How can an output signal LEAD the input?

[Electric guy]

I'm working on a bio-potential amplifier (EMG/ECG) which is pretty
simple - just a couple of leads connected to the inverting and
non-inverting inputs of an op-amp (forget the exact part but it's
specialized differential amplifier with gain set by external resistor)
followed by a second (ordinary) op-amp for an extra X100 gain. Total
gain is adjustable from about 1,000 to 10,000.

We've had some problems with the differential amplifier blowing out
due to the use of cautery equipment (electro-surgery) or
defibrillators. A pair of 1N914 diodes wired across the inputs didn't
help.

When the differential amplifier blows, some sort of leakage current
back-feeds from the amp, through the leads, and into the heart
(causing fibrillation). This is not a human/clinical application.

So we try putting capacitors (0.1 uf) in series with each of the two
input leads. The theory is that if the amp blows it won't be able to
back-feed any current through the leads.

But I've noticed that at the low frequencies of ECG signals (the QRS
complex resembles a single cycle of 10 to 15 hz sinusoid) that the
output of the amplifier actually LEADS the input by about 8 ms.

While feeding in burst sine waves of higher frequencies, the effect
disappears at higher frequencies (say, 50 to 100 hz).

I was wondering what sort of mathematics can explain this.

 

[Rene Tschaggelar]

I'm working on a bio-potential amplifier (EMG/ECG) which is pretty
simple - just a couple of leads connected to the inverting and
non-inverting inputs of an op-amp (forget the exact part but it's
specialized differential amplifier with gain set by external resistor)
followed by a second (ordinary) op-amp for an extra X100 gain. Total
gain is adjustable from about 1,000 to 10,000.

We've had some problems with the differential amplifier blowing out
due to the use of cautery equipment (electro-surgery) or
defibrillators. A pair of 1N914 diodes wired across the inputs didn't
help.

There are differential amplifier with +-500Volts common mode.

When the differential amplifier blows, some sort of leakage current
back-feeds from the amp, through the leads, and into the heart
(causing fibrillation). This is not a human/clinical application.

I get the impression that blowing the amplifier should be avoided then.
Again, high input impedances lower the risk of doing damage.

So we try putting capacitors (0.1 uf) in series with each of the two
input leads. The theory is that if the amp blows it won't be able to
back-feed any current through the leads.

But I've noticed that at the low frequencies of ECG signals (the QRS
complex resembles a single cycle of 10 to 15 hz sinusoid) that the
output of the amplifier actually LEADS the input by about 8 ms.

There is no timemachine this way.
A high pass such as the caps differentiates the signal.

Rene

 

[James Meyer]

But I've noticed that at the low frequencies of ECG signals (the QRS
complex resembles a single cycle of 10 to 15 hz sinusoid) that the
output of the amplifier actually LEADS the input by about 8 ms.

I was wondering what sort of mathematics can explain this.

Mathematics can't explain it. Drugs might. Quit sniffing the
anesthetics and see if the phenomenon goes away.

Jim

 

[Electric guy]

you didn't specify the part number your using for the first stage
on the OP-AMP ? and jugging from the effects your describing i
think you must be working with supply voltages far about the
norm of 5 to 10 volts.

The part is INA114AP (or INA114BP - not sure). Supply voltage is +/-
12 volts.

http://focus.ti.com/lit/ds/symlink/ina114.pdf

The example circuits don't really show any input protection methods.

What I'm doing is pretty much identical to Figure 10, page 12 in the
above PDF file - except I'm not using the leg-drive circuit.

So I'm trying a pair of 0.1 uf caps in series with the input leads (1
cap on each input lead).

With the cap's in place, and observing both the input and output
signals on a scope, the output signal seems to lead the input signal.

Specifically - the input signal is a single cycle of a sine wave with
a frequency approx. 10 to 15 hz and a repetition rate of 1 hz. It's
not that the *entire* signal is leading on the output. Both the input
and output signals start at the same time. It's the position of the
min and max points (peak and valley) on the output signal that is
leading the peak and valley of the input signal by 8 ms.

Basically, the input and output signal are time-matched at the very
start of the signal, for the rest of the signal the output is leading
the input by 8 ms.

I will look into the protection circuit you describe.

 

[legg]

I'm working on a bio-potential amplifier (EMG/ECG) which is pretty
simple - just a couple of leads connected to the inverting and
non-inverting inputs of an op-amp (forget the exact part but it's
specialized differential amplifier with gain set by external resistor)
followed by a second (ordinary) op-amp for an extra X100 gain. Total
gain is adjustable from about 1,000 to 10,000.

We've had some problems with the differential amplifier blowing out
due to the use of cautery equipment (electro-surgery) or
defibrillators. A pair of 1N914 diodes wired across the inputs didn't
help.

Whether for human applications or not, Your circuit construction and
isolation methods should comply with those of BF type patient-applied
parts.

The only time this is not the case, is when you are dealing with
inanimate or dead matter.

When the differential amplifier blows, some sort of leakage current
back-feeds from the amp, through the leads, and into the heart
(causing fibrillation). This is not a human/clinical application.

So we try putting capacitors (0.1 uf) in series with each of the two
input leads. The theory is that if the amp blows it won't be able to
back-feed any current through the leads.

But I've noticed that at the low frequencies of ECG signals (the QRS
complex resembles a single cycle of 10 to 15 hz sinusoid) that the
output of the amplifier actually LEADS the input by about 8 ms.

Examine one-shot events and re-evaluate. High pass filters (ie
insuficiently capacitively-coupled low frequency amplifiers) will give
this impression for repetative signals.

You need a recording or very high persistence display for this.

RL

 

[Jamie]

I'm working on a bio-potential amplifier (EMG/ECG) which is pretty
simple - just a couple of leads connected to the inverting and
non-inverting inputs of an op-amp (forget the exact part but it's
specialized differential amplifier with gain set by external resistor)
followed by a second (ordinary) op-amp for an extra X100 gain. Total
gain is adjustable from about 1,000 to 10,000.

We've had some problems with the differential amplifier blowing out
due to the use of cautery equipment (electro-surgery) or
defibrillators. A pair of 1N914 diodes wired across the inputs didn't
help.

When the differential amplifier blows, some sort of leakage current
back-feeds from the amp, through the leads, and into the heart
(causing fibrillation). This is not a human/clinical application.

So we try putting capacitors (0.1 uf) in series with each of the two
input leads. The theory is that if the amp blows it won't be able to
back-feed any current through the leads.

But I've noticed that at the low frequencies of ECG signals (the QRS
complex resembles a single cycle of 10 to 15 hz sinusoid) that the
output of the amplifier actually LEADS the input by about 8 ms.

While feeding in burst sine waves of higher frequencies, the effect
disappears at higher frequencies (say, 50 to 100 hz).

I was wondering what sort of mathematics can explain this.

the LEAD effect your seeing maybe an ambient capacitance around
your components.
you didn't specify the part number your using for the first stage
on the OP-AMP ? and jugging from the effects your describing i think
you must be working with supply voltages far about the norm of 5 to 10
volts.
you may want to try a MosFET with clamping diodes for your input built
into your probe .
the probe would have the third wire to supply the voltage for the amp.
more like an active FET probe.
if you put a neon lamp at the head end then pass the signal through a
high value resister like a 470K ohm then have some small amount of
capacitance as a by pass with clamping diodes i think you will do fine
for suppressing the high voltage surge.
the neon will keep it from going over the 70 volt average.

the clamping diodes after the resistor will keep the 70 volts down in
the .7 range.
using a cheap TL082 is a good selection for this also.
i think i remember it having some static protection on the inputs.

 

[John Larkin]

I'm working on a bio-potential amplifier (EMG/ECG) which is pretty
simple - just a couple of leads connected to the inverting and
non-inverting inputs of an op-amp (forget the exact part but it's
specialized differential amplifier with gain set by external resistor)
followed by a second (ordinary) op-amp for an extra X100 gain. Total
gain is adjustable from about 1,000 to 10,000.

We've had some problems with the differential amplifier blowing out
due to the use of cautery equipment (electro-surgery) or
defibrillators. A pair of 1N914 diodes wired across the inputs didn't
help.

When the differential amplifier blows, some sort of leakage current
back-feeds from the amp, through the leads, and into the heart
(causing fibrillation). This is not a human/clinical application.

So we try putting capacitors (0.1 uf) in series with each of the two
input leads. The theory is that if the amp blows it won't be able to
back-feed any current through the leads.

But I've noticed that at the low frequencies of ECG signals (the QRS
complex resembles a single cycle of 10 to 15 hz sinusoid) that the
output of the amplifier actually LEADS the input by about 8 ms.

While feeding in burst sine waves of higher frequencies, the effect
disappears at higher frequencies (say, 50 to 100 hz).

I was wondering what sort of mathematics can explain this.

The series cap differentiates low-frequency signals, making a rounded
pulse appear to happen sooner than the true peak.

John

 

[Rich Grise]

The part is INA114AP (or INA114BP - not sure). Supply voltage is +/- 12
volts.

http://focus.ti.com/lit/ds/symlink/ina114.pdf

The example circuits don't really show any input protection methods.

What I'm doing is pretty much identical to Figure 10, page 12 in the above
PDF file - except I'm not using the leg-drive circuit.

So I'm trying a pair of 0.1 uf caps in series with the input leads (1 cap
on each input lead).

With the cap's in place, and observing both the input and output signals
on a scope, the output signal seems to lead the input signal.

Specifically - the input signal is a single cycle of a sine wave with a
frequency approx. 10 to 15 hz and a repetition rate of 1 hz. It's not
that the *entire* signal is leading on the output. Both the input and
output signals start at the same time. It's the position of the min and
max points (peak and valley) on the output signal that is leading the peak
and valley of the input signal by 8 ms.

As Rene Tschaggelar said in his reply yesterday, you're simply high-pass
filtering the signal, increasing the slew rate so it peaks sooner.

HTH!
Rich

 

[ChrisGibboGibson]

legg wrote:

[snip]

Your circuit construction and
isolation methods should comply with those of BF type patient-applied
parts.

The only time this is not the case, is when you are dealing with
inanimate or dead matter.

Well....

OP did say that.....

Which quite often can lead to death.

So once this happens he *is* "dealing with inanimate or dead matter"

Is that a good enough get out for him?

;-)

Gibbo

 

[N. Thornton]

I'm working on a bio-potential amplifier (EMG/ECG) which is pretty
simple - just a couple of leads connected to the inverting and
non-inverting inputs of an op-amp (forget the exact part but it's

We've had some problems with the differential amplifier blowing out
due to the use of cautery equipment (electro-surgery) or
defibrillators. A pair of 1N914 diodes wired across the inputs didn't
help.

When the differential amplifier blows, some sort of leakage current
back-feeds from the amp, through the leads, and into the heart
(causing fibrillation). This is not a human/clinical application.

vivisection.

go learn about static protection and apply it. 2 diodes wont do much.

 

[legg]

legg wrote:

[snip]

Your circuit construction and
isolation methods should comply with those of BF type patient-applied
parts.

The only time this is not the case, is when you are dealing with
inanimate or dead matter.

Well....

OP did say that.....

Which quite often can lead to death.

He describes the damage to the monitor as being correlated to the
simultaneous application of either electrosurgery or defibrillator
equipment.

Monitors of suitable isolation class intentionally do not exibit
excess LF leakage under external influence, nor is this permitted
under single-fault abnormal conditions (ie failure).

I believe inaxx isolation amplifiers data carries a disclaimer
indicating that they are not inherently safe in this application. I do
not doubt that they could possibly be so configured, if the
application requirement is understood and the resulting circuitry and
power supply interface is designed and tested for the application.

RL