some difficult. is there any way to make it easier?

i tried to use ".load" to load dc values from the dc results of a close-loop circuit, but failed. i don't know why

please help. thx.

Back to top

IP Logged

Andrew Beckett

Senior Fellow

Offline

Life, don't talk to
me about Life...

Posts: 1742
Bracknell, UK

Re: can i use .LOAD to set all operating points in ac analysis?
Reply #1 - Jan 2^nd, 2009, 7:43am

Often you want to do this in order to find the loop gain, so better is to use spectre's "stb" (stability) analysis. This allows the loop to remain closed, and consequently avoids the problem with having to break the loop between analyses.

If you're doing this for other reasons, spectre has a "switch" element (see "spectre -h switch" - or sp1tswitch/sp2tswitch etc from analogLib if using ADE) which is an ideal switch which can be in different positions in different analyses.

Other simulators (e.g. hspice) have an "AC" resistance - you can have a different value of a resistor during the DC and AC analyses.

But stb analysis is generally the right way to do this.

Best Regards,

Andrew.

Back to top

IP Logged

hzfeiyun

Junior Member

Offline

Posts: 16

Re: can i use .LOAD to set all operating points in ac analysis?
Reply #2 - Jan 5^th, 2009, 1:25am

Thanks for your points. Could u express more about AC resistor in HSPICE simulation? Do u mean a ideal inductor? Wink

Andrew Beckett wrote on Jan 2^nd, 2009, 7:43am:

Often you want to do this in order to find the loop gain, so better is to use spectre's "stb" (stability) analysis. This allows the loop to remain closed, and consequently avoids the problem with having to break the loop between analyses.

If you're doing this for other reasons, spectre has a "switch" element (see "spectre -h switch" - or sp1tswitch/sp2tswitch etc from analogLib if using ADE) which is an ideal switch which can be in different positions in different analyses.

Other simulators (e.g. hspice) have an "AC" resistance - you can have a different value of a resistor during the DC and AC analyses.

But stb analysis is generally the right way to do this.

Best Regards,

Andrew.

Back to top

IP Logged

Andrew Beckett

Senior Fellow

Offline

Life, don't talk to
me about Life...

Posts: 1742
Bracknell, UK

Re: can i use .LOAD to set all operating points in ac analysis?
Reply #3 - Jan 5^th, 2009, 3:26am

I may come across as an old fuddy-duddy, but the English word is "you" not "u". It makes it very hard for some of us to read if you abbreviate like this...

Anyway, this is covered in the HSPICE manual. It is not an ideal inductor, but a different resistor value for DC and AC analyses. Here's an example from the manual:

Quote:

The Rxxx resistor, from node 98999999 to node 87654321, with a resistance of
1 ohm for DC and time-domain analyses, and 10 gigohms for AC analyses.
Rxxx 98999999 87654321 1 AC=1e10

Regards,

Andrew.

Back to top

IP Logged

hzfeiyun

Junior Member

Offline

Posts: 16

Re: can i use .LOAD to set all operating points in ac analysis?
Reply #4 - Jan 7^th, 2009, 10:34pm

Thank you. I just got it in HSPICE manual.

Andrew Beckett wrote on Jan 5^th, 2009, 3:26am:

I may come across as an old fuddy-duddy, but the English word is "you" not "u". It makes it very hard for some of us to read if you abbreviate like this...

Anyway, this is covered in the HSPICE manual. It is not an ideal inductor, but a different resistor value for DC and AC analyses. Here's an example from the manual:

Quote:

The Rxxx resistor, from node 98999999 to node 87654321, with a resistance of
1 ohm for DC and time-domain analyses, and 10 gigohms for AC analyses.
Rxxx 98999999 87654321 1 AC=1e10

Regards,

Andrew.

Back to top

IP Logged

hzfeiyun Junior Member Offline Posts: 16	Re: can i use .LOAD to set all operating points in ac analysis? Reply #5 - Jan 13^th, 2009, 2:41am but, how to set ac source without affecting .op results in HSPICE. If I set v1 1 0 ac=1, it means set dc v(1) to 0v during .op analysis.
Back to top	IP Logged

Andrew Beckett

Senior Fellow

Offline

Life, don't talk to
me about Life...

Posts: 1742
Bracknell, UK

Re: can i use .LOAD to set all operating points in ac analysis?
Reply #6 - Jan 13^th, 2009, 3:43am

The ac value of the source does not affect the DC operating point. The DC value (0) does of course affect the DC operating point.

What's the problem? You can specify the DC value of the source, and the small-signal magnitude of the source for AC analysis. I don't think I understand what you're asking...

Regards,

Andrew.

Back to top

IP Logged

hzfeiyun

Junior Member

Offline

Posts: 16

Re: can i use .LOAD to set all operating points in ac analysis?
Reply #7 - Jan 14^th, 2009, 2:22am

Does "v1 node1 node0 ac=1" mean setting dc value of v1 to 0v as default?

I am trying to get the dc value of node1 from .op analysis of a close-loop circuit and automatically apply this dc value to this node1 during ac analysis of the open-loop circuit, for example, a simple two-stage op-amp.

Back to top

IP Logged

Andrew Beckett

Senior Fellow

Offline

Life, don't talk to
me about Life...

Posts: 1742
Bracknell, UK

Re: can i use .LOAD to set all operating points in ac analysis?
Reply #8 - Jan 14^th, 2009, 3:07am

The DC operating point around which the AC analysis linearises the equations will have 0V between node1 and node0 with your example source. Then, the response of these linearised equations to a 1V (the AC source) stimulus across node1 and node0 will be computed, and you can then see the results of the transfer functions of these linearised equations to other nodes in the circuit.

If you're hoping to use a voltage source to break/close a loop, then that won't work. You can use AC resistors for that (or the "switch" component in spectre).

There's a very good coverage of this in Ken's book http://www.designers-guide.org/Books/#Kundert-1995 - you can see the first few pages of the AC chapter if you follow the links on that page.

It's still not terribly clear to me what you're trying to do here.

There are also a lot of posts in this forum talking about analysing stability - you might want to use the Forum search function.

Regards,

Andrew.

Back to top

IP Logged