But it was a DC amplifier, with a gain of 20, with a maximum offset of125mV

Possible but you don't know until it is tried. TS is probably doing this right now.

 

On
https://www.pcb.com/products?m=482c05
footnote 1 says
"[1] With >= 1M ohm input impedance of readout device."

Not particularly clear what it means, but enough to sew the seeds of doubt in my mind.
It sort-of-implies that it should be driving a >1Meg input impedance, but maybe not.
If I had to design an interface without full knowledge of the "preamp". I'd choose a low-bias current op-amp.
I'm not even sure it is a preamp. What preamp would only have 5mV output?
On the other hand - it seems to have a 4mA output current to the sensor. Piezos don't need a bias current so what's that for? I've seen electrostatic transducers (Polaroid used to make them) which do need a bias current. Am I reading too much into this and seeing problems that don't exist?

Some of the experiments we're running use the transducers at well under 1% of their capabilities. The full-scale outputs would be on the order of 5V. Hence, the 5mV output.

 

Below is the LTspice simulation of an example op amp/comparator circuit:

View attachment 224777

We had most of the parts on hand to build this but I had to order the LT1212--they should arrive from Digikey by the weekend.

BTW, what is the purpose of the 10k resistor at the output of the 393 between +5V and Out? I've seen it in the examples on the 393 datasheet before but I don't get what purpose it serves. Seems like the voltage at Out would be unaffected... (Typical ME trying to do EE stuff...)

 

what is the purpose of the 10k resistor at the output of the 393 between +5V and Out?

The output of that comparator is a transistor open-collector (see below) so you need a pull-up resistor to get a voltage output.

 

What is the timing resolution you are trying to resolve (?) ? My thought is that the LM393 is a relatively slow device, both in responding to a small change at the input and also in the speed of the output stage, particularly when turning off to make a positive-going edge. Do you have a feel for how fast the output edge has to be to capture the timing accuracy you need? An alternative to the LM393 output comparator is a fast opamp with near rail-to-rail performance. The dual version of the part Wally used is the LT1211. 1/2 for the input amplifier and 1/2 for the output comparator looks good to me, and you get a faster rising edge with no pull up resistor.

ak

 

From the graph in post #8 would estimate the rise time at about 150us. well above the 1.3us response time of the LM393 but would add some hysteresis.
SG

 

The output of that comparator is a transistor open-collector (see below) so you need a pull-up resistor to get a voltage output.
View attachment 225092

Ah! Thank you!

 

The dual version of the part Wally used is the LT1211. 1/2 for the input amplifier and 1/2 for the output comparator looks good to me, and you get a faster rising edge with no pull up resistor.

Not necessarily.
Op amps can take a long time to come out of saturation due to the charge-up of the compensation capacitor.

 

What is the timing resolution you are trying to resolve (?) ? My thought is that the LM393 is a relatively slow device, both in responding to a small change at the input and also in the speed of the output stage, particularly when turning off to make a positive-going edge. Do you have a feel for how fast the output edge has to be to capture the timing accuracy you need? An alternative to the LM393 output comparator is a fast opamp with near rail-to-rail performance. The dual version of the part Wally used is the LT1211. 1/2 for the input amplifier and 1/2 for the output comparator looks good to me, and you get a faster rising edge with no pull up resistor.

ak

We're just using this to trigger a series of timers. I need to make 4 of the circuits that Crutschow suggested. As long as the outputs from each circuit have similar rise times to trigger the TTL inputs (within a few microseconds of each other), the slope of the rise won't affect the measurement we're trying to make.

 

Just occurred to me today that I didn't follow up. The circuit suggested by Crutschow worked out great. Thanks for the help!

 

Thanks for the follow up.
Glad to know it worked.