You are obviously right about the voltage source connection. I believe series connection is the correct one:
Revisiting the voltage divider, let's assume I will be using 2x9V alkaline batteries in series for a total voltage of 18-19V. The TL062BC has a typical input bias current of 65pA and maximum of 200pA. I believe this determines the current that the bias voltage point should be able to provide without the bias voltage moving significantly (for example, maintaining a 0.1% maximum deviation). Given @ron68 's comments, it seems that I can use 1M or 2M resistors for the voltage divider. With 2M, the current that flows through the divider is 4.5uA and with 1M it's 9uA. I make the assumption that the op amp can draw from these currents while at the same time having the voltage bias practically completely stable. So let's assume I go for 2M resistors.
About the capacitors used, I believe the bulk decoupling capacitor of 100u from 0V to 18V is a given. Then we have two options for the most part. a) placing an electrolytic of say 47uF between the bias voltage point and ground or b) placing film capacitor(s) in parallel with the bias voltage resistors.
The capacitors have two purposes the way I see it: 1) noise filtration for the power supply and 2) bias voltage point stabilization by providing current locally to restore the bias voltage when it deviates from mid supply. Your comments about removing the two film capacitors makes me think that point 2) is a non issue, as the resistors alone can maintain a stable bias voltage. But what about noise filtration? Even with a battery it should be a good idea.
Lastly, after reading on forums oriented more to electric bass electronics, it seems that it's important that the input impedance that a passive pickup sees is between 750k and 1.5M. So 1M input impedance is a happy medium. Low input impedance is obviously sub-optimal for passives, but it seems that "the higher the better" does not apply for the op amp's input impedance due to the pickups' characteristics. So, if Rbias sets the input impedance to 1M, it is imperative that it stays in the circuit. It is a bit counter intuitive at first glance that the resistor has 4.5V to its left and also to its right but I guess DC operation is different than AC operation. It is also not clear to the untrained eye (like mine) exactly how this Rbias makes the pickup see a 1M input impedance the way it is connected. The first instinct is that it would just impede current flow from the bias voltage point to the op amp input, but I think this is wrong intuition.
So, given 18V, 2M voltage divider resistors and the need to set the input impedance of the op amp to 1M (meaning Rbias should probably stay in the picture) there's the issure of what to do with capacitors. I believe the bulk capacitor is a given, but would you add anything for noise suppression (and the unlikely need for helping with bias voltage stabilization) and how? For example would film capacitors help? Would a classic electrolytic of 47u from bias voltage point to ground help with low frequency filtering and as a charge reservoir (although most probably unnecessary in this role)? What would you recommend I do with capacitors?
Last edited:
Hi,
It's better to decrease the capacitor of bias resistors if you go with 2+2Meg, otherwise it will take too much time to the bias level to settle after turning on/off.
From DC point of view, bias node see impedance of 1Meg (2//2 = 1).
With 47uF, using the basic time constant RC formula (time to charge the capacitor to 63%), we would have 1,000,000*0.000047 = 47sec.
It's too much. I would go with 4.7uF, so time constant of 4.7sec.
From audio AC point of view, bias node impedance is close to zero (ground), so R3 is virtually grounded in one end.
The AC audio impedance seem at the input is R3 (1Meg).
I'm disregarding op amp input impedance.
Once connected to the bass pickups, things get quiet, but when disconnected it might be susceptible to noise due to the high impedance of 1Meg (turn volume down before connecting/disconnecting). Just make sure the following stage (power amp, pedals etc) can receive all the voltage produced by this pre-amp, cause during a cable failure this pre-amp will easily go rail-to-rail (+/-7.5V dual battery or +/-3V single battery).
You may test later on, changing the R3 to a lower value, if needed, so as to achieve a good sound and good S/N.
It's better to decrease the capacitor of bias resistors if you go with 2+2Meg, otherwise it will take too much time to the bias level to settle after turning on/off.
From DC point of view, bias node see impedance of 1Meg (2//2 = 1).
With 47uF, using the basic time constant RC formula (time to charge the capacitor to 63%), we would have 1,000,000*0.000047 = 47sec.
It's too much. I would go with 4.7uF, so time constant of 4.7sec.
From audio AC point of view, bias node impedance is close to zero (ground), so R3 is virtually grounded in one end.
The AC audio impedance seem at the input is R3 (1Meg).
I'm disregarding op amp input impedance.
Once connected to the bass pickups, things get quiet, but when disconnected it might be susceptible to noise due to the high impedance of 1Meg (turn volume down before connecting/disconnecting). Just make sure the following stage (power amp, pedals etc) can receive all the voltage produced by this pre-amp, cause during a cable failure this pre-amp will easily go rail-to-rail (+/-7.5V dual battery or +/-3V single battery).
You may test later on, changing the R3 to a lower value, if needed, so as to achieve a good sound and good S/N.
@ron68 for a quick startup I 'm willing to do something like this:
This is more or less what you propose, the more traditional filtration scheme I tend to find online, with a bit faster charge up time.
Ce_1 is there for noise filtering, as even with 2M voltage divider resistors the bias voltage would not need much stabilization with the TL072 drawing current from it.
A small note is that the buffer will be permanently connected to the source (pickups) so there's little potential for noise due to disconnection. A potential malfunction and wide voltage swing at the output is cause for concern however, especially at 18V.
My only question is if this can be improved in any way by film capacitors, as per the initial recommendation by @Mark Tillotson.
Perhaps they were suitable for a different application and in this case things will be fine with just a small electrolytic between bias point and ground?
This is more or less what you propose, the more traditional filtration scheme I tend to find online, with a bit faster charge up time.
Ce_1 is there for noise filtering, as even with 2M voltage divider resistors the bias voltage would not need much stabilization with the TL072 drawing current from it.
A small note is that the buffer will be permanently connected to the source (pickups) so there's little potential for noise due to disconnection. A potential malfunction and wide voltage swing at the output is cause for concern however, especially at 18V.
My only question is if this can be improved in any way by film capacitors, as per the initial recommendation by @Mark Tillotson.
Perhaps they were suitable for a different application and in this case things will be fine with just a small electrolytic between bias point and ground?
Last edited:
Hi,
Beware the TL06x op-amps in small-signal circuits. Yes, they draw less power than their TL07x counterparts, but the 06x are much noiser.
I have two Boss GE7 graphic equaliser pedals. Swapping the 06x in them out for 07x markedly reduced noise in both cases. YMMV, of course, but at least fit a socket so you can compare the two in your pre-amp.
Cheers,
Ant.
Beware the TL06x op-amps in small-signal circuits. Yes, they draw less power than their TL07x counterparts, but the 06x are much noiser.
I have two Boss GE7 graphic equaliser pedals. Swapping the 06x in them out for 07x markedly reduced noise in both cases. YMMV, of course, but at least fit a socket so you can compare the two in your pre-amp.
Cheers,
Ant.
@Ant Moore thanks for reporting your findings. Yes, it is a well known fact documented in the chips' specifications. But the TL062 is very appealing for an onboard buffer because of the limited energy reservoir. Eventually I will test and measure both, for sure.
Hi!
That will work fine with only the electrolytic in between the 2 x 1Meg resistors. RC will be 1.1sec. This is key to reject any power supply noise or variations.
For battery, it is more important to avoid low frequencies oscillations due to interaction with the load.
If you find issues with that, just increase the 2.2uF bias capacitor.
The other capacitor (100uF) will not play much importance, although it doesn't hurt if installed, since the opamp will reflect in the output what it sees in the +Input. Opamps have very high common mode power supply rejection.
And as mentioned, if noise is an issue, just swap them and check for your application if this is relevant or not.
In this particular case, where G=1, I think you'll be fine.
That will work fine with only the electrolytic in between the 2 x 1Meg resistors. RC will be 1.1sec. This is key to reject any power supply noise or variations.
For battery, it is more important to avoid low frequencies oscillations due to interaction with the load.
If you find issues with that, just increase the 2.2uF bias capacitor.
The other capacitor (100uF) will not play much importance, although it doesn't hurt if installed, since the opamp will reflect in the output what it sees in the +Input. Opamps have very high common mode power supply rejection.
And as mentioned, if noise is an issue, just swap them and check for your application if this is relevant or not.
In this particular case, where G=1, I think you'll be fine.
@stocktrader200 Thanks. Do you think post #43 makes sense for a 18V single supply with two 9V alkalines in series? Is something missing or redundant? What would you change there to properly bias the non inverting inputs of a TL072? (you may have noticed I changed course midway this thread, from dual supply to single supply, for reasons I will elaborate on later).