Why do most Craig Anderton's projects use bipolar power supplies?

[SpringbokUK]

Hello everyone

I'm sort of new here, I've asked like maybe 3 newb questions many years ago and since then I've just been cloning pedals and not really progressing.

But anyways to the point.
I have owned Craig Anderton's Electronic Projects For Musicians for about 6 years now and I've never built anything from it because most of it is using bipolar power supply not unipolar. Is there a specific reason for this? Most manufacturers use unipolar as far as I am aware. Bipolar does seem inconvenient but I imagine there are positives as well as negatives (that was not a pun) to using bipolar power supplies.

If anyone could perhaps advise me on how to perhaps convert his projects to unipolar that'd be ace. I like my stuff running off one supply.


Nice one
Liam

[mth5044]

Can you post examples?

The typical way of transitioning from bipolar to single supply would be to make the negative voltage to 0V, the 0V to 1/2*supply then have the positive voltage be the supply.

So a +/-9V (-9V, 0V, +9V) would then go to 0V, 4.5V, 9V. That doesn't always work, and if it does work, it might not be a simple whitewash of the schematic. Alternatively you can use charge pump / inverter chips to get a -V supply, but these are typically limited in the current they can provide.

Your best bet is to post a schematic you're interested in and get applicable help.

[Mark Hammer]

I'm tempted to answer it as the late great Professor Irwin Corey - World's Foremost Authority - might have.

However, I will answer you more informatively.

1) Both the first and second edition were published well before the introduction of power bricks, pedalboards, and even use of external wallwarts.  So there was no anticipation of the need to make anything compatible with standard supplies.

2) As the book frequently illustrates, Craig was anticipating use of modules in a rack system, similar to contemporary Eurorack synth modules.

3) There is more headroom with a +/-9V supply than a simpler 9V supply with Vref.  Given the limitations of many op-amps of the time, this was helpful.  It is also worth noting that the brunt of the projects were NOT distortion-oriented ones, so clean was often the objective.

4) Finally, I suspect te contents of the book reflected Craig's design knowledge at the time.  He was, after all, nearly 40 years younger then than he is now.

[SpringbokUK]

I didn't have examples really as i was just trying to cover all grounds.
Yeah see I thought this over. When I look at modern schematics the op amps are always biased at the voltage input. And that actually baffled me for a while as to why engineers were doing this and then I read Craig's book again and noticed the difference and sort of understood but the penny sometimes falls from a great height doesn't it? And the penny usually hits the ground when I'm doing something menial and come to a revelation. What I'm saying is it takes me a while to grasp things. And yeah I've seen charge pump circuits I've used one when I cloned a triple wreck. Perhaps a good charge pump circuit would benefit a lot of circuits. The more power the better!

Haha never heard of this irwell chap although I am in my early 20s and 'ignorance is bliss'  still plagues me.
Yeah see I thought it may have something to do with headroom but I never like to speculate until I have all the answers. And I was actually wondering when this book was published as it does seem like a product of its time and yeah that's a long long time ago.

I think both your responses have added up to a nice solution. I think I need modern books maybe haha.
I was actually nervous about asking questions as the last time I asked a question here I was sort of made to feel a little stupid.

But hey thanks guys!

Prepare to see my other newb questions floating around the forum.
Someone has to ask them though don't they.

[Rob Strand]

I have owned Craig Anderton's Electronic Projects For Musicians for about 6 years now and I've never built anything from it because most of it is using bipolar power supply not unipolar. Is there a specific reason for this? Most manufacturers use unipolar as far as I am aware. Bipolar does seem inconvenient but I imagine there are positives as well as negatives (that was not a pun) to using bipolar power supplies.

It was the design culture of the era.   Opamps were considered as devices with bipolar supplies.   Later on "single supply" opamps came about like the LM324 but they aren't great for audio.  While these were used in car stereo's they majority of audio tasks kept with the standard opamps and bipolar supplies.  Many projects in electronics magazines had bipolar supplies.  The effects pedal industry was probably the biggest user of the single supply configuration; active guitar electronics also used it.  The single supply idea become more widespread through the 80's.

EDIT: While I'd seen the idea earlier, the first time I realized to coolness of single supplies was in the late 70's when I back-engineered an MXR distortion+.

[PRR]

> I never like to speculate until I have all the answers.

The best time to speculate is BEFORE you have all the answers!!

Start speculating when you have half the question.

The *original* (long before your time) use of Operational Amplifiers was to do computations including both + and - quantities. Say you make whiskey at a certain rate for some time, then tap-off whiskey into trucks at a faster rate in less time. Is the barrel more or less full at the end? While you can work that out by hand, a real distillery has multiple stills, multiple shipping dates, variable leakage (crap casks or employee sampling), each still is fed by other processes which must not run dry or run over...

All this (and far more) can be worked out by feeding currents onto capacitors. The 'easy' way to get a precise current is a precise resistor and a high-gain amplifier, an OpAmp. Ultimately many of these problems must cover + and - results. (Also the basic operation is inverting, so half your partial results are electrically '-'.)

So Operational Amplifier "MEANT" bi-polar supplies so-that inputs and answers could be plus or minus.

OpAmps are not needed for Audio. In audio the "answer" is ALWAYS swinging both ways around Zero. So we do need to swing both ways, but we do NOT need to preserve "zero" through the system. We just have input and output caps and bleed resistors which will restore a "zero" reference. We got opamps in audio because '741 and friends got SO cheap and SO easy to design with that fashion rushed this way. We often do work audio opamps with bipolar (look in nearly any large analog studio console). OTOH we do not have to. Broadcast consoles were rigged with single supply (so they could run on a truck battery in case of disaster) and a half-supply reference made for the signals to swing both-ways from (and in/out caps).

As said, older opamp designs borrowed heavily from classic opamp systems with bipolar supply. Hey, this is new, let's do it the most common way. But single supply was "obvious", and after replacing _two_ batteries a few times, the advantage of single-rail became clearer. Not without problem: not long after Craig's book I designed a whole system around single 30V, got to the last part (the LED meter), and discovered I could not get the half-ref meter signal down to the groundy LEDs without another little bit of board and chip.

[reddesert]

I found this application note enlightening, about problems that come up when biasing op-amps to provide a voltage reference in single-supply application: http://www.analog.com/en/analog-dialogue/articles/avoiding-op-amp-instability-problems.html  For example, the difference between Figure 2 (Vsupply/2 bias voltage decoupled from op-amp input by a bias resistor, and stabilized by a capacitor) and Figure 1 is moderately subtle, and many pedal circuits are already designed the superior way, but it is not hard to see how problems could arise, especially in high gain circuits or when the power supply has limited capacity and can suffer noise induced by variable current draw from the circuit.

Back in the Dark Ages, like the 1970s to early 1980s, wall warts were actually relatively expensive and low quality (unregulated, sometimes broke or overheated, low capacity). Some of the larger, more sophisticated early pedals had their own AC cord and transformer inside the box. When you have your own transformer, access to bipolar DC power is easy.

[ElectricDruid]

One reason that no-one has mentioned yet for why the Anderton schematics use bipolar supplies is "because it's much easier to design op-amp circuits with bipolar supplies". Sure, you can make do with single supplies if you learn the tricks for designing that way, but pretty much all op-amp circuits work more easily and with less parts on a "proper" bipolar supply. Most of the extra stuff you do on a single supply is just to make that single supply look like a bipolar one from the op-amps point-of-view. In that sense, there aren't any single supply op-amp circuits, just some that are pretending!

Tom

[poiureza]

Yes I concur.  Just how difficult is it to add an ICL7660 in your circuit and benefit from bipolar PS ?
There are so many circuits that benefit from increased headroom.

[antonis]

Just how difficult is it to add an ICL7660 in your circuit and benefit from bipolar PS ?
There are so many circuits that benefit from increased headroom.

"Headroom" is considered as the magnitude of power supply, so why bother with ILC7660 instead of doubling your single PS just from the start of your design..??

[SpringbokUK]

All interesting stuff

I think I need to study op amps for some time. I guess using bipolar ps is convenient for sole use because then I could power all the circuits I've built that require it but as it is fashion these days to power our pedals using unipolar ps then I guess it isn't convenient for everybody else?

Think i'll have a good ol search for some new books!

[neilmcnasty]

The simple answer:
Unipolar is the standard used for pedals and similar.
Bipolar is being used in studio gear, so that the total voltage becomes 18v (the difference between +9v to -9v), which (amongst other) gives more headroom, an the fact that the (half voltage) bias point then becomes 0v (ground).

Sent from my C6603 using Tapatalk

[PRR]

> Bipolar is being used in studio gear, so that the total voltage becomes 18v (the difference between +9v to -9v)...

Except, being free of battery legacy, the default pick in much opamp work is +/-15V, the rails used in Analog Computers, picked so that max signal could be a nice round 10V.

Some studio gear pushes the ratings, +/-18V and even +/-24V. I've used +/-17V because it is easy to get from a 12V AC winding.

Some admits that a few volts more/less makes little difference, and takes +/-12V as a convenient supply.

[Rob Strand]

Some studio gear pushes the ratings, +/-18V

I've seen plenty of those too.

I suppose a lot of datasheets were spec'd at +/-15V.  Without any extra info it's not a bad idea to follow the datasheet.

[PRR]

> datasheets were spec'd at +/-15V.  Without any extra info it's not a bad idea to follow the datasheet.

The performance ratings are usually given on +/-15V because this is the traditional supply. (Many chips now also have performance on lesser supplies.)

Most of the classic analog chips were a 40V process, but they learned to call it "36V process" (+/-18V) to cover customer slop. That is by far the most common data sheet limit.

There was one console maker who knew most of these chips aimed at 40V, and apparently got a maker to "select" TL072 good for +/-20V supplies. We know this because 30+ years later these systems have more failed chips than most. Granted, 30+ years is far-far-far past what anybody expected in those days: consoles were often replaced in 5 years as record production got more and more complex.

[Rob Strand]

The performance ratings are usually given on +/-15V because this is the traditional supply. (Many chips now also have performance on lesser supplies.)

Yes, it's good because you get more info but bad because the datasheets grow. (And if you haven't had your morning coffee you can read the wrong table.)

Most of the classic analog chips were a 40V process, but they learned to call it "36V process" (+/-18V) to cover customer slop. That is by far the most common data sheet limit.

Yep.

There was one console maker who knew most of these chips aimed at 40V, and apparently got a maker to "select" TL072 good for +/-20V supplies. We know this because 30+ years later these systems have more failed chips than most. Granted, 30+ years is far-far-far past what anybody expected in those days: consoles were often replaced in 5 years as record production got more and more complex.

It's like most things the closer you push the boundary the sooner it will fail.  A lot of defense and medical stuff is limited to less than 80% of the maximum ratings.  Usually for power but the same thing applies for other stress parameters.   A lot of flyback supplies fail because the switches sail too close to the wind (sometimes because of the snubbers aren't done right, others for efficiency.)