Impedance explained?

[cordura21]

thanks to everybody who is participating in this very interesting thread.

I wish there was an "impedance knob" in my amp, so you can adjust it for whatever you're plugging in. There are impedance matchers at Radio Shack, so I guess that wouldn't be so expensive.

Now that I participate of this thread, I understand other articles better. I recommend to read this one, from tape.com by Bruce Bartlett:
http://www.tape.com/Bartlett_Articles/impedance.html

There it says that if you plug a low Z mic into a high Z mixer input (phone jack) you'll get a weak signal because the mixer will be expecting a relatively high Z signal, so the gain will be low.
I thought that bridging was ok with a 10 times impedance relationship OR MORE. But from what the article says, if the relationship is very big, then you will have weak gain, resulting (I pressume) in a bad signal to noise ratio.

Maybe that's the diminishing returns that somebody was mentioning earlier in the thread. I hope my reasoning is correct.

Cheers, Andrés

 

[skippy]

The *gain* doesn't change materially: it's simply that a high-Z input amp is usually expecting a different input voltage level than the low-Z input amp, right from the start of the design phase.

High-Z sources usually produce a greater voltage swing than low-Z sources, by their nature. An amp that is intended for use with a high-Z source will therefore usually exhibit a lower overall voltage gain than one used with the low-Z source. A guitar amp may only give you 40dB of voltage gain from input to output, whereas a mic pre can run at 60dB of voltage gain all day, every day. Low-Z sources are actually better at outputting _current_ than voltage.

The mixer example you give bears that out, although two things could actually be happening. Usually, 1/4" phone jack inputs will be line level inputs: they'll be expecting -10dBV or +4dBm levels at a relatively low input impedance, and they may only have 20dB of make-up gain available. If you put a low-Z mic signal in, it may only be swinging around -50dBV- which means that even with the gain knob maxed, you'll get insufficient signal. Same thing happens if you put a low-Z mic into the input of your guitar amp. The voltage swing from that low-Z source is just too low, so you won't get much output even cranked to 11.

That's where an "impedance matching" transformer comes in: you trade off some current for some voltage. Most folks know that transformers change voltages: if you have a transformer with 10 times as many turns in its secondary (output) winding as in its primary (input) winding, you get a 10:1 voltage stepup. What most people don't know is that transformers also change impedances based on their turns ratio. If that 10:1 step-up transformer has a 100Kohm resistor across its secondary as its load (like the input of your guitar amp), the primary will *act* as if it has a 10Kohm resistor across it (as if the transformer was replaced by a resistance equal to the load scaled by the turns ratio).

So if we run our low-Z mic into a 10:1 stepup transformer, the mic thinks it is seeing 1/10th of the input impedance of the amp, call it 10Kohms. It's a low-Z mic, so it likes that- it can easily produce enough current to swing the exact same voltage into 10Kohms- it doesn't care about the loading. And the amp sees 10X the voltage swing after the stepup, or an extra 10dB of "voltage gain" before we even get into the input connector. So it has more input voltage swing, and it likes *that*. All's right with the world, because our impedance transformation traded current that we didn't need for voltage that we did.

This works because the low-Z mic can produce a fair amount of current, so the *power* (voltage times current) transferred across the transformer stays nearly constant. There are limits, and you don't get something for nothing: you can't hook a 10,000:1 step-up transformer across your Telecaster and launch small creatures into orbit, or power whole cities while shredding. But for reasonable translations between voltage level and source/load impedances, transformers are your friends. Connect that same transformer _backwards_, and you could use it to make your guitar happily drive a low-Z mic preamp (we usually call this a "direct box", and we do this all the time). That 1:10 scaling would make the resulting load seen by the guitar be 100Kohms instead of the usual 10Kohms of the mic preamp, and in that case we'd trade off voltage we didn't need for current that we did- to control the loading of the pickup.

I don't want to dig too deeply into signal transfer through transformers, because it reminds me way too much of my unpleasant and now long-ago academic life (we called the inrtoductory course into Electricity and Magnetism "S&M", and with good reason). But hopefully that will help a little more, and give you an idea how impedance matching works...

It just blows my mind that anybody is actually still reading this thread!

 

[cordura21]

I am still here, wrestling with the complexity of the subject AND fighting my translation errors too. But it's surely worth it. I am gonna have to read it all over again to try to get it all at once , but it's no Tolkien book so I think I'll get it.

Thanks for it,
Cheers, Andrés

 

[guinsu]

Hi, I'm sorta new here, but I had to post after seeing the great replies here (I think I finally understand impedance after the explainations I saw here).

Anyway my question was sort of a practical application of what I saw here. I have a Shure SM-57 mic and a Fostex VF-16. The SM-57 has an impendance of 150 Ohms, the Fostex inputs are 10k (balanced) and 50K(unbalanced). Now I am curious whether the input impedance is too high for my mic? One reason I am asking is that I went through the manual for my Tascam 4 track and the balanced inputs were only 5k. Does this matter at all?

I also noticed the output impedance on the line and monitor out on the Fostex are 10k Ohms and 100 Ohms on the Tascam. That seems like a big difference, does that difference matter when I am hooking it into a tape deck/amp/normal piece of home audio equipment?

Thanks for any help...

Tim

 

[skippy]

The input impedance of the mic pres on the Fostex are just fine at 10Kohms. If it's a mic pre, it wil have enough gain to be able to handle your SM57. Really, it's a matter of gain: if the preamp is intended to handle low-Z mic level signals, its input impedance is really up to the designer, and can vary widely over the range of about 1Kohms to 100Kohms based upone the designer's particular religion. If there's enough gain there, the actual input impedance seen by the mic is a secondary issue.

The output impedances of the Fostex (at 10Kohms) sound like a mistake, or a misprint. It's damned *hard* to make active electronics have an output impedance that high! On the Fostex web site, they spec those outputs as "requiring a load impedance of 10Kohms or more"- which is an entirely different kettle of fish. That means that their actual output impedance is probably something well *under* 1Kohms, but the web site doesn't say.

What they're saying there is that their output buffers are too feeble to drive a real, studio-style 600ohm terminated load (which we haven't talked about at _all_, because they are really ancient history for most home studio uses). That's fine: the chances are that you'll never see one of those in a home studio, or even in most professional studios these days. Most modern prosumer recording gear (stuff that you and I would buy) is designed around low-Z outputs driving bridging (high-Z) inputs, so you're in fine shape.

Where you'd run into trouble with the VF16 would be in driving old, transformer input studio gear (like old passive EQs, or other stuff that was designed to present a 600ohm load). That stuff is incredibly uncommon these days even in pro rooms. If you *did* drive something like that, you'd get a lot of distortion, very little headroom, and very little output level (as the low-Z input would load your poor -10dBV output buffers nearly to death).

That gets into the old +4dBm versus -10dBV problem. Lots of prosumer gear these days is designed with -10dBV outputs, like the ones on your VF16. They can swing 0.1V into a 10Kohm load all day: that's what "-10dBV" means. But the +4dBm spec is built around driving 1 milliwatt into *600*ohms as its load: a real tractor-pulling exercise, by comparison. Good +4dBm drivers have _muscle_. Something with +4dBm outputs can push 600ohms all day long, where a -10dBV output would be right down on its knees, clipping, distorting, and being generally unhappy.

Fact is that you seldom _need_ to do that any more, because so much modern gear (even pro +4dBm gear) has high-Z, bridging *inputs*. You almost never have to actually *drive* a real, honest-to-Gawd 600ohm load these days. It's gotten to the point now where usually, the only practical difference between +4dBm and -10dBV gear is 12dB in the voltage levels, because the inputs are all bridging anyway. But the basic thing to remember is that +4dBm outputs were designed from the beginning to be able to push significant power into *low impedance*, terminated loads, and -10dBV outs _aren't_.

So: don't sweat connecting the VF16 to _any_ normal piece of consumer or prosumer home studio gear. They can pretty much all handle it, and most pieces will even offer that 12dB of make-up voltage gain needed to bump your -10dBV levels to the +4dBm levels of pro-inclined gear.

However, if you rescue that moldy old Fairchild limiter from the defunct AM radio station in your town, and try to use it: you'll have grief with its 600-ohm inputs, until you buy a bump box with active electronics and real +4dBm outs to drive it...

And before anyone asks: yes, the voltage difference between -10 and +4 really is approximately 12dB. Really. Why? Measurements specified in dB units are ratios with respect to a reference. The -10dBV is with respect to 1V across 10Kohms, the +4dBm is with respect to 1 mW into 600 ohms, or .7745V. Different reference impedances, different reference voltage levels: apples and oranges.

That's why the difference is approximately 12dB, not the 14 you'd probably expect. If you want to get pedantic about it, the voltage expressed by 0dBm (.7745V) is actually -2.218487dBV, not that anybody is _ever_ gonna care but nerds like me...

 

[camus]

Mind if I butt in here with a dumb question? I've been wanting to use some of my "guitar rack" effect units (Peavey Valverb) as external fx connected to my mixer, but I notice their recommended load impedances for their outputs is pretty high, around 33Kohms. The line inputs on my mixer are rated at about 10Kohms, the stereo returns at 8.6Kohms. How do I get around this? Or should I skip using them as external fx units?

 

[skippy]

Sounds like they optimized the output stage of that unit to drive the extremely high-Z input of a guitar amp, rather than going direct into a board. You'll be loading it a little more heavily than they had in mind, but I don't think that that will be a problem: it's a case of active electronics driving the mismatch, as opposed to having the pickups themselves trying to drive the mismatch. I doubt that you'll encounter any problems at all- but if you do, they will most likely be in running a little short on headroom when trying to drive really high levels into the board. And even so, you may actually like the sound of it as it breaks up! One size does not fit all.

That's a small enough mismatch to be worth trying, and if you like the sound you get, ignoring. The chances of doing harm to the equipment with this small a mismatch are essentially zero. It's not like you're trying to drive 600 ohms with it: the output buffer stage would get right pissed about *that*, most likely. 10Kohms should be a walk in the park.

Another thing you might want to try, just for the sonic adventure of it, might be a matching transformer. The cheapo Rat Shack matching transformer might be _perfect_ for this: it'll add some nice nonlinearities and lumpy distortions to the signal that you might actually really like, and you could use it to convert that slightly low input impedance of the line input on your board into something that looks more like the input of a guitar amp.

You don't _need_ to do this: it's just a suggestion for an amusing, and just maybe useful, experiment. When you're trying to _create_ a tone (as opposed to trying to perfectly preserve a signal you've captured) sometimes lousy, lossy transformers like that one are just the ticket. But I'd certainly try it without first, and if you like the tone you get, then use it in good health.

 

[camus]

Thanks a bunch for the suggestions, Skip.....you da man!!! I guess I'd better sign up for that crash course in EE I've been wanting to take for years now, sigh...