Transformer for Balanced to Unbalanced Adapter?

[merlinb]

Your simplest option would be a two-opamp differential amp, using a JFET opamp (e.g. TL072) and 10M resistors for R5/R6. As shown, this circuit needs a bipolar supply, but you could always do a single-supply version. Rg is an optional attenuation resistor (gain control).

[johnyradio]

Thx!

yes, i figured out it needed to be a differential amp.

i'm seeking a lower-parts-count method than yours. found these. My only concern is ensuring they have high enough input impedance for the piezos, which i hope might simply be a matter of selecting the right parts?







The last one requires a constant current source (the linked circles at the bottom going to ground). I've read the simplest constant-current source is a JFET with its gate attached to its source, plus a resistor.

[merlinb]

Three transistors is three parts with 9 legs in total. A dual opamp is one part with 8 legs. Which one is simpler?
My suggestion would need one IC, 8 resistors, 4 caps, battery, bare minimum. Is that really too many parts?

(None of the circuit snippets you posted will have enough input impedance. Modifying them to have enough input impedance will result in circuits with a LOT more parts than the one I posted!  ).

[johnyradio]

these seem like all good circuits worth trying. This article describes yours and mine:
http://www.circuitstoday.com/differential-amplifier

The article also states "Differential amplifiers have high common mode rejection ratio (CMRR) and high input impedance" implying that all three circuits have high input impedance.

you're right, your 2-opamp circuit is certainly simple! but i believe i count more total solder-points on the 2-opamp circuit than any of the others. Not by a lot of course, just about 7. (not counting the external connections or battery, as that would be the same for all methods).

(Your single opamp diff amp won't have enough input impedance. At least, not without being extraordinarily noisy).

For purely my educational interest, how would you calculate the input impedance of the above circuits?
How would you increase the input impedance of the single op-amp circuit?

many thanks

[R.G.]

Whatever you do, you need to ensure that both inputs are high impedance. By "high impedance" I mean "1M resistors look like short circuits compared to the inputs". In practice, that means FET input opamps and both inputs going to a "+" input, not through a resistor to a "-" input.

For simple opamps, the series resistor into the "-" input is the input impedance, for all intents and purposes. For the "+" input, the resistor to the bias voltage is in parallel with the input pin, which in a good FET input opamp is so big that you may as well say it's infinite. There are multiple many caveats and gotchas hidden under these points, just so you know.

This is the classical instrumentation amplifier:
http://en.wikipedia.org/wiki/Instrumentation_amplifier
Two "+" inputs, one for each side of the input signal, and these two amps then used as the inputs to a differential amplifier. Not shown is any biasing resistors needed. If your piezos don't leak at least some DC, you will have to add high value resistors to some DC reference voltage.

Again, your specific circuit situation matters. What signal level will you get from the piezos? Piezos tend to have large outputs compared to wound inductive pickups, so they typically need much lower gains. If you have high signal levels (~1v) then you have lower issues with noise, and things get easier, including being able to use the simpler two-opamp diffamp.

If you have large common mode interference issues, it pushes you to the three-opamp version. TI has a three-opamp instrumentation amplifier circuit in a single 8-pin package with laser trimmed accuracy for a few bucks if you really want smaller.

None of the circuits you posted are particularly high impedance or high accuracy. As merlin says, go for a decent IC implementation unless you know your application doesn't require it.

Again, you haven't said what you intend to do other than connect a balanced piezo pickup, so advice is merely speculation.

[johnyradio]

thanks, R.G., that's extremely helpful.

"the series resistor into the "-" input is the input impedance"

then wouldn't a large enough resistor provide a high-impedance input?

"If your piezos don't leak at least some DC, you will have to add high value resistors to some DC reference voltage."

--So, if the piezos DO leak some DC, then i do NOT need a DC reference V?

"Piezos tend to have large outputs compared to wound inductive pickups, so they typically need much lower gains. If you have high signal levels (~1v) then you have lower issues with noise, and things get easier, including being able to use the simpler two-opamp diffamp."

--ok, so the 2-opamp version is an option, unless i'm amplifying extremely quiet sounds?

"If you have large common mode interference issues, it pushes you to the three-opamp version. "

I've read that "Common Mode Rejection Ratio is a measure of how well the opamp rejects signals applied to both inputs simultaneously.Common mode rejection depends on the available open loop gain." So, that depends on the specs of the opamp itself, and not on my input or my circuit, correct?

"TI has a three-opamp instrumentation amplifier circuit in a single 8-pin package with laser trimmed accuracy for a few bucks if you really want smaller."

Cool! This chip appears to be 3 opamps on a chip. Would you recommend it?
http://www.analog.com/static/imported-files/data_sheets/AD620.pdf

"Again, you haven't said what you intend to do other than connect a balanced piezo pickup, so advice is merely speculation."

ok, thanks for specifying, because i did not know what info you needed. I will have two applications:

• amplifying acoustic musical instruments
• amplifying very quiet sounds with very close micing

many thanks for your detailed answers.

[R.G.]

I'd look at this: http://www.ti.com/lit/ds/symlink/ina217.pdf

The INA217 has 60M + input impedance both differential and common mode, goes down to a gain of one (which some of them don't), and runs from 9V to 36V, which some of them don't.

If you're SURE you'll never get over 5V or 12V or whatever for power supply voltages, you can go to one of the other instrument amps that specify lower power supplies; you could also put a 78L05 in there to regulate it down if you're not sure. But I like the INA217. It's high impedance and quite low noise on its own. Good other specs.

It's an 8-pin DIP, simple. The microphone amp application circuit in the datasheet will give you some ideas, although it is intended for 600 ohm use. For piezo use, you'll need biasing resistors in the 1M to 10M range, and those should be metal film. Simple resistor biasing and an integrated instrument amp gives up the possibility of low-noise biasing by bootstrapping, but you may be OK. They're a bit pricey, about $6.80 at Mouser, but they make for a simple, high performance setup.

[PRR]

>> "the series resistor into the "-" input is the input impedance"
> then wouldn't a large enough resistor provide a high-impedance input?

AND loss of signal AND high hiss-voltage.

> amplifying very quiet sounds

Then low-low-Hiss is a key design goal.

INA217 is optimized for 200 ohms NOT MegOhms.

Piezos are a lot more like Capacitor mikes. Which "always" have a "head amp" inside the shell. Piezos can do short cables without a head-buffer. Cap-mike is 10-100pFd, large piezos are 100pFd-1000pFd. Compare with 30pFd/foot of cable, a piezo will drive a length of cable that would overwhelm a capacitor-mike capsule.

But still.... capacitor mike head-amps should be your inspiration.

And parts-count be darned. Do you want a bad solution super-quick? Most of the ideas so far will be disappointing enough that you will want to build something else.

And IMHO that dual-piezo is not "balanced" for all practical applications. But whatever.

Use two JFETs or two parts of TL072 to buffer each disk separately. Then to a differential amplifier (may be '217, or the 1-opamp plan) to extract the difference and gain-up.

> amplifying very quiet sounds

There is no magic bullet. It is not clear to me that the piezos have any advantage over, say, AKG 414, with 14dB SPL self-noise level at output which won't be hissed-up by a decent mike-amp like INA217. Of course there is the $2 versus $1,500 difference. However an EV-635A or even a SM-58 into an INA217 will give a hiss-level which few rooms reach.

BTW, that NEETS-Modules "differential amp" is bogus on the face of it. It shows +/-2V peak signal directly across a BJT's B-E junction. Which saturates nearer 0.020V. It could only be linear for VERY small voltages, under 10mV. So someone at NEETS wasn't thinking, or mis-copied a flawed plan, for poor tutorial purpose.

[R.G.]

AND loss of signal AND high hiss-voltage.
INA217 is optimized for 200 ohms NOT MegOhms.

Yep, hence my comments on bootstrapping being good to eliminate issues with bias resistors. An IA can be set up to get good bootstrapping, but it gets trickier.

The real bottom line is that overall it is difficult to amplify low signal, high impedance sources with low noise. It's a difficult task.

At least we got him off transformers, which would be nearly impossible, and certainly impractical.