The piezoelectric material responds to the vibration of the bridge with small changes in the electrical field it's part of. An amplifier can convert these small fluctuations into sound.

One big problem with piezo pickups is noise. Some people say that shielding the circuit doesn't help much, but I've found that it does somewhat, although piezo noise tends to be more from lack of a proper ground than from lighting fixtures and such. The noise is akin to a low-frequency hum that sounds to me like it's DC current that has no way to leave the circuit. Isolating the piezo side of the circuit with a transformer unit usually helps. A DI (direct box) can be used for that purpose. So can a preamp unit. Some preamps seem to act as buffers that eliminate piezo hum noise even if you run a 1/4-inch unbalanced output instead of an XLR (balanced a.k.a. properly-grounded) output. But not all do, so you have try different things. A DI, however, is probably a better bet for ground loop suppression.

Grounding the piezo to piece of metal, like I did here, helps a tiny bit, but that's still a floating ground, not an "earth" ground, and is effective really only when the player is touching the strings, which in turn touch the tailpiece, which in turn touches the piezo circuit. In other words, the player is the ground. Update 4-19-2008: Contrary to previous: shielding does seem to make a big difference, at least with the MSI piezo tab I used for this project. Grounding through the instrument to the player doesn't make much difference. DI ground loop isolation is more effective. Doesn't mean it's applicable to all piezos, though - Fishman pickups are different.

Another big problem with piezo pickups is that they can sound tinny and weak. That's often because of impedance mismatches. Piezo pickups have extremely high impedance (AC resistance, denoted by Z) in Megohms, and when they're plugged into an ordinary guitar or bass amplifier, they can sound screechy, because most guitar/bass amps aren't designed to handle inputs with impedances that high. Z matching can be done with a direct box in many cases. I've been using a passive DI in tandem with a preamp unit - more about that below.

A piezo transducer pretty much begs for a preamp of some sort, especially if you're going to be using a cable more than a meter long. There are manufacturers who claim that their piezo pickups don't need a preamp, but I've found that piezo signals always sound way better when they're processed by a preamplifier before reaching the amplifier. This is partly because of the impedance issue mentioned above - many preamp units offer very high input Z (4 to 10 MOhms), and output Z matched more to an amp or a board. In addition, many preamp units have both an instrument-level 1/4" output (Z ~= 100k to 220k Ohms, sometimes balanced, sometimes not), and an XLR output (Z ~= 600 Ohms) for direct connection to a mixing board.

This is the signal chain I've been using lately:

1. Cello into SansAmp Para DI preamp (Z ~= 4.7 MOhms, seems to be perfect for this pickup).

2. XLR out from Para DI into the XLR output of a Radial passive DI box (600 Ohms to 600 Ohms).

3. 1/4" input from DI into a Roland DB-700 bass combo amplifier (100kOhms to amp).

***Note that the DI box is wired backwards. This is known as reamping the input signal.

Although I could just run the preamp's 1/4" output straight to the amplifier, I've found that I get a cleaner and much stronger signal through the backwards DI box. The DI also deals with the ground-loop problem mentioned above.

There are quite a few DIY folks out there who have done a lot of interesting things with instrument amplification. Just type something like "piezo impedance circuit" into Google and you'll be swamped with information. Although I tend to use amps with high-impedance inputs designed for acoustic instruments, I may build an impedance-matching buffer preamp in the near future anyway - perhaps something like the Mint Box Buffer.

Pending: finishing on tailpiece, the real piezo pickup, end pin brace.