Back to Basics: Sample/Track and Hold Amplifiers

Last week we went into details of ADC and how source impedance can effect timing. Now there will be cases where you are looking to sample two independent signals, say bio-med signals, on 2 separate analog channels on an MCU with a single ADC.

With a time-multiplexed SAR ADC, the internal sampling capacitor first charges from channel A. The mux flips to channel B. If the source is not low impedance or acquisition time is short, some charge from A rides into B. You see a ghost voltage of the previous channel. One fix is to wait after each switch so it settles. That reduces ghosting, but it creates timing skew between channels. Meaning it’s non-time synchronised.

Track and hold Amplifier

Sometimes this skew is not acceptable. In bio-med you might want ECG and PPG at the same instant. This is where sample or track-and-hold stages help. You place a T/H per channel and drive all holds with one edge. In track the output follows the input. On hold the switch opens, a small capacitor freezes the value, and a buffer drives the ADC. The ADC then converts sequentially, but the samples come from the same instant.

There are few parameters to keep in mind while selecting one, Aperture delay and Jitter set the exact instant you grab the signal. Lower jitter means cleaner SNR and less channel-to-channel timing error. Droop is how the held voltage slowly sags because of leakage. Too much droop changes readings. Hold step is the small jump when the switch opens due to charge injection. Big steps look like offsets or spikes. I usually aim for low jitter, low droop, and small hold step, so both channels look clean and truly simultaneous.

So next time when you want absolutely 2 signals at the same time look into sample and hold Amplifiers. TI and Analog devices have a few in their portfolio.

If you liked the post, Share it with your friends!

Advanced Tech: PicoLeaf

I only recently learned about the PicoLeaf tech from Murata. It’s a cool bit of tech worth discussing. Last time we discussed piezos and PicoLeaf takes this piezo tech to the next level.

PicoLeaf is a flexible piezoelectric film sensor that can detect bend, twist, press and vibration. It’s thin(0.2mm) and tiny, so it wraps around curved parts. It’s even sensitive enough for µm-level motion, which is great for HMI and biosignal acquisition use without large parts. The film uses polylactic acid. Yes, PLA, the same stuff you use for FDM 3D printing. Murata orients the PLA to make it piezoelectric. It can be mounted on a device with a double-sided adhesives.

Pico Leaf

Like any piezo, it produces charge in proportion to strain rate. We can read it with a charge (I/V) amplifier (with a mid-supply bias) and then add gain and filtering to pass it to an ADC. All the taps/presses get captured nicely via the microcontrollers ADC at reasonable sampling rates. It can detect displacement direction(press/release) and displacement velocity based on signal amplitude. Its non-pyroelectric also, meaning no value drift because of heat. Power consumption is zero as its passive, so only power needed is for the driver portion in µA range.

Pico Leaf

The thin form factor opens up many applications. Since PLA transmits light, it can even sit under clear panels, making it useful for industrial touch displays that require a firm push. Sensitivity may be lower than capacitive touch, since the datasheet does not specify minimum activation force. It is well suited for flexible gloves to detect finger bends or sign gestures. If the claims are correct, it could be a breakthrough in robotic hands for grip sensing. With all the humanoids that planned in the next few years, PicoLeaf would fit right in. Murata’s tests shows it survives 500k bend cycles, which is great.

What I would love to see is whether it can also operate in reverse like a conventional piezo, producing mechanical motion when voltage is applied. With this form factor, that would be a killer feature.

If you liked the post, Share it with your friends!
1 2 3 86