Decrease a 0.20V signal down to 0.06V from analog sensor going to microcontroller for ADC reading without losing sensitivity range

Thread Starter

fe7565

Joined Aug 13, 2016
9
I am trying to interface a sensor with 200mV-350mV span to replace a sensor with a 60mV-150mV span. Do not want to lose the sensitivity of the (voltage => frequency) signal.

The signal is DC and AC (depending in operation mode) and has 2-10mAh. It's going to an AVR or Cortex microcontroller's ADC input for reading.

Basically, I would like the new sensor with the 200mV resting signal to replace another sensor with a 60mV resting signal and able to still have a close to 150mV sensing span.

Thank you for your assistance!
 

ericgibbs

Joined Jan 29, 2010
9,320
hi fe,
One option would be a OPA amp to lower the to 200/350mV to 60/150mV
Please post more details of the signal input, , you say AC/DC, frequency.. etc.
E

Edited:
 
Last edited:

Thread Starter

fe7565

Joined Aug 13, 2016
9
Thank you for the reply. The signal is a doppler voltage (can be +/- in amplitude) and proportional to the frequency shift caused by movement in front of the sensor. The sensor output signal frequency can be 0-65kHz.

Please note that the circuit below has an op amp that (I think) cannot handle such frequencies, and it will be replaced with a 50 MHZ or higher bandwidth-product dual Op Amp. Also the two 2,2nF capacitors for the gain stages are either eliminated or replaced with one in the very low pF range.

1578389184565.png
 
Last edited:

Thread Starter

fe7565

Joined Aug 13, 2016
9
My apologies, there are several corrections:

The original sensor: V supply 5 V and DC offset is 60mV (it transmits this level of voltage at rest) and sensor range-span reacting to movement is +/- 60-150 uV (micro-Volts!)

The new sensor: V supply 3V and DC offset is 1.8V (it transmits this level of voltage at rest) and sensor range-span reacting to movement is +/- 120-360 uV (micro-Volts!)

The dual Op Amp to be used will have a combined gain of about 12,000-15,000.
 

Alec_t

Joined Sep 17, 2013
10,708
If you just reduce the value of one of the feedback resistors from 1 Megohm (not 1 milliohm as shown!) to 500kohm you will get the same output as from a 60-180uV sensor. Close enough?
Rather than modify the circuit, couldn't you make software adjustments in the Arduino?
 
Last edited:

Thread Starter

fe7565

Joined Aug 13, 2016
9
Thank you all for the responses. The sensor is the HB-100 10.525 Ghz radar doppler with a 60-150uV span and a 0.06V DC offset output feed to the and the new sensor is a 24 GHZ sensor with a 1.8V DC offset output to the MCU via the OP AMP circuit.

I rather use the existing OP AMP circuit that came with the HB-100 sensor and instead just try to condition the new sensor's signal (120-360uV with a 1.8V DC offset) to a similar envelope as the original sensor is.

I think the main issue here is being able to bring the new sensor's 1.8V DC offset down to the 0.6V DC offset of the old sensor without significantly affecting the uV level sensing span/range.

Of course, the reading will have to be adjusted in the code since the old sensor is is 10GHZ and the new one is 24GHZ...and the values will double.

Attached old and new sensor data.
 

Attachments

Sensacell

Joined Jun 19, 2012
2,544
The DC output offset is totally irrelevant - the amplifier is AC coupled by the input capacitors which totally block any DC offset.

Your only concern with this circuit is adjusting the overall gain to provide the appropriate sensitivity.
No need for a higher GPBW - the LM324 is a bit noisy, there are better modern units.

Where do you get the idea that the sensor will output 65 kHz? That would be one very fast-moving object!
 

Thread Starter

fe7565

Joined Aug 13, 2016
9
The DC output offset is totally irrelevant - the amplifier is AC coupled by the input capacitors which totally block any DC offset.

Your only concern with this circuit is adjusting the overall gain to provide the appropriate sensitivity.
No need for a higher GPBW - the LM324 is a bit noisy, there are better modern units.

Where do you get the idea that the sensor will output 65 kHz? That would be one very fast-moving object!
I looked up the concept of DC and AC coupled: https://www.quora.com/What-does-DC-coupled-amplifier-and-AC-coupled-amplifier-exactly-mean-What-are-the-differences-between-them-and-when-to-use-which-one

Makes sense, you are correct, thank you for pointing it out. Although the circuit that I may end up using (posted above) is AC coupled, I am not sure if the original circuit that came with the HB-100 sensor with its 0.06V resting signal is AC or DC coupled. I will check the first few traces after the sensors to see what components were used. Do not have the circuit diagram for it.

I tried to interface (to the original circuit that the HB-100 came with, which is a chronograph) another 10GHZ sensor that has a 1.6V DC offset resting voltage output, but did not work. When I used a third 24GHZ sensor with also a 0.06V DC offset resting output voltage, it worked. Of course the reading registered around twice the actual speed because the third sensor was 24GHZ vs 10GHZ)

The 15KHZ-65KHZ OP AMP envelope estimation came from the maximum velocity envelope, which is around 1,200 fps but will never reach. Supposed to measure the velocity of airsoft/pellets propelled by high-pressure compressed air in the 400-900fps range.

(Some of the terms and concepts I used are probably incorrect, my apologies)
 

ericgibbs

Joined Jan 29, 2010
9,320
hi fe,
This LTSpice simulation is of your original circuit and signal, but using better OPA's also 100pf across the Rfb's
Check it for accuracy and if possible post a 'wave form' shot of the operating input signal, I can then refine the sim.
E
AAA 110 08.52.gif
 
Last edited:

Alec_t

Joined Sep 17, 2013
10,708
My understanding of the new sensor datasheet is that a signal in the range 120-360uV is expected from a target of area 0.5m^2 at a distance of 5m and approaching (or moving away from) the sensor directly. You will get a much lower amplitude signal from an airsoft pellet, unless the pellet is extremely close to the sensor. This means you may need an amplifier with more than two gain stages, and perhaps sophisticated noise filtering.
Where will you position the sensor in relation to the line of flight of the pellets? Note that the signal will be reduced even further if the pellet is not moving directly towards (or away from) the sensor.
A pellet moving at 900fps towards the sensor will result in a Doppler shift of ~43kHz, so the LM324 is most unlikely to give enough amplification.
 
Last edited:

Thread Starter

fe7565

Joined Aug 13, 2016
9
My apologies for my tardiness in replying, I only had a chance to revisit this project a couple of days ago.

@ericgibbs Thank you very much for the simulation circuit. It's a great way to visualize it. This coming weekend, I will try to take a reading of the sensor output with a pellet at around 380 fps using Audacity. I have an old oscilloscope, but very wobbly about using it. I also have a DS0201 mini-oscilloscope, but not sure if is fast enough.
The signal range for the op-amp circuit based on my calculations is between 15khz-65khz (pellet speed range). I read that for a good op-amp circuit the frequency response should be 4-5 times higher in order to keep the noise in check.

@Alec_t Based on my research you are correct. The signal would be affected by the pellet's path and distance to the sensor. The pellet in my case would be about 10-50cm from the sensor and would be moving directly away from the sensor without any angles. So the angle calculation is not a factor in the doppler shift. Byt the target area where the pellet has to pass through for reading is quite small so alignment of the pellet's path with the sensor's reading angles (horizontal and vertical) are very important.
 
Top