Re: How to Simulate Op Amp Supply Current in Ltspice

On Tue, 17 Mar 2009 13:57:50 -0700, Jim Thompson
<To-Email-Use-The-Envelope-Icon@xxxxxxxxxxxxxxx> wrote:

On Tue, 17 Mar 2009 13:10:57 -0700, D from BC
<myrealaddress@xxxxxxxxx> wrote:

On Tue, 17 Mar 2009 12:55:03 -0700, Jim Thompson
<To-Email-Use-The-Envelope-Icon@xxxxxxxxxxxxxxx> wrote:

On Tue, 17 Mar 2009 13:45:16 -0400, "Brett Holden"
<brettholden@xxxxxxxxxx> wrote:


"D from BC" <myrealaddress@xxxxxxxxx> wrote in message
news:5n9ur4hveljebg51ttrf0shtdvhsvjpmqb@xxxxxxxxxx
On Mon, 16 Mar 2009 23:33:31 -0400, "Brett Holden"
<brettholden@xxxxxxxxxx> wrote:


"Helmut Sennewald" <helmutsennewald@xxxxxxxxxxx> wrote in message
news:gpkrvk$prn$00$1@xxxxxxxxxxxxxxxxxxxx
"D from BC" <myrealaddress@xxxxxxxxx> schrieb im Newsbeitrag
news:t4frr4ploerlri4kog4njr1tig28j1v8rc@xxxxxxxxxx
How do I simulate op amp supply current in Ltspice?

I think I pick level 3a but I don't know the parameter name for the
supply current.


D from BC
myrealaddress(at)comic(dot)com
BC, Canada
Posted to usenet sci.electronics.design

Hello D from BC,

I think you talk about the quiescent current.
Just add a resistor between V+ and V- for every opamp in your schematic.

If you want do that in the model file, you have to save it with a
different name.
This would require a modified symbol too.

Best regards,
Helmut

Many manufacturers do not accurately supply current. That includes Linear
Technology. I wanted to monitor current using a LT1097 (if memory serves)
low power amplifier as a comparator and "measured" something like 13ma
into
the V+ and V- terminals.

When I complained about this to Mike Englehardt's department I got the
excuse "we don't model supply current" and also, "don't use an opamp we
have
comparators".

I thought their reply was pretty lame.



Bummer :(
Oh well.. I'll have to make do..
Thanks


D from BC
myrealaddress(at)comic(dot)com
BC, Canada
Posted to usenet sci.electronics.design

Pardon me, it was about a year ago and my memory is foggy. I don't want to
be unfair to LTC.

They probaby said "... don't always model supply current for opamps". Not
always. Or more probably it was something like "... not modelled for use as
comparator". I forget exactly, so it was wrong for me to try and quote what
I don't accurately remember.

In all fairness, I did not breadboard the circuit. For all I know the LT1097
(and a handful of other opamps I tried) might actually draw +/- 13ma when
used as a comparator.

But, I decided to use a comparator. Supply current can often be measured
just like Helmut stated, but not always.


Power supply currents are tricky to model for simulation. It's easy
enough to include load currents, but currents flowing in compensation
capacitors can have some pretty wild behavior.

The only real way to do a behavioral model properly is to start with a
device-level netlist, then match behavioral to device-level.

I get calls all the time about modeling with two consistent results:

(1) The semiconductor company won't share the "real" netlist, wanting
modeling from the data ***... what a crock, and...

(2) They expect it to be done for virtually free.

Thus I haven't ever done any commercial modeling... just for my own
internal use where I already have access to an accurate netlist, and
process library.

I'm considering model generation on the side, _as_I_can_get_to_it_ ...
free... for non-commercial use.

Might be good advertising and get me some design jobs ;-)

What would people like to see modeled?

...Jim Thompson

A LMV431 shunt regulator?? :)
It can be modeled with an op amp that draws about 100uA of quiescent
current.


D from BC
myrealaddress(at)comic(dot)com
BC, Canada
Posted to usenet sci.electronics.design

Ask and yee shall receive... actually done back in 2002 ;-)

.SUBCKT LMV431 R A K
** by Jim Thompson, © 2002-2009 **
R_R1 N_1 N_2 0.25
E_BUF1 N_3 0 R 0 1

E_BUF1 N_3 A R A 1

R_R7 N_3 A 1Meg
V_V1 N_2 A 1.24V
D_D1 A K D1N4148
F_F1 K A VF_F1 1
VF_F1 N_3 N_1 0V
.ENDS LMV431

...Jim Thompson

First thing I did was use it to build a follower... more current out
than in. Cool trick!

Regards,
Bob Monsen
.

Quantcast