Balancing an unbalanced circuit

[tedsorvino]

If not 1% then what - 5% ? 1% tolerance Metal Film is inexpensive now and basically the standard thing for non-precision commercial application.

IMO the right value of resistor (or any passive component) is more important than its tolerance (and its extra capabillities) and paying mouser (or any other specialist e shop out of my city) something like 0.7 euro per resistor (+ shipping) for a cheapo pedal thingy is not for me. I know it will sound "better" than most behringers or mooers etc (and it will cost less) anyways.
If I find 10 for 1 euro (or something normal) of course the best tolerance make my BOM.
But anyways...resistors tolerances is not a debate I would enter and precision maybe "ideal" but not something I give a lot of thought.
Music and sound is absolutely subjective.
But just to put things back in order thanks for your posts Newmarket. I do appreciate your input, knowledge and advice.

 

[Matt Syson]

As an observation, having an 'impedance balanced output' compared to unbalanced is pointless unless the input it will be working into is properly and well balanced. The output impedance of a TL071 or 2 is not very low at all frequencies although as a buffer, (as used here) it is probably better than if it was being used as a stage with gain. Also the TL072 running with only a 9 volt supply would run out of voltage headroom so investigating a dual op amp that is 'specified' as 'rail to rail' inputs and outputs may be beneficial. As noted previously the impedance of the 10uF capacitor is significant at powerline frequencies 50/60 Hz although the mismatch between any 2 'freshly manufactured capacitors will be relatively small. So basically there are many things 'wrong' with this scheme but it obviously basically works so probably best to put the lid on the box and play music instead.
Matt S

 

[ruffrecords]

Yes indeed. And in general that's what I was initially asking. Guitar input devices transformed for good use as line level devices

OK, understood. The reason I asked is because in my view this is not a particularly good design. I can understand converting effects pedals like fuzz or phase to balanced line in/out but there are plenty of much better compressor designs than the one you posted.

And a couple of comments on other posts. Yes, 10uF output capacitance is OK for feeding a guitar amp input bit rather low for feeding a line input. If you can it would be worth increasing this to 100uF.

Buying small numbers of 0.1% resistors is rather expensive. The important thing about the two resistors in an impedance balanced output is not the exact value but that they are the same value. So what I do is take ten 1% resistors and measure them until I find two of the same or very nearly the same value. It does not even matter that the meter used to read them is not accurate as long as it is consistent. I use this process all the time for matching phantom power feedd resistors.

Cheers

Ian

 

[ccaudle]

IMO the right value of resistor (or any passive component) is more important than its tolerance (and its extra capabillities)

I have no idea what "extra capabilities" of a resistor would be, but tolerance is specifically how close the value is to the "right value" you selected.
In a guitar pedal close tolerance is not likely to matter, but in a filter circuit the tolerance of the resistors and capacitors will determine how closely the response matches the design response, and in a stereo or surround sound device the tolerance will determine how closely the channels match each other.
In a balanced interface the tolerance determines how well the interface can reject common mode noise (to a certain extent, because imbalances in the source and cable affect that as well, which is why high common mode impedance is important).
You did start this thread asking about balanced interfaces, and component matching is very important there.

As Ian mentioned, close tolerance does not necessarily mean buying 0.1% resistors if you have the time to match devices yourself. And the advantage of the That Corp or TI devices is that they do the resistor matching for you in package so you don't have to worry about it.

But to say as a blanket statement that device tolerance isn't important just indicates that you have not thought through different scenarios other than your fairly narrow single channel device use case.

 

[JohnRoberts]

8 pages of posts and still veering off into the weeds.

JR

 

[tedsorvino]

I have no idea what "extra capabilities" of a resistor would be, but tolerance is specifically how close the value is to the "right value" you selected.
In a guitar pedal close tolerance is not likely to matter, but in a filter circuit the tolerance of the resistors and capacitors will determine how closely the response matches the design response, and in a stereo or surround sound device the tolerance will determine how closely the channels match each other.
In a balanced interface the tolerance determines how well the interface can reject common mode noise (to a certain extent, because imbalances in the source and cable affect that as well, which is why high common mode impedance is important).
You did start this thread asking about balanced interfaces, and component matching is very important there.

As Ian mentioned, close tolerance does not necessarily mean buying 0.1% resistors if you have the time to match devices yourself. And the advantage of the That Corp or TI devices is that they do the resistor matching for you in package so you don't have to worry about it.

But to say as a blanket statement that device tolerance isn't important just indicates that you have not thought through different scenarios other than your fairly narrow single channel device use case.

By extra capabilities CCaudie, I mean special audio - guitar capacitors vs simple capacitors or special op amps vs simpler op amps etc. I mean all this specialist b@#$^t. And yes the TI sounds exactly the same as the THAT to me. And you 're surely right on what you mention in all things, since it's your opinion and I do respect it, but I just don't care too much about it.
By the way ...please open a decent 1960s compressor or pre amp or guitar amp and i guess you will not find anything like a precision resistor. And if the device under examination doesn't sound fantastic then we surely have different tastes. Especially if we are talking about fairly narrowly but super coolly used r'n'r devices.

 

[tedsorvino]

OK, understood. The reason I asked is because in my view this is not a particularly good design. I can understand converting effects pedals like fuzz or phase to balanced line in/out but there are plenty of much better compressor designs than the one you posted.

And a couple of comments on other posts. Yes, 10uF output capacitance is OK for feeding a guitar amp input bit rather low for feeding a line input. If you can it would be worth increasing this to 100uF.

Buying small numbers of 0.1% resistors is rather expensive. The important thing about the two resistors in an impedance balanced output is not the exact value but that they are the same value. So what I do is take ten 1% resistors and measure them until I find two of the same or very nearly the same value. It does not even matter that the meter used to read them is not accurate as long as it is consistent. I use this process all the time for matching phantom power feedd resistors.

Cheers

Ian

I know Ian, that many better things exist in the compressor world. And I ve built some of those cool designs. Like a great dbx style based on the 1252 ic, or the (not so great) op amp version of 1176, or simpler ross compressor (with a wampler dan armstrong tone and mix control), or a couple of diamond opticals (like a cheapo LA - 3A version).
But this shure thingy is a great limiter ruff and ready soft drive sounding thing.

 

[tedsorvino]

As an observation, having an 'impedance balanced output' compared to unbalanced is pointless unless the input it will be working into is properly and well balanced. The output impedance of a TL071 or 2 is not very low at all frequencies although as a buffer, (as used here) it is probably better than if it was being used as a stage with gain. Also the TL072 running with only a 9 volt supply would run out of voltage headroom so investigating a dual op amp that is 'specified' as 'rail to rail' inputs and outputs may be beneficial. As noted previously the impedance of the 10uF capacitor is significant at powerline frequencies 50/60 Hz although the mismatch between any 2 'freshly manufactured capacitors will be relatively small. So basically there are many things 'wrong' with this scheme but it obviously basically works so probably best to put the lid on the box and play music instead.
Matt S

Cheers Matt Syson. Just to mention that a lot of designs use the very simple and cheap TL072 in the guitar pedal world. Sometimes fed through power pump ic supplies. But in general (apparently) 9v seem adequate. Strangely enough TL071 is not used so much. For single op amps some LM53something or even TL061 are more frequently used. Can't tell why...I guess matter of habit-"tradition" or just taste and tone.

 

[ruffrecords]

Cheers Matt Syson. Just to mention that a lot of designs use the very simple and cheap TL072 in the guitar pedal world. Sometimes fed through power pump ic supplies. But in general (apparently) 9v seem adequate. Strangely enough TL071 is not used so much. For single op amps some LM53something or even TL061 are more frequently used. Can't tell why...I guess matter of habit-"tradition" or just taste and tone.

I think Matt is alluding to the problem of headroom. A 9V supply is probably adequate for a guitar input device but a line level device will really need a higher value supply in order to maintain a decent headroom.

Cheers

Ian

 

[ruffrecords]

I know Ian, that many better things exist in the compressor world. And I ve built some of those cool designs. Like a great dbx style based on the 1252 ic, or the (not so great) op amp version of 1176, or simpler ross compressor (with a wampler dan armstrong tone and mix control), or a couple of diamond opticals (like a cheapo LA - 3A version).
But this shure thingy is a great limiter ruff and ready soft drive sounding thing.

No problem. If you like the way it sounds that is of course fine. SO, to complete the answer.,to balance the input you can repurpose the original input op amp in the exact same way they do it in rhw LA3A.

Cheers

Ian

 

[ruffrecords]

I have no idea what "extra capabilities" of a resistor would be, but tolerance is specifically how close the value is to the "right value" you selected.
In a guitar pedal close tolerance is not likely to matter, but in a filter circuit the tolerance of the resistors and capacitors will determine how closely the response matches the design response, and in a stereo or surround sound device the tolerance will determine how closely the channels match each other.
In a balanced interface the tolerance determines how well the interface can reject common mode noise (to a certain extent, because imbalances in the source and cable affect that as well, which is why high common mode impedance is important).
You did start this thread asking about balanced interfaces, and component matching is very important there.

As Ian mentioned, close tolerance does not necessarily mean buying 0.1% resistors if you have the time to match devices yourself. And the advantage of the That Corp or TI devices is that they do the resistor matching for you in package so you don't have to worry about it.

But to say as a blanket statement that device tolerance isn't important just indicates that you have not thought through different scenarios other than your fairly narrow single channel device use case.

I think I understand what Ted is getting at. If you try to buy 1% resistors you will probably get values from the E96 series and these are always expensive. However, if you shop around, you will find regular E24 resistors, which many people will erroneously tell you are only 5% tolerance, but which are actually accurate to1% and these are cheap. In other words, a 1% expensive 6K81 resistor is no better than a cheaper 1% 6K8. The error is in assuming the tolerance has to be indicated by the number of digits in the value.

Cheers

Ian

 

[Matt Syson]

A TL061 is low current consumption so handy for battery powered circuitry. You would need to lok at data sheets for low voltage operation as some usual' types don't really wake up and become linear amplifiers until the total supply is about 7 Volts (hence the use of charge pumped supplies in some units). There are hundreds of different attempts to make a 'perfect' operational amplifier and none actually make the idealised concept a reality.

 

[tedsorvino]

No problem. If you like the way it sounds that is of course fine. SO, to complete the answer.,to balance the input you can repurpose the original input op amp in the exact same way they do it in rhw LA3A.

Cheers

Ian

Thanks Ian for all the clarification. Such to the point answers and "matter-of-fact-ly" really help online communication and support.
Could you please be more specific on what you mean be "repurpose" and what is the way they do it on "rhw (???) LA3A" - is there any schematic for that?
And after all how many volts would I need per rail? I guess more than 9v.

Thanks

 

[Newmarket]

As Ian mentioned, close tolerance does not necessarily mean buying 0.1% resistors if you have the time to match devices yourself. And the advantage of the That Corp or TI devices is that they do the resistor matching for you in package so you don't have to worry about it.

Might not be an issue but there is the technical point that high precision (low tolerance) resistors tend to come with low tempco eg 10/20 ppm (0.1%) cf 100/200 ppm (1%).
Better still it's usually the ratiometric accuracy that is often the important thing (eg look at your typical '4 Resistor; Diff Amp). This is where the on-chip laser trimmed resistors excel since they are on the same substrate and so highly thermally coupled. There is the issue that you are basically stuck with the values chosen and they might (emphasise "might") be higher than you want.
There is the option to use discrete "dual" resistor parts where there are two resistors on the same device. I designed these in previously to high precision electronics. IIRC the absolute tolerance is spec'd at 10ppm but ratiometrically the tolerance under the specified test conditions is 2ppm.
And a wide range of values is available (allowing for supply chain issues).

 

[Newmarket]

By extra capabilities CCaudie, I mean special audio - guitar capacitors vs simple capacitors or special op amps vs simpler op amps etc. I mean all this specialist b@#$^t. And yes the TI sounds exactly the same as the THAT to me. And you 're surely right on what you mention in all things, since it's your opinion and I do respect it, but I just don't care too much about it.
By the way ...please open a decent 1960s compressor or pre amp or guitar amp and i guess you will not find anything like a precision resistor. And if the device under examination doesn't sound fantastic then we surely have different tastes. Especially if we are talking about fairly narrowly but super coolly used r'n'r devices.

Just ignore the "guitar capacitor" nonsense.
TI parts do "sound" the same as THAT stuff in reality. The whole point of a balanced interface is to avoid noise problems caused by "Ground Problems" and/or external interference, When this is an issue is when you might experience a difference in performance. Or if there is a fault in the wiring / cabling.
re "60s Compressor" etc - of course not - high performance modern parts didn't exist.
The fact is that carbon type resistors as you might find in there are noisier (non-ideal resistor characteristics) than the metal film types now available. If you like the noise then fine - but that is not "fidelity".
But it's not as objectionable as the noise that a balanced interface is designed to reject as that tends to specific frequencies rather than broadband/pink/white noise (basically flavours of "Hiss").

 

[Newmarket]

Cheers Matt Syson. Just to mention that a lot of designs use the very simple and cheap TL072 in the guitar pedal world. Sometimes fed through power pump ic supplies. But in general (apparently) 9v seem adequate. Strangely enough TL071 is not used so much. For single op amps some LM53something or even TL061 are more frequently used. Can't tell why...I guess matter of habit-"tradition" or just taste and tone.

Yes- TL072 is widespread. There are top grade mixing desks where TL072s are everywhere except where a NE5534 (historically expensive) is used. But we have better now so it's better to use those.
TL06x series offers lower power option. It's often considered important for pedals - battery life, But there's a noise penalty.

 

[ruffrecords]

Thanks Ian for all the clarification. Such to the point answers and "matter-of-fact-ly" really help online communication and support.
Could you please be more specific on what you mean be "repurpose" and what is the way they do it on "rhw (???) LA3A" - is there any schematic for that?
And after all how many volts would I need per rail? I guess more than 9v.

Thanks

I recommend you check out thie first part of this article by Rod Elliot:

Audio Designs With Opamps - 3

in fact his whole web site is a wealth of useful audio information.

Cheers

Ian

 

[abbey road d enfer]

please open a decent 1960s compressor or pre amp or guitar amp and i guess you will not find anything like a precision resistor.

Just because these circuits do not require precision resistors. Even a 10% deviation leads to less than 1dB gain variation, worst case, which is not an issue.
But active balanced interfaces do require precision matching. CMRR is not a matter of "sounding good", it's purely a mathematical thing.
Actually, for many years, British console manufacturers had CMRR trimmers in their balanced output stages.

 

[tedsorvino]

I recommend you check out thie first part of this article by Rod Elliot:

Audio Designs With Opamps - 3

in fact his whole web site is a wealth of useful audio information.

Cheers

Ian

Yes indeed it is a classic (like the geofex and beavis audio sites). I had visited it sometimes in the past for other info (power supplies i think) and it was really informative.