High(low) impedances by TTl logic

[gotilio]

Can someone tell me what does high or low impedances mean for the inputs and outputs for TTL logic,and how you can see that a port is high(low) impedant.
And what if you connect other things to such ports?

 

[Kenneth Mann]

Impedance consequences

The typical TTL device output consists of two (NPN) transistors, connected from power to ground, with the emitter of the "upper" transistor going (through a diode) to the collector of the "lower" transistor, and its emitter to ground. The collector of the "upper" transistor is connected through a very small resistor, to the power (Vcc). The "output" is generally between the diode and the "lower" transistor. This output is "low impedance", which allows one unit to drive the inputs of several devices (fan-out). Those inputs are of relatively "high impedance", allowing several of them to be driven by a single output without overloading it.

When the output is in a "high" state ("1", "ON"), the upper transistor is turned ON and the lower transistor turned OFF. This means that current flow is from Vcc, through the output, to the input(s) of the following device(s), to ground. These inputs are of moderately high impedance, so that they don't draw an especially large current. In this condition (ON), the inputs serve as current sinks, and since they draw little current, several can be driven by a single output. The result is a "1" voltage from the output to the connected inputs.

When the output goes to a "low" state ("0", "OFF"), the upper transistor is turned OFF and the lower transistor turned ON. This means that current flow is now inward via the output, and goes to ground through the lower transistor. That current now must come from the inputs that follow (via their Vcc's). Again, since the inputs are of relatively high impedance, they will not overload the output that is connected to them, even though that output is almost a dead short to ground. (And, it doesn't take much current to control the inputs.) The result in this case, is a "0" from the output onto the connected inputs.

This type of output is referred to as a "Totem Pole" output. There is one caution in this design configuration; the low impedance of the output means that two such outputs must never be connected together. To understand this, consider the case in which two are connected and one goes to "ON" while the other goes to "OFF". The output that is ON is connected directly to Vcc, while the output that is OFF is connected directly to ground, creating a near dead short, and resulting (usually within minutes) in barbecued IC. (Besides, the logic may also be wrong, if not planned for.)

There are, however, cases in which we want to be able to connect outputs together. One such condition occurrs when it is desired to create a "wire-OR" case (which I suspect is rarely used these days). This is simply the case in which, rather than using an OR gate, outputs are tied directly together. I don't recommend it; it sometimes confuses the logic - mainly because it is rarely used. It can definitely create a circuit tracing nightmare (An invisible OR-gate) for a technician who must do maintenance. To do this, ICs with "Open-Collector" outputs are used (along with pull-up resistors).

A more common case comes up when it is desired to create a "Bus". This entity, which is very common in computer design and with similar units, simply allows an output to be "disabled" by putting that output into a "high-impedance" state. Several outputs may be connected each line in a bus, however at any time, only one can be enabled, thus keeping our rule intact. To accomplish this, we use what are called "Tri-State" gates, etc. These are simply IC devices, each of which has an input that allows its output to be disabled. All this input does is turn off Both transistors in the output of that device (no matter what the other inputs to that device are doing). As a result, if we follow our rule, and disable the outputs of all ICs that go to the bus line except the one chosen to be active, it will be the only one seen by the bus line, and there will be no conflicts. The high impedance output allows an IC to be 'hidden'. Then when we want it, we simply enable its output, and disable all the rest. I hope that this is understandable.

KM

 

[rayjohn01]

Cant really understand why anybody would wish to look at TTL these days when Cmos does a far better job and is far more understandable. Threshold voltages in TTL are poorly defined , temperature dependant , not at 1/2 supply , the technology takes gobs of current most wasted , it's only claim is speed and yet at high speed it is sometimes hard to apply due to the levels of current switched.
High speed Cmos wastes very lttle power , has thresholds which even if not specified are within a few % of 1/2 supply giving the greatest noise margin , supplies do not have to be accurate -- in undefined circumstances you can often operate at 1 volt
at reduced speed , and today the speed is such that most casual users would never use it.
There are many adherents to TTL -- but it was a tecnology which grew out of the Bipolar transistor era when the King was the Opamp , also the initial experiments with MOS tecnology ( nmos ,pmos, resulted in VERY hard to understand and design circuits)
But that ceased with CMOS --- today no-one would consider making a microprocessor in any thing but Cmos -- except under extrordinary circumstances such as the extreme of speed and power be damned.
Since processors reach speeds of ~ 4 Ghz I would suggest that most mere mortals are not even equipped with the know how as to use it . You cannot build Bipolar devices with the dimensions of a simple Cmos device they are just too complex for that ( with hidden problems ) .
Just a little while ago I had made some comment on a Forum about feature sizes in IC Chips , and he came back with a comment that I was full of SHT and that feature sizes were only 3 microns , in fact feature sizes are at 20 nMeters in some special applications --- He is also an advocate of TTL .
I do not know what problem you are dealing with Gotilio -- because you did not say
if it's homework -- then I guess you have to give an answer and the other contributor gave a good answer -- if it's by choice , then I would say ditch TTL it is an extremely limited and old technology .
Ray.

 

[gotilio]

It's a chapter in a book that I have to study,but I dindn't understand the part about the impedances.
thanks for the explanation!

 

[rayjohn01]

Exactly -- and I wonder why -- the real truth about impeances in TTL is extremely complex -- they change value for any level of voltage applied -- they are totally non-linear -- so the answers can only be given in the simplest of cases -- so I think you can see why I question why anybody would ask you this --- to really understand this topic you would have to study the precise action of transistors ( both npn and pnp )
under a variety if bias conditions -- and not just at DC but at speed .
To me the way you pose the question it is unanswerable -- and I question whether the question poser knows what the heck he is talking about .
I think that you need clarification about the question , impedance as such is usually considered as some simple linear proerty of a linear circuit -- including inductance capacitance and resistance , TTL hardly qualifies .
Ray.

 

[Kenneth Mann]

Exactly -- and I wonder why -- the real truth about impeances in TTL is extremely complex -- they change value for any level of voltage applied -- they are totally non-linear -- so the answers can only be given in the simplest of cases -- so I think you can see why I question why anybody would ask you this --- to really understand this topic you would have to study the precise action of transistors ( both npn and pnp )
under a variety if bias conditions -- and not just at DC but at speed .
To me the way you pose the question it is unanswerable -- and I question whether the question poser knows what the heck he is talking about .
I think that you need clarification about the question , impedance as such is usually considered as some simple linear proerty of a linear circuit -- including inductance capacitance and resistance , TTL hardly qualifies .
Ray.

There is and has for years, been a generally accepted practice of using the term "impedance" with respect to digital circuitry in its "English" context as that which 'impedes' the flow of current, and thus defines the level of an output signal. Granted, it does not fall within the general linear design context of resistance, capacitance or inductance, but then it is not needed in that context. If you have a better term, I'm sure the world will welcome it.

In the mean time, I propose a compromise: I won't become a TTL advocate if you'll not put forth that which simply confuses people.

With great respect,
KM

 

[rayjohn01]

z's

Then with great respect you will realize that the sink impedance of TTL is dependant on the current sunk as it is a saturation resistance unless arranged to be caught by some diode arrangement.
And the source impedance ( at a transistor emmitter is dependant on that transistor feed arrangements and its beta.
Generally then you will not talk of impedances but sink or source voltages under load.
Ray.