How much headroom is needed?

[bioroids]

Hi!

I was wondering, why not use a 5v regulator and run a pedal, say a chorus or phaser, directly of the 5v supply?
The standard opamps would be biased at 2.5v, and say you dont use a rail-to-rail but a standard opamp, which can get to, say, 1.5v of the rails. That gives about 2v peak to peak for a signal without clipping. 

Do you think it is enough for you regular guitar signal? I think it is plenty of headroom, but I haven't tested it yet, so any kind of opinions would be nice to hear!

regards!

Miguel

[GibsonGM]

It could work, depending on how much gain you're playing with.   Maybe do a dry run with just your run of the mill OA amplification circuit, and see what happens?  You could switch out resistors and try gain=5, 8, 10, etc.  and see how far you can go on 5v....I'd think a gain of near 10 would start you into clipping.  Next issue would be, how much gain is required for making a nice phasing effect, etc?

[slacker]

It would probably work Ok, you might notice the difference if you hit them with a really hot signal , but in that case there's most likely already some distortion going on so a bit more's not going to make much difference.
The PT2399 delay chip runs on 5volts and I've never heard anyone complaining about them not having sufficient headroom.

[bioroids]

I'm trying it with a simple tremolo and it seems to sound clean with my SG. I'm sure 2volts is plenty of headroom for a clean guitar signal, but
I'm not so sure about the signal coming from a distortion pedal. Though if that were the case, a little more distortion wouldn't be so noticeably.

Thinking again, a distortion pedal will increase the RMS level, but will chop the peaks. A constant signal at 2v rms will be really too loud, and that would only make sense if you are trying to overdrive an amp. So I think there is more danger on the uncompressed and peaky clean signal. Which could be made more peaky using an envelope filter, but who uses that effect anyway ?

Thanks for your answers!

Miguel

[Mark Hammer]

The amount of headroom needed would partly depend on where the effect is in the signal path, as well as the headroom and supply requirement for the  specific devices in the effect.

For example, the CA3080 chips used in so many devices, like compressors, some phasers, and some noise gates, takes a dim view of input signals greater than about 100mv, so the input is always attenuated before it reached the tender inputs of the 3080.  The CA3094, LM13600/700 and CA3280 are better, but not by volts and volts. 

The FETs used in many phasers are selected for their ability to be appropriately biased for a reasonable period off a 9v battery, such that the amount of sweep possible from them is relatvely assured for a while.  In the case of the venerable Phase 90, use of a 5v supply would mean that the zener used to set the max bias voltage (effectively acting like a regulated supply) would need to be changed (although with a max 5v supply, you probably wouldn't need the zener).  Of course, what the FET needs to be biased at for "cruising speed" is separate from the potential sweep it was chosen for in response to LFO voltages greater than that bias voltage.

In other cases, like time-based effects that run off 2nd generation MN32xx Matsushita/Panasonic chips, the BBD itself can be down-regulated to be biased and run perfectly well off 5v, but it is not done so in every case.  Indeed, in some cases the bias for the BBD is derived from a bias voltage introduced via a fixed resistive divider pair somewhere earlier in the audio path.  Bottom line is that , as with some phasers, reducing your supply voltage to 5v can seriously screw up your device biasing.

All of this is not to say that use of a lower supply voltage is to be avoided at all times.  Rather, so much in the design of many effects is predicated on the presence of a 9v supply, that adapting to its absence may be more trouble than its worth.   Not impossible, just a poor ROI.

[bioroids]

Mark, you are absolutely right.
But if you are designing an effect from scrath, then you can do anything you want 

I'm designing a low part count PWM-controlled AVR-based phaser, that's why I want to run everything from the 5v supply. It seems like the headroom is Ok. Details should come soon if I finally get it to work!

Regards

Miguel

[Mark Hammer]

Mark, you are absolutely right.

As are you mi amigo.   If it's your design, then you can assume whatever you want about the parts/levels and adapt as needed.

As regards the initial question about how much headroom is needed, I guess it needs to be said that more headroom is generally needed earlier in the signal chain where the signal is "purest" and the most dynamics present in the signal.  Generally speaking, each successive effect in series tends to remove just a little of the dynamics.  By the time the signal has been run through a compressor, a distortion, a FET/OTA-based phaser that has to reduce the signal level to avoid clipping, and one of those flangers that includes diodes to provide Q&D limiting of regeneration, there isn't a whole lot of dynamics left.  Just one of the many reasons why I keep blathering on about the sensing part of noise gates needing to be up front and the gating part itself at the end.

So, yeah, while I wouldn't wish to run an initial clean booster or EQ off 5v, I would imagine many effects could be run off it.  Certainly my SSM2166-based Q&D compressor is quite happy to function off 5v.