I use an op amp to create a virtual ground out of a 9V supply, so I get +/-4.5V (a bit less since there is also a protection diode).
The 9V supply is split in half with 2 resistors, the bottom one bypassed by a 100n ceramic cap.
The questions are:
1) do I need a capacitor at the output of the op-amp ?
2) what parameters should I be looking for when choosing the op amp that will make the virtual ground?
Many thanks
akis
The 9V supply is split in half with 2 resistors, the bottom one bypassed by a 100n ceramic cap.
The questions are:
1) do I need a capacitor at the output of the op-amp ?
2) what parameters should I be looking for when choosing the op amp that will make the virtual ground?
Many thanks
akis
1/ For most opamps I would say yes, and use two caps ? one from opamp output to Vcc and one from output to what is now the 'minus' rail. The caps could be as high as 47uf or 100uf.
2/ Ah ! If battery life is important then you should be designing for low power throughout and so something like a TL061 (cooking grade and easily available) could be very workable, or there are many new micro power opamps available. FET input allows the opamp bias network to be high impedance.
2/ Ah ! If battery life is important then you should be designing for low power throughout and so something like a TL061 (cooking grade and easily available) could be very workable, or there are many new micro power opamps available. FET input allows the opamp bias network to be high impedance.
Rail splitter
I have used a rail splitter for battery ps rather than an op amp design.
Try a TLE2426CLP. One of my op amp sites for reference:
Virtual Ground Circuits
I have used a rail splitter for battery ps rather than an op amp design.
Try a TLE2426CLP. One of my op amp sites for reference:
Virtual Ground Circuits
"9 V" battery V can be a lot less - many rechargeables start at 7.2 V, all drop off as discharged
output C can be an issue for op amps - you need to isolate most op amps from the C near their unity gain frequency
and of course the op amp has to be able to supply the current your circuit has flowing in the active ground
the tle2426 is very limited in output current - OK if it works for your application
output C can be an issue for op amps - you need to isolate most op amps from the C near their unity gain frequency
and of course the op amp has to be able to supply the current your circuit has flowing in the active ground
the tle2426 is very limited in output current - OK if it works for your application
Hi guys, thanks for replies, the circuit is for a 9V "stomp box" (effects pedal) I just needed a virtual ground for the inputs of the audio op-amp and there is hardly any load imposed on it. The pedal is powered from a PSU so we are guaranteed 9V all the time. For the audio op-amp I used a LM8333N and for the vground I used an MC33072. The effects pedal goes in the FX loop so it handles rather large signals and has a 50K input resistance (FX send) and expects similar resistance from the next stage (FX return). It is all in audio range and because it is for a guitar it is almost guaranteed to be less than 5 KHz.
Hi,
Stomp boxes have been designed for ages without the need for
an op-amp derived "virtual ground", its just over complication.
Hardly any load ? Standard AC coupled single voltage rail design.
"Virtual ground" is a modern buzzword and implies sourcing
and sinking substantial currents as though a real ground.
Prevalent as a fashion in headphone amplifiers.
rgds, sreten.
Stomp boxes have been designed for ages without the need for
an op-amp derived "virtual ground", its just over complication.
Hardly any load ? Standard AC coupled single voltage rail design.
"Virtual ground" is a modern buzzword and implies sourcing
and sinking substantial currents as though a real ground.
Prevalent as a fashion in headphone amplifiers.
rgds, sreten.
Last edited:
Hi,
I have then misused the term "virtual ground" since I am not sourcing or sinking any currents worth mentioning.
The single sided version would require an extra one resistor and two caps , the "virtual ground" version required 2 resistors and one op amp. I should have done it using the single sided version, on retrospect.
In case anyone cares:
It is a "solo volume boost with mid-boost" as requested by my son to use during his solos. It goes on his pedal board and in the FX loop of his amp. It provides 2x boost on the mids (between 660hz and 960 hz) and an additional volume boost of 5x, for a total of 10x if you are in the centre frequencies. There are two pots to control mid boost and volume boost, an LED to show when on and a true bypass foot switch. On their leftmost position (pots completely off) it's like a bypass, so you can never attenuate. The difference is that the pedal is inserted into the FX loop, so as not to be in front of the amp, and not make a difference to the sound - and in any case on the "modern" high gain channel, it would not make much difference. Another way to do that is to ask the sound engineer to boost your volume when you are playing your soloes -but my son does not trust him to do it. So it is a pedal developed in haste, but it works rather nicely, with 0 distortion or altering of the sound except if you make it clip. However the FX return of my son's amp (Mesa Boogie Recro-Verb 25) clips at +/-2V and my pedal clips at +/-3.5V or thereabouts. I started designing this using an ebay module that takes 5V-12V DC and outputs +/-15V DC (sounds fantastic on paper), so as to have lots of headroom to deal with very large signals coming in from the FX Send, may be not this amp but another, but the modules have not arrived in time (from China) so I had to rewire the circuit hence the "virtual ground" solution. The most difficult part of the construction was to drill the holes onto the aluminium enclosure, Hammond 1590B, for the 1/4" phono jacks which had to be drilled perfectly as there was no room for play not even 1mm.
I have then misused the term "virtual ground" since I am not sourcing or sinking any currents worth mentioning.
The single sided version would require an extra one resistor and two caps , the "virtual ground" version required 2 resistors and one op amp. I should have done it using the single sided version, on retrospect.
In case anyone cares:
It is a "solo volume boost with mid-boost" as requested by my son to use during his solos. It goes on his pedal board and in the FX loop of his amp. It provides 2x boost on the mids (between 660hz and 960 hz) and an additional volume boost of 5x, for a total of 10x if you are in the centre frequencies. There are two pots to control mid boost and volume boost, an LED to show when on and a true bypass foot switch. On their leftmost position (pots completely off) it's like a bypass, so you can never attenuate. The difference is that the pedal is inserted into the FX loop, so as not to be in front of the amp, and not make a difference to the sound - and in any case on the "modern" high gain channel, it would not make much difference. Another way to do that is to ask the sound engineer to boost your volume when you are playing your soloes -but my son does not trust him to do it. So it is a pedal developed in haste, but it works rather nicely, with 0 distortion or altering of the sound except if you make it clip. However the FX return of my son's amp (Mesa Boogie Recro-Verb 25) clips at +/-2V and my pedal clips at +/-3.5V or thereabouts. I started designing this using an ebay module that takes 5V-12V DC and outputs +/-15V DC (sounds fantastic on paper), so as to have lots of headroom to deal with very large signals coming in from the FX Send, may be not this amp but another, but the modules have not arrived in time (from China) so I had to rewire the circuit hence the "virtual ground" solution. The most difficult part of the construction was to drill the holes onto the aluminium enclosure, Hammond 1590B, for the 1/4" phono jacks which had to be drilled perfectly as there was no room for play not even 1mm.
Three resistors and a capacitor is all you need.
An active virtual ground will yield no benefits for this application.
An active virtual ground will yield no benefits for this application.
2 x 9V is a great solution for any effects pedal that might require more headroom.
Most guitar effects pedals are 9V and many have a dual battery/PSU option and my son did not want to have an "odd" box that requires a special PSU. He does not use batteries as they are so unreliable.
The op-amp to buffer a rail-splitter so as to get +/-4.5V but it was an overkill, did not really need the op-amp to buffer it, just smaller resistors and a large capacitor would have sufficed. However the cost of the op-amp is around 60p which is also more or less the cost of electrolytics these days. And since it is only one off it does not really matter.
I don't understand, you wrote that you had designed this to run with a +/- 15 V DC module. This would already mean that you have a ground and would only need to regulate the rails down to 4.5V.
A +/-15V DC-DC module (using the 9V supply as input). That tiny module which I had hoped to fit inside the stomp box has yet not arrived and as my son had a live performance I had to hack something together in a hurry, hence the rail-splitter and op-amp buffer.
Wait hold on, will this device and its source or load connect to the same PSU?
If so, abort the idea.
If so, abort the idea.
Isn't this a recipe for instability? AFAIK, one can't slap a cap beyond 100-1000pF or so onto the output of an op amp without torturing the amplifier, essentially making a zero out of the amp's output resistance and the load capacitance, which then prevents the feedback loop from closing at a favorable phase margin. The only way (that I know - please correct me if possible!) to stabilize this is to add series resistance to the amp output, which defeats the whole point.
I'd also question the need for this - ideally, the amplifier and it's closed loop output Z will provide a low impedance for the 'fake' ground. Sure, the magnitude of this fake ground impedance will rise at 6dB/oct, which is probably what prompted the urge to throw a cap across it, but this is usually not possible, unless one does some fancy tricks to isolate the rail splitter amp from this capacitive load.
Rail splitters can work, but probably only as low current nodes to define specific DC voltages. Am I confused?
Small capacitive loading (in the pf and nf region) can certainly be an issue for many opamps. Using large caps is brute force, but gets the job done with many opamps. The TL061 I mentioned (as a cooking grade device to try) actually has an inbuilt 50 ohm resistor in series with the output.
Although we think of opamps as 'perfect', the output impedance will be pretty lousy at HF if we were to rely on just the opamp alone as a means of defining a clean ground.
Although we think of opamps as 'perfect', the output impedance will be pretty lousy at HF if we were to rely on just the opamp alone as a means of defining a clean ground.
Monty--IIRC, doesn't a very large cap on the output become the dominant pole? It's the middle ranges that can kill.
E.g. Analog Devices: Analog Dialogue: Ask The Applications Engineer - 25
(If I'm mistaken, please correct me!)
E.g. Analog Devices: Analog Dialogue: Ask The Applications Engineer - 25
(If I'm mistaken, please correct me!)
- Status
- Not open for further replies.