May impedence mismatch hurt my audio interface?

[lamed]

Hi,

I'd like to experience first-hand the quality loss you get when pluging something in the wrong kind of input (mainly as a way to train my hears a little bit), but I'm not 100% certain that it's safe for my audio interface. I know some of you guys are electrical engineers so I thought this would be the best place to ask.

If I plug my guitar (passive pickups with rather high output) on a line level input, is there any chance to damage the interface? What if I send a line level signal in an Hi-Z input? If I had some buffer or guitar preamp in between, will this make it less safe?

As a bonus question : I'm intending to run an amp sim plugin for this test, what kind of sound may make the quality loss most evident?

Thanks in advance for your advices.

[thetragichero]

kinda asking to paint with a broad brush but generally no it shouldn't hurt anything
feeding the interface a tone of extra voltage through the input could (say taking the speaker output of an amplifier... don't do that)

[lamed]

I really wanted to try this because some years ago, I remember not noticing a uge difference when switching the input of my tascam recorder between inst and line level. It bugged not to hear the huge difference everyone was talking about. I was then using a Sansamp stomp box in front of the recorder.

I was pleased to notice immediatly a huge difference on my computer. An overdrive in front of the interface makes it a little less noticable but I still get it easily. Of course, if I had an amp sim after the signal loss, the difference becomes absolutely awful. Lots of highs are lost and I get a lot of mud.

I thought it was a good idea to feel the difference rather than just knowing the theory. Might help me to uderstand the problem quicker if I run into an impedence match problem someday.

Thanks for answering my question. I was a bit worried to do something bad to my small, rather cheap, but really usefull interface.

[reTrEaD]

Jun 25, 2020 2:48:24 GMT -5 lamed said:

If I plug my guitar (passive pickups with rather high output) on a line level input, is there any chance to damage the interface?

No damage should occur. Passive guitar pickups (even those categorized as 'high output') produce a signal level that's considerably less than what a line level input is designed for. There would be two possible issues.
1 - The signal from the guitar would be a bit weak and require more gain applied after the fact. This could result in a bit more noise.
2 - Many of the inputs labeled 'Line Level' have a medium impedance (somewhere between 1k and 10k). This will load the guitar pickups more than a high impedance input. The result will be significant loss of high frequencies.

Jun 25, 2020 2:48:24 GMT -5 lamed said:

What if I send a line level signal in an Hi-Z input?

The potential for damage depends on the amplitude of the line level signal and how well the Hi-Z input is designed to tolerate signals that are of greater amplitude than intended.

If the input is somewhat larger than intended, clipping and distortion will occur but the input circuits won't be damaged. If the input is massively larger than intended, the input circuits could be damaged.

Proceed with caution any time you apply a signal that could be significantly larger than the input circuits were designed for. If possible use a reduced volume on the device that is driving the input and gradually increase the volume on the driving device, stopping when you notice any clipping or distortion.

Jun 25, 2020 2:48:24 GMT -5 lamed said:

If I had some buffer or guitar preamp in between, will this make it less safe?

A unity gain buffer would not be an issue. A preamp or overdrive pedal could be dangerous.

Jun 27, 2020 9:24:25 GMT -5 lamed said:

I really wanted to try this because some years ago, I remember not noticing a uge difference when switching the input of my tascam recorder between inst and line level. It bugged not to hear the huge difference everyone was talking about. I was then using a Sansamp stomp box in front of the recorder.

The stompbox you were using might have masked the problem with plugging a guitar into a line level input. Most stompboxes are capable of driving medium impedance inputs without a loss of high frequencies.

[ashcatlt]

I literally plug guitar pedals into line inputs and even microphone inputs all the time. In fact, I’d rather use a guitar pedal than a passive DI because it won’t have the transformer step down attenuation that makes the somewhat too low signal from the guitar even lower. If you’re plugging into a mic input, probably make sure phantom power is off, but otherwise it just works. But yeah the pedal presents a Hi-Z load to the pickup, and its output is low enough - and importantly non-inductive - Z that driving a lower impedance input isn’t a problem.

On most interfaces the “instrument” input (whether its a separate jack or just a button that affects existing jacks) adds something like 9-10db or gain over the line input. That can actually be too much for really hot guitars especially on interfaces that have low input headroom to begin with. It’s a common complaint with some popular interfaces from Focusrite and others. In those cases it might preferable to go guitar>pedal>line input just to avoid clipping on the way in. If you need gain later, it’s easy to get it in your DAW.

But the OP wanted to know why we advise against plugging a passive guitar into a low impedance input. It’s not so much about level. We already said we can get gain if we need it. The problem is the inductance of the pickup over the low-Z load. You can definitely test it safely by plugging the guitar straight into the line input. It works exactly like and for the same reasons as turning down the T pot on your guitar.

[lamed]

Here are some technical details about the gear I used. Maybe they'll lead to some more interesting comments (I learn a lot of things everytime I come here).

The interface in a Focusrite Scarlet 4i4 (gen 3). The inputs are 'combo' mic/line/inst plugs. Here are the line/inst modes spécifications :

Line Inputs 1-2 (Variable Gain)

Frequency Response
20Hz - 20kHz ± 0.1dB
Dynamic Range
110.5dB (A-weighted)
THD+N
<0.002%
Maximum Input Level
22dBu (at minimum gain)
Gain Range
56dB
Impedance
60kΩ


Instrument Inputs

Frequency Response
20Hz - 20kHz ± 0.1dB
Dynamic Range
110dB (A-weighted)
THD+N
<0.03%
Maximum Output Level (0 dBFS)
12.5dBu (at minimum gain)
Gain Range
56dB
Impedance
1.5MΩ

The stomp box I tried was a MXR GT-OD (Input Impedance 1 MΩOutput Impedance <7.5 kΩ).

The thing that amazed me the most was how good some amp plugins reacted to the physical stomp box, once I got the right pedal level/interface input settings  (inst mode sounded really better, even with the OD). Cheap hardware amp sims are still crap, but with enough CPU power, amp simulation have come a long way. High end VST plugins come with a big perk... They may be expensive, but women don't see them. See, I'm not Newey. There's not enough gear in the house to hide just one new pedal... Software, on the other side, is nearly invisible

[ashcatlt]

So, I think there's a typo on their spec sheet. I went and looked at their website (perhaps where you got it from) and see that it's the same kind of weird.

The Line Input has:
Maximum Input Level
22dBu (at minimum gain)

while the the Instrument Input says:
Maximum Output Level (0 dBFS)
12.5dBu (at minimum gain)

which doesn't make any sense. There's nothing coming out of that hole unless it's super broken. That's an input, and should read like the first.

One might note that the difference is 9.5db. Somebody said something about that... And that's the difference in how the signal hits the amp sim. But like you can get that gain any number of ways. Most amp sims have input gain that is separate from any of the actual amp controls which are specifically meant to help calibrate what's coming into your interface against what the amp sim "expects" to see. Nobody every gives us any straight informationn on what those expectations might be. You'd think it would be easy to say like "-18dbFS will hit the amp the same as a 1V P2P guitar signal" or whatever, but frankly it kind of doesn't matter at this point. If you were looking to accurately reamp to a real analog amp, then it could be more important to know about the input and output specs of your gear and maintain unity so that you're hitting it the same as if you'd plugged into it to begin with, but with plugins, seriously just turn the knobs til it sounds good. If the "amp" controls don't do it, try that input gain, and if it doesn't have one of those, use some other pluging before the amp sim.

But that's all about level. Get out of the habit of equating impedances with signal levels. Yes, line level signals tend to be a little louder than guitar signals which are a sometimes louder than mic levels but all of these things are so widely variable, and as long as you're not getting too much noise or unwanted distortion, it's all kind of relative and can be corrected with gain or attenuation at about any point.

Impedance is a different issue all together and while it's often true that different holes intended for different inputs will have both different impedance AND gain/sensitivity, that's not necessarily universal and it's not universally necessary. For most of what we do, we want to pass as much voltage as possible from our source to our receiver. In order to do that, we want the input impedance of the receiver to be significantly larger than output impedance of the source. Why?

Well, because it's a voltage divider. The source impedance is the "top resistor", and the receiver is the "bottom resistor", and the amount of voltage the receiver is going to see is dependent on the ratio between them in the same way as like the two halves of a volume pot. As the top resistor gets big with respect to the bottom one, the voltage across that bottom one goes down. Conversely, if the bottom is much bigger than the top, there is much less voltage "lost" in the source, and more of it is sensed across the receiver. So like what's the ideal ratio for this? Well, if we want to get out exactly what we put in, we'd need either zero on top or infinity on the bottom or both. That is the theoretical ideal. In the real world we can't have anywhere close, and we often have to compromise for all kinds of different reasons. We kind of decide how much we can afford to lose and do our best. A very common rule of thumb is to shoot for an input impedance 10 times that of the output so that the loss will be less than 1db and barely noticeable.

Now if you look at the output impedance of your pedal (<7K) and the input impedance of your line input (60K), you'll see that we're really pretty close to that rule of thumb. In fact, I'd tend to think that <7K is a kind of conservative high estimate. So fine. We're not losing enough to really worry about.

But wait now, if we go measure the resistance of our guitar pickups, most single coils will actually be significantly smaller than 7K. A lot of humbuckers actually sit right around 8K, and even a really hot wound HB won't be much more than double this. If the pedal works, why doesn't guitar?

Because the pickupo isn't a pure resistance, but rather an inductance, and its impedance depends on the frequency of the signal it's trying to pass. That reading you get from your meter is what it looks like to a 0HZ DC signal, but it gets bigger for higher frequencies, and frankly tends to be really big or even huge right around what we normally consider to be the upper midrange of the audio spectrum. So if we want anything more than a "warm fart" (as the late ChrisK used to say) out of the guitar, you really want for the input impedance to larger than that really big or even huge impedance that the pickup presents to such high frequencies. Now, the cable capacitance throws a whole new wrench in the works, but in this case it means that increasing the input impedance only really helps up to a certain point, so we don't have to go super nuts. 500K is generally good enough for most magnetic pickups, 1M is pretty standard everywhere because it's plenty for almost everything, and going much further just doesn't give us much benefit.

Hope some of this helps.

[lamed]

Jun 27, 2020 17:05:02 GMT -5 ashcatlt said:

Hope some of this helps.

Definitely helpful. Thanks for taking the time to write a detailed explaination. I remember from school the mathematics behind resistance and direct current, but don't know much about impedance and alternative current in general. All I knew was that impedance was en extension of the concept of resistance to ac and the the signal frequency was the beast that made the equations more intimidating.

Your post shed a lot of light on the difference between a pickup and a pedal output despite their similar levels. I also learned why pluging my guitar in a line input took a lot of trebles out, just as rolling the volume down does.

Maybe it should be sticked somewhere on top of a forum. I, for sure, will come back to reread it a couple of times.

[Yogi B]

Jun 27, 2020 17:05:02 GMT -5 ashcatlt said:

I went and looked at their website (perhaps where you got it from) and see that it's the same kind of weird.

My first reaction was that is was just weird backwardly thought out wording: that the maximum digital output (0 dBFS) occurs with an input of 12.5dBu and anything above that is therefore guaranteed to be harshly (digitally) clipped. However, looking at the spec tables for Focusrite's other interfaces (which don't suffer the same weirdness) it does just appear to be a typo, yes.

---

Speaking of inaccurate specifications, I generally take a pedals specified input and output impedances with a pinch of salt. For example, I've seen people equate input impedance to the value of the input pull-down resistor, and while it does have some bearing on the input impedance it may not be the whole picture. Additionally, I've seen people simply ignoring volume controls when looking at the output impedance.

Having looked up a schematic for the MXR GT-OD their stated input impedance looks to be in the right ballpark, but the output impedance looks like double what I would expect (at least by the way I've learnt to do that calculation -- thus I'm bracing myself for a correction from someone more clued-in than I).

For reference, here is a partial schematic containing the input and output sections:

\gdef\ohm#1{\operatorname{#1\Omega}{\!}}Assuming a test frequency of 1kHz and the capacitors' impedance at this frequency is negligible, the input impedance is roughly equal to R_1 + (R_3 \parallel R_4) = 910\ohm{k} which is approximately the stated 1 megohm. (This is an upper estimate due to ignoring the input impedance of the transistor, but as the darlington has very high gain that should be quite large; thus when combined in parallel with the rest will have only minimal effect on the overall input impedance.)

[EDIT] I've since found a schematic for the essentially identical MXR ZW-44 that has my R4 as 2.2 megohms, which would make the stated 1 megohm spot on. [/EDIT]

For the output impedance -- assuming the opamp's own output impedance is negligible and again so is the cap's impedance -- the output impedance is equal to the value of R14 plus the upper half of the volume control, both in parallel with the lower half of the volume control. This value is at a maximum when the resistance of both branches are equal, and that is true when both are equal to half the total value (R_{14} + R_\text{Volume}) / 2 = 14.7\ohm{k} / 2 = 7.35\ohm{k}. This is approximately the stated 7.5k, but remember that we have two branches with that resistance value; combining those in parallel further halves the actual maximum output impedance to be 3.675k.

[ashcatlt]

Jun 28, 2020 16:08:04 GMT -5 lamed said:

I also learned why pluging my guitar in a line input took a lot of trebles out, just as rolling the volume down does.

Glad to help, but I said that plugging into lowZ input is like turning down the Tone pot because it’s an inductor which gets big a high frequencies over a smaller static resistance. We lose treble when turning down the Volume pot more because of the large static resistance being over the cable capacitance which looks small for high frequencies. It’s a minor point that I bring up only to help you see the filters in the circuit. As far as sound goes, there is no real difference between and LR and RC lowpass with the same cutoff frequency of course.

Edit to add -
Now I completely ignored the V and T up to this point for simplicity’s sake. It is important to remember that they are there in parallel with the rest of the stuff. Two resistances in parallel can never total quite as much as the smaller of them, and will never be less than half the smallest. One thing that means is that as the bigger value gets bigger, it’s has less and less effect on the total. We get to a point of diminishing returns. Maybe look at our 10 to 1 rule of thumb again. With most guitars you’ll hear a big difference between a 10K and 100K input. 100K to 500K will be significant. 500K to 1M will be subtle but real. 1M to 10M you’d really have to clean out your ears and get weird about it.

@yogi B - Yeah we ignored those capacitors. Mostly we assyoume that they are big enough to ignore, but that can be dangerous. There’s no good reason for a guitar pedal to pass a 2 Hz signal, and if you choose a cap that will pass 2Hz into a 1M load that you assume will be the normal usage, it’s like way overkill. Even passing 20Hz isn’t really necessary in guitar work, so if we plug the thing into a 100K load, we won’t notice any bass loss, but if we plug into 10K, isn’t our cutoff now up to 200Hz? That should be audible, and would be one of the signs one might look for to tell you the two units aren’t compatible.

Gonna say that again here: For most active gear the signs of too low inputZ is the opposite of a passive magnetic guitar pickup. You expect to lose bass first because the capacitor looks bigger at low frequencies where the inductive pickup looks bigger at higher. This is true of raw piezo elements too, since they act like capacitors.

[lamed]

but I said that plugging into lowZ input is like turning down the Tone pot because it’s an inductor which gets big a high frequencies over a smaller static resistance. We lose treble when turning down the Volume pot more because of the large static resistance being over the cable capacitance which looks small for high frequencies.

Oh, I missread "V pot" for "T pot". I'll read again and think a bit about that to set things straight in my head.

[ashcatlt]

The V pot normally wired leaves the load seen by the pickup (the R in the LR lowpass) alone or even increases it some. It loses treble because the R in the RC filter created with the cable gets bigger.

The T pot does change R in the LR. I’m kind of skimming past the tone cap for simplicity’s sake. It’s “common wisdom” around here that the cap doesn’t much affect the tone control action until you turn it way down. Most of what it does is set a lower limit on the cutoff frequency of that LR circuit.

[ashcatlt]

It should maybe be mentioned that this whole time, we’ve been talking about relatively low voltage gear that doesn’t have a whole lot of current available so we can really just plug anything into anything without having to worry about actual physical damage. That really was the OP here, and the answer (as we said before) is that it won’t actually hurt anything, but it might not sound good*.

But if we start talking about beefier circuits with large voltages and piles of current - power amp outputs and the like - we need to be a lot more careful, and we need to know more about the circuits to say how much of a mismatch it might handle. Somebody around here once said that a decent rule of thumb is a factor of 2 in either direction should be safe-ish. So like an 8Ohm speaker output** can be expected to survive running a speaker load anywhere between 4 and 16 total. But then as I understand it, a solid state amp can handle higher impedance speakers and even usually survive an open connection (no speaker) but could blow itself up trying to drive a much lower impedance while a tube amp can (still probably shouldn’t) drive a dead short but will explode if you unplug the speaker.





*Sometimes that’s exactly what we want For example, the infamous FuzzFace and most of its authentic clones have an extremely low input impedance specifically because when you plug in a passive guitar it kills most of the top end before it gets distorted.

**Note that this is not necessarily the Output Impedance of that jack, but rather the expected load impedance. Theoretically if we want the most possible power transfer, we’d want a perfect match between the source and load, but that’s not usually possible or very good for other reasons.

[cooltone]

Serendipity! I think it is because many have turned to recording that items like this have come up.
Anyway, I have just finished the design and build of a buffer/splitter unit. It works quite well.
The front end is a hi impedance input (1M) and a low impedance input (130k with 3db cut) into a fet source-follower. It has dual outputs for routing to your amp and usb interface, so you can set your guitar as you want while recording the dry signal. The unit is sonically transparent with minimal noise.
Uses a battery, no switch, but it takes less than 30uA (about 1.5 years between battery changes. Component cost is less than £15.
Happy to share details if you want.