Understanding Impedance: Exploring Practical Applications and Manipulations"

[edmondng]

I know the formulas, to some extend what impedance is. But i guess i would like to know what it really means in real world and how to use it/match it or manipulate it.

Please help correct my understanding or increase it. :)

High impedance is like high resistance and vice versa. You want low impedance to ground to remove noise for example. Which is why people put caps to ground because it removes high frequency noise/oscillation (ie: high f and fixed C makes low Z)

Now on some ic's, take op amp for example it has high input impedance. Is this to limit current and preserve the signal?

I've always thought that even when we connect two device together, we want to match the impedance to maximize signal transfer. If one is at 1000Ohm and the other at 50 Ohm, we don;t want to use resistor (because we lose the signal in terms of heat?) but use caps/inductor provided we know the freq? (how does this work over a variable frequency?)

The other thing is, why high impedance and low impedance on some equipment, ic's etc. Someone mentioned that BJT's have low input impedance (I suppose because they use current to turn on/off unlike mosfet which i would think have high input impedance?)

The other day i was trying to measure some voltage on my thermocouple, and the reading that come out did not correlate with the spec of thermocouple. It was indeed increasing/decreasing but was not linear with temperature. In fact, the numbers were all off (ambient temp came up with voltage reading when referred to spec equivalent to 60C etc..) Now i think this has something to do with impedance because i just connected directly to the meter. What is the proper way then to match it?

Are impedance most significant when it comes to small signal? When to consider and how to use it. Just wanting to increase my understanding.
Thanks

 

[berkeman]

I'll try to address a few of your questions...

** High input impedance is important when the signal source has a moderate to high output impedance, and you don't want to load the source and distort the signal. This is why DVMs and oscilloscopes have high input impedance probes.

** You want to match the source impedance and load impedance when you want to transfer maximum power from the source to the load. Usually the source impedance will be real (like an antenna at resonance, or the characteristic impedance of a transmission line, etc.), so you match it with a real input impedance at your load (like a wideband resistor structure). If the output impedance of your source is complex, then you match it best with the complex conjugate load impedance (same real part, but opposite imaginary part). The broadband impedance matching that you are asking about can only really be done with real impedances, I believe. Like matching a transmission line's characteristic impedance to a load.

** BJTs do have a lower input impedance than FETs, in general, because you are supplying a base current, and $$Z_{IN} = \frac{\Delta V}{\Delta I}$$

 

[ranger]

The other thing is, why high impedance and low impedance on some equipment, ic's etc. Someone mentioned that BJT's have low input impedance (I suppose because they use current to turn on/off unlike mosfet which i would think have high input impedance?)

The BJT is designed to be a current controlled device. Its like saying we have to forward bias the diode junction to bring the device into conduction. When we do this, the input impedance is very small. In a bjt amplifier, the input impedance looking into the amplifier (base) from the signal source is dependent on the resistor biasing network and beta for that that particular transistor.
However MOSFETs have very high input impedance becuase there is an insulation (SiO₂) barrier between the gate and the rest of the FET. So because of this no current flows into the gate. Instead the FET is controlled by the electric field to vary the width of the channel between the source and drain (hence the name - Field Effect Transistor).
The JFET is similar, although it has a reversed biased gate-source region. Becuase this region is reversed biased in normal operation, the input impedance is very high. If the gate-source region is forward biased, current flows into the gate and it defeats the purpose of the JFET.

 

[edmondng]

thanks berkeman and ranger. so here's how i understand impedance

source = put out signal
load = receive signal

if load = high impedance, and source = low impedance, then its is ok because the load will have the most voltage transfer resulting in strong signal

if load impedance = source impedance, also ok because it matches so max transfer.

if load impedance < source impedance, then it is bad. The source signal is too small to transfer to the load (low frequencies roll off and you lose signal) so you need to increase the impedance of the load by maybe putting resistor in series/parallel.

Now if high input(load) impedance is good, why not make all input devices with high input impedance? like wise why does some source have high output impedance (i guess they must have)
so with a 50ohm source, i can just connect to the scope that has 10kOhm impedance and be fine (the 1M is probe to ground while 50ohm is actual impedance to scope using BNC)?

 

[berkeman]

if load impedance < source impedance, then it is bad. The source signal is too small to transfer to the load (low frequencies roll off and you lose signal) so you need to increase the impedance of the load by maybe putting resistor in series/parallel.

Now if high input(load) impedance is good, why not make all input devices with high input impedance? like wise why does some source have high output impedance (i guess they must have)
so with a 50ohm source, i can just connect to the scope that has 10kOhm impedance and be fine (the 1M is probe to ground while 50ohm is actual impedance to scope using BNC)?

Don't confuse power transfer applications with signal measurement applications. They are two different animals, with two different optimum solutions in terms of impedances.

And even in small-signal applications in 50 Ohm systems, you are maintaining 50 Ohms for Zout, Zin and Zo (characteristic impedance of any transmission lines) in order to avoid reflections of high frequency signals, which will mess up the measurements.

And if you have a 50 Ohm source that you want to watch with your oscilloscope, you need to terminate the source in 50 Ohms at the 'scope, or you will not get a correct amplitude measurement. Like, if you have a signal generator with a 50 Ohm output, and use a 'scope probe to watch that output, you will see a signal that is 2x what the sig gen is set for. Quiz Question -- why is that?

 

[edmondng]

Like, if you have a signal generator with a 50 Ohm output, and use a 'scope probe to watch that output, you will see a signal that is 2x what the sig gen is set for. Quiz Question -- why is that?

signal gen 50ohm and put scope probe you get twice? Maybe the scope, signal gen or probe is broken :)

I've not seen that before. But if its twice wouldn't that mean the impedance is twice like 100ohm?

 

[berkeman]

I've not seen that before. But if its twice wouldn't that mean the impedance is twice like 100ohm?

You'll see it a lot as you work more with sig gens and 'scopes. Getting 2x what the sig gen says it's outputting wouldn't mean the load impedance is 100 Ohms. Think of the voltege divider formed by the 50 Ohm output impedance of the sig gen and the Z Ohm impedance of the load. The sig gen output amplitude setting/readings on its display assume that it is driving a 50 Ohm load, so it is factoring in a divide-by-two of its open circuit output voltage. So with no load (or a high Z load like a 1Meg 'scope probe), you see 2x the sig gen display's reading. If you connected a 100 Ohm load, you'd see 2/3 * 2, or 4/3 the display reading. Make sense?

 

[edmondng]

are you trying to say the scope probe sees 2x but the scope (because it has 50ohm impedance) sees the true signal?

so if my source generator is 600ohm impedance and scope is 50 ohm, i would see (50/650)*2 of the signal strength? Which would probably mean than my input signal has degraded, but if my input impedance is always much higher than source then it would never be a problem?

i have to think this over in my sleep tonight. I haven't seen my signal on scope being 2x what i set on the generator although i do have to tweak the amplitude to get what i want. Are you also trying to say what we see on the scope is not what an actual circuit/parts are seeing?

 

[berkeman]

I'm referring to 50 Ohm output impedance signal generators with digital displays of output amplitude (you set the desired output amplitude with a keypad or scroll wheel, typically). Those are calibrated assuming a 50Ohm load.

For 600 Ohm output signal generators, they probably just show you what the open-circuit voltage is approximately going to be. And your measured output value will be off by the voltage divider formed with the load impedance.

 

[Averagesupernova]

I'm referring to 50 Ohm output impedance signal generators with digital displays of output amplitude (you set the desired output amplitude with a keypad or scroll wheel, typically). Those are calibrated assuming a 50Ohm load.

Not necessarily. Alot of signal generators sample the output to provide negative feedback to help stabilize the output. I'm not saying that the output will not change with loading, but it may not double as you describe. It really depends on the design of the device.