Buffered mults

[raisinbag]

I'd like to make up a buffered mult on strip board for shits and giggles. I can't seem to find a decent schematic to learn from. I am guessing you basically run signal through an opamp at unity gain and then?????????

[sduck]

Boy am I glad you asked. I was thinking of building up a few also, but my powers of searching on this website are weak, so hopefully someone will post a schematic. Or I'll have to try doing the dreaded search...

[roglok]

this discussion is just going on over at e-m: http://electro-music.com/forum/topic-56272.html

[Barcode]

Here's another thread. This one has some layouts posted.

http://www.electro-music.com/forum/viewtopic.php?t=49356

[raisinbag]

Boy am I glad you asked. I was thinking of building up a few also, but my powers of searching on this website are weak, so hopefully someone will post a schematic. Or I'll have to try doing the dreaded search...

Ya I already did the dreaded search and found nothing frutful. I'll go check out the posts you other guys made and check back with more questions (when I'm off the ski hill, yipeeeeeee!)

[russma]

For the uninitiated, how is a buffered mult different than an un-buffered one?

I'm guessing it's something to do with maintaining the voltage level across all outputs?

[hpsounds]

viewtopic.php?t=46106

Hédi K.

[raisinbag]

Well from my understanding, you are running the signal through an opamp (or a few opamps) when buffering. When not buffering you just split the signals. Whole purpose of buffering is to retain signal strength and I think also takes care of impedance issues.

[burdij]

Here you go:



Place a temporary load, the 100K resistor with the dotted line, on one of the outputs. Put a steady voltage signal into the input. Adjust the potentiometer so that output voltage measured at the top of the load resistor equals the input voltage.

[sduck]

Thanks all!

[burdij]

I made a slight mistake in the schematic which has already been corrected. The potentiometer should be a 10K which will give a X1 +- 5% gain adjust range.

[raisinbag]

Yay Thanks that is exactly what I was looking for!!!!!!!!!!!!!!!

[EMwhite]

How critical is the value of the 27pF cap (will 33 be ok?);

What about the 95.3K resistor. Naturally, I don't have that in my parts bin so how about 30+30+30+3.3+3.3... new at this so please bear with me.

Is the point that you can affect between a 90-something K resistance and a 105 or so amount of resistance thus boosting or cutting the voltage in order to dial in unity?

Final stupid question of the evening, how to wire the 10K pot? I've seen the 1,2,3 (ground,wiper/center,+xV) wiring, should two of the terminals be connected together, if so which? Thanks.

[Barcode]

What about the 95.3K resistor. Naturally, I don't have that in my parts bin so how about 30+30+30+3.3+3.3... new at this so please bear with me.

I'm guessing you have loads of 100k resistors laying around like me. And yes they will work.

[Luka]

i built the one listed in this thread - the schematic provided by Big G resolved all my concerns with standard buffering circuits and works great

viewtopic.php?t=19142&highlight=buffered+mult

[glitchpop]

Great thread

I have a few more noob questions for you burdij once you cover EMwhites

1. IC is listed as TL0XX. Does that mean any opamp will do?

2. +-12V is not shown on the schematic is that assumed to be connected to appropriate pins? Or is it/ will it work unpowered?

Thanks!

[burdij]

How critical is the value of the 27pF cap (will 33 be ok?);

The capacitor gives the circuit stability when the gain of the amp is set near X1. The value is not too critical. I would say anything between 22 and 100pF would be fine. If you go higher than that, you will be affecting the high frequency response.

What about the 95.3K resistor. Naturally, I don't have that in my parts bin so how about 30+30+30+3.3+3.3... new at this so please bear with me.

95.3K is the nearest value to 95K in 1%. A 82K and a 12K in series would work in 5% values.

Is the point that you can affect between a 90-something K resistance and a 105 or so amount of resistance thus boosting or cutting the voltage in order to dial in unity?

Yes, that is the reason. A common use of a buffered mult would be to distribute note CV values to multiple oscillators. The value of the gain needs to be as close to X1 as possible in order to maintain the 1V/Octave scaling. If the gain is not close to that value, you would be only able to tune all oscillators to be in tune at one value. The oscillators will stay relatively in tune with each other but all would go out of tune the further from that "in tune" value they got.

Final stupid question of the evening, how to wire the 10K pot? I've seen the 1,2,3 (ground,wiper/center,+xV) wiring, should two of the terminals be connected together, if so which? Thanks.

You could use only two of the three pot connections, the wiper and one of the end connections. I usually connect the wiper to one of the end connections and use that in conjunction with the other end connection in this sort of application. That way, there is always a DC path returning to the negative input from the output which prevents bumps in the gain of the amp as the wiper scraps along the surface of the resistance element. If the path opens briefly, the gain of the circuit will head for infinity and produce noticeably more noise in the output. I usually also arrange the connections so the higher resistance occurs in the clockwise direction which would produce higher gain but that is only a cultural bias toward right-handedness.

[burdij]

1. IC is listed as TL0XX. Does that mean any opamp will do?

I was suggesting using an op amp such as the TL062, TL064, TL072, TL074, TL082, TL084 series but any quality, low offset, high speed opamp will work. The capacitor assures stability so the choice of device shouldn't matter too much. I wouldn't use "old timey" op amps like 741s though as they just don't have very good frequency response.

2. +-12V is not shown on the schematic is that assumed to be connected to appropriate pins? Or is it/ will it work unpowered?

The op amp power pins do need to be connected to a power source. Whether it is +-12V or +-15V doesn't matter. You should use the same power as your other synth modules are powered with.

[qfactor]

The op amp power pins do need to be connected to a power source. Whether it is +-12V or +-15V doesn't matter. You should use the same power as your other synth modules are powered with.

Pardon, but could this buffer work off a 9v battery? Am looking at attempting a "portable" buffered mult!

[roglok]

i built the one listed in this thread - the schematic provided by Big G resolved all my concerns with standard buffering circuits and works great

viewtopic.php?t=19142&highlight=buffered+mult

so... how does burdij's single op-amp design with output compensation in the feedback path compare to Big G's circuit? I was under the impression that each output needs its own buffer.

[burdij]

i built the one listed in this thread - the schematic provided by Big G resolved all my concerns with standard buffering circuits and works great

viewtopic.php?t=19142&highlight=buffered+mult

so... how does burdij's single op-amp design with output compensation in the feedback path compare to Big G's circuit? I was under the impression that each output needs its own buffer.

You can tie as many 101K loads (the equivalent load of the output resistor and one input load of the next stage) to the output of the opamp in the circuit shown as you like as long as the total load at the maximum output swing stays below the maximum current output of the op amp used. For a TL074, the maximum output swing is +-13.5 V with a 10K load. Ten parallel output loads of 101K is roughly 10K ohms. I don't know if I would use that many as the total capacitance of all the cables plugged into the outputs would start to become a significant value.

[burdij]

The op amp power pins do need to be connected to a power source. Whether it is +-12V or +-15V doesn't matter. You should use the same power as your other synth modules are powered with.

Pardon, but could this buffer work off a 9v battery? Am looking at attempting a "portable" buffered mult!

If you use two 9V batteries, one for positive and one for negative, the circuit should work properly. Remember that lowering the supply voltages will also lower your output voltage swing "headroom". For a TL074, the output voltage will swing to within about 1.5 volts of the power supply voltage values and, as mentioned in the previous post, will also be dependent on the load on the opamp output since the amplifier has some finite output impedance. You could look around for an opamp designed to do "rail-to-rail" output swing like an LT1677 for this application to fully utilize the available power supply range.

[roglok]

i built the one listed in this thread - the schematic provided by Big G resolved all my concerns with standard buffering circuits and works great

viewtopic.php?t=19142&highlight=buffered+mult

so... how does burdij's single op-amp design with output compensation in the feedback path compare to Big G's circuit? I was under the impression that each output needs its own buffer.

You can tie as many 101K loads (the equivalent load of the output resistor and one input load of the next stage) to the output of the opamp in the circuit shown as you like as long as the total load at the maximum output swing stays below the maximum current output of the op amp used. For a TL074, the maximum output swing is +-13.5 V with a 10K load. Ten parallel output loads of 101K is roughly 10K ohms. I don't know if I would use that many as the total capacitance of all the cables plugged into the outputs would start to become a significant value.

Thanks, that sounds great! Still I don't understand why most buffered multiple designs are built around multiple op-amp buffers when it only requires one. What I understood from your reply is that with your design it's possible to use a single op-amp (e.g. TL071) for up to 10 outputs and still have an accurate 1:1 output scale. Are other designs overcomplicating the matter then?

[Eric G]

Here you go:

I don't see the point with all those components. One non-inverting OPamp will do fine, and skip all the resistors. Also skip the 1k resistors and you need no calibrations.
The 1k resistors are only for short-circuit protection, and if you happen to short and destroy an OPamp (I never did) then just replace it.

[CJ Miller]

When I saw the thread title, I thought it said: "Buttered Muffs"...