20V Clean Boost, will it work?

[lukasthielmann]

Hi,

I've been reading this forum for awhile now it seems that every google search leads me here. Rod Elliot's site has been incredibly helpful also http://sound.westhost.com/dwopa.htm

I've been trying to learn how circuits actually work (rather than just copying other people's work). I thought a good place to start would be a super simple opamp boost. I know it's nothing revolutionary but I've learnt a lot through the process.

Schemaic is here https://drive.google.com/file/d/0B0umcThVkof-ZENuWXZVQWZwblU/view?usp=sharing

I have a few  questions though.

1) Will this circuit work?
2) In the power supply section will the 100nF cap act as a bypass cap for the opamp? (bypassing is one thing about opamps I still don't get)
3) Is it better to keep the over all resistance high or low in the opamp feedback loop? (keeping the ratios for gain constant obviously)
4) Gain is set at 3.4Av (10dB), is that about right or does it need to be higher/lower?

Why 20v?
Mainly because I wanted to have heaps of headroom and I had a 22v DC wall-wort that was just sitting there.

Thanks heaps,

Lukas

[GibsonGM]

Hi Lukas, welcome to the forum

Some good questions there.   Without getting too technical (as I don't know how much background you have), we can say that decoupling caps act like "little batteries" to supply current when a circuit draws "a lot", when it spikes, bringing the supply down a little...the decouplers 'smooth out' what would be a rippling voltage.   This makes circuits operate "better" (sometimes it is critical to their operation).    They help provide CLEAN power.

I think you should review some opamp circuits, there seems to be a few things you should change in your circuit for it to work well. 
We use a BIAS VOLTAGE (often called "Vref") to make the opamp work in the middle of its supply range, rather than "from zero to supply" - if we don't, it will only amplify positive signals and clip the negative parts!!   So, check this out:  http://www.ti.com/lit/an/sloa030a/sloa030a.pdf

Don't get too hung up on the math, just 'get the idea' and look at some known schematics such as a Tube Screamer.   It's ok to copy the power section from a circuit like that to get Vref, as long as you know that it's a voltage divider that is decoupled.  Pretty standard.

I believe that fairly high R values in the loop are common and 'good' - to avoid loading the incoming signal, for one thing.  Look at other designs, and design notes like the one I linked to for more info.    Don't expect your wall wart to give you exactly 22V unless it's regulated already...you might get some 'odd' results from it, and your regulator, like ending up with 18V or something.

The gain is something YOU set (don't know what you want to get out of it).  I might make R4 a 1Meg pot (max gain about 11), make gain variable!  Take some AC measurements at output, make sure you're not putting out too much voltage, ha ha....     I suspect you'd only get 5 to 6v out from a guitar signal in - so if you really want to USE the higher voltage there, add another stage with some 'tone' shaping in between.   As-is, with a bias added, you'd swing around about 10V....so you'd drop to 5V and go up to 15v.  10V P-P, that can punish an amp, ha ha.

Play around with this, see what you get!  With the addition of a reference/bias voltage, I believe you will have a working circuit.

[Gus]

First what peak to peak input voltage can the following device take?

You need a Vref, look at other gain stages like this that run on a single supply

The input cap need to be on the other side of the 1 meg

It is hard to make something different with an opamp boost most(all) of the time it is textbook circuits for boost

Look at places like the TI web site for app notes and circuits

search this forum for opamp based boosts

You have 100 k in series with the output 250K volume control 240/340 max output level about .7.  Think about how this will be loaded by the following device(s).

EDIT I guess we posted about the same time with some of the information

Also I like to have a 3VDC min drop on the 317 for regulation

[lukasthielmann]

Wow thanks,

It's late in NZ so I'll have a proper read tomorrow.

Had a quick look over the TS scheme. Could I split the 1MOhm pull-down into two 500k's and put them after the input cap to make a voltage divider? That way there is 10v going into the non-inverting input.

Thanks,

[tubegeek]

Re: splitting the pull-down:

Almost, but that's not quite how you want to do it.

You DO need a voltage divider, connected after the input cap, so your thinking is correct there.

But if you remove the pull-down before the cap, you lose its function which is to prevent a "pop" when switching the effect in and out.
(If your booster is always-on, no need for the pull-down.)

Also the three nodes of the voltage divider are +20 and ground at the two ends and V_bias in the middle. You weren't clear in your post whether you were planning to connect to +20 or not - but it was late in NZ so that was probably why.

An op-amp draws so little input current that this next advice is partly more educational than practical, but here goes.

Ideally the bias voltage is what's called a "stiff" supply: low-impedance so that it will not be affected as much by loading.
Also the input bias node must be VERY clean because any noise introduced here will be amplified by everything afterwards. Larger resistors contribute more noise by virtue of their own properties.

So for those two reasons, you'll want to go somewhat lower on those two resistors. More "wasted" current but better V_bias supply. Maybe 5K-50K instead of 500K each. I'd bet Rod Elliot has this explained somewhere on the ABSOFRICKINLUTELY AMAZING ESP site.

(The fact that you've located that site early on in your stompboxology career gives you immense power, use it for good instead of evil and you'll be a wizard in no time. Thank Ohm for Rod Elliott!)

The "noiseless bias" concept is to put a filter cap between that middle voltage-divided node and ground, and then add a resistor in series with that node to the input you'd like to bias. The purpose of that one, is to prevent the actual SIGNAL from being bypassed to ground. It's also to isolate each item you want to feed the V_bias from each other, if there is more than one. Search the phrase "noiseless bias" on this site and you'll get a fuller explanation.

[lukasthielmann]

Wow thanks everyone!

I think I get it now. I've got tomorrow off so I'll breadboard it and see how it works!

[Gus]

Something posted before to the left of input is a simple guitar cable sim this also has a lowpass filter at the end

[tubegeek]

Here's the obligatory link to GEOFEX:

Mod your Mu-amp/Minibooster

[samhay]

Take another look at the LM317 data sheet too. Your R2 wants to be significantly bigger - 220R(ish).

[Ben N]

This old thread may also be of interest (especially Dave Barber's input): http://www.diystompboxes.com/smfforum/index.php?topic=33578.0

[PRR]

> trying to learn how circuits actually work (rather than just copying other people's work).

Keep copying. No shame in that. You learn a lot from a GOOD copy. Copy musical composers from Mozart to McCartney; every dot is where it is for a reason, and copying is the best way to meet every dot and ponder why it is there.

On your plan: what is the idle DC voltage on opamp "+in" pin? Can not say. What should it be? Which resistor is bigger on a LM317? Why?

Feedback impedances.... Lower is less hiss until it gets down below the source impedance. For e-guitar you should be below 50K, maybe 10K, on the lower resistor. However small resistors will load the opamp, so for TL07x the total R4+R5 should be above 2K. And these resistors versus desired bass response set the value of C6, which may sometimes become awkward for practical reasons.