[HBQRP] Who's an op amp expert?

[Dale Botkin]

Supply voltage is non-negotiable, that's coming from another part of the 
board.  And I selected that part for its ability to run at 3.3V and 
below, so it should be fine.  I actually do have a 50K pot in place now 
and have played with the bias voltage; I can clip the tops or bottoms 
off of the output wave form as I expected, but nothing else.

I tried Sean's ideas as well...  not surprisingly there was no change, 
other than reducing the input voltage a bit with a smaller coupling 
capacitor.  I made the feedback resistor 51K and the series resistor 
10K.  Now the gain is (roughly) 5...  Hmmm.  Very close to the same 
result, but why does it look like the output signal is being clipped?  
That would only happen if the output voltage is hitting the rails, but 
it says...  oh crud...  320 mV...  which is....

Sean nailed it -- it was indeed a "forehead slapper".  After last 
weekend my scope sat in the bag until I pulled it out this afternoon and 
plugged it in.  I never did check to see if all the settings were the 
same as when I last used it.  Perhaps if the scope channel and the probe 
were both still set for 10X it would be easier to interpret the 
results.  Channel 2 was set for a 1X probe, but the probe was set for 
10X as is my usual habit.  The amp's been working perfectly all along.  :)

73 - and thanks, guys!
Dale - N0XAS

On 2/18/2014 7:07 PM, Rick Bennett wrote:
> I am certainly no expert either, but a couple thoughts.  You might try a higher supply voltage.  Also might need to play with the bias voltage on Pin 3.  Maybe put pots in place of R9 and/or R10 and try different bias levels.
>
> Rick
>
> ----- Original Message -----
> From: "Sean Pepin" <smpepin at gmail.com>
> To: hbqrp at mailman.qth.net
> Sent: Tuesday, February 18, 2014 4:41:30 PM
> Subject: Re: [HBQRP] Who's an op amp expert?
>
> On 02/18/2014 03:26 PM, Dale Botkin wrote:
>> Trying to troubleshoot what should be a dead simple op amp circuit and
>> I could use a hand.
> Hmm. I doubt I can help, but I have some suggestions.
>
> R9 and R10 are pretty small. The only draw on the virtual ground at pin
> 3 of U2A is the opamp's input bias current, which is rated in picoamps
> on this guy. You can safely use 100k for both. I would bypass this node
> with 1n, just to avoid any oddball stability issues. Doing this will
> help save power if that's a concern. None of this should affect your
> issue though.
>
> Reduce C7 to 100n temporarily. It may be kind of large when considering
> the feedback network is made of very high value resistors. This is the
> closest thing to a relevant issue I can spot.
>
> Reduce the value of the feedback network resistors to something like 15k
> and 22k. Output current can more easily charge C7 and this lowers the
> impedance of U2A pin 2, reducing the possibility for interference.
>
> Add a small AC coupled load. The opamp is supposed to be able to source
> something like +/- 10mA, so it's no sissy. Something like 100n in series
> with U2A pin 1 and then 100k to ground.
>
> Hope that helps. I have a hunch that it might be a forehead slapper
> though. I can't see anything that'd keep this from working.
>