Pedal impedance question

clarkefan

I have to admit I've never understood impedance.

I only know that high impedance is good, low output impedance is also good, and lastly that the input impedance of one pedal should be at least ten times the output impedance of the one before it.

Ok, I've learned the rules but I don't understand. For instance, what happens to the sound of a pedal if its output impedance is say 100k ohms and the input of the next one is say 500k ohms? Or other pedals further downstream for that matter?

Any "for dummies" explanations or links gratefully received

ICBM

Generally if the following impedance is too low compared to the pedal’s output, the sound will start to get ‘strangled’ - you will lose volume and usually top end, sometimes bottom end and possibly dynamic response.

Pedals further downstream will also matter if the ones in between are mechanically bypassed (true or half) and turned off. If they’re half-bypass the input impedances add up in parallel. A buffered pedal ‘breaks the chain’, and then later pedals don’t matter.

If the pedal’s output impedance is too high, the capacitance of the following cable(s) can become important as well, with similar results. Having the volume pot as the last component in the circuit with no following buffer does this if the pot value is fairly high.

Does that make sense?

ecc83

This is where the  archaic, "shit engineering" of the whole guitar electronics chain causes problems that should have been removed decades ago.  In a properly engineered chain of audio devices (a pre amp/mixer/outboard FX/tape in/out e.g) each device has an adequately low (100 Ohms or lower) output impedance and a high input impedance ("Z") of 10k Ohms often higher.

Passive guitars need to see around 1,000,000 Ohms if HF is to be preserved, the "magic meg" . But, since "The Industry" is all over the shop with crap "True Bybass" designs you can never be sure what each box will do to the sound. One good point is that no matter what silly design concepts peeps use, nothing will break. Might sound dim or "strangled" as IC said but it won't smoke.

Dave.

Sporky

Assuming you've got some basic electronics knowledge this might help (if not we can get you there). Bear in mind that this is a very simple model and ignores the cable between the two.

http://www.monkeyfx.co.uk/pictures/pedalimpedance.PNG

You can crudely model the two pedals this way. The output of the first pedal is basically a signal generator (a voltage source in most pedals) in series with the pedal's output impedance (R1). The input of the second pedal is just its input impedance (r2).

The output of the signal generator gets shared between the two impedances, in proportion to their values. So, taking some arbitrary values, if R1 is 10K and R2 is 10,000,000K, about one millionth of the signal is "lost" to the internal resistance of the first pedal. This is fine.

If, however, R1 is 10K and R2 is also 10K, half the signal is lost to the internal resistance of the first pedal. This is probably not fine.

In your example, if R1 is 100K and R2 is 500K, one sixth of the signal is lost.

If it's ten times (so R1 is say 10K and R2 is 100K) then one eleventh of the signal is lost - in engineering we quite often say that a difference of less than ten percent isn't a difference worth caring about (this may not be the case for a bridge or a nuclear reactor).

As ecc says, a guitar usually wants to see an input impedance of about 1M (1000K), as a typical guitar pickup has an impedance of about 5-10K. So really you want the output impedance to be about a thousandth of the input impedance it's driving. Good practice says that when designing a stage its input impedance should be so high that anything can drive it satisfactorily, and its output stage should be so low that it can drive anything satisfactorily. Good practice is, sadly, rather rare in guitar electronics. 

mistercharlie

Does a buffer at the front of the line fix this completely, or is it only good until the next badly-designed pedal?

Sporky

A buffer, by definition, has a very high input impedance and a very low output impedance (and unity gain, and flat frequency and phase responses). Really every pedal should have a buffer at the input and a buffer at the output to isolate the effect circuit from the surrounding stages.

So a single buffer at the start of a signal chain will present a high input impedance to the guitar, which goes a long way to reducing treble loss and signal level loss from the guitar to the first pedal (or the amp). Essentially it makes it easier for the guitar pickups to drive the signal down the cable.

However, a buffer at the start of a chain can't do anything about the rest of the chain - if you had two pedals in the middle with "bad" impedances in series you'll still get the loss between those two.

It can certainly help with long true bypass chains though, as a proper buffer will have no problem driving several/many metres of cable.

It's worth noting, though, that a lot of the more expensive "buffers" aren't buffers. If it's got a level control it's not a buffer except at one particular setting. If it's got a tone control it can't be a buffer as it can't have flat phase response. That may not mean that it doesn't sound fine, of course, but "buffer" is a defined engineering term.

ICBM

It's also worth mentioning (for Viz ) that what a buffer can't fix is another buffer - either before or afterwards - which is a *bad* buffer and actually alters the tone... it will impose that change no matter how good a buffer you have elsewhere.

The Digitech Whammy has a good example of a 'bad' buffer. Boss pedals do too, to a lesser extent - they don't alter the tone, but they do reduce the volume very slightly - almost unnoticeably on their own, but if you have several of them in series it adds up and is enough to give the impression that it just doesn't sound as good - it's easy to correct with a small level boost, but that shouldn't really be necessary!

clarkefan

Wow thanks guys, very educational!

So to ask some obvious questions, why aren't all input impedances 1meg and outputs 1k?

And, can pedals be modded with appropriate, say, resistors to adjust if necessary? 

Sporky

clarkefan said:

Wow thanks guys, very educational!

So to ask some obvious questions, why aren't all input impedances 1meg and outputs 1k?

And, can pedals be modded with appropriate, say, resistors to adjust if necessary? 

You can't mod with resistors to increase the input impedance in a useful way.

If you put them in parallel then you drop the resistance. If you put them in series then you just drop signal across the new resistor, not across the effect's input. You can add a buffer to the input and/or output of a pedal though.

Not all input impedances are super-high because that often means adding a buffer at the input, which (a) upsets some mojo-fiend customers, and (b) means more components, which means more cost and more space needed on the board(s). Not all outputs are low impedance for the same reason - the effect circuit itself simply may not work that way.

If you have a look at a Pete Cornish board (as an example), he quite often uses Boss pedals ('cos that's what the customer likes) but puts his own buffers between them. So an effect circuit with "bad" impedances may still sound great - it just means that it may not play nicely with others.

There's a whole load of circuitry that well engineered pedals have that others don't. Input and output buffers, good polarity protection (not just a diode!), modules that reduce hum, power supply boosting and/or stabilisation and so on. Are they necessary? Depends how reliable you want/need the pedal to be.

peteri

Just to add, I had a Gigrig G2 - sold it for a variety of reasons, but one thing which contributed (because it just annoyed me) was that the input impedance was much higher than normal (from memory 2MOhm).

On most of my guitars, this contributed to making the high end more 'present' and likely 'better'.

But on a couple (old Tele's) it made them sound brittle and too much top end.

I could have fixed it with a buffer, but to me that wasn't the point of the unit and it seemed just wrong and a little bit of a swizz.

So - high input impedance isn't always good!

ICBM

clarkefan said:

So to ask some obvious questions, why aren't all input impedances 1meg and outputs 1k?

Boss ones are.

(Or the vast majority of them are, there may be one or two that are slightly worse, I can't remember - but they're all good enough.)

So are most pedals from the 80s-90s, before the days when boutique true-bypass pedals became fashionable...

Anything with a large footswitch designed as part of the enclosure* and with a soft action is likely to operate a buffered switching circuit, which if not 1M/1K, will be at least in that ballpark.

(*With the exception of things like the current Maxon pedals which are designed to look like Ibanez 9-series switches but operate mechanical switches underneath.)

Sporky

peteri said:

So - high input impedance isn't always good!

I don't think it's fair to blame a well-designed input stage for that.

If the guitars had well-buffered outputs then they'd work fine. But we're stuck in a ridiculous situation where bad design decisions made 60 years ago mean we're all stuck with unreliable (in engineering terms) systems where guitars sound different depending on what they're plugged into.

ICBM

Sporky said:

If the guitars had well-buffered outputs then they'd work fine. But we're stuck in a ridiculous situation where bad design decisions made 60 years ago mean we're all stuck with unreliable (in engineering terms) systems where guitars sound different depending on what they're plugged into.

To be fair, 60 years ago miniaturised transistor buffers didn't exist - transistors had only just become commercially available at all, and were not very reliable at that point - and putting a valve-driven one in a guitar has a fair few problems . I can forgive Leo Fender and Ted McCarty for not using them.

Sporky

You are right - not fair to blame the people 60 years ago.

But you'd think we'd have moved on.

We're not still driving 50s and 60s cars, after all. Though I do think fins are due a comeback.

Danny1969

I watched a great documentary about early Japanese transistor manufacturing the other day ...... in the early days out of every 1000 manufactured around 4 worked properly with acceptable HFE and most were OC  

The ouput impedence of a typical drive pedal would be whatever value pot they used for volume I would have thought as a typical opamp can drive 100 Ohms or lower quite well 

Sporky

Danny1969 said:

The ouput impedence of a typical drive pedal would be whatever value pot they used for volume I would have thought as a typical opamp can drive 100 Ohms or lower quite well 

Wouldn't it be the pot in parallel with the op-amp's own output impedance - so a lot lower?

ICBM

Sporky said:

You are right - not fair to blame the people 60 years ago.

But you'd think we'd have moved on.

To be perfectly honest, I do still prefer passive instruments - including basses and electro-acoustics. I know this is not "good" electronic engineering, but it has the advantages of very high reliability and long technology life - you can always add a buffer immediately afterwards if you want, which is then far easier to bypass or replace as necessary than putting it (and the battery) in the instrument itself.

Danny1969

Sporky said:

Danny1969 said:

The ouput impedence of a typical drive pedal would be whatever value pot they used for volume I would have thought as a typical opamp can drive 100 Ohms or lower quite well 

Wouldn't it be the pot in parallel with the op-amp's own output impedance - so a lot lower?

Yes thinking about it on full bore it would, half way though would be half pot value in series though  ... I think

Most designs have a 2 to 4K resistor in series on the output too so the outputs never driving into anything low 

ICBM

Danny1969 said:

Sporky said:

Danny1969 said:

The ouput impedence of a typical drive pedal would be whatever value pot they used for volume I would have thought as a typical opamp can drive 100 Ohms or lower quite well 

Wouldn't it be the pot in parallel with the op-amp's own output impedance - so a lot lower?

Yes thinking about it on full bore it would, half way though would be half pot value in series though  ... I think

Yes. This is the reason that pedals which have the volume pot last in the circuit with no following buffer are very subject to loading from long cables when the volume is set lower.

clarkefan

ICBM said:

clarkefan said:

So to ask some obvious questions, why aren't all input impedances 1meg and outputs 1k?

Boss ones are.

(Or the vast majority of them are, there may be one or two that are slightly worse, I can't remember - but they're all good enough.)

Is this why several Boss pedals on a board causes a small loss in volume, the loss you've mentioned several times?  The addition of multiple really small losses as described earlier by Sporky?

If I understood Sporky right each "1k output into 1meg input" scenario would mean a 1/1001 = 0.099% loss in signal, which is really small.  And all other brands would be equally affected.

Sporky

At a guess you need about a 5-10% loss to hear anything.