Using an oscope on a battery powered circuit

Thread Starter

BackyardBrains

Joined Aug 26, 2010
33
I have a 9V battery powered circuit that we use to help kids learn about the brain. Students place two pins into a living cockroach leg, and my circuit amplifies the electrical "spikes" in the leg. A photo of the finished product (without a cockroach leg)


It works, but the speaker is not nearly loud enough. I am trying to debug it carefully to see what the outputs of each of the stages are. I am suspecting there might be a DC offset that is being amplified, and clipping the signal. I only have a Tektronix oscilliscope that plugs into the wall. When I hook up the probe to the circuit with a leg on it, it distorts the signal badly.

Are there tips to see what is happening within the circuit? Also, any kind suggestions for improvement would be very much appreciated. Schematic is attached.

Thank you, Greg
 

Attachments

Audioguru

Joined Dec 20, 2007
11,248
You are shorting pin 3 of the AD623 to Vcc/2 instead of feeding both pin 2 and pin 3 though matched 100k resistors from Vcc/2 to balance them so the AD623 can cancel mains hum.
Maybe the ordinary input wires should be shielded audio cable.

Your speaker is very small and might have an impedance too high for much power. It is a little earphone.
 

Thread Starter

BackyardBrains

Joined Aug 26, 2010
33
You are shorting pin 3 of the AD623 to Vcc/2 instead of feeding both pin 2 and pin 3 though matched 100k resistors from Vcc/2 to balance them so the AD623 can cancel mains hum.
Maybe the ordinary input wires should be shielded audio cable.

Your speaker is very small and might have an impedance too high for much power. It is a little earphone.

Thanks so much. Can you explain this again? You say I should have a large resistor (100k) between Vcc/2 and pin2 and another 100k between Vcc/2 and pin3?

For background. I wasn't sure why I needed to short pin3 to VCC/2. I did that through trial and error. That was the only way it would work. :) But I thought theoretically by reading the specs, that I shouldn't need that.

I can try what you suggest. What is the thinking behind it?

Greg
 

thatoneguy

Joined Feb 19, 2009
6,359
Try switching the probes to x10 to increase input impedance, you may need to change volts/div depending if the scope detects the swtich or not.

Put ground clip(s) on battery ground, then set channel 2 to invert and add. Then channels 1 and 2 can be used across the input and output of an amp to easily and clearly measure gain/distortion added by that stage.

--ETA: Datasheet I'm looking at, shows the 100k resistor Audioguru mentioned.
 

Audioguru

Joined Dec 20, 2007
11,248
The instrumentation amplifier IC is designed to cancel interference (mains hum) that appears the same at both inputs. The interference is the same level only if the resistors in series with each input are exactly the same (1% or better). Your circuit had the inputs extremely unbalanced with no bias voltage on one and picked up a lot of interference.
 

Thread Starter

BackyardBrains

Joined Aug 26, 2010
33
Put ground clip(s) on battery ground, then set channel 2 to invert and add. Then channels 1 and 2 can be used across the input and output of an amp to easily and clearly measure gain/distortion added by that stage.
Thanks. This was a brilliant suggestion. I've been debugging all night!

One thing I noticed. When I put the LM386 on a power supply (a separate battery), everything works and I can increase the gain so that it is loud and stable. When I connect the LM386 power and ground to the rest of the circuit, it becomes unstable at > 25% the gain of the dual-power solution.

Strange days... but at least I can measure now. Any ideas as to why this would be? Could it be the virtual ground (VCC/2) that I am using in the prior stages, or is it just simply not enough current to the LM386.

Thanks for your help!
 

Audioguru

Joined Dec 20, 2007
11,248
Your virtual ground circuit does not have a filter capacitor to ground at the junction of R1 and R2 so its voltage is probably jumping up and down when the LM386 draws current.
 

Thread Starter

BackyardBrains

Joined Aug 26, 2010
33
Your virtual ground circuit does not have a filter capacitor to ground at the junction of R1 and R2 so its voltage is probably jumping up and down when the LM386 draws current.
Thanks for the suggestion. I had originally placed C1 and C2 there for that reason. Should those caps be before U4A (pin3 not pin2) ? Also, I had a mistake on those particular component values (amended in the attachment). C1 and C2 I changed a while ago to 10uF.
 

Attachments

Audioguru

Joined Dec 20, 2007
11,248
Your C1 and C2 are not filters. The output of the virtual ground opamp bounces around all over the place since its input is not filtered.

The opamp for the virtual ground doesn't do anything and can be removed. If the junction of R1 and R2 has a filter capacitor to ground then this point can easily feed the "half the supply voltage" reference to the high resistance inputs of the opamps.
 

Thread Starter

BackyardBrains

Joined Aug 26, 2010
33
Your C1 and C2 are not filters. The output of the virtual ground opamp bounces around all over the place since its input is not filtered.

The opamp for the virtual ground doesn't do anything and can be removed. If the junction of R1 and R2 has a filter capacitor to ground then this point can easily feed the "half the supply voltage" reference to the high resistance inputs of the opamps.
Thanks. I will try this out. What values would you recommend? And why? I've seen a range of values being used for this, and was never fully sure why people chose a value. Was it trial and error during development?

That is good news that I can remove the U4A op-amp. I wanted to use another op-amp in a later stage anyway... was about to switch to a quad chip.
 

Audioguru

Joined Dec 20, 2007
11,248
Look at the max input bias current of the opamp that needs Vcc/2 and the max input reference current of the intrumentation amplifier and add them. The divider current should be at least 15 times more. Use a suitable filter capacitor to filter out battery voltage fluctuations at the divider caused by the supply current to the LM386 amplifier with your load at the lowest expected frequency.
 
Top