Opamp and circuit help please.

Hello,

I have been trying to make a 40kHz ultrasonic transmitter and receiver
without much success until the recent "Opamp problem" post and the
subsequent replies. It answered many questions for me.
My desire is to have the receiver battery powered and be reliable.

I have put a schematic on the net at:
http://www.geocities.com/talionis.geo/Temp/temp.html

Unable to locate a suggested LM6132 low power opamp locally I am
trying a LM833 in its place. With the LM833 (GBW of 10MHz min - 15MHz
max, slew rate of 7V/uS) I set both halves of the opamp with a 1K and
33K for a gain of 33*33=1089.
The square root of a desired 1000 stage gain = 31.6 so a standard
value 33K negative feedback resistor was chosen for each stage. The
GBP would then be 10*33*40000=13.2MHz which is within the upper and
lower limits of the LM833. Correct so far?

There is a LM393 comparator in place of an LM358 opamp as with a
transistor based receiver circuit I built with a CA3140 opamp as a
comparator (1.5m range) I found that a drop in battery voltage caused
it to stop working. Re-adjusting the threshold trimmer fixed this but
I don't want to have to keep doing that as the battery depletes.

I've made the D1, D2 diodes schottky types, not quite the right ones I
know. 1SS106 barrier diodes would be better but are unavailable.

The yellow dots on the schematic are some things I'm not very sure
about. What are the functions of R4, C2 and C14 and are their values
correct?
Do I need to add anything to have a long shielded cable to TX1?

I have a basic 555 transmitter I made for the transistor based
receiver circuit and using it on the above LM833 circuit it didn't
work at all. Changing the opamps resistor values to Part-A 10k/1M and
Part-B 10k/100K from a printed off rangefinder circuit it did but at
only 2-3cm.
I am not sure why the transmitter worked on the second setup and not
the first. I read where it'd be better to stay away from large value
resistors to decrease noise with such high gain. It's not the best
transmitter but it must have been roughly tuned to 40kHz to work at
all. I do not have a scope to measure it but I have seen a frequency
meter kitset that I think might just do to tune it or its replacement.

The transmitter on the schematic I haven't yet built.

Any help always appreciated.

Andrew.

 

[John Popelish]

Rubicon wrote:

Hello,

I have been trying to make a 40kHz ultrasonic transmitter and receiver
without much success until the recent "Opamp problem" post and the
subsequent replies. It answered many questions for me.
My desire is to have the receiver battery powered and be reliable.

I have put a schematic on the net at:
http://www.geocities.com/talionis.geo/Temp/temp.html

Eliminate C14 and tie pins 3 and 5 together.

Unable to locate a suggested LM6132 low power opamp locally I am
trying a LM833 in its place.

The LM6132 has a wider input common mode range and output voltage
swing but this may not be a problem if you expect the final amplitude
to be well below the supply rails.

With the LM833 (GBW of 10MHz min - 15MHz
max, slew rate of 7V/uS) I set both halves of the opamp with a 1K and
33K for a gain of 33*33=1089.
The square root of a desired 1000 stage gain = 31.6 so a standard
value 33K negative feedback resistor was chosen for each stage. The
GBP would then be 10*33*40000=13.2MHz which is within the upper and
lower limits of the LM833. Correct so far?

Sounds reasonable.

There is a LM393 comparator in place of an LM358 opamp as with a
transistor based receiver circuit I built with a CA3140 opamp as a
comparator (1.5m range) I found that a drop in battery voltage caused
it to stop working. Re-adjusting the threshold trimmer fixed this but
I don't want to have to keep doing that as the battery depletes.

Then you will need to replace the supply with a stable reference as
the source of the decision voltage at TR1. You may be able ot build a
single voltage reference that is shared by all functions that need
one.

I've made the D1, D2 diodes schottky types, not quite the right ones I
know. 1SS106 barrier diodes would be better but are unavailable.
I like the 15 mA dual Schottky, BAS70-04.

http://rocky.digikey.com/WebLib/Zetex/Web%20Data/BAS70.pdf

The yellow dots on the schematic are some things I'm not very sure
about. What are the functions of R4, C2 and C14 and are their values
correct?

I doubt R4 is necessary. C2 has to have a much lower impedance around
40 kHz than the parallel combination of R1,5 to clean the amplifiers
of supply feedback. Of course R 1 and 5 could be higher values before
the two amplifier bias currents cause much of a bias point shift.
With the LM6132 they could easily each be a meg ohm (reduces battery
current a bit). The LM833 has a worst case bias current of 1 uA, so a
pair of 1 meg resistors supplying that to two stages would shift the
bias point a volt. I would probably use 100k resistors with that
amp. This gives a 50k divider impedance. At 40k, a capacitor with
500 ohms impedance (100 times reduction of any 40 kHz getting in
through the positive supply) would be .008 uf. So 10 uf is pretty
generous. I would probably use a .01 or .1 uf ceramic (not Y5V or Z5U
because of their microphonics) or film.

As I said earlier, eliminate C14.

Do I need to add anything to have a long shielded cable to TX1?

Put the 100 ohm resistor below the transducer, and tie the shield to
the positive rail.

I have a basic 555 transmitter I made for the transistor based
receiver circuit and using it on the above LM833 circuit it didn't
work at all. Changing the opamps resistor values to Part-A 10k/1M and
Part-B 10k/100K from a printed off rangefinder circuit it did but at
only 2-3cm.

You may want ot start thinking about how to transform these two gain
stages into band pass filters with gain to reduce interference.

I am not sure why the transmitter worked on the second setup and not
the first. I read where it'd be better to stay away from large value
resistors to decrease noise with such high gain.

Depending on your construction it may have more to do with stray
capacitance across the feedback resistors, which reduces gain. Lower
values of resistance reduce this effect.

It's not the best
transmitter but it must have been roughly tuned to 40kHz to work at
all. I do not have a scope to measure it but I have seen a frequency
meter kitset that I think might just do to tune it or its replacement.

Connect a signal diode (1N4148, etc.) in series with your digital volt
meter and adjust the frequency for maximum positive voltage at the
drain of MOS1.

The transmitter on the schematic I haven't yet built.

Any help always appreciated.

Andrew.

--
John Popelish

 

[Jamie]

its simple.
look for a single chip DC-DC inverter.
design your circuit to operate like at
5 volts.
they have Converters that will maintain
5 volts even when the input voltages gets
down as low as 1 volt.
if you don't want to do that then use a
high efficiency type.
below is a link to a reg that should handle the operation of
at least your reciver and comparator.

http://www.micrel.com/product-info/products/mic2954.shtml


Rubicon wrote:

Hello,

I have been trying to make a 40kHz ultrasonic transmitter and receiver
without much success until the recent "Opamp problem" post and the
subsequent replies. It answered many questions for me.
My desire is to have the receiver battery powered and be reliable.

I have put a schematic on the net at:
http://www.geocities.com/talionis.geo/Temp/temp.html

Unable to locate a suggested LM6132 low power opamp locally I am
trying a LM833 in its place. With the LM833 (GBW of 10MHz min - 15MHz
max, slew rate of 7V/uS) I set both halves of the opamp with a 1K and
33K for a gain of 33*33=1089.
The square root of a desired 1000 stage gain = 31.6 so a standard
value 33K negative feedback resistor was chosen for each stage. The
GBP would then be 10*33*40000=13.2MHz which is within the upper and
lower limits of the LM833. Correct so far?

There is a LM393 comparator in place of an LM358 opamp as with a
transistor based receiver circuit I built with a CA3140 opamp as a
comparator (1.5m range) I found that a drop in battery voltage caused
it to stop working. Re-adjusting the threshold trimmer fixed this but
I don't want to have to keep doing that as the battery depletes.

I've made the D1, D2 diodes schottky types, not quite the right ones I
know. 1SS106 barrier diodes would be better but are unavailable.

The yellow dots on the schematic are some things I'm not very sure
about. What are the functions of R4, C2 and C14 and are their values
correct?
Do I need to add anything to have a long shielded cable to TX1?

I have a basic 555 transmitter I made for the transistor based
receiver circuit and using it on the above LM833 circuit it didn't
work at all. Changing the opamps resistor values to Part-A 10k/1M and
Part-B 10k/100K from a printed off rangefinder circuit it did but at
only 2-3cm.
I am not sure why the transmitter worked on the second setup and not
the first. I read where it'd be better to stay away from large value
resistors to decrease noise with such high gain. It's not the best
transmitter but it must have been roughly tuned to 40kHz to work at
all. I do not have a scope to measure it but I have seen a frequency
meter kitset that I think might just do to tune it or its replacement.

The transmitter on the schematic I haven't yet built.

Any help always appreciated.

Andrew.

 

[John Popelish]

Rubicon wrote:
(snip)

You advised to "Put the 100 ohm resistor below the transducer, and tie
the shield to the positive rail" I'm not 100% sure what you mean here.

The mosfet has a very fast fall time that has nothing useful to do
with the emitted frequency but causes noise emission from the
transducer lines. If you move the resistor to the drain lead of the
transducer, you slow the rise and fall times on both transducer leads,
reducing RFI. Since then, the transducer is connected directly to the
positive rail you might as well just use that as the shield return.
Still probably best to put a twisted pair inside the shield, and not
connect the transducer end of the shield to anything.

I understand that with an opamp as a comparator without negative
feedback a little swing in input voltage causes the opamps output to
saturate to almost V+ or V-. Which would be best to use here for a
sensitive ultrasonic receiver, an LM358 opamp configured as a
comparator or an LM393 comparator? My liking of the LM393 is due to
its lower power consumption.

Agreed, as long as the long term state does not pass current through a
pull up resistor. If you include that current, the opamp may win.

I'm looking into bandpass filtering. I had read somewhere that
ultrasonic transducers were a mechanical filter and that additional
electronic filtering wouldn't be needed. Wrong I take it. I have done
a little searching and I think that the type of filter required here
is the active bandpass type with a high Ho and low Q.

Agreed. The filter is there just to keep wide band electrical noise
from being amplified as much as the transducer signal is.

--
John Popelish

 

John,

Thanks for the help there. I've made the alterations as you suggested
and its range with the basic transmitter is now approx 20cm using my
original 1K/33K per stage resistor ratios and trying the LM358 instead
of the LM393. A significant percentage increase in range to me but its
been suggested by a "noddy" that if I'm that close I'd might as well
lean forward and turn it on myself.
I may well have damaged the combined Tx/Rx transducers in my first
attempts but I'm not sure about that.


You advised to "Put the 100 ohm resistor below the transducer, and tie
the shield to the positive rail" I'm not 100% sure what you mean here.
Tie the long cables shield wire to V+ and then to the transducers
negative pin while the cables core goes from the mosfets drain, 100
ohm resistor and to the transducers positive pin? Doesn't sound right,
to me the shield wire is negative. Could you please clarify?

I understand that with an opamp as a comparator without negative
feedback a little swing in input voltage causes the opamps output to
saturate to almost V+ or V-. Which would be best to use here for a
sensitive ultrasonic receiver, an LM358 opamp configured as a
comparator or an LM393 comparator? My liking of the LM393 is due to
its lower power consumption.

I'm looking into bandpass filtering. I had read somewhere that
ultrasonic transducers were a mechanical filter and that additional
electronic filtering wouldn't be needed. Wrong I take it. I have done
a little searching and I think that the type of filter required here
is the active bandpass type with a high Ho and low Q.

Not yet looked at the TR1 stable supply reference but I will.

Thanks for the tuning tip

Regards,

Andrew.





On Sun, 08 Aug 2004 14:13:47 -0400, John Popelish <jpopelish@rica.net>
wrote:

Rubicon wrote:

Hello,

I have been trying to make a 40kHz ultrasonic transmitter and receiver
without much success until the recent "Opamp problem" post and the
subsequent replies. It answered many questions for me.
My desire is to have the receiver battery powered and be reliable.

I have put a schematic on the net at:
http://www.geocities.com/talionis.geo/Temp/temp.html

Eliminate C14 and tie pins 3 and 5 together.

Unable to locate a suggested LM6132 low power opamp locally I am
trying a LM833 in its place.

The LM6132 has a wider input common mode range and output voltage
swing but this may not be a problem if you expect the final amplitude
to be well below the supply rails.

With the LM833 (GBW of 10MHz min - 15MHz
max, slew rate of 7V/uS) I set both halves of the opamp with a 1K and
33K for a gain of 33*33=1089.
The square root of a desired 1000 stage gain = 31.6 so a standard
value 33K negative feedback resistor was chosen for each stage. The
GBP would then be 10*33*40000=13.2MHz which is within the upper and
lower limits of the LM833. Correct so far?

Sounds reasonable.

There is a LM393 comparator in place of an LM358 opamp as with a
transistor based receiver circuit I built with a CA3140 opamp as a
comparator (1.5m range) I found that a drop in battery voltage caused
it to stop working. Re-adjusting the threshold trimmer fixed this but
I don't want to have to keep doing that as the battery depletes.

Then you will need to replace the supply with a stable reference as
the source of the decision voltage at TR1. You may be able ot build a
single voltage reference that is shared by all functions that need
one.

I've made the D1, D2 diodes schottky types, not quite the right ones I
know. 1SS106 barrier diodes would be better but are unavailable.
I like the 15 mA dual Schottky, BAS70-04.

http://rocky.digikey.com/WebLib/Zetex/Web%20Data/BAS70.pdf

The yellow dots on the schematic are some things I'm not very sure
about. What are the functions of R4, C2 and C14 and are their values
correct?

I doubt R4 is necessary. C2 has to have a much lower impedance around
40 kHz than the parallel combination of R1,5 to clean the amplifiers
of supply feedback. Of course R 1 and 5 could be higher values before
the two amplifier bias currents cause much of a bias point shift.
With the LM6132 they could easily each be a meg ohm (reduces battery
current a bit). The LM833 has a worst case bias current of 1 uA, so a
pair of 1 meg resistors supplying that to two stages would shift the
bias point a volt. I would probably use 100k resistors with that
amp. This gives a 50k divider impedance. At 40k, a capacitor with
500 ohms impedance (100 times reduction of any 40 kHz getting in
through the positive supply) would be .008 uf. So 10 uf is pretty
generous. I would probably use a .01 or .1 uf ceramic (not Y5V or Z5U
because of their microphonics) or film.

As I said earlier, eliminate C14.

Do I need to add anything to have a long shielded cable to TX1?

Put the 100 ohm resistor below the transducer, and tie the shield to
the positive rail.

I have a basic 555 transmitter I made for the transistor based
receiver circuit and using it on the above LM833 circuit it didn't
work at all. Changing the opamps resistor values to Part-A 10k/1M and
Part-B 10k/100K from a printed off rangefinder circuit it did but at
only 2-3cm.

You may want ot start thinking about how to transform these two gain
stages into band pass filters with gain to reduce interference.

I am not sure why the transmitter worked on the second setup and not
the first. I read where it'd be better to stay away from large value
resistors to decrease noise with such high gain.

Depending on your construction it may have more to do with stray
capacitance across the feedback resistors, which reduces gain. Lower
values of resistance reduce this effect.

It's not the best
transmitter but it must have been roughly tuned to 40kHz to work at
all. I do not have a scope to measure it but I have seen a frequency
meter kitset that I think might just do to tune it or its replacement.

Connect a signal diode (1N4148, etc.) in series with your digital volt
meter and adjust the frequency for maximum positive voltage at the
drain of MOS1.

The transmitter on the schematic I haven't yet built.

Any help always appreciated.

Andrew.


--
John Popelish

 

Jamie,

Thankyou for the advice and the link. I have it up on screen at the
moment.

Regards,

Andrew.




On Sun, 08 Aug 2004 12:48:29 -0700, Jamie
<jamie_5_not_valid_after_5_Please@charter.net> wrote:

its simple.
look for a single chip DC-DC inverter.
design your circuit to operate like at
5 volts.
they have Converters that will maintain
5 volts even when the input voltages gets
down as low as 1 volt.
if you don't want to do that then use a
high efficiency type.
below is a link to a reg that should handle the operation of
at least your reciver and comparator.

http://www.micrel.com/product-info/products/mic2954.shtml


Rubicon wrote:

Hello,

I have been trying to make a 40kHz ultrasonic transmitter and receiver
without much success until the recent "Opamp problem" post and the
subsequent replies. It answered many questions for me.
My desire is to have the receiver battery powered and be reliable.

I have put a schematic on the net at:
http://www.geocities.com/talionis.geo/Temp/temp.html

Unable to locate a suggested LM6132 low power opamp locally I am
trying a LM833 in its place. With the LM833 (GBW of 10MHz min - 15MHz
max, slew rate of 7V/uS) I set both halves of the opamp with a 1K and
33K for a gain of 33*33=1089.
The square root of a desired 1000 stage gain = 31.6 so a standard
value 33K negative feedback resistor was chosen for each stage. The
GBP would then be 10*33*40000=13.2MHz which is within the upper and
lower limits of the LM833. Correct so far?

There is a LM393 comparator in place of an LM358 opamp as with a
transistor based receiver circuit I built with a CA3140 opamp as a
comparator (1.5m range) I found that a drop in battery voltage caused
it to stop working. Re-adjusting the threshold trimmer fixed this but
I don't want to have to keep doing that as the battery depletes.

I've made the D1, D2 diodes schottky types, not quite the right ones I
know. 1SS106 barrier diodes would be better but are unavailable.

The yellow dots on the schematic are some things I'm not very sure
about. What are the functions of R4, C2 and C14 and are their values
correct?
Do I need to add anything to have a long shielded cable to TX1?

I have a basic 555 transmitter I made for the transistor based
receiver circuit and using it on the above LM833 circuit it didn't
work at all. Changing the opamps resistor values to Part-A 10k/1M and
Part-B 10k/100K from a printed off rangefinder circuit it did but at
only 2-3cm.
I am not sure why the transmitter worked on the second setup and not
the first. I read where it'd be better to stay away from large value
resistors to decrease noise with such high gain. It's not the best
transmitter but it must have been roughly tuned to 40kHz to work at
all. I do not have a scope to measure it but I have seen a frequency
meter kitset that I think might just do to tune it or its replacement.

The transmitter on the schematic I haven't yet built.

Any help always appreciated.

Andrew.