Impedance....
Wall-power examples.
I have a 60 Watt 120V lamp. Obviously it must suck 0.5 Amperes at 120V to eat 60 Watts. The ratio 120V/0.5A is the working impedance: 240 ohms.
(I say "working" because an incandescent lamp will read lower when cold; pay no attention to that.)
I have a heat-blower motor which, at start-up, pulls 60A at 120V: 120V/60A= 2 ohms.
Those are load impedances.
What are source impedances?
There is a hydro-power dam upriver here. If it were wired for 120V it might _safely_ put out 4,000 Amps: a 0.03 ohm load. Say I go there, measure 120V exactly, then plug-in my 60W lamp and measure 119.9997V. A 0.5A load causes a 0.0003V drop. 0.0003V/0.5A= 0.0006 ohms output impedance.
Back at my house, at the end of a long feeder wire, that same 60W lamp and 0.5A load causes 0.1V drop. 0.1V/0.5A= 0.2 ohm output impedance of my 400-foot wire.
It is usually "best" to use low-Z sources and hi-Z loads. For example, with my 0.1 ohm feeder wire, one 60W lamp causes a negligible 0.04% drop and 0.04% power-waste in my feeder. "Good power practice" allows 2% drop, I can have forty-eight 60W lamps all at once.
My feeder wire is not "best practice". When my blower kicks-on, the 60A draw in 0.1 ohm feeder drops my voltage by 60A*0.1ohm= 6 Volts, a 5% drop. This causes lamps to dim. That's why "best practice" calls for 2% drop. However in this case, a 2%-drop feeder would cost me $10K++. I have decided that my lamp-dim is more acceptable than the cost of no-dim feeder. The "cost of sag" is often a real trade-off.
Audio:
In power situations we may assume a single 50 or 60Hz frequency. Audio runs 20Hz-20KHz. Many real sources and loads are not the same voltage/current ratio at all frequencies.
In domestic power, the "best practice" says 2% max sag because the eye is good at detecting lamp-dim (and because many lamps' output falls faster than voltage). The ear is no so sensitive, and hardly notices 10% "dimming" (1dB level drop).
So the common rule for audio is to make your loads >10 times higher than your sources.
Imagine 1.000 Volt coming out of a 1 ohm source. If we load with infinity, we get 1.000V. Load with 100 ohms, get 0.99V. Load with 10 ohms, get 0.9V (actually 0.90909...V). Load with 1 ohm, get 0.5V. The ">10 rule" means you have 90% of the most you could get, and 90% is as good as 100% for most audio purposes.
When impedances vary with frequency, pick the worst-case.
A gitar pickup with cable is 5K resistance at low/mid frequency but rises to 100K at some medium-high frequency (and then falls off). Take the 100K worst-case. You want to load with 10 times higher: 1Meg.
Many audio chips have output impedances below 1K. Most line-level box-loads are designed for >10K input impedance. In current custom, often 22K.
Back when tubes ruled, you could not expect a source (such as preamp) to be less than 10K output impedance. Most line inputs were over 100K input impedance.
If you hang a modern recorder on an old tube hi-fi preamp: 22K load on 8K source, you only get 73% of the un-loaded voltage. This may be acceptable. Maybe it is the only preamp you have. Maybe it has cherished flavor (but be sure the flavor is not spoiled by the heavy load).
Just for completeness:
Looong telephone lines, so long that the wire parasitics matter, have impedance of 100-900 ohms (depending on construction). And a very long line can echo, and this can be minimized with a "matching" resistor. So there is a vast body of gear with 150/600 ohm nominal impedance. Also when tubes were COSTLY, you did not waste them making low-Z outputs. You transformed to nominal impedance, then did the math to figure what really happened. If an output is true 600 ohms, and an input is true 600 ohms, the voltage drops 50%; however tube-cost is minimized.
Since the cost of a tube fell below a week's-pay, most gear has been designed for low output Z and high input Z. The few situations where you "must" match can be loaded with resistors. Most "pro line inputs" are over 10K actual impedance.
The "oatmeal-1" output stage has internal impedance a bit under 1K, low-enuff for 10K loads. It won't deliver much more than 1mA output current, but 1mA in 10K load is 10V peak, and musical instrument interfaces should rarely need more than 1V or so on the line.
Topic: Alembic F2-B Pre-Amp Restoration project (Read 62983 times)