Amplify 1mV output accelerometer

Thread Starter

Derek Carson

Joined Jan 18, 2019
4
Hello all. I’ve seen a number of similar threads but not exactly sure which direction to go. I’m trying to amplify a piezoelectric accelerometer which outputs 1mV/g. But I’m not concerned with high frequency or even varying frequency. This is to simply validate the accelerometer is outputting 1mV when it is tested using a 1g accel shaker with a constant frequency of 159.4Hz. I’d like the gain to be 100 at minimum but most likely 1000 would be preferred. I’ve looked at a few pre-built packages with instrumentation amplifiers such as the AD620 and AD623 but not sure if that’s they way to go.
 

Jon Chandler

Joined Jun 12, 2008
1,073
1 mV/g isn't much signal to work with, unless you're doing shock testing.

Is it a piezo without internal electronics? If so, you'll probably need a charge amp or charge converter, connected with special low-noise cable designed for the purpose.
 

Thread Starter

Derek Carson

Joined Jan 18, 2019
4
Thank you for the responses. The accel is a Dytran 3143D10.

For context, I am already able to acquire the signal with our current data acquisition setup that has been fully qualified. But the workflow of the higher-level system occurs in such a manner that the accel is not used on that system until near the end of the build. So what I'm attempting to do is devise a cheaper/more mobile method of performing a basic functionality check of these accelerometer assemblies, which are comprised of the 3143D10 + cable fabricated for this accel. For the purpose of catching off-nominal behavior earlier in the system build. Which very rarely happens, but it's a considerable problem when found later, so it's worth the up-front testing.

For this I have a 1g rms accel shaker set at 159.4Hz, and would like to amplify that signal enough to acquire with a USB data acquisition device where the output will be displayed through LabVIEW.
 

Attachments

MisterBill2

Joined Jan 23, 2018
19,052
OK, it seems that what you need is an amplifier with a gain of 100 and no noise or instability. If you can obtain a KISTLER charge amplifier that will be the solution. Such a device may be available as used equipment because I used them back in 1974 in a sound and vibration laboratory.
But a simple amplifier with response down to DC, a very high input impedance, and a gain of 100 is what you require.
 

Thread Starter

Derek Carson

Joined Jan 18, 2019
4
OK, it seems that what you need is an amplifier with a gain of 100 and no noise or instability. If you can obtain a KISTLER charge amplifier that will be the solution. Such a device may be available as used equipment because I used them back in 1974 in a sound and vibration laboratory.
But a simple amplifier with response down to DC, a very high input impedance, and a gain of 100 is what you require.
Thank you MisterBill. I shall definitely look into that. Much appreciated!
 

MisterBill2

Joined Jan 23, 2018
19,052
The reason for suggesting a gain of 100 is that it should keep the signal within the low distortion capabilities of the computer sound-card input.
What is unknown is if there are any input bias voltage requirements of that accelerometer.
For the amplifier circuit and selection assistance I suggest looking at the applications circuits of Maxim semiconductors or Analog Devices semiconductors. Application notes usually are more reliable than other sources because they want their products to provide the best possible performance.
 

Jon Chandler

Joined Jun 12, 2008
1,073
If you can obtain a KISTLER charge amplifier that will be the solution....
The Dytran 3143D10 accelerometer has ICP-type* integrated electronics and a low impedence output. A charge amp isn't needed (or useful) in this case

Just to be sure.... the accelerometer is provided with constant-current power, yes?

*ICP is a registered trademark of PCB Piezotronics (I have been warned by PCB that I must say this). Different manufactures call this by different names but the implementations are all similar. An integral amp in the accelerometer converts the high impedance charge output of the accelerometer crystal to a low impedance voltage output. Power is provided by a constant-current source on the signal line which is biased by the accelerometer to ~12V. The AC accelerometer output rides on this 12v level; the DC bias voltage is removed at the instrument end with a DC blocking cap.

[Just in case anybody wants to argue (here? Imagine that.), yes in post #3, I did question if a charge amp might be needed. This was before the accelerometer model was stated in post #5, with a link to the datasheet.]
 

MisterBill2

Joined Jan 23, 2018
19,052
The Dytran 3143D10 accelerometer has ICP-type* integrated electronics and a low impedence output. A charge amp isn't needed (or useful) in this case

Just to be sure.... the accelerometer is provided with constant-current power, yes?

*ICP is a registered trademark of PCB Piezotronics (I have been warned by PCB that I must say this). Different manufactures call this by different names but the implementations are all similar. An integral amp in the accelerometer converts the high impedance charge output of the accelerometer crystal to a low impedance voltage output. Power is provided by a constant-current source on the signal line which is biased by the accelerometer to ~12V. The AC accelerometer output rides on this 12v level; the DC bias voltage is removed at the instrument end with a DC blocking cap.

[Just in case anybody wants to argue (here? Imagine that.), yes in post #3, I did question if a charge amp might be needed. This was before the accelerometer model was stated in post #5, with a link to the datasheet.]
OK, with no idea of just which type of acceleromter was being used, the charge amp was a fair guess, although the good news is that it was incorrect. So now what needs to be known is what that regulated constant current would be, and what the intended load resistor value should be. What has changed is that the amplifier no longer needs to have a very high input impedance.
 

Thread Starter

Derek Carson

Joined Jan 18, 2019
4
The Dytran 3143D10 accelerometer has ICP-type* integrated electronics and a low impedence output. A charge amp isn't needed (or useful) in this case

Just to be sure.... the accelerometer is provided with constant-current power, yes?

*ICP is a registered trademark of PCB Piezotronics (I have been warned by PCB that I must say this). Different manufactures call this by different names but the implementations are all similar. An integral amp in the accelerometer converts the high impedance charge output of the accelerometer crystal to a low impedance voltage output. Power is provided by a constant-current source on the signal line which is biased by the accelerometer to ~12V. The AC accelerometer output rides on this 12v level; the DC bias voltage is removed at the instrument end with a DC blocking cap.

[Just in case anybody wants to argue (here? Imagine that.), yes in post #3, I did question if a charge amp might be needed. This was before the accelerometer model was stated in post #5, with a link to the datasheet.]
Yes, the accelerometer excitation from the final data acquisition chassis is 3.6mA. Thank you for your response.
 
Top