Tremolo circuit 9v or 18v

[jethrofloyd67]

Hey guys, I've been working on a tremolo design for a couple months now and have finally got it just how I want it.
There's just 1 more thing, I would like the option of 9v or 18v as far as the audio path for headroom options.
but the LFO was designed using 9v (battery) and uses LEDs, would it mess up the finely tuned harmony I've achieved if I sent it 18v?
I dont have an 18v adapter to try it out and want to make sure it wont mess it up.

Thanks in advance 

[R.G.]

There's just 1 more thing, I would like the option of 9v or 18v as far as the audio path for headroom options.
but the LFO was designed using 9v (battery) and uses LEDs, would it mess up the finely tuned harmony I've achieved if I sent it 18v?

Depends on the circuit. Some yes, some no. Back in one of the electronics classes, we had to derive sensitivity equations - things like how much X changed in relation to changes in Y, where X could be a gain, DC bias point, impedance, practically anything, and Y might be a component value, power supply, bias voltage, or transistor gain. The point of this was to teach us to think in terms of "if I change the power supply/resistor value/bias/transistor type, how far does this change what I'm interested in?

Roughly speaking, you're asking "what is the sensitivity of some thing(s) about my LFO to doubling the power supply voltage?", and that's impossible to determine without the circuit schematic. Unless they're designed specially for it, discrete single-ended circuits tend to have a sensitivity of 1 (i.e. change Y by 50%, X changes 50%) or worse, sometimes much worse. One of the reasons opamps exist is that they're designed to make things like opamp gain, distortion characteristics, power supply voltages, etc. not matter at all in most cases, and to have a sensitivity of exactly 1.0 to the programming resistors. Other opamp circuits are not as immune to external effects.

I dont have an 18v adapter to try it out and want to make sure it wont mess it up.

If I were you, I'd buy a second 9V battery and try it that way. And I'd also try it with 7V (weak battery), 8.5V (nominal 9V battery under load), 12-14V (two weak batteries), 18V and 19V, just for starters, especially if you plan to sell these (and most of the people asking questions like this are so planning) and especially if you don't want unhappy customers or a bad rep.

[jethrofloyd67]

its based around the LFO in the Tremulus lune, I can post a schem if that helps.
I've nearly plugged in 2 batteries but worried about burning out the LEDs, I'm an overly cautious fellow.
I do plan on selling them, and still have a good bit of testing to do.
its no biggie to add a 9v regulator for the LFO section. option 2 is to add a charge pump with toggle for 9v or 18v, and just send the LFO straight 9v

[armdnrdy]

If your goal is headroom, but you would like to be able to run it on standard 9 volt pedal voltage, I would tend to lean more towards a design that incorporates a charge pump to create +9/0/-9 bipolar voltage.

You gain headroom and don't have to worry about trying to find quantities of 18VDC wall adapters to power the pedals you are offering for sale.

[merlinb]

I can post a schem if that helps.

ALWAYS POST THE SCHEMATIC.

[samhay]

If your goal is headroom, but you would like to be able to run it on standard 9 volt pedal voltage, I would tend to lean more towards a design that incorporates a charge pump to create +9/0/-9 bipolar voltage.

+1 and you can leave your LFO running on +9/0 single supply.

[midwayfair]

Imagine going into class one day, and everyone's talking about a book that you've not only never read, but it's brand new and you've never even heard of it. You've read books before, so you know what a book is, but everyone starts asking you questions about the plot and the characters and you're expected to give an intelligent answer to the questions.

[jethrofloyd67]

Ok sorry, here's the schem

I hadn't thought of a bipolar supply, haven't given them much of a chance but I guess as long as I'm using a charge pump its worth trying
just for money/space saving reasons is there any down side to just putting a 9v regulator between DC in and the LFO and letting folks power it how they want?
I like the idea of putting voltage option inside, but I'm not keen on redo-ing the whole layout or making a daughter board (I'm just a tad bit lazy  )

[jethrofloyd67]

found this bipolar charge pump, so I just change the gnd connections in the audio circuit to -9v and connect the LFO as usual?

[armdnrdy]

I would use the more dependable LT1054CP.

It's not going to be quite so easy to change the +9 / GND circuit to +9/0/-9

Take a look at this circuit. This is a panner; but if you delete one of the outputs...basically it's a tremolo.

Maybe you can get some ideas how the bipolar supply is implemented in the audio section.

[jethrofloyd67]

I already have a few 7660scpa chips, saw on another post about them being more dependable than 1044's, if the 1054 is even better then I'll get those next time.

so it seems that only the source and emitter need the -9v, which does make sense.
but am I right in thinking that the 2.2M voltage divider before the jfet may not be necessary now, I'm new to this bipolar concept.



also thanks for that schem cause it reminded me of an idea for a phaser/filter effect, I just didnt know how to make it work before knowledge of the bipolar supply...
now I shall rule the world! 
..or at least have some fun with a cool effect 

[armdnrdy]

I am by no means an expert either! I modify and design by the seat of my pants! (A lot of research for even the basics, and help from this forum)

When it comes to charge pumps....I keep one on hand. The 1054 has a large output current and no history of failure like the MAX1044.

They are a bit more costly and smaller suppliers don't stock them.

I usually purchase 1/2 a dozen or so just to keep on hand if I need to place an order for other items from a different supplier.

[midwayfair]

Increasing the voltage for your J201 won't give you more headroom. It will more than double the gain, but the input voltage limit is still 0.8V and will never be higher. Using a lower gain FET will give you more headroom. Better, use a MOSFET and it'll actually be able to take advantage of the higher voltage and still have super high input impedance.

Other things ...
You can do the trem/vibe switch a lot simpler, use only one LDR, a cheaper and smaller SPST, and not lose bass at lower depth settings in vibe mode. See here.

Your input cap is incredibly overkill. A 22nF is more than sufficient for human hearing range with that input impedance, and most of the bass is going to be obliterated in your vibe stage the way it's set up anyway. Output cap is similarly oversized. 22nF is more than big enough with the 500K pot.

[jethrofloyd67]

I found the 1054 on Smallbear, I just already bought a batch of 7660s

midwayfair, in testing I was getting break-up with the 5089 not the jfet, but maybe that's irrelevant to your point. I, like armnrdy, design by trial and error and crosschecking with online sources. So I'm sure its not the most elegant design but I was working with what I have. I haven't used mosfets before and would have to order some which is financially a complete no-go right now. I've been working like mad to finish this and am dying to put it on my board (never had a real trem before), is it worth the wait to redo this with a mosfet?

the switch I'm using is on-on-on so in the middle I get vibe and trem in series instead of parallel which sounded better to me.
I didn't notice a loss of bass with the vibe mode, though I wasn't really listening for it either. would it be better to switch the 10nf and LDR like the Blue Warbler?

also, I may not quite understand the science/math behind it, but to me a 22nf sounds far too small, any pedal I've made with caps under 1uf totally kills the deep lows that I love.
I'd rather go with overkill and have all the fatness intact than risk having another bass killer.

[jethrofloyd67]

I just read the rest of the Blue Warbler doc, didn't realize there was a whole explanation if I had kept scrolling 
I think I understand a bit better now, but correct me if not.

by switching the 10nf and LDR for the vibe and putting the output cap in series with the LDR I'll get better bass response and consistency in volume between trem and vibe modes.
I could drop the jfet altogether and just use the 5089 or mosfet if possible (the jfet is only there cause I was experimenting with things that ended up not working and I just left it there)

[PRR]

The audio path is designed single-supply, so (now we know) it does not make sense to go bipolar supply.

The audio path will *first* overload in the '5088 stage. It has gain of 2, so when it is slamming a 18V rail the FET stage may only need a 9V rail. You could feed the FET either 9V or 18V, whatever wires easiest.

The first opamp stage will do the same rate at any supply voltage. However it is direct-coupled to the second opamp stage, which has resistor values surely highly optimized for the prototype 9V supply. And there's no "headroom" issue in the wobbulator stages.

So take your 18V for the '5088 and regulate-down a 9V supply for FET and opamps and LEDs.

It gets messy if you want to power "either 9V or 18V".

I believe most of those "22uFd" caps could be much-much smaller. Like 100X smaller.

[jethrofloyd67]

It sounds like the jfet isnt really doing much for the circuit except a high input impedance, perhaps just drop it to save parts?

There's probably more going on here than I understand, but my past experience is telling me that 22nf will cut too much lows for my taste.
But I guess it wouldn't be hard to just socket them for testing.
What cap values would you recommend for use with bass?

My original idea was to use a charge pump to give the audio circuit 18v and give the LFO 9v straight from the DC plug or use a regulator.
How does it get messy if I wanted the option of either 9v or 18v for the audio circuit?

I appreciate you bearing with my lack of understanding!

[PRR]

> sounds like the jfet isnt really doing much for the circuit except a high input impedance, perhaps just drop it to save parts?

1) Parts is cheap.

2) You NEED a high impedance input for guitar.

3) FET gives a gain near 2, which over-compensates the '88 cathodyne's less-than-unity gain, and phase shifter shiftiness, which you fine-trim at the output pot.

> past experience is telling me that 22nf will cut too much lows

Cap value is meaningless except RELATIVE to the resistance in the circuit.

Which is apparent when you ask how the 10nF caps do anything interesting. We might like to vary the caps but there is no good way. We get the same effect by varying the resistors they work agains.

It is difficult to talk about a plan I can't point-at and has no part-numbers.

The input cap drives FET bias network. This almost looks like 2.2Meg twice, or 1.1Meg. In fact the Drain-Gate resistor is "Millered" by the FET gain. Which is about 2. So 1Meg||2.2Meg is 660K or so.

You need a reactance chart. Or calculator. I hear you can find them online. Me, I remember that 0.01ufd against 1Meg gives 17Hz low-cut, which is plenty ample for hi-fi. By simple proportion, 0.01uFd against 660K would be like 25Hz. Which is still pretty darn deep. OTOH 22uFd would be 22/0.01 or 2 thousand times deeper!! 0.01Hz, less than a cycle per *minute*. In fact at turn-on it will take near a minute to charge-up to barely-working.

The first and second caps analyze similar-enuff that I would just use 0.01u both places. I really do not think that is too short for Bass, but if you like 0.022u better do that. The output is nominally 500K pot so 0.02 or 0.05u seems fine. However it also feeds an external load. If an Inst In it should be high-Z like 500K. If it is any lower, it loads-down the phazer network and you'd need a buffer. So 0.1uFd is probably ample.

Also a 22u is sure to be an electrolytic. They leak. Possibly enough to spoil bias made with MegaOhm resistors. Now, or in a decade when the electrolytic leak increases. (Or after a ride in a hot trunk.)

Also a 22u, you have to know which side will be more positive. In this case it is fairly certain on the plan (sometimes murkier on the board backward and upside down). However a 0.01u will surely be a non-electrolytic and you don't have to care which way is which.

I also remember that 10uFd is good for 17Hz against 1k resistance. So for you to feel a bass-lack with 22uFd means you were working with few-hundred Ohm resistance. You find these between stages of a BJT amp. Especially at Emitters which act-like 27 Ohms or thereabouts.

[jethrofloyd67]

2) You NEED a high impedance input for guitar.
I guess I thought since this would be a few pedals in and not see the guitar directly that it isn't as crucial.

3) FET gives a gain near 2, which over-compensates the '88 cathodyne's less-than-unity gain...
by over-compensates are you suggesting lowering the gain?

Cap value is meaningless except RELATIVE to the resistance in the circuit.
This is the part I don't understand very well, math was never my best subject....
I'll have to look up some charts or calculators, obviously still have much to learn

also, am I better off feeding the LFO 9v straight from the plug and the circuit from the charge pump or feed the whole board from the charge pump and use a regulator for the LFO?
or does it matter either way?

I re-arranged things at the end based on the Blue Warbler and now have this

[midwayfair]

I also suck at math. But I can use the interwebs like a champ! I also used to vastly overestimate the necessary size of coupling caps, because I wasn't examining the resistances in the circuit.

Here's my favorite high pass filter calculator:
http://sim.okawa-denshi.jp/en/CRhikeisan.htm

Your input cap shown is 22u. Your input impedance is ~1M (use that for the "R"). You can also use this for output frequencies. Plug in the values you have. Then take into account the following numbers:

Human hearing is 20Hz to 20,000Hz in absolutely perfect hearing. PRR says he can't hear much above 2K. Most of us can't hear anything above 17,000. Our guitar amps rarely reproduce much of anything above 12-15K even with the treble and presence at max (and who does that?). Look up the frequency of the low E on a bass guitar and compare it to the lowest frequency in your coupling caps. Also, you'll need to approximate the cutoff frequencies of your coupling caps at the output to make sure it's balanced with what goes in. There's no sense in preserving 20Hz at the input only to find that your output cutoff is 150Hz.

Incidentally, you keep saying, "in my experience." From this I gather that your experience doesn't include ... breadboarding this circuit. It's a bad idea to prototype by building on a circuit board. Use the breadboard. Whatever extra time you spend breadboarding it will pale in comparison to the time spent desoldering components when you want to change something.