EarthTonesAudio "Buster" (buffer and boost in-one)

[earthtonesaudio]

Inspired by the MI Audio Boost 'n' Buff, maybe even a workalike (I haven't seen a schem for that one, so who knows). Simple op-amp boost with gain variable from 1 to 11, SPST makes it just a buffer (gain of 1). Low output impedance at all settings, high input impedance, easy to mod for your needs.

[mnordbye]

Looks like my next quick-and-easy breadboard project. Have you had good results with it?

[biggy boy]

What would be a good IC to use?
Maybe a TL071?

Glen

[earthtonesaudio]

Yep, works fine, nothing fancy.  Full audio bandwidth (and beyond), just a little bit of gain and good buffering.  I did not include pull-down resistors, as this is not intended to be true bypassed.

In theory, if you use a TL071, have the gain up full, and switch to "buffered" mode, you lose some output voltage swing capability.  This means instead of swinging 6V p-p, the output will only swing maybe 5V p-p.  However, if your input signal is that huge already you probably have little need for either boost or buffer.  In practice, it basically just works fine.  If this issue still bugs you, multiply R4 and R1 by anywhere from 2 to 10, and that "problem" goes away completely.

[waltk]

Alex,
Thanks for sharing this!  ...made one; works great.  If you have a minute for a couple questions:

Should the V+ go straight to pin 7 on the opamp, or after the diode?  I wasn't really thinking about this when I made my layout, but later thought maybe it should have been after.
Is a 1N4001 OK for the diode?

Here's the layout I used, and a couple photos.  I also made a layout that combined the Buster with a TDA7240 amplifier, and ended up with a clean 20 watt amp that sounded great.  So the Buster makes a nice little preamp.  The whole thing was hardly bigger than a Ruby and LOUD.

[earthtonesaudio]

Thanks for sharing your build report, I'm glad you like it!

The diode is just there for reverse polarity protection.  As such it should be before the op-amp power connections.  However, if you're never going to plug in the power supply backwards, it's not necessary.  As for types, any (non Zener) diode should work.  Schottky or Ge will drop a little less voltage than Si, but with this circuit it's not critical.

[waltk]

The diode is just there for reverse polarity protection.

Thanks, that's what I suspected...  I etched some extra PCBs.  If you you want one, just PM me with your mailing address.

[waltk]

Alex,

One more question, if you have a minute... I tried using the same design with half of a dual op amp (swapping pins 6 for 1 and 7 for , and got a strange effect.  It worked normally with the gain knob set from 0 to about 90%, then it distorted heavily ('til about 93%), then cut out completely.  I don't see any obvious reason why.  Is it a problem to leave the other three pins of the op amp unconnected?  Can you think of any other reason it might do this?

[AM]

How noisy is this compared to other boosters like the Mosfet etc. Ben pointed me to this discussion through another thread. It looks very interesting and probably very close to what I need. Low parts count, clean and simple enough. Has anybody tested/compared it with any of the other standard boosters for noise levels?

Also a couple more questions regarding some parts substitutions:
1. I'm trying to avoid electro caps. I have some 1uF metal films. Would replacing the 10uF C2 with 1uF have a strong effect on the circuits output impedance etc?
2. Are C3, C4 (10uF) just for filtering? Could they be both omitted if I use a 9v regulated power supply?
Thanks

[earthtonesaudio]

Alex,
One more question, if you have a minute... I tried using the same design with half of a dual op amp (swapping pins 6 for 1 and 7 for , and got a strange effect.  It worked normally with the gain knob set from 0 to about 90%, then it distorted heavily ('til about 93%), then cut out completely.  I don't see any obvious reason why.  Is it a problem to leave the other three pins of the op amp unconnected?  Can you think of any other reason it might do this?

Hi Walt,
My best guess is that there's some DC offset at the inputs, causing it to act like a comparator when you put the gain up high.  Instead of DC coupling R4 to 1/2 supply, put a large capacitor in-line with it, going to either V1/2 or ground.  As for the other op-amp, it's better to tie the inputs to something rather than leave them floating.  The best thing for an "extra" op-amp in this circuit would be to use it as a buffer for V1/2, allowing you to use larger R5/6 resistors, smaller C3, and reduce power consumption.
What op-amp are you using?

How noisy is this compared to other boosters like the Mosfet etc. Ben pointed me to this discussion through another thread. It looks very interesting and probably very close to what I need. Low parts count, clean and simple enough. Has anybody tested/compared it with any of the other standard boosters for noise levels?

Also a couple more questions regarding some parts substitutions:
1. I'm trying to avoid electro caps. I have some 1uF metal films. Would replacing the 10uF C2 with 1uF have a strong effect on the circuits output impedance etc?
2. Are C3, C4 (10uF) just for filtering? Could they be both omitted if I use a 9v regulated power supply?
Thanks

Hi AM,
Sorry, I haven't done a head-to-head with other boosters.  There's not much gain so it shouldn't be too bad. 
1. Using 1u for the output is fine, but keep in mind that whatever resistive load follows it will make a high pass filter.  If the load that follows is 10k, a 1u cap will start cutting lows at 16Hz.  If the load is 1k, the lows are reduced starting at 159Hz, etc.  As you can see, my choice of 10uF is probably a bit of overkill. 
2. Yes and no.  They are just for filtering, but C4 is for power supply noise and C3 is for power supply plus any thermal noise created by the 10k/10k voltage divider.  If using a clean supply you could omit C4 and reduce the value of C3 to perhaps 10-100nF.

[Thomeeque]

Is it a problem to leave the other three pins of the op amp unconnected?

Avoiding Noise and Power Problems with Unused Op Amps

Abstract: If connected improperly, an uncommitted op amp leads to problems such as increased power
consumption and added noise.

But in this case adviced V1/2 buffer is probably best idea..

Btw. I would advice to "double" C3 and C4 electrolytics by ceramic 100nF caps to improve higher frequencies blocking.

T.

[Thomeeque]

C3 is for power supply plus any thermal noise created by the 10k/10k voltage divider..

..plus it blocks whatever would get there via R4 as well..? I'm not 100% sure, but IMO you could get into some oscillations this way otherwise (as the same source is used as reference for positive input).

Edit: Very often you can see this "gain dividers" (R4/R1 here) anchored via decoupling cap directly to circuit's ground instead of connecting to the reference voltage source.

[Gus]

Alex  some things I see that maybe I would do different

Where is the series resistance in the output leg?.  Stability issues with capacitive loads, you don't know what people will plug this into.

Do you need 10Ks in the Vref can you go higher to reduce current drain on the battery?  You do have a 1meg input resistor.

[soggybag]

Nice idea.

What software did you use to draw the schematic?

[earthtonesaudio]

Hi Gus and Thom,

Yeah, there is plenty of room for tweaks and improvements to this.  Decoupling the feedback network from the voltage reference is probably a better way than DC coupling as I drew it, since it will be more tolerant of errors at the amp's input.  If you use a single op-amp package, there are offset adjustment terminals on most models, so that's another option.
Some small series output resistance is a fine idea, but I personally have never needed it.

Some other things maybe worth adding:
Antiparallel diodes across the inputs
A small value cap from output to the (-) input.
Buffered 1/2 supply reference

It could easily become a rather complex build, but I guess it's up to the individual to decide when enough is enough. 

Nice idea.

What software did you use to draw the schematic?

I used DIYLC (DIY Layout Creator sofware) on a mac.

[waltk]

My best guess is that there's some DC offset at the inputs, causing it to act like a comparator when you put the gain up high.  Instead of DC coupling R4 to 1/2 supply, put a large capacitor in-line with it, going to either V1/2 or ground.  As for the other op-amp, it's better to tie the inputs to something rather than leave them floating.  The best thing for an "extra" op-amp in this circuit would be to use it as a buffer for V1/2, allowing you to use larger R5/6 resistors, smaller C3, and reduce power consumption.
What op-amp are you using?

Thanks for the quick reply! My understanding of how this works is kind of limited, so I appreciate your patience with the dumb questions.  I tried a TL072 and a 5532 with the same result, but I might also try others.  I was thinking that a rail-to-rail op amp might allow for increased gain.

If I wanted to do a quick test putting a large cap in-line with R4, what value would I use?  By in-line, you mean in series right?

If I use the other op-amp for a buffer, what could I increase R5/6 to, and what would be a good value for C3?  Also (to confirm my understanding of your answers to AM), could C4 be eliminated?

Would you make any other changes based on Gus' comments?

How noisy is this compared to other boosters like the Mosfet etc. Ben pointed me to this discussion through another thread. It looks very interesting and probably very close to what I need. Low parts count, clean and simple enough. Has anybody tested/compared it with any of the other standard boosters for noise levels?

I didn't attempt any measurements, or look at it on a scope, but just by ear it sounds very clean - comparable to the AMZ MOSFET boost.  Because the gain is reasonably low, you don't get any clipping, so the output cleanliness depends on the specs of the op-amp.

Abstract: If connected improperly, an uncommitted op amp leads to problems such as increased power
consumption and added noise.

I checked out the link.  Thanks for passing that along , and your other comments.

[waltk]

My best guess is that there's some DC offset at the inputs, causing it to act like a comparator when you put the gain up high.  Instead of DC coupling R4 to 1/2 supply, put a large capacitor in-line with it, going to either V1/2 or ground.  As for the other op-amp, it's better to tie the inputs to something rather than leave them floating.  The best thing for an "extra" op-amp in this circuit would be to use it as a buffer for V1/2, allowing you to use larger R5/6 resistors, smaller C3, and reduce power consumption.
What op-amp are you using?

OK. So I now I have R4 coupled to ground through a 10uF cap.  That helped a little, but I also rewired the dual opamp so the extra half is used as a buffer before the one with the gain control (not sure I did what you had in mind, but at least the inputs aren't floating, and it's still working).  It sounds great until I turn the gain all the way up, then I get some oscillation.  I can live with that because nobody is forcing me to turn the gain all the way up.  I'm still wondering what values would be good for R5/6 and C3.  Anybody?

[earthtonesaudio]

I'm just now seeing your posts, Walt. 

Let me try to answer everything really succinctly. 

-A rail-rail op-amp will give you more headroom, but not more gain. 
-Cap in series, correct.  Smaller values will roll off more bass.  Check out the Screamer article by Jack Orman here: http://www.muzique.com/lab/fatt.htm  ...Note that the resistor and capacitor coming off the (-) input and going to ground (or Vref) both have an effect on the frequency response.
-I was suggesting to use the extra op-amp as a buffer for the Vref voltage, not the signal.  Basically you would configure the extra amp as a noninverting buffer, connect the input to the 1/2 supply point, and use the output for biasing the signal op-amp (R2).  Doing this puts less current demands on the R5/6 point, so you can use larger resistors.  Using larger resistors lets you use a smaller cap while keeping the same noise filtering at that point.  So R5/6 could be increased to 100k/100k and the cap reduced to perhaps 0.1uF. 
-For the oscillation, you might try Gus' suggestion of adding some resistance in series with the output.  100 ohms is a good starting value, increase if necessary, but if you make it 1000 ohms with no improvement, then something else is the problem.

I'll try to look a little closer at this and see if I can find the cause of the problems you're experiencing, but I won't have a chance until after next week.  Good luck!

[waltk]

Alex - Thanks for the reply.  I'll try to educate myself a bit more about this, and implement your suggestions.  I'll let you how it goes, and look forward to your input after next week...

[waltk]

Did I get this right?