Hi guys, i wanted the input of some more knowledgeable folks than myself about a project I have in mind.
How hard would it be to convert circuits for 5 volt operation While keeping the same sounds? I'm asking because i want to build digitally controlled effects and most digital pots are rated for 5v (I know there are several higher voltage ones, but they tend to be very expensive)
My end goal is to have a pedalboard and to power all effects on 5 volts but to keep their original sounds.,
Is it just a matter of adjusting the gain resistors of amplyfying stages in proportion the modified voltage levels? I dontknow if this wouldchange impedances or frequency responses of effects...
Or could it be as simple as reducing my guitar's input signal in proportion to the change in source voltage to achieve the same practical headroom? (But at lower output volumes)
I'm particilary binterested. In doing this with fet based effects.
I'm asking about this before building the circuits because i have all the parts in the shopping cart, but if this is not feasibleid have to change a lot of them
Anyways any thougths, suggestions or recomendations are welcome. Thanks!
I think it might vary from circuit to circuit as to how much you have to change. I'd imagine it might do something to the noise floor too? There are circuits that take the voltage down and then back up, like FV1 based pedals as the chip runs on 5V, possibly stuff with pt2399's and the like.
One thing that will be affected will be purposeful clipping. Diodes have clipping thresholds that you have to reach to get the distorted sound. If you reduce the voltage swing then try to run it through LED clipping... it would sort of be like rolling off the volume of your guitar I suspect, maybe without the tonal changes sometimes associated with it.
Could you set up buffers/gainstages around each pot to bring the voltage down and up? Not sure how the noise would be. If you use SMD components and make a dozen PCB's at once you could sprinkle them through the circuits without having to change the meat and potatoes?
I'm talking out my arse, see posts below for better ideas
Some things already DO run off 5V, and not just digital pedals. For instance any time based effect that runs off an MN32xx-series BBD generally converts the 9V supply down to 5V to power the chips. If you were lucky enough to score an SSM2166 dynamics control chip for a compressor, that also runs off 5V.
There is nothing wrong, either, with using a 9V supply and down-regulating part of the supply down to 5V for those pedals that can use it, and retaining the 9V capability for those that need it.
So for instance, in for example i feed my circuit on 5 volts (0.55 times 9 volts) and i attenuate my input signal to 0.55 times the original signal, could i leave the gain stages intact? And for replacing leds for clipping, could i for example in a red led (around 1.8-2 forward voltage) for something like a series silicon + schotty diode? (Around 0.9 to 1 volt i think?)
Another question, could you look a this datahseet? These are cheap digipots, and from what i understand from the datasheet the log is 5 volts, but across the terminals it can take +/- 5 volts, so i could power an effect it with a 9 volt baterry without damaging the pot??
X9C102, X9C103, X9C104, X9C503 Datasheet - Intersil
PDFhttps://www.intersil.com › documents
(//http:///%3EPDFhttps://www.intersil.com%20%E2%80%BA%20documents)
A general rule of thumb is if you change the voltage and/or the components you will change the circuit characteristics and the sound. I think 9V was chosen because of the 9V battery but I can't see why everything couldn't be done on 5V it is just that most circuits have been designed for 9V.
The most challenging aspect with your run-of-the-mill guitar effects is going to be how to manage any pots that are referenced to somewhere other than ground. Gain pots on opamps that are floating at Vref are one in particular that comes to mind.
Also as mentioned, the reduced headroom may complicate diode clipping thresholds, though should be manageable.
Quote from: Phoenix on January 11, 2018, 04:24:42 AM
The most challenging aspect with your run-of-the-mill guitar effects is going to be how to manage any pots that are referenced to somewhere other than ground. Gain poots on opamps that are floating at Vref are one in particular that comes to mind.
Also as mentioned, the reduced headroom may complicate diode clipping thresholds, though should be manageable.
I don't understand what you mean by that. Could you please explain it in more detail? Can't I just bias the wiper of the digipot at 1/2 vdd and put a large output cap after the wiper t avoid clipping?
For most effects it will be pretty easy using rail-to-rail opamps (specially designed for 5V operation). But a few effects do rely on the extra headroom, or use opamp clipping which will alter the sound when squashed down to 5V levels.
another thing i was considering was replacing the potentiometers with led+LDR combos sealed in epoxy glue (1 for rheostats and 2 for voltage dividers). This method would eliminate all noise, bias and bandwidth related issued.
I was thinking of programming a microcontroler to vary the LED's brightness using pwm (this idea would need a one time calibration procedure to determine the duty cycle of each individual LDR combo needed to reach the desired resistance range, say from 0 to 1Mohm, or 0 to 1k)
what do you think of this approach? I've seen around the forum that some people do this to control modulation based effects, has any of you tried to replace pots with this arrangement?
what material have you used to insulate the leds+LDRs from light besides heatshrink (I'm inclined to use black heatshrink + some epoxy outside to make a more stable mecanical conection)
Another issue might be that LDRs are very temperature dependant, but maybe a temperature sensor inside the pedal could help me account for this factor?
¿What do you guys think?
Quote from: joakinrox on January 11, 2018, 05:20:06 AM
I don't understand what you mean by that. Could you please explain it in more detail? Can't I just bias the wiper of the digipot at 1/2 vdd and put a large output cap after the wiper t avoid clipping?
The easiest possible option would be to simply drop digital pots in as replacements in a stock circuit, given that typical effects units DON'T feature rail-to-rail opamps, and as such have swing of only about 5Vpp even when operated from 9V (discrete transistor designs can be quite different of course, swinging more-or-less depending on the design). However, see the case below, where the digital pot is powered from the 5V regulator, but is biased at Vref, which is 4.5V. Even if the opamp only swings 5Vpp, that's 2V to 7V, so those 7V peaks will exceed the maximum 5V rail of the digital pot.
You can't always just lower the operating voltage of the entire circuit to 5V, because the circuit will not operate correctly.
Of course, you can try substituting rail-to-rail opamps like Merlin suggests, which may work for most opamp circuits.
You do also occasionally find digital pots which have one of their terminals connected to ground, but they're not too common.
The small range of values available is also probably a significant limiting factor.
(https://s9.postimg.org/dt4auoyrj/Digital_pot.png)
.
One thing you will have to watch like a hawk is that an awful lot of these PIC-controlled digital pots have a max "wiper" current of only 1mA. Exceed that rating and cue magic smoke.
Some devices work on input current rather than voltage, giving you more headroom than you may think. The THAT4316, for example, runs on 5V.
I am confused about what you really want. In your first post you want digital pots for a microprocessor.
Quote from: joakinrox on January 10, 2018, 07:24:08 PMHow hard would it be to convert circuits for 5 volt operation While keeping the same sounds? I'm asking because i want to build digitally controlled effects and most digital pots are rated for 5v
Yet in this post you come up with LDRs
Quote from: joakinrox on January 11, 2018, 06:11:47 AManother thing i was considering was replacing the potentiometers with led+LDR combos sealed in epoxy glue (1 for rheostats and 2 for voltage dividers). This method would eliminate all noise, bias and bandwidth related issued
If it just the elimination of crackle from pots you could use a FET as the variable resistor and a big cap/filter on the gate to block crackles.
Thinking about lower voltage operation I think it would be better to run off 3V from two batteries. It would give plenty of headroom for 0.7V diode drops and reduce the gain required of the original signal that has to leave the box at 1V anyway.
Quote from: joakinrox on January 10, 2018, 07:24:08 PM
How hard would it be to convert circuits for 5 volt operation While keeping the same sounds? I'm asking because i want to build digitally controlled effects and most digital pots are rated for 5v (I know there are several higher voltage ones, but they tend to be very expensive)
My end goal is to have a pedalboard and to power all effects on 5 volts but to keep their original sounds.
As others have said, some circuits will be harder to convert than others, and reducing clipping thresholds proportionally is an obvious snag. But nothing's impossible.
One possibility you haven't mentioned for controlling circuits at this kind of level would be a multiplying DAC. If you feed the analog signal to the Vref input, you can control how much of it passes with the digital input.
Also consider that the audio part of the circuit (or most of it) could still work at a standard 9V, and only a few elements within it might need to work at 5V. This is how I did the Flangelicious flanger, which uses a PIC for digital control and runs the BBD at 5V, but the rest of the circuit is standard stuff at 9V.
Tom
Another vote for the multiplying DAC here. The DAC0808 and DAC1208 are available just about everywhere. The usual audio pot is a logarithmic item but if you have enough steps the control granularity for a linear DAC should not be a problem.
But if you need to run a circuit at a fixed voltage, if you bias the non-inverting input to an op amp to 4.5 volts as is done in a lot of effects, the circuit will drive the inverting input to within a few millivolts of that voltage as long as it remains in the linear range (not using op amp clipping). You can place a digital pot on either the input or feedback side with the inverting input driven to 4.5 volts so you don't have to accommodate the full input swing but you may have to reference the digital pot control lines to the 4.5 volt level.
Quote from: amptramp on January 13, 2018, 05:05:42 PM
Another vote for the multiplying DAC here. The DAC0808 and DAC1208 are available just about everywhere. The usual audio pot is a logarithmic item but if you have enough steps the control granularity for a linear DAC should not be a problem.
But if you need to run a circuit at a fixed voltage, if you bias the non-inverting input to an op amp to 4.5 volts as is done in a lot of effects, the circuit will drive the inverting input to within a few millivolts of that voltage as long as it remains in the linear range (not using op amp clipping). You can place a digital pot on either the input or feedback side with the inverting input driven to 4.5 volts so you don't have to accommodate the full input swing but you may have to reference the digital pot control lines to the 4.5 volt level.
I investigated about the multiplying dacs, and from what I gather they seem to have a buffer in the "wiper" right? I think this would make them ideal to replace "output volume" pots found in most effects,
But for potentiometers on places like tonestacks, wouldn't the prescense of a "buffer" in the wiper change the frequency response of a tonestack circuit?
Yes; in most if not all cases you would loose the pot-to-other-components interaction so the "tone curve" would be compromised.
On a slightly different but related note I was thinking about building a pedal power supply that runs off a 5v usb mobile charger, on the grounds that:
- they've become very common and cheap (and rated up to 2.1a in some cases)
- i saw a micro-usb breakout board on bitsbox and though it might be unteresting to try.
So I built a voltage doubler as per the attached. I was kind of hoping that the natural voltage drop in the circuit would bring me to 9v (hence the rather optimistic label on the schematic) but it's closer to 9.8v. I'm assuming there's not enough headroom for a 9v regulator, so would welcome any advice on my options to get to the 8.8-9.2v range without clobbering current availability?
(http://photos.app.goo.gl/P3JwSUDVfJSSjkBU2)
(https://s10.postimg.org/87t7rf8qt/usb_shcem.jpg) (https://postimg.org/image/87t7rf8qt/)
Use 1N4001 (or other silicon) diodes in the charge pump rather than 1N5818 and you'll get more voltage drop, bringing you closer to 9V.
Be aware that the charge pump is limited in output current - maybe 100 mA ? - so even if you have a higher amperage USB charger, the output of the 9V pump will be limited.
I think you have to take each of your FX one at a time as you work through this. It's too complex to give a general answer.
The most general I can give you is that, yes, you can do this. You can probably re-design most circuits to work with a digital pots in some capacity.
As for whether you are concerned about pushing the button on ordering digital pots -- this is going to be a trial-and-error R&D exercise. It's going to take time and cost money. If you are worried about the cost of your first order of digital pots being too much, then don't even start. It will cost you a lot more than that by the time you have succeeded.
If you can afford to fail on the risk that the digital pots won't work out the way you expected, then order the damn things and get to work figuring out how to make this happen.
As for digital control of analog circuits, you will probably find there are a lot of different ways to skin this cat. You may find that you will need to employ different strategies in different places to get the desired result. One circuit at a time, one pot at a time...
Quote from: reddesert on January 23, 2018, 01:35:37 PM
Use 1N4001 (or other silicon) diodes in the charge pump rather than 1N5818 and you'll get more voltage drop, bringing you closer to 9V.
Be aware that the charge pump is limited in output current - maybe 100 mA ? - so even if you have a higher amperage USB charger, the output of the 9V pump will be limited.
Thanks, that's really helpful. Looks like max current is 160mA from the data sheet, so i'd probably need to run 2 or three to form a pedalboard supply (i'm thinking one unit per 2 outputs, running all of them in parallel with the input voltage).
Quote from: Transmogrifox on January 23, 2018, 09:58:50 PM
I think you have to take each of your FX one at a time as you work through this. It's too complex to give a general answer.
The most general I can give you is that, yes, you can do this. You can probably re-design most circuits to work with a digital pots in some capacity.
As for whether you are concerned about pushing the button on ordering digital pots -- this is going to be a trial-and-error R&D exercise. It's going to take time and cost money. If you are worried about the cost of your first order of digital pots being too much, then don't even start. It will cost you a lot more than that by the time you have succeeded.
If you can afford to fail on the risk that the digital pots won't work out the way you expected, then order the damn things and get to work figuring out how to make this happen.
As for digital control of analog circuits, you will probably find there are a lot of different ways to skin this cat. You may find that you will need to employ different strategies in different places to get the desired result. One circuit at a time, one pot at a time...
Yeah, I think that's the approach I'm gonna have to take. So far I think I'm going to try the following:
a)-Multiplying DACs for Volume Output Pots or Gain pots that are wired as a voltage divider before the first stage (As their output is buffered)
b) -Digital potentiometers for voltage dividers stuck on places like tonestacks.
c) -1 LDR + Led combo controlled by PWM for variable resistors (like for example a feeedback resistor of a opamp)
d) -2 LDRs controled by 2 different LEDS in PWM mode to "make" voltage dividers with resistances higher than 100k (as pots larger than 100k can't be found for 9 volt operation or are too expensive)
The hardest part of this will probably be calibrating the LDRS to ensure consistency. I'm thinking of two aproaches:
-Try mapping various curves of their resistance at different duty cycles at several temperature conditions and switch to a determinate curve based on temperature. This would mean using a temperature sensor (LM75?)
-Measuring the resistance of each LDR with the analog pins of the uController, and adjusting the duty cycle of the leds dynamically. this would probably involve some kind of algorithm to take the slow response time of the LDRs into account (that I'm frankly not sure I'd be capable of implementing)
I agree completely with Transmogrifox - all extremely good advice.
I'd just like to add one thing; I expect you've got a microprocessor of some sort in this system, since you're talking about adding memories and DACs and PWM and so on.
So given that there's a processor in the circuit, don't underestimate what you can move into the digital domain directly. For example, rather than use a uP to control two digital pots to control rate and depth of an LFO, why not generate the LFO signal in the uP directly and output it from there to the rest of the analog audio circuitry? You don't have to mess with your analog signal path, but you can save yourself considerable circuit complexity by rolling some of it up into code inside the processor. If one of the circuits is a PWM phaser, why not generate the LFO-modulated PWM signals directly in the processor? If it's a flanger, generate the LFO-modulated biphase clock in the processor. Or the do the same/similar for chorus.
Obviously if all the circuits are Fuzzface variants this approach won't be so useful...;)
HTH,
Tom
Guys, I found some Intersil pots (quite cheap, around 1 buck a piece);
https://www.intersil.com/content/dam/Intersil/documents/x9c1/x9c102-103-104-503.pdf
I read on the datasheet that these pots can handle +/- 5 volts with respect to the VSS pin, so could i power the pot at this to handle a 0-9 volts signal?: by powering VSS at 4.5Volts and VCC at 9 volts? this would make 0 volts = -4.5 and 9 volts +4.5 right?
but now, if i do this i would have to shift my control signals as well right? I think 0-5 volt control signals would not work if my chip is powered from 4.5 to 9 volts right?
Quote from: joakinrox on January 28, 2018, 04:47:43 PM
Guys, I found some Intersil pots (quite cheap, around 1 buck a piece);
https://www.intersil.com/content/dam/Intersil/documents/x9c1/x9c102-103-104-503.pdf
I read on the datasheet that these pots can handle +/- 5 volts with respect to the VSS pin, so could i power the pot at this to handle a 0-9 volts signal?: by powering VSS at 4.5Volts and VCC at 9 volts? this would make 0 volts = -4.5 and 9 volts +4.5 right?
but now, if i do this i would have to shift my control signals as well right? I think 0-5 volt control signals would not work if my chip is powered from 4.5 to 9 volts right?
You'd have to power the rest of the logic from your 4.5V to 9V power supply too, yes. If you're shifting your ground, you need to be consistent. Otherwise it won't work, and would probably fry something.
What you've effectively done is create a +/-4.5V bipolar supply (as we often finish up doing in pedals, actually). You're then running the logic on the "upper half" - 0V to 4.5V.
HTH,
Tom