Plumbing questions

Google Groups
1

Now I’ve got my nice new shiny scope I’ve been playing around with the
piezos I used as a drum triggers.

Firstly, it turns out that piezos have an orientation - hit them on one
side and you get an initial large +ve spike followed by an irregular
decaying AC signal. Turn them over and you get a large initial -ve
spike. The ones I’ve got have random orientations - the black/red leads
tell you nothing about what the orientation is.

I was using this circuit:
http://leucos.lstilde.org/wp/wp-content/uploads/2009/05/piezo_circuit.png and
hindsight, with even my limited electronics knowledge, it seems like a
load of tosh - full article is at
http://leucos.lstilde.org/wp/2009/06/piezo-transducer-signal-conditioning/

Here are what I think are the problems, do they seem valid?

  1. The 2nd BAT85 (schottky) from the left serves no purpose as it is in
    parallel with the 5.1V zener.

  2. The RC pair are supposed to be providing smoothing but they are in
    parallel, a configuration that as far as I’ve been able to find out just
    adjusts the phase of the current through the capacitor. I think what is
    actually needed is a RC low-pass filter.

  3. The circuit is intended to protect the input of the MCU not just
    against +ve overvoltage (the zener) but also -ve voltages (the 2 BAT85s,
    one of which is redundant). However, by observation it doesn’t do that.
    I’ve simplified the circuit so it just has a single BAT85 to protect
    against -ve voltages and no other components.

The first image below (pos.png) shows what happens with a large +ve
spike and is as expected - the input voltage (red) initially goes +ve
and the post-diode voltage (yellows) follows, then when the input
voltage goes -ve the diode prevents its output voltage from going below
0V, as expected - well, mostly, it still goes to -1.6V, not sure why.
However if we flip the piezo over and generate a leading -ve pulse we
get a -ve output from the diode that’s way beyond what we’d expect
(neg.png), -22.4V in this case. This puzzled the hell out of me until I
looked at the BAT85 datasheet. The max reverse voltage is specified as
30V, here it’s having -74V stuffed across it. The BAT85 is completely
unsuitable for the job, I’ve seen spikes of over 100V from the piezo, so
a diode with a VR of 200V is probably what’s needed.

Does all that seem correct?

Alan Burlison

pos.png800×600 8.64 KB

neg.png800×600 8.56 KB

2

Now I’ve got my nice new shiny scope I’ve been playing around with the
piezos I used as a drum triggers.

Firstly, it turns out that piezos have an orientation - hit them on one
side and you get an initial large +ve spike followed by an irregular
decaying AC signal. Turn them over and you get a large initial -ve spike.
The ones I’ve got have random orientations - the black/red leads tell you
nothing about what the orientation is.

I was using this circuit: http://leucos.lstilde.org/wp/**
wp-content/uploads/2009/05/piezo_circuit.pnghttp://leucos.lstilde.org/wp/wp-content/uploads/2009/05/piezo_circuit.pngand hindsight, with even my limited electronics knowledge, it seems like a
load of tosh - full article is at http://leucos.lstilde.org/wp/
2009/06/piezo-transducer-**signal-conditioning/http://leucos.lstilde.org/wp/2009/06/piezo-transducer-signal-conditioning/

Here are what I think are the problems, do they seem valid?

  1. The 2nd BAT85 (schottky) from the left serves no purpose as it is in
    parallel with the 5.1V zener.

It is possible it is there for reverse protection if the first one breaks
down…

  1. The RC pair are supposed to be providing smoothing but they are in
    parallel, a configuration that as far as I’ve been able to find out just
    adjusts the phase of the current through the capacitor. I think what is
    actually needed is a RC low-pass filter.

It is a timing circuit, you charge the cap up, and then it’ll discharge
through the R.

  1. The circuit is intended to protect the input of the MCU not just
    against +ve overvoltage (the zener) but also -ve voltages (the 2 BAT85s,
    one of which is redundant). However, by observation it doesn’t do that.
    I’ve simplified the circuit so it just has a single BAT85 to protect
    against -ve voltages and no other components.

The zenner will take time to break down, thus you may find it’ll take tome
to clamp the voltage.

You may also be seeing the effect of the clamp from the zenner because the
output of the pizeo is being held at Vclamp plus the forward voltage drop
of the input diode.

The first image below (pos.png) shows what happens with a large +ve spike
and is as expected - the input voltage (red) initially goes +ve and the
post-diode voltage (yellows) follows, then when the input voltage goes -ve
the diode prevents its output voltage from going below 0V, as expected -
well, mostly, it still goes to -1.6V, not sure why. However if we flip the
piezo over and generate a leading -ve pulse we get a -ve output from the
diode that’s way beyond what we’d expect (neg.png), -22.4V in this case.
This puzzled the hell out of me until I looked at the BAT85 datasheet.
The max reverse voltage is specified as 30V, here it’s having -74V stuffed
across it. The BAT85 is completely unsuitable for the job, I’ve seen
spikes of over 100V from the piezo, so a diode with a VR of 200V is
probably what’s needed.

Yeah, and you need something fast response. Anything schottky class should
be good for that.

The diode will take time to start conduction.

Note, you should take a look at the input diagram for the atmega pins, as
they have diodes to the VCC and GND rails. If your inputs go to far above
or below these start to conduct. The aim is to minimise this effect so the
diodes don’t die and take the next stage of the chip with them.

Does all that seem correct?

–
Alan Burlison

–
You received this message because you are subscribed to the Google Groups
"HAC:Manchester" group.
To post to this group, send an email to hacman@googlegroups.com.
To unsubscribe from this group, send email to hacman+unsubscribe@**
googlegroups.com hacman%2Bunsubscribe@googlegroups.com.
For more options, visit this group at http://groups.google.com/**
group/hacman?hl=en-GB http://groups.google.com/group/hacman?hl=en-GB.

Ben Dooks, http://www.fluff.org/ben/ bjdooks@googlemail.com

3
  1. The 2nd BAT85 (schottky) from the left serves no purpose as it is in
    parallel with the 5.1V zener.

It is possible it is there for reverse protection if the first one breaks
down…

The diodes that I’ve blown up have all failed closed-circuit, doesn’t
that mean it would still be useless?

  1. The RC pair are supposed to be providing smoothing but they are in
    parallel, a configuration that as far as I’ve been able to find out just
    adjusts the phase of the current through the capacitor. I think what is
    actually needed is a RC low-pass filter.

It is a timing circuit, you charge the cap up, and then it’ll discharge
through the R.

Ah right, so it does have a purpose then, thanks. However, looking at
it in practice it doesn’t seem to be doing very much - tap.png is a
single sharp tap on the piezo and has the RC in place, there’s not much
smoothing going on, and certainly not the 10msec he says there is in the
webpage.

The zenner will take time to break down, thus you may find it’ll take tome
to clamp the voltage.

The first set of waveforms were just the schottky - no zener in the circuit.

You may also be seeing the effect of the clamp from the zenner because the
output of the pizeo is being held at Vclamp plus the forward voltage drop
of the input diode.

Yes, in the full circuit I see a significant buildup of voltage across
the piezo - see pulses.png below. It’s taking more than 0.5 second for
the -ve voltage across the piezo to decay. Methinks a resistor across
the piezo is in order…

Yeah, and you need something fast response. Anything schottky class should
be good for that.

The diode will take time to start conduction.

Is that the forward recovery time? The problem with most of the
schottkys that I’ve looked at is they have fairly low maximum reverse
voltages, I need something that will cope with 200V or so and those
mostly seem to be normal signal diodes.

Note, you should take a look at the input diagram for the atmega pins, as
they have diodes to the VCC and GND rails. If your inputs go to far above
or below these start to conduct. The aim is to minimise this effect so the
diodes don’t die and take the next stage of the chip with them.

OK, so the signal conditioning is important then :slight_smile:

Alan Burlison

tap.png800×600 9.23 KB

pulses.png800×600 10.1 KB