inverting / non-inverting op-amp circuits

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi,

As part of a dual electronic filter network I want to build exactly mirrored input buffer stages. Thus 1 half of the rest of the network will be the exact inverse of the other. This mirrored signal will then form the input of a bridged power amplifier.

I'm a little confused about how to do this as the rules for gain for inverting and non-inverting op amps are different:

non-inverting is Vout/Vin =1+Rf/R1,
and inverting is Vout/Vin = Rf/R1

where the Vs are input and output voltage, Rf is the feedback resistor and R1 is the input resistor.

So if I want my inverting amp to have exactly the opposite output voltage of my non-inverting amp the values of Rf and R1 will have to be different, just not sure how to work out how different. Can anyone help?

Cheers!
 
Just a little algebra will do it. Choose R1 to be the same for both halves, to simplify things. Since the gain for both halves must be equal, you can substitute the two equations to give:

1 + Rf1/R1 = Rf2/R1

where Rf1 is Rf for the non-inverting and Rf2 is Rf for inverting configuration.

Thus:

Rf2 = Rf1 + R1



You might also want to know that there are other, more symmetrical ways of doing this ("phase splitting").
 
phase splitting

Thanks for the info Mr. Evil, and for the cool link. Was wondering about phase splitting and what it entails. As I was falling asleep last night I realised that an easier way of explaining what I want toi do is converting an unbalanced to a balanced signal. There are probably hundreds of schematics out there waiting to be found, but low-noise/distortion would be good.

Any good links?

Cheers
 
It's actually very simple :) All you need is the following order of building blocks:

1. Input buffer/L+R summing circuit
2. Low-pass filter (2nd- or 4th-order)
3. Balanced output/bridging adaptor

These are such basic and fundamental building blocks that there is tons of info easy to find out there. They should all be pretty easy to hook up together.

Maybe take a look at the ESP site for some pointers or even complete projects. If you are still stuck, as me nicely and I may draw you a quick sketch :)
 
Hi,

Yes, I see that, and have found quite a few schematics for the bits. As it happens, each question answered throws up several new ones. I currently have two concerns:

1. The sheer number of op.amps in the signal path. With the wm8 and all the other modules I'm now pushing 10, each adding noise and distortion. I'm wondering if I can limit this somehow by combining some of the modules.

2. I just learned of 90 deg. phase shift in Butterworth networks. Is putting one of these in the path going to furtther screw up the overall response?

Cheers!
 
1. Don't be overly worried about this. There is a lot of nonsense talked about op-amps in the signal path. A typical recording studio console has around 100 op-amps in the signal path, none of which are particularly esoteric.

A respected guy on here actually did listening tests with cascaded op-amps and even using the cheapest ones he could find he couldn't notice hardly anything even with 20 op-amps.

That's not to say don't try to eliminate op-amps where you can. It saves on cost if nothing else. Just don't cut your nose off to spite your face.

2. Yes. But that's the nature of all crossover networks. If it bothers you, ditch your tweeters as well and go fullrange :angel:

There are alternatives to Butterworh that may be do-able anyway, which have more favourable phase characteristics, at the expense of attenuation.
 
Can someone look at this schematic and tell me if I'm being stupid?

Have now abandoned Butterworth in favour of Linkwitz Riley as it fits the fallof of my front pair much better (and no phase shift :)). There is a phase shift control network here :
http://sound.westhost.com/project103.htm

but it looks as if this won't be necessary if I place the sub carefully. Does anyone have experience using these shift contrtollers over a limited frequency range? My sub will initially only operate over 1 octave from 20 to 40 Hz and the phase shift module would be fairly easy to add to the other controls if it were likely to help.

I'm attching a pdf of where I'm at so far.

Questions:
Are the BB OPA134 ops. a complete waste of money at these frequencies?

Is my money saving power supply (robbing the power amp supply) a good idea or should I invest in a dedicated +/-15V. supply?

In the Linkwitz Riley circuit are there any considerations to the values of the components other than the frequency and availability ( ie. any inherent advantage/disadvantage to large or small resistor or cap values)?

Thanks
 

Attachments

  • sub_schematic.pdf
    30.6 KB · Views: 79
Your thinking that simply changing one filter from Butterworth to LR is a little different from the reality, I'm afraid. The phase relationship is only constant with LR filters if both high-pass and low-pass filters are matched, i.e. share the same Q of 0.5 and -6dB point. I doubt that your main speakers have a roll-off that will be exactly this, but it is not impossible.

A phase shift control can be eliminated by placement in theory, but in practice it may be that the place where phase shift is small results in horrid sound due to room modes. As long as your sub is within a couple of metres of your main speakers you probably only need to choose between normal and inverted phase (0 or -180 degrees).

No problem with using those op-amps for a sub. There are other cheaper options that may work just as well in that application, but it's not like you are going for silly esoteric ones, or will be needing lots of them.

Using the amp rails is a great idea. Check out ESP P102.

EDIT: Just seen your edit. Yes there is consideration to be given that you do not choose values which result in too low resistor values as you will load the op-amp too much. Too high resistor values will be noisy. Avoid caps over 1uF as they are hard to get in non-polarised types.
 
Rolloff

I phoned the Mfg. of the front pair earlier. They say that -6 dB is pretty accurate at 40Hz and rolloff at 24dB/8ve. That way I'm hoping to avoid a high pass altogether.

Q. is a mystery for both speakers as I only have it for the raw driver for the sub, and System audio don't publish full specs. Do you know how I calculate Q for the complete speaker if I know the volume of the enclosure and Thiel smalls for the driver?

Project 102 suggests that emitter current on the output transistors will also effect base (and zener ) current. 100mA doesn't sound like enough to drive the Wm8 + the rest of the ops.

Thanks for all your help!
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.