Where does clipping take place (for sherlocks holmses:))?

  • Thread starter Thread starter lockesilver
  • Start date Start date
Break out anything you want. The fact remains that the inputs and outputs he's using aren't matched to each other. Let's get the guy recording first, then if you want to prove how much you think you know about electronics, we can start another thread. :rolleyes:

Nah, let's do it in this thread. You are confusing *level* with *impedance*. They are two totally separate things.

For your reference:

Mic, typical: -60dBV to -30dBV level, < 250 ohm output impedance
Line, typical: -10dBV to +2dBV level, < 600 ohm output impedance (often more like 100 ohm, so let's use that)
Instrument, -30dBV-ish, ~10K ohm output impedance -- this is highly dependent upon the instrument in question.

Now, for inputs, these are the typical impedance:

Mic: 1K
Line: 10K
Instrument: 1M

Because of that, some people who don't really understand what they are talking about think that plugging a mic or line output into an instrument input is an "impedance mismatch". It is not. Impedances want to be bridged (meaning load impedance should be much greater than source impedance); this is to prevent loss of signal (which is voltage). If we want to determine the signal loss, we can use this simple formula:

dBloss = 20 * log (Zload / (Zload + Zsource))

Here is a handy table for you to convert loads and sources, using my typical figures above:

Source Mic Line Instr
Mic -1.9dB -0.2dB -0.0dB
Line -0.8dB -0.1dB -0.0dB
Instr -20.8dB -6.0dB -0.1dB

So we can easily see that a) impedance mismatches *only* occur with high source impedances and low load impedances, as with the guitar into anything but an instrument input. It actually gets even worse for the guitar because its source impedance is complex, which means that the loss varies by frequency, so not only does the guitar lose most of its signal, its frequency response changes too. That's bad.

But none of that is happening with a low output impedance source like a mic amp. There could be issues of noise or headroom, but those are not caused by differences in impedance, they are caused by design decisions. For example, if you label an input "mic", probably the designer assumes that the signal level incoming is low, and they should apply an amount of minimum gain to said signal. If the incoming signal is actually hot, the gear will therefore clip. But there is no law of physics that says that a mic input must have minimum gain; it could be zero. It could have an input pad . . .

As for noise, if you have a line input, you might assume that the incoming signal is hot, so you might be happy with input noise that is 30dB worse than you could tolerate for a mic input. You might also assume that no gain is required, because the signal is supposed to be line level already. So when a poor soul plugs a mic directly into that input, the result is a quiet yet noisy signal, which they notice when they add gain at a subsequent stage. But there is no law of physics that says you can't design a very quiet line input with gain if required. It just isn't generally done due to cost and lack of need.

The next thing that happens is that poor soul will ask what happened, and somebody will say "impedance mismatch". But that's wrong, it's just a noisy input with no gain. How can there be an impedance mismatch when the source impedance of a typical mic and typical line output are ALMOST EXACTLY THE SAME!!!

There is no mismatch; there is just a level difference of up to 60dB. Level and impedance are not the same thing.

For the record, if he has the Behri output into an input labeled with a "mic/guitar" switch, the problem is the switch on mic probably has minimum gain, which is clipping the Cakewalk. Has nothing to do with impedance. I'd be willing to bet that switching to "guitar" would help matters, but you'd probably say that would make the "mismatch" worse, right? Read above again.

Any, yes, RCA inputs, likely much less gain, the correct answer. BUT NOT BECAUSE OF IMPEDANCE.
 
BUT NOT BECAUSE OF IMPEDANCE.
Thanks for that in-depth analysis!

In more Caveman terms, one must strive to apply gain at the input and output so that the output of each stage ends up in the, "sweet spot" for input level for that device. Sometimes you just have to work with what you've got.

As mentioned, even a mic has some input gain control: proximity.
 
I use the big input labeled "Mic/Guitar" (set to Mic). The unit also has a small input called "MIC (plug-in powered)/digital in".

I didn't realize that unit had a mic preamp in it. So you're going from the +4 output of a preamp into another preamp expecting a mic-level signal. Yeah, it's gonna clip. Connect your preamp to a line input.
 
I'm glad that your focus is on assisting the OP rather than self-aggrandizement. Oh, wait. Strike that. Reverse it.

So, plugging a line level output into a mic input isn't an impedance mismatch? I believe you're arguing semantics more than anything.

Dude, try driving a 4 ohm load with a 16 ohm tap off of an amp's output transformer. By your definition, that's not an impedance mismatch, but it time it will definitely cause the xfrmr to fail.
 
So, plugging a line level output into a mic input isn't an impedance mismatch? I believe you're arguing semantics more than anything.

No, it's not an impedance mismatch if it's a balanced low impedance line out into a balanced low impedance mic input. It's a voltage (level) mismatch.
 
RE: Where does clipping take place?

I want to know where all the stuff that's clipped ends up? I've been doing it since the 60's - that's a lot of stuff. Where is it now? Is there a big pile somewhere? Who legally owns it? Can you eat it?
 
No, it's not an impedance mismatch if it's a balanced low impedance line out into a balanced low impedance mic input.
Yeah, but I'm referring specifically to the OP's situation.

For the sake of argument - fuck yeah, I'm wrong. I give, check out the big brain on Brad and all that.

Now - let's get the OP to connect his Behringer to the line input on the Cakewalk unit and see if he can get his stuff working. You know, fix the problem before arguing tertiary minutiae.

After that - we can debate who's the Alpha Geek or whatever. And the good part? The OP will be able to write AND record a song about it. :cool:
 
I want to know where all the stuff that's clipped ends up? I've been doing it since the 60's - that's a lot of stuff. Where is it now? Is there a big pile somewhere? Who legally owns it? Can you eat it?
It's where dark matter comes from.
 
So, plugging a line level output into a mic input isn't an impedance mismatch? I believe you're arguing semantics more than anything.

Absolutely not, and not a semantic difference, but a real physical difference. If the mic amp's gain can be set to unity and/or it has adequate headroom, this is not a problem at all. I do it every day testing out amplifier circuits with a converter loop.

Dude, try driving a 4 ohm load with a 16 ohm tap off of an amp's output transformer. By your definition, that's not an impedance mismatch, but it time it will definitely cause the xfrmr to fail.

Tube power stages are not like small-signal solid state circuits; in that case you want to maximize power transfer which does require impedance matching. This is not at all comparable to the small-signal circuits we are discussing where you want to minimize signal (voltage) loss.

First, you almost never want to match mic impedances. If you have a variable input impedance mic preamp, set it at 150 ohm. That should be pretty close to matching most mics. What happens? The level will drop something like 6dB. Not a big deal, you can makeup the gain. What happens next depends on the type of microphone. Dynamic microphones will have some change in frequency and transient response. That can be desirable, although 150 is going kinda low, most people enjoy something closer to 500 ohm, but try it and see what you like.

Most condenser microphones on the other hand are resistive sources, so their frequency response doesn't change, you just lose some level. Unfortunately, you will also experience an increase in distortion. Read some datasheets that specify the difference in max SPL handling, between 1K and 150 it's often quite significant.

The same thing will be experienced with some line amps driving heavy loads; distortion increases. Take a typical opamp output using +/-15V rails; with peak-to-peak voltage it would have to source/sink up to 100mA into a 150 ohm load. Most typical opamps used for line outputs in prosumer equipment cannot offer more than 40mA or so. Also, even if they could, you would have a problem with thermal dissipation; 30Vp-p is roughly 10VRMS, into 150 ohm is almost 0.7W per channel. That exceeds the rating for a SOIC-8 for one channel and the rating for a DIP-8 for two. Even with the signal limited to 40mA, that still gets uncomfortably close.

So for modern solid-state circuits, bridging, not matching, impedance is the rule. This, by the way, is also the case for guitar circuits; most people will say it's very bad to plug a guitar directly into a line input, even though that would result in a fairly close impedance match. Until you turn down one of the guitar's volume pots, but that's a topic for another day . . .

As an aside, your guitar amp mismatch situation will likely NOT cause the transformer to fail, read this:

http://www.geofex.com/tubeampfaq/TUBEFAQ.htm#matchspkr

or this:

http://www.thegearpage.net/board/archive/index.php/t-282689.html

or this:

http://www.audiocircle.com/index.php?topic=60694.0

Now if you get more extreme than 4 or 16 ohm, then I would start to worry (although I understand some guitar amp's output jacks are normalled to a short-circuit). But you have to realize something about speakers; they have a complex impedance curve. If a 16 ohm load really destroyed a transformer at its 8 ohm tap, then you could never connect any speaker to it. Have a look at a typical guitar speaker's impedance curve, here is one I picked at random, but this is normal:

http://eminence.com/pdf/legend-1058.pdf

See how it's 8 ohm at 200Hz but 16 ohm at 2kHz? Nothing will melt if you play the highest note on your guitar repeatedly, except maybe your ears . . . I don't know about its resonant frequency, maybe people doing tube amp guitar cabs are in the habit of using notch filters (& Zobels), maybe they are not. Good question for Harvey Gerst, although there is some discussion here:

http://www.diyaudio.com/forums/tube...vantages-using-zobel-series-notch-filter.html
 
Now - let's get the OP to connect his Behringer to the line input on the Cakewalk unit and see if he can get his stuff working. You know, fix the problem before arguing tertiary minutiae.

Ah, but this is the essence of understanding gain staging. It's about gain! Any solid-state circuit will have a low output impedance and can therefore be plugged into any input. But if the levels are wrong, correctly bridging impedance will not save you . . .

Remember we started with an OP who didn't specify which input he was using, and said that the Cakewalk gain knob was set "awfully loud". He has given us more information which we now understand better, but the problem always has been too much gain and never anything to do with impedance.
 
Nah, let's do it in this thread. You are confusing *level* with *impedance*. They are two totally separate things.

For your reference:

Mic, typical: -60dBV to -30dBV level, < 250 ohm output impedance
Line, typical: -10dBV to +2dBV level, < 600 ohm output impedance (often more like 100 ohm, so let's use that)
Instrument, -30dBV-ish, ~10K ohm output impedance -- this is highly dependent upon the instrument in question.

Now, for inputs, these are the typical impedance:

Mic: 1K
Line: 10K
Instrument: 1M

Because of that, some people who don't really understand what they are talking about think that plugging a mic or line output into an instrument input is an "impedance mismatch". It is not. Impedances want to be bridged (meaning load impedance should be much greater than source impedance); this is to prevent loss of signal (which is voltage). If we want to determine the signal loss, we can use this simple formula:

dBloss = 20 * log (Zload / (Zload + Zsource))

Here is a handy table for you to convert loads and sources, using my typical figures above:

Source Mic Line Instr
Mic -1.9dB -0.2dB -0.0dB
Line -0.8dB -0.1dB -0.0dB
Instr -20.8dB -6.0dB -0.1dB

So we can easily see that a) impedance mismatches *only* occur with high source impedances and low load impedances, as with the guitar into anything but an instrument input. It actually gets even worse for the guitar because its source impedance is complex, which means that the loss varies by frequency, so not only does the guitar lose most of its signal, its frequency response changes too. That's bad.

But none of that is happening with a low output impedance source like a mic amp. There could be issues of noise or headroom, but those are not caused by differences in impedance, they are caused by design decisions. For example, if you label an input "mic", probably the designer assumes that the signal level incoming is low, and they should apply an amount of minimum gain to said signal. If the incoming signal is actually hot, the gear will therefore clip. But there is no law of physics that says that a mic input must have minimum gain; it could be zero. It could have an input pad . . .

As for noise, if you have a line input, you might assume that the incoming signal is hot, so you might be happy with input noise that is 30dB worse than you could tolerate for a mic input. You might also assume that no gain is required, because the signal is supposed to be line level already. So when a poor soul plugs a mic directly into that input, the result is a quiet yet noisy signal, which they notice when they add gain at a subsequent stage. But there is no law of physics that says you can't design a very quiet line input with gain if required. It just isn't generally done due to cost and lack of need.

The next thing that happens is that poor soul will ask what happened, and somebody will say "impedance mismatch". But that's wrong, it's just a noisy input with no gain. How can there be an impedance mismatch when the source impedance of a typical mic and typical line output are ALMOST EXACTLY THE SAME!!!

There is no mismatch; there is just a level difference of up to 60dB. Level and impedance are not the same thing.

For the record, if he has the Behri output into an input labeled with a "mic/guitar" switch, the problem is the switch on mic probably has minimum gain, which is clipping the Cakewalk. Has nothing to do with impedance. I'd be willing to bet that switching to "guitar" would help matters, but you'd probably say that would make the "mismatch" worse, right? Read above again.

Any, yes, RCA inputs, likely much less gain, the correct answer. BUT NOT BECAUSE OF IMPEDANCE.

Great explanation man! I have to say, I really learn a lot an enjoy reading your posts. Any time that you've posted I immediately read it. You really contribute a lot to this forum. Thanks and CHEERS~! :D
 
Remember we started with an OP who didn't specify which input he was using, and said that the Cakewalk gain knob was set "awfully loud". He has given us more information which we now understand better, but the problem always has been too much gain and never anything to do with impedance.
And until I actually asked, everyone was shooting in the dark.

Knowledge is great. Displaying knowledge for the sake of displaying knowledge is different from troubleshooting. The OP needed troubleshooting, not chest-thumping.

A person can have all the knowledge in the world, but that doesn't mean they can actually fix stuff.

Most of the time, problems get solved more quickly by asking questions than showing the person with the problem how much you know. I may not have exhibited your level of detailed knowledge, and may have misused a term or two, but answer this: Who concentrated on getting to the OP's issue?

Sorry, I guess we just have different philosophies when it comes to troubleshooting and problem solving. It's all good though.
 
And until I actually asked, everyone was shooting in the dark.

Knowledge is great. Displaying knowledge for the sake of displaying knowledge is different from troubleshooting. The OP needed troubleshooting, not chest-thumping.

I tried to keep my posts to troubleshooting. I knew impedance wasn't the problem. You kept saying it was. From the beginning I said it was a level problem, from the beginning you said it was impedance. Exact words out of your mouth, and you didn't even explain what impedance was to OP, nor detail how to fix it.

And then you said:

You're experiencing a severe level mismatch, due to something that's sometimes known as an impedance mismatch.

Level is not impedance, period. Let's try again: what is the impedance of a 120VRMS signal? Who knows? It could be anything. A few milliohms if that is your wall socket. Probably a rather larger figure if that is the plate of your power tube. Get it now?

If I had never explained in detail why you were wrong, you would still be ignorant of this very basic principle of electronics.


Then you said:

Let's get the guy recording first, then if you want to prove how much you think you know about electronics, we can start another thread.

What I *think* I know? I don't know much, but I know enough. Challenge me like that and you're going to get served. And so you did.

A person can have all the knowledge in the world, but that doesn't mean they can actually fix stuff.

Yeah, um, I design & build stuff for a living :rolleyes:

Who concentrated on getting to the OP's issue?

I did, until you distracted me by being constantly and obstinately incorrect.
 
Great explanation man! I have to say, I really learn a lot an enjoy reading your posts. Any time that you've posted I immediately read it. You really contribute a lot to this forum. Thanks and CHEERS~! :D

Well I do post a lot of absolute garbage unrelated to physics . . . :drunk:
 
Sorry to interrupt an interesting discussion about impedance.

Guys, thank you all.

It was as simple as that - I was using the wrong input all this time.
How easy it is to overlook the obvious.

Since the input I used has a mic amp built in, I was running the signal through double amplification which simply had to result in clipping.

Well, it was really stupid of me. I thought

a) the word 'digital' on the small input meant it's intended for some digital sources (and everything else in the chain was analog so...)

b) that the 'small' jacks were used mostly in audiophile equipment (why should it then be used in my super modern professional Behringer-tube powered recording studio)

Yeah, a fool's logic can get pretty advanced.
It's working like a charm now (I still plan to get rid of the preamp and upgrade, to a mixer most likely).

I'm not sure whether I would EVER think of trying the other input. The shooting in the dark and the impedance was actually informative. I guess I'm at the stage of learning new things all the time which is kind of exciting - and ironically the more goes wrong the more I learn. :)

Case solved.
 
Well there are a lot of different kind of connectors, many with multiple incompatible uses. Here is a very partial list:

1/8" (3.5mm) miniplug/minijack - usually seen as tip-ring-sleeve (TRS), can be stereo (as with headphones), a powered microphone connection (one of two incompatible "standards", stereo as with camcorders, portable recorders, or mono with PC mics), and sometimes but not too often as a line input, which could be balanced (TRS) or unbalanced (TS)

1/4" phone plug/jack - guitars use unbalanced TS, headphones are stereo (TRS), line inputs can be TS or TRS, depending on whether or not they are balanced. A 1/4" line output is *usually* nominal +4dBu level, but it doesn't have to be. Used in the past for high-impedance (read 25K) microphones (TS), this is pretty rare in new gear now. Very rarely used as for a phantom-powered mic input (TRS), although some people wire TRS patchbays that way . . .

XLR - the 3-pin version is mostly used for professional (balanced, low-impedance) microphones; it's also used for professional digital audio transmission ("AES"). 4-pin and higher XLRs can be used for just about anything from power supplies to stereo microphones to anything else you can think of . . .

RCA - usually unbalanced line-level, nominal -10dBV level (~12dB difference from +4dBu). But generally used in consumer equipment for any connection, including phono outputs (much lower level) and even small powered speakers on things like boomboxes. Pros tend to think of RCAs as amateur-grade, because they aren't latching, they aren't balanced, -10dBV is less than 11 (I mean +4dBu), the connectors often aren't physically robust and probably a dozen other reasons. Mostly we just like to laugh at audiophiles who spend a lot of money on an unbalanced connection :D But for short runs of cables that don't get patched constantly, they work fine. Importantly, also used for S/PDIF, which is the consumer version of the AES digital transmission format. *Usually* S/PDIF jacks will be a different color, sometimes black, sometimes yellow. Confusingly, yellow RCAs are also used for video . . . but analog audio is nearly always white and red (left and red).

hmmm . . . . what else . . .

TOSLINK - an optical connector that is usually either S/PDIF format (stereo) or ADAT format (eight channels up to 48kHz, four channels up to 96kHz). Can also be used for a clock signal.

BNC - most often used in audio gear for clock signal. Used in video or RF gear for tons of stuff . . .

Speakon - these are pro power connectors, from amp to speaker.

mini-XLRs - most often used for wireless bodypacks, the format varies by manufacturer. Shure uses a four pin with power and different terminations on each pin.

I'm sure I've forgotten something . . . banana plugs . . . eh, I hope nobody is using those . . . uh, I do :o
 
, also used for S/PDIF, which is the consumer version of the AES digital transmission format. *Usually* S/PDIF jacks will be a different color, sometimes black, sometimes yellow
But remember,you can't use an RCA cable for SPDIF.
 
But remember,you can't use an RCA cable for SPDIF.

That depends on the characteristic impedance of the cable, you just have to check if it's 75 ohm . . . and also its length. The rule of thumb for a cable to act as a transmission line is 1/10 wavelength, which for S/PDIF (running about 3mHz) is 10m. Really, S/PDIF is rather tolerant of peccadillos . . .

AES, on the other hand is designed for longer cable runs, if so you should seek out 110 ohm cable.
 
Back
Top