Re: Old email on virtual ground circuit - Re: [buildcheapeeg] Re: High resolution EEG schematic

From: Joerg Hansmann (info_at_jhansmann.de)
Date: 2002-03-12 19:59:01


Hi,

----- Original Message -----
From: Sar Saloth <sarsaloth_at_yahoo.com>
To: <buildcheapeeg_at_yahoogroups.com>
Sent: Monday, March 11, 2002 9:04 PM
Subject: Old email on virtual ground circuit - Re: [buildcheapeeg] Re: High resolution EEG schematic

...
> >ACK.
>
> I am very new to this news-group type stuff. What does ACK stand for? By
> your usage it is more like [NACK] than [ACK], should I assume it is like
> ARGH!, then it would make sense to me. (;.<>

Probably some aspect of your answer seemed Ok to me, so I could agree
with it. ;-)

...
> > > been made. Nevertheless, I can see some low level noise that shouldn't be
> > > there, and when I shock it it does indeed oscillate until powered down.
> >
> >Looks like a very low phase margin.
>
> Wouldn't it be a non-linearity if the size of the impulse was an issue? If
> it was part of the linear response, couldn't it be set off even by a small
> signal?

IMO both is true:
Nonlinearity (phase and gain as a function of the signal magnitude)
and a pulse of sufficient size can push a circuit with low phase margin
over the edge.
If the circuit had much phase margin, it would have no problems
even when recovering from saturation.

Look at v(out2) from about 0.1ms to 3ms in "VirtualGnd-InputResponse.pdf"
(that you have posted recently) :
There is a nearly linear ramp because the opamp is at its output current
limit.

To investigte this assumption further this, I have put a sinewave source
with 50uVp-p in series with the step input stimulus.
As you can see on picture "VirtualGnd-Tuned_non_linear_gain01.png" the
AC-amplitude during the ramp-time is zero, meaning that the amplifier is
saturated.

...
> >I thought 500 mOhms were quite normal for a 47uF 10V tantalum type.
>
> I meant that I didn't realize ESR values were so large. I have gone back
> and found that I usually didn't use tantalums in those situations,

BTW:
I have tuned the ESR compensated circuit for usage with the
LT1114 (see picture "VirtualGnd-Tuned02_bode_plot.png")
by increasing the ESR of the 47uF C by about factor 10 (5000mOhms,
IIRR this could be a normal aluminium electrolytic capacitor )
and inserting a damping R in series with C11 to remove a HF resonance
at about 10Mhz (a series resonator formed by C11 and its ESI of about
10nH).

> so what
> I thought was universally stable was only in my circuits.
> I apologize for the false claim.

No problem.

...
> I think that is what you said (that the overshoot curves in the data sheet
> were for capacitors without ESR or what you called ESR compensated
> circuits). Maybe you were just on too great a roll with the ACK.

ACK ;-)

...
> > > This is getting me thinking, if the OpAmp output has a part in
> > > it, that might explain why the same circuit with an LTC27L4 OpAmp seemed
> > > stable until there was a significant perturbation - enough to cause the
> > > gain to drop

Gain drop normally stops oscillation or limits it to a maximum value.

However oscillation will occurr in circuits with low phase margin
when the gain recovers after saturation.

> due to loading the amplifier, which is when it went into
> > > oscillation.
> >
> >If you provide the LTC27L4 spice model I could do some simulations.
>
> Oops, I meant TLC27L4, almost identical model to your TLC272.

If you want to do simulations with the TLC272 I can provide
*.sub and *.sym files I have modified for use with LTSPICE.

> Anyway, thanks for the education. I am posting the circuits you posted,
> but with my circuit "tuned" and with a load added to look at load regulation.
>
> Your original circuit is superior here also!

Not really, because it has been designed for TLC272.
Better use the version tuned for use with LT1114 (in the attachment).

> |:.( I guess the only
> place my circuit would be better is if the circuit required too many
> distributed ceramic capacitors so that your circuit would become
> unstable. Is that right?

I do not really see evidence for this assumption...

Are 100 x 100nF ceramic capacitors enough ?
(see "VirtualGnd-Tuned_100x100nF".asc, .png)

> btw, the FilterCad UI is very nice. Is there a commercial full-featured
> product with a similar UI?

I don't know. My experience with SPICE like simulators is
rather limited.
BTW: What features do you think are missing most in LTSPICE (SwitcherCAD) ?

> Thanks for the education.

Only information, not education ;-)

Regards,

Joerg



picture

picture





picture

This archive was generated by hypermail 2.1.4 : 2002-07-27 12:28:40 BST