diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Software Tools (http://www.diyaudio.com/forums/software-tools/)
-   -   Spice oddity. Is it me or a glitch ? (http://www.diyaudio.com/forums/software-tools/210865-spice-oddity-me-glitch.html)

Mooly 16th April 2012 10:57 AM

Spice oddity. Is it me or a glitch ?
 
2 Attachment(s)
Can't figure this one out.

The circuit (which is just a doodle) initially has a DC servo and simulates just fine. This is picture #1 below. I then record the DC voltage at the servo output and sustitute the servo with a fixed voltage source. Again the simulation is fine.

Now for the strange part.

Not wanting the servo I start to delete it component by component using the "scissors" to remove each bit. When I get as far as deleting the opamp or V1 the main circuit suddenly fails to simulate. The main amp output just rises to the positive rail. It also ends up with a B-E voltage of 30 volts across Q1 (Emitter most negative). I can delete the 1 Meg, and the 0.1uF and V2 and their traces and the circuit is still fine. When I come to delete either the opamp or V2 it fails.

I can get as far as in picture 2. Up to this point the circuit simulates correctly. If I now delete anything else to do with the servo it fails. I even altered the voltage V1 to zero and that made no difference either.

Is it me :) or is there a glitch somewhere ? I can't see where I am going wrong.

AndrewT 16th April 2012 11:04 AM

3.1V of error correction is enormous. That made me look at the standard (without DC servo) circuit.
You have no route for the input offset current to return to source or ground.
If that 3.1V is out by a tiny amount, the output offset will rocket.

Mooly 16th April 2012 11:07 AM

Hi Andrew,
It's single ended so the 3.1 volts is OK. It a bias voltage rather than than an error correction.

I can't see where the interaction come in between the opamp and V1 in picture 2 and the main circuit. It's OK up to this point.

AndrewT 16th April 2012 11:11 AM

can you remove the 3.1V error correction (you can call it bias if you want) and replace it with the normal ground referencing resistor to allow the input bias current to flow to source/ground.
The amplifier should then bias itself to a reasonable output offset and you can then apply error correction to that if needed.

Once that is done, maybe then Spice will allow you to delete the last of the DC servo components and any definitions that go with them.

BTW,
may I remind you I am not a prepackaged simulator software user.
The (very simple) modeling I do, is on my own software.

Mooly 16th April 2012 11:12 AM

2 Attachment(s)
This is what happens. Look at the difference in the pictures (opamp).

AndrewT 16th April 2012 11:16 AM

You still have that forced error correction.
That rise in voltage looks like charging of C1 (?)

Let the input find it's own bias for the input offset current.

Mooly 16th April 2012 11:16 AM

The input of Q1 has to be biased to around 3 volts or so. I could rig a fixed bias divider up to try.

I've to go out shortly but keep the ideas coming :)

AndrewT 16th April 2012 11:19 AM

Is it a +3V or -3V on the base of the input transistor?
Let it do it's own biasing with the normal ground referencing resistor.

Mooly 16th April 2012 11:21 AM

Quote:

Originally Posted by AndrewT (Post 2987272)
You still have that forced error correction.
That rise in voltage looks like charging of C1 (?)

Let the input find it's own bias for the input offset current.

I agree it does look like C1 charging. Then there is the 30 volts across B-E of Q1. That doesn't sound right in that I would expect it to break down or zener at around 8 volts although I guess that's maybe a "model issue".

Mooly 16th April 2012 11:23 AM

Quote:

Originally Posted by AndrewT (Post 2987276)
Is it a +3V or -3V on the base of the input transistor?

It's minus 3


All times are GMT. The time now is 10:43 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio


Content Relevant URLs by vBSEO 3.3.2