0
\$\begingroup\$

I'm trying to build Sparkfun's sound sensor and testing the first stage only. Here's the first stage:

enter image description here

If I scope before C2 and at Pin 1, I get the following when I scream at it:

enter image description here

where Ch1 is the mic input and Ch2 is the opamp output. As you can see, the gain is just 1x where it should actually be 100x.

Here's my breadboard:

enter image description here enter image description here enter image description here

Am I missing something here? Shouldn't I get a 100x gain? I'm using a LM324 (with 5V) rather than a LMV324, but I think they're essentially the same.

Thanks

\$\endgroup\$
  • \$\begingroup\$ In the photo it looks like the 1k0 is touching the metal body of the mic. and the LMV324 is a rail to rail o/p and the lm324 ain't. \$\endgroup\$ – JIm Dearden Jun 4 '15 at 13:23
  • \$\begingroup\$ @JImDearden Thanks. Well, the 1k and the mic aren't actually touching - sorry about the bad angle. If the LM324 isn't a rail to rail, is there something that could be done? \$\endgroup\$ – Kar Jun 4 '15 at 13:34
  • \$\begingroup\$ What I would do first is to check the DC voltage on the opamp's inputs and output. They should all be around half the supply voltage. If that's OK then the circuit should just work (excluding trivial errors). \$\endgroup\$ – Bimpelrekkie Jun 4 '15 at 14:04
  • \$\begingroup\$ What are the mystery leads connected to? Is their ground and Vcc ON THE IC? Is CH1 set to 10x ? What is the scope vertical reading actual eg 50 mV/divn or 500 mV/divn? Is the mic polarity OK, Is the mic resistor in the ground lead and did you intend that. || NBNBNB If there is no gnd or VCC on IC then Vout ~= Vin as it is a francy piece of wire. \$\endgroup\$ – Russell McMahon Jun 4 '15 at 14:05
  • \$\begingroup\$ Looks a bit like you're shorting out your power supply through your 'scope ground leads. Seems like one is connected to GND and the other to Vcc. \$\endgroup\$ – brhans Jun 4 '15 at 14:06
4
\$\begingroup\$

As shown in the photos you have neither power supply pin connected.
Most ICs work very badly in that state :-).

If there is no gnd or VCC on IC then Vout ~= Vin as it is a fancy piece of wire with a 10k to ground and supply. In that case what you see is about what you'd expect. .

Note that pin4 is Vcc and pin 11 is ground.

Not being rail to rail is fine. Maybe increase R5 to about 12K to move DC point away from top rail but not essential.


It's not 100% certain what your stage-off connections go to but if clip at top right is scope-ground it's apparently connected to V+. If psu has same ground they may get annoyed.

It's potentially OK to do but you seem to have Mic R1 in gnd side not Vcc side. This is OK as long as Mic polarity is OK> Is Mic polarity OK?

It's not obvious what your scope shows. Y scale is 50 mV/div and CH2 probe is 10X. That may mean 50 mV/div or 500 mV deep-ending on Rigol philosophy. I've never used a Rigol. CH1 is presumable 10x as well but I think you and it say not.

LM324 input common mode range is about gnd to Vcc-1.5 V so here = about 0 to 3.5V Vin DC point as shown is 2.5V so you have 1v of input swing positively which should be plenty. Vout swings near but not to Vcc - see data sheet.

That should do for now.
Try powering it and report back.


Checklist:

What are the mystery leads in the photos connected to?

Is there ground and Vcc ON THE IC?

Is CH1 set to 10x ?

What is the scope vertical display actual eg 50 mV/divn or 500 mV/divn?

Is the mic polarity OK

Is the mic resistor in the ground lead and did you intend that.

If clip at top right is scope-ground it's apparently connected to V+. If psu has same ground they may get annoyed.

\$\endgroup\$
  • \$\begingroup\$ I actually forgot the ground it during the picture taking! I've grounded it now, but still the same. \$\endgroup\$ – Kar Jun 4 '15 at 14:01
  • \$\begingroup\$ But is there power on the IC - Vcc and ground ? \$\endgroup\$ – Russell McMahon Jun 4 '15 at 14:02
  • \$\begingroup\$ Yeah, there was VCC. It's a white wire actually (sorry about the bad choice of colour!) Does the mic have polarity? \$\endgroup\$ – Kar Jun 4 '15 at 14:06
  • \$\begingroup\$ Having grounded the op amp and fixed my probe, I think I'm getting the proper gain! :) Thanks \$\endgroup\$ – Kar Jun 4 '15 at 16:49
  • \$\begingroup\$ Having power on the opamp usually helps :-). You say "fixed your probe" -> was ground clip connected to Vcc as it appeared to be? \$\endgroup\$ – Russell McMahon Jun 4 '15 at 18:36
2
\$\begingroup\$

I'm going to suggest that you are far better off configuring your preamp as a non-inverting amplifier instead of inverting.

One of the problems is that although most electret mic capsules have about 1k output impedance, some don't. The gain of your amplifier depends on the total impedance from the inverting input to (virtual) ground.

Another problem is that because the output impedance of the capsule is about that 1k value, the impedance seen as a load by the capsule is set by R2 (currently 1k). That causes a further gain reduction.

If you were to capacitively-couple R2 to ground and feed the mic input into the (+) input, most of these problems vanish.

Pick the value of the coupling capacitor from R2 to ground to set your desired low-frequency break point. Note that this capacitor will be fairly large because R2 is such a low value.

I'd also increase the value of R5 & R6 so as to increase the input impedance of the amplifier. Anywhere from 22k to 100k is good. Note that the input impedance is the value of R5 in parallel with R6.

In general, I usually like my load impedance to be at least 10 times the source impedance (bridging connection). Assuming that the output impedance of the mic capsule is about 1k, I'd like to see the input impedance of the amplifier to be at least 10k. That would require 22k resistors for R5 & R6. You can safely go higher if you wish.

\$\endgroup\$
  • \$\begingroup\$ All good advice - but I think he has a major show stoppr apart from the "fine details". Having the scope ground plugged into Vcc - as it appears to be - could well be enough. Stay tuned ... :-) \$\endgroup\$ – Russell McMahon Jun 4 '15 at 16:45

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.