[Moved]: VIH, VIL, VOH and VOL in a demux, mux, decoders, etc

[AMSA84]

Hi guys,

I am trying to measure these parameters but I don't know how. I just know how to do it for inverter but for this case I don't know.

Can someone give me a tip?

Regards.

 

[erikl]

Re: VIH, VIL, VOH and VOL in a demux, mux, decoders, etc

Sweep input(s) over the supply voltage range for b.c. & w.c. PVT corners to find these input value limits for predefined output value limits, usually predetermined by the process used, respectively the Digital Logic standard output value limits to be compatible with.

 

[AMSA84]

Re: VIH, VIL, VOH and VOL in a demux, mux, decoders, etc

Hi erikl, thank you for your answer.

Allow me to ask you what you mean with b.c. and w.c.? I guess that the w.c. has to do with worst corner (which makes sense) but regarding the b.c. I don't havea clue.

Regarding the configuration for the simulation, the testbench as simply as if we were doing the DC sweep for the inverter, right? For example, what I did was put 2 of the 3 inputs to zero and swept only one of them. Then I plotted the Vout = f(Vina). The Vil and Vih are the values in the vin axis that we get when we plot the dc sweep? The same for the vol and voh?

Regards.

 

[erikl]

Re: VIH, VIL, VOH and VOL in a demux, mux, decoders, etc

... but regarding the b.c. I don't have a clue.

The contrary of w.c.!

Regarding the configuration for the simulation, the testbench as simply as if we were doing the DC sweep for the inverter, right?

Right!

For example, what I did was put 2 of the 3 inputs to zero and swept only one of them. Then I plotted the Vout = f(Vina). The Vil and Vih are the values in the vin axis that we get when we plot the dc sweep? The same for the vol and voh?

From one single simulation run you get just one single set of Vi and Vo values. Vil you get from a b.c. corner run, Vih from a w.c. corner run. Correspondingly Vol from a b.c. run, Voh from a w.c. run, specified always with a given load. Also, specify Vi and Vo values for typ. case (tt section of a lib). [/QUOTE]

 

[AMSA84]

Re: VIH, VIL, VOH and VOL in a demux, mux, decoders, etc

HI erikl, thank you very much for your explanation. I did that.

Let me ask you now one more thing, if I want to use the calculator to evaluate these points do you what functions to use? I can't think of a function that can notice the change on the output sweep voltage to grab the respective values that I want. Thought in the cross function but I don't think that it works.

Regards.

 

[erikl]

Re: VIH, VIL, VOH and VOL in a demux, mux, decoders, etc

... if I want to use the calculator to evaluate these points do you what functions to use? I can't think of a function that can notice the change on the output sweep voltage to grab the respective values that I want. Thought in the cross function but I don't think that it works.

I wouldn't know any calculator function to produce a table for a data sheet ;-). See here an HCMOS example of a data sheet table for VIH, VIL, VOH and VOL values: View attachment logic-levels.pdf

I hope you understand what b.c., typ. and w.c. values mean. These are not data adequate for curves!

I always used to create such data sheet tables for my circuits manually.

 

[AMSA84]

Re: VIH, VIL, VOH and VOL in a demux, mux, decoders, etc

Thanks erikl.

Now allow me to ask you another question. It has to do with the input capacitance and the output capacitance of a decoder or a mux.
Normally the spec regading the iput capacitance of the mux that we see in the datasheet are measured in which conditions?

 

[erikl]

Re: VIH, VIL, VOH and VOL in a demux, mux, decoders, etc

... the input capacitance of the mux that we see in the datasheet are measured in which conditions?

In the given conditions (mostly VDD and Temperature). The input capacitance of logic gates isn't so important, if it is given at all in a dataSheet, it is mostly found near the end of the electrical characteristics, s. e.g. this dataSheet, p6/12. Here C_in isn't given as result of a real analysis, but as an overestimated value, to be always on the safe side, meaning this value will not be tested at the go/no go test.

 

[AMSA84]

Re: VIH, VIL, VOH and VOL in a demux, mux, decoders, etc

I see. Thanks for the answer erikl.

 

[AMSA84]

Re: VIH, VIL, VOH and VOL in a demux, mux, decoders, etc

Hi erikl,

Sorry to dig this post again, but I have an issue here.

I have ran the simulation to determine the voh vol and vih vil and I did both in transient and in the dc analysis both with sufficient small step size.

Now, the problem is that I get different values for each of the simulations.

Are not supposed to be the same?

 

[KlausST]

Re: VIH, VIL, VOH and VOL in a demux, mux, decoders, etc

Hi,

Now, the problem is that I get different values for each of the simulations.

... different in microvolts is OK
... different in volts is not OK

Klaus

 

[AMSA84]

Re: VIH, VIL, VOH and VOL in a demux, mux, decoders, etc

So, for example:

To get the vih and vil using the transient simulation I put a triangular waveform at one of the input that I want to test those two parameters and then I measure the crossing point between the output and the triangular waveform. I think I can do this like that.

Now if I do the same using the dc sweep, in the same ports, I bascially get for the VIH a difference of 0.001V and for the VIL is the same.

For the output levels I get a more significant difference. For the VOH I get 3.725V vs 3.8V in DC sweep and for the VOL I get in transient 26mV and for the DC sweep I get simply 20nV which is bascially zero.

Maybe I should try to tight the maxstep in the transient? I am using 1mV for the DC sweep.

Regards.

 

[dick_freebird]

Re: VIH, VIL, VOH and VOL in a demux, mux, decoders, etc

For a triangle waveform to produce the same result as a
DC hysteresis sweep, the delay of the path in question
must be << the ramp time -through the region of interest-
(e.g. the linear window of an inverter chain). A 1uS total
rise time and a 10nS prop delay will give you 1%-ish error
if the whole range is linear, 10%-ish if only (say) 2V to 3V
is the linear input range. That's for switching VIH/VIL; you
are possibly going to find other issues such as input level
affecting on resistance or leakage or IDD and depending on
what your VIH/VIL is asserted to encompass, maybe the
answer (all told) is outside the logic-only VIH/VIL you'd
extract from the sim-set.

 

[AMSA84]

Well, the rise time and fall time of the triangular waveform are both set to 1ms. I think that's enough right?

 

[KlausST]

Hi,

Well, the rise time and fall time of the triangular waveform are both set to 1ms. I think that's enough right?

Calculation example:
With 1V/ms trianlge and a delay of 1us you get a voltage error of +1mV rising and -1mV falling. A total difference of 2mV.

Klaus

 

[AMSA84]

Where the delay comes from? Why it has to be taken into account?

 

[dick_freebird]

You are complaining about a DC sweep which has no time sense,
not dead-matching a transient simulation which does. That's
one point of fundamental discrepancy. A switching event that
appears sharp at DC might be pretty slow at tran because of
low overdrive "in the moment" and so on. Meanwhile the input
continues its travel and you pick off its value when the output
criterion is satisfied, which happens later and variably so as
well (prop delays responding to PVT). Output transition time
!= input "now it will transition, eventually" time.