variable amplifier 1

[xray]

I'm currently trying to put together a variable gain for volume control of a medical listening device (20-2kHz). The preamp is a low voltage op amp LM6582 wired in an inverting configuration -(Rf/Ri). The adiuo amplifier is a low voltage LM4860 1W essentialy an inverting configuration. The problem occurs when placing a pot in the feedback path. The signal is attenuated nicely until about 35% of the maximum value of the pot. The signal at 35% goes to the rail 3v. The feedback path will handle a nominal valued resistor that is equal to the maximum value of the pot, so the device is not near its maximum gain. These devices have not been wired together and remain seperatley tested until I get a nice volume control out of on or the other. I am looking to use these two devices for this solution. I may have to use a voltage divider of the electret microphone signal. Any thoughts sent my way would be helpful.

 

[Merlin1xx1]

okay i think i know what you're problem is, if i'm understanding what you said correctly.
now you said you're using a pot in the feedback loop of the amp? instead use a pot as Ri, this is a method that i have done before with a 741 and it worked fine all the way up to the rail voltage. if you have a dedicated Rf and Ri and then a pot in the feedback, then you are effectively attenuating the feedback, right? well you can only go so far with that before the amp doesn't see any feedback and BAM , slam the rails. if thats how you've got it set up then this may help...otherwise i've wasted your time.

 

[xray]

Merlin;
Thanks for the reply I will place the pot in with the input resistor and give that a try. I still am not sure about the output going to the rail at about 35% of the pot's nominal value. The pot's nom value is 100k and truns down to zero. If the pot is the only device in the feedback path what causes the railing at about 35%. Theory: -(Rpot/Ri) if Rpot is zero Ohms voltage follower. If Rpot is maximum 100k and Ri is 10k we see a gain of 10 180deg out of phase.

 

[Merlin1xx1]

xray,
okay, i think i misunderstood exactly how the circuit was setup initially. from what i gather now your Ri is 10k and your Rf is the pot. this makes sense. so the amp hits the rails at 35% of the pots max value...so 35k, at this point the amp is set up for a gain of 3.5. two questions,
1) What are your supply voltages for the amp
2) How large is the input signal(max)?

is it possible that at a gain of 3.5 you are already within 2V of the supply? maybe your problem lies within the supply voltage rather than the resistor network. from what i gather you have a supply of roughly 5V correct? so if you are hitting the rails at a gain of 3.5 your input is somewhere near 860mV RMS. if this is correct then your problem lies at the supply voltage, if this is not correct then i would suspect your problem to be with the IC and not the circuit.

 

[Merlin1xx1]

one other thing, i was trying to find some info on the LM6582 but can't find anything, not even on national's website, is that the correct part number?

 

[xray]

Merlin thanks for time and effort in my inquiry. I have forgotten the c in the device it is an LMC6582 by National. I have narrowed the railing of the signal to the pot in the feedback path. the pot is a three lead device and I have been grounding the third lead. The grounding of the third lead affects possibly the feedback, when the pot is used as a two lead device the operation is flawless. When investigating this I have seen the un-used portin or the pot looks similiar to a load off of the op amp. I think that the op amp can drive this 65%( 65k) load without substantial output current, therefore I'm not sure about why it rails. The problem is in using a three lead pot in the feedback path, so using the three lead pot as a two lead pot in the feedback solved the problem. Merlin thanks for checking in on my problem.

 

[Trex77]

Xray... don't use the pot as a two teminal unit... that will create too much noise during adj... What works better is to tie the two end-points into the ckt (output of opamp, and one of the opamp inputs). Then short the wiper to the output forming a Reostat. During adj the value will never go over the 100K, with wiper only you can go to an open ckt. So, a better level or gain adj can be made with less noise. I agree, don't ground the third teminal... instead tie it to the opamp output as I have detailed.

Also, the capacitance of the feedback loop will change with the position of the wiper, becoming less with more ohms and becoming greater with less ohms.... just what you want to occur to keep the noise BW within reasonable limits.

 

[jbartos]

Suggestion: Many manufacturers of integrated circuits provide parameters/data for interfaces, including potentiometers/rheostats. These parameters/data should not be deviated without consulting the manufacturer since a higher current or voltage may increase aging or damage the integrated circuit/chip.