Single ended signal over TL072 to a differential line

[Rüdiger Leibrandt]

BUT HOW?!

I find hundred of publications online taking care of the conversion from
differential to single ended signals , like
forums.ni.com/ni/board/message?board. id=100&message.id=656
or
groups.google.de/group/sci.electronics.design/
browse_thread/thread/c91d76458a73fec3/9a1b51038ea535e8
but I need it the other way round - I need to drive a weak signal over a
highly exposed area to my measuring-station.

Anyone knowing where I may find something that tells me how I get the
starting side of the sensorline working?

All hints and help appreciated!

 

[martin griffith]

BUT HOW?!

I find hundred of publications online taking care of the conversion from
differential to single ended signals , like
forums.ni.com/ni/board/message?board. id=100&message.id=656
or
groups.google.de/group/sci.electronics.design/
browse_thread/thread/c91d76458a73fec3/9a1b51038ea535e8
but I need it the other way round - I need to drive a weak signal over a
highly exposed area to my measuring-station.

Anyone knowing where I may find something that tells me how I get the
starting side of the sensorline working?

All hints and help appreciated!

You dont say how large the signal is, or the bandwidth, or cable type

Assuming twisted pair cable, and say 20K B/W you could start by
having a pseudo differential signal. Connect the output of your 072 to
one of the lines via a 50 ohm resistor, connect the other cable to
ground with another 50 ohm resistor, this will look like a balanced
signal to a differential receiver and have good common mode rejection,
maybe about 6dB worse than a fully balanced output


martin

 

[Rüdiger Leibrandt]

You dont say how large the signal is, or the bandwidth, or cable type

Assuming twisted pair cable, and say 20K B/W you could start by
having a pseudo differential signal. Connect the output of your 072 to
one of the lines via a 50 ohm resistor, connect the other cable to
ground with another 50 ohm resistor, this will look like a balanced
signal to a differential receiver and have good common mode rejection,
maybe about 6dB worse than a fully balanced output


martin

Ai - sorry!
input signal is "a few" millivolt ( usually - I will need to have potis to
fine-tune the amplification I think ) but can go up to as much as 5 Volts
in the seldom extreme. The lines are an ethernet Category 5 cable, 8
meters, running straight into a ( not defined ) differential input of a
PCI-datasampling-card in a PC. The sensor is a small K-Band transceiver (
aka radar from InnoSent ).
I do not know where to add the resistors to shape and tune the output of the
combined opamps. I'm lost if and where to add the poti's to control the
amplification and so on. I hope that there might be a website with an
example circuit which I could adapt to my needs.

 

[martin griffith]

Ai - sorry!
input signal is "a few" millivolt ( usually - I will need to have potis to
fine-tune the amplification I think ) but can go up to as much as 5 Volts
in the seldom extreme. The lines are an ethernet Category 5 cable, 8
meters, running straight into a ( not defined ) differential input of a
PCI-datasampling-card in a PC. The sensor is a small K-Band transceiver (
aka radar from InnoSent ).
I do not know where to add the resistors to shape and tune the output of the
combined opamps. I'm lost if and where to add the poti's to control the
amplification and so on. I hope that there might be a website with an
example circuit which I could adapt to my needs.

http://es.geocities.com/mart_in_medina/P_CAD.pdf
will show you the basics


martin

 

[Winfield]

http://es.geocities.com/mart_in_medina/P_CAD.pdf
will show you the basics

martin

If he uses a TL072 dual opamp (good choice you made), he
might as well use the 2nd-half to make a full-differential output.

 

[martin griffith]

If he uses a TL072 dual opamp (good choice you made), he
might as well use the 2nd-half to make a full-differential output.

Yeah, I know, but he did say single ended, I had thought of mentioning
diff drivers etc. But there is too much "Nah you dont want to do it
that way" around here, when a 072 will do the job.

Anyway, lets see how the OP gets on.


martin

 

[Rüdiger Leibrandt]

http://es.geocities.com/mart_in_medina/P_CAD.pdf
will show you the basics


martin

Okay, Thank you for that first example. My boss gave me this one here:
http://www.aries.homeftp.net/publish/Electronics/concept.jpg
Digital Media is not really that kind of study teaching one how to design
line drivers I am afraid...
Maybe you can interpret the missing bits or the concept he uses there.

 

[Winfield]

Okay, Thank you for that first example. My boss gave me this one here:http://www.aries.homeftp.net/publish/Electronics/concept.jpg
Digital Media is not really that kind of study teaching one how to design
line drivers I am afraid...
Maybe you can interpret the missing bits or the concept he uses there.

You'll need feedback components for the two driver ICs, and don't
forget
the cable-matching resistors Martin pointed out. The receiver end
needs
input-protection stuff, and a real diif amp, see Martin's part
selection -
you likely don't want a power opamp for a receiver!

 

[martin griffith]

Okay, Thank you for that first example. My boss gave me this one here:
http://www.aries.homeftp.net/publish/Electronics/concept.jpg
Digital Media is not really that kind of study teaching one how to design
line drivers I am afraid...
Maybe you can interpret the missing bits or the concept he uses there.

have a look at TI's DRV134 data sheet.




martin

 

[martin griffith]

Okay, Thank you for that first example. My boss gave me this one here:
http://www.aries.homeftp.net/publish/Electronics/concept.jpg
Digital Media is not really that kind of study teaching one how to design
line drivers I am afraid...
Maybe you can interpret the missing bits or the concept he uses there.

a bit more here
http://es.geocities.com/mart_in_medina/P_CAD2.pdf


martin

 

[Rüdiger Leibrandt]

You'll need feedback components for the two driver ICs, and don't
forget
the cable-matching resistors Martin pointed out. The receiver end
needs
input-protection stuff, and a real diif amp, see Martin's part
selection -
you likely don't want a power opamp for a receiver!

The single opamp on the right I added just to get the concept clear - the
PCI-Card has variable impedance differential inputs, the software can set a
whole range of details like input capacitance, input impedance and so on. I
just have my headaches with the part on the K-Band transceiver's side, how
to get the input-levels configured correctly - I never before worked with
differential line drivers other then resoldering ethernet-sockets.

 

[martin griffith]

On Wed, 18 Apr 2007 12:42:52 +0200, in sci.electronics.design Rüdiger

differential line drivers other then resoldering ethernet-sockets.

Cut off the connectors and use a terminal strip. It has the added
advantage that nobody will steal the cable for expanding the network


martin

 

[Eeyore]

have a look at TI's DRV134 data sheet.

Also see my response in another group. It's similar but uses the TL072.

Graham

 

[Rüdiger Leibrandt]

Okay - the last word after a lot of datasheet reading is now a MAX435 - it's
a wideband Transconductance Amplifier.
I just need to find out how to give it a variable gain without ruining the
impedance of it's output, and then I am fine.

Thank you all very much for the help you provided to me!

 

[john jardine]

Okay - the last word after a lot of datasheet reading is now a MAX435 - it's
a wideband Transconductance Amplifier.
I just need to find out how to give it a variable gain without ruining the
impedance of it's output, and then I am fine.

Thank you all very much for the help you provided to me!

Well a postscript then
Looks like the Maxim thing is obsolete.
You'd have been better off with an AD8132.

 

[Rüdiger Leibrandt]

Well a postscript then
Looks like the Maxim thing is obsolete.
You'd have been better off with an AD8132.

We have 30 of those Maxim IC's laying idly around - someday we would have
thrown them away otherwise.... Thats active trash-reduction

 

[joseph2k]

Ai - sorry!
input signal is "a few" millivolt ( usually - I will need to have potis to
fine-tune the amplification I think ) but can go up to as much as 5 Volts
in the seldom extreme. The lines are an ethernet Category 5 cable, 8
meters, running straight into a ( not defined ) differential input of a
PCI-datasampling-card in a PC. The sensor is a small K-Band transceiver (
aka radar from InnoSent ).
I do not know where to add the resistors to shape and tune the output of
the combined opamps. I'm lost if and where to add the poti's to control
the amplification and so on. I hope that there might be a website with an
example circuit which I could adapt to my needs.

Having read this thread so far, it sounds like going for 4 to 20 mA SCADA
transmitter is the way to go.