Inferno #05 30 апреля 2004

Железо

Iron - We investigate the chip K561TL1. .

 K561TL1


   This 4-element 2, NOT with Schmitt triggers on the inputs. 
The latter distinguishes the work This mikruhi of all known 
logic AND-NOT same K155LA3.


   Schmitt trigger changes its state
(And, consequently, the output signal) is sharply - with 
gradual change in the input signal within certain limits - to 
the border switch. Thus, the switch logic 0 will trigger when 
the input signal drops to a level equal to or below Ul0 - 
border switching low level, and switch to a logic 1 will, if 
the input signal rises to the level of equal to or greater than 
Ul1 - border switching high level. 

   Level Ul1, of course, higher level
Ul0, and if you draw a graph of the
between input and output voltage of Schmitt trigger, it will 
loop - the so-called "hysteresis" (as opposed to simply element 
with the feature as a line). Loop is sometimes portrayed in the 
logic elements with a Schmitt trigger. When the signal level 
greater than Ul0, but less Ul1 trigger is not switches, while 
maintaining its previous state. Thanks to this single area of 
​​the interference gradual change in the signal are excluded. Ie

on the negligible level of interference that
usually act on any signal (at
the tens of mV), they can not change
trigger status (their amplitude is much
smaller than the width of the blank areas). If only
input signal is not too close
to the boundary shift, and thus the probability of switching 
efficiency was high. And in the case of switching under the 
influence interference, it would be equally probable (with a 
useful point of view) and a single, because for another 
switching signal must overcome the entire width of blank areas 
- many greater level of interference.


   Since the range of food mikruhi of 3
to 15, the boundaries shift also depend on food (almost 
linearly), and it is believed that Ul0 approximately equal to 
one third of the Ucc (Supply voltage), t.e.nizhe half, and

Ul1 approximately equal to two thirds of the Ucc - ie
more than half the power. Although usually well
denote the permissible limits of CMOS levels
- Technology, not the border Schmitt trigger.
Sources of information are not quite correct ...

   Personally, I'm interested in current consumption
(When did amigovskuyu mouse Speccy).
Clearly, it depends on the input voltage, and I spent several 
experiments for determine the maximum engorgement

K561TL1. Data kept in mind, and after assembling the mouse just 
forgot. When going to write an article - repeated the 
experiment, but with the foreign analogs MC14093BCP, and 
suspected the difference (although the schema mouse 
irrelevant). When there was K561TL1, decided to completely 
razobratsya with this issue. As a result, screen appeared and 
the text. 

   At the same time decided to write on the screen of the 
schedule Log RADIO 6 / 90, although suspected

it is wrong (in low power boundary Ul1
dropped below half power). Still there
unsavory information about inverting Schmitt trigger. It 
appeared that a 4 Schmitt trigger with the logic-2, NO, and

would be expected to single
switching, if applied to one input 0 (and
the second did not react - according to the logic element). 
Although in itself trigger Schmitt is not associated with any 
logic - This simple device, with input and output.




         As a result, found out the following.


   Like all mikruhi 561 series, this does not consume the input 
current (of course, not counting leakage currents) and consumes 
negligible Supply Current (I did not to measure this shameful

value) if the input signals are close to 0
or supply voltage. Of course, the outputs
I does not load, and connect to them
Digital multimeter M832 (input impedance 1 Mohm - close to the 
load currents leakage can be ignored).


   The same device I sometimes controlled
food mikruhi (putting him on a laboratory unit equal to 3.5 V), 
but mostly I they took it for the input voltage, namely voltage 
supplied from the divider - the multi-resistor 10 k ohms, 
shunted capacitor from interference. This allowed us to get 
closer to the borders of trigger close enough to assert their

value.

   If there was a divider, as I took it for electricity? A
very simple: just 7 feet
sockets through another device on the negative supply. This 
device has been TS20-05. When disadvantage of using (a set of 
jacks), he has a high sensitivity (collected

on an operational amplifier) ​​and almost no effect on the 
measured circuit. Where would all testers believe? I can only 
say that the maximum error was less stress 5%, and the currents 
are less than 8%. Nutrition mikruhi was 5 V + - 0.02 V. The 
temperature in komnote 22 degrees (possible slight drift 
parameters mikruhi temperature). 

   At first, I decided to check one input. Gave him the voltage 
from the divider and the other, to eliminate the influence by 
connecting plus food. When approaching the limits

switch (k Ul0, if the trigger was switched to 1, and Ul1, if 0) 
the current consumption increased (t.k.rosli switching 
currents), reaching a maximum on the boundary shift, and 
immediately after the switch drops sharply to a small value - 
t.k.vhodnoe stress significantly different from 0 and supply 
voltage. 

   I spun the divider in the opposite direction,
closer to the opposite boundary, the
which the process was repeated: first, the current increased, 
then the trigger was switched, and the current drops sharply. 
Several times carefully selecting the borders, I wrote them 
down voltage and associated power consumption. Then put the 
import mikruhu and repeated the experiment. The result was very 
different to current (Had to change the range of the tester).

Then I switched to the unused inputs
plus to minus, the outputs of the permanently switched to the 
unit, and repeated the experiment with the same input and both 
mikruhami. According to the magazine, radio, the input should 
not was switched trigger (once the item is already

switched the output to 1).

   Yes, of course, the output was always 1. But!
All my measurements were repeated without change. Toggle 
flip-flop has been seen by a sharp decline in the current 
conventional boundaries. 



   Maximum current consumption by a trigger
Schmitt input element (one trigger /
input, respectively), with appropriate switching threshold, 
measured at nutrition Ucc = 5V, outputs unloaded and

other inputs are connected to plus or minus supply:



     Domestic chip K561TL1

Ul0 = 2.14V Icc = 0.3mA, Ul1 = 3.17V Icc = 0.41mA



  Import (MOTOROLA) analog MC14093BCP

Ul0 = 1.98V Icc = 0.11mA, Ul1 = 3.21V Icc = 0.11mA



   I am sure that this is the current trigger
Schmitt, instead of an AND gate, NOT, etc.
since an element of the trigger ready to come
logic levels are equal to 0 or voltage
nutrition, and, as always in such cases, through current is 
absent. Of course, when switching element uses a certain 
current, but only a very short time - a time switching, it does 
not even measure a simple way. In the rest of the time (and the 
majority) can be assumed that the element does not take energy 
(in the absence of stress, of course). 


   I measured the current consumption of a trigger (and them as 
inputs, 8), and one can simply multiplying by 8 to calculate 
the maximum current consumed by a chip with no load,

at 5V. But the possible spread of parameters of different 
inputs, and I decided it vyyasnit.Dlya by connecting all the 
inputs together, to the divisor, and repeated the experiment 
for both chips. 

   By switching threshold approached very carefully, t.k.u some 
inputs are could be known to me earlier values.

I watched as the value of current increases and
near the boundary drops sharply back. I write this voltage 
switching and carefully turn the border - another sharp 
decline, off the record. Further movement abroad leads to a 
gradual decrease current consumption.


   Consequently, all the triggers already switched. Move back 
to the other border. All repeats, triggers switch 2 - 3 sharp 
current fall in a narrow range of input voltage. Theory 
multiplication by 8 is essentially confirmed. And imported 
chips was so large (about our) scatter Ul0 inputs, I (arrow 
tester - ammeter) watched all 8 downturns - Switching triggers 
on the inputs. Repeated the experiment several times and 
adjusted the measured values. 



   Maximum current consumption without the chip load for Ucc = 
5V and the voltage on all inputs is equal to: 


     Domestic chip K561TL1

Ul0 = 2.16V Icc = 2.5mA, Ul1 = 3.17V Icc = 3.4mA
Specified limits at which triggers another
pereklyuchilis.A not completely switched
(All) at: Ul0 = 2.13V, Ul1 = 3.19V.



  Import (MOTOROLA) analog MC14093BCP

Ul0 = 2.04V Icc = 0.83mA, Ul1 = 3.21V Icc = 0.88mA
Specified limits at which triggers another
pereklyuchilis.A not completely switched
(All) at: Ul0 = 1.91V, Ul1 = 3.25V.


   According to the results of experiments we can say that
Each input circuit is provided with an independent
Schmitt trigger, the most consuming
energy on the boundary shift (in our
mikruh border Ul1, in this foreign analogues in any border), 
and not counting load on the output circuit, this current

consumption is fundamental. It's a little
denied information from the magazine radio,
incorrectly describes the elements mikruhi.

   Analyzing the switching boundaries can be
say that they (at Ucc = 5V) lie on both
side of half the supply voltage,
and Ul0 little more than a third supply, and Ul1
slightly less than two-thirds power. This indicates an 
inaccuracy in the schedule of the journal (Low value Ul1). 
Likely edited by simply copied the information sent to them 
from one of the readers are not checked. The man could not take 
into account the error measurement introduced by the device 
into a voltage divider (if devices were more simple and less 
sensitive). However, it is likely that produced when

mikruhi differ from the present, which I
doubt.

   However, the difference between reduced chips - is obvious. 
Despite the similar values ​​Ul1, they have markedly different

Ul0, and the asymmetry of power on the borders of our
mikruhi to the symmetry of the current import
indicates a different scheme Schmitt trigger.
The large difference in current may indicate
the difference in chip manufacturing technology (as well as 
letters of BCP in the name of analogue). But perhaps the reason 
the block, the standard the structure of our mikruhi. Her 
flip-flops (they output stage) are calculated on a standard

CMOS load, rather than one entry inside the chip. Therefore, 
these voracious and import trigger recalculated taking into 
account the follow-unit load. 

   Despite the relative engorgement, it is impossible to say 
that our chip huzhe.Ona even more sensitive to the signal 
(voltage Ul0 Ul1 and closer to each other than imports), and a 
small plus to use a mouse. Incidentally, the maximum current 
consumption chipset slightly more than that (since the border 
operation consistent with the very sensitive input, the rest 
were not at the limit, and their power values ​​were below the 
limit), but indeed the maximum currents, probably not

max: 2.6mA (at Ul0), 3.5mA (at Ul1) -
K561TL1; 0.9mA (Ul0 or Ul1) - MC14093BCP.
And with the spread of performance, no more
3 - 4 mA at 1 mA, and our imports have mikruhi. In the 
calculation of universal schemes, without reliable information 
(not a specific, proven mikruhu) I usually double the known 
(measured) value of consumption current.


   I believe that the current import mikruhi on
boundaries are almost equal (this follows from the measured 
values). Despite the difference in 0.05 mA, the threshold for 
triggers Ul0 was with a spread of nearly 0.13 V, which explains 
the difference between the current at Ul0 and Ul1. Disorder, 
clearly permissible, t.k.nebolshoy about 5V, but it's a little 
minus.A here more symmetry currents and voltage limits - A 
small plus. Mikruhu characterizes our little spread on the 
borders - a sign of high-precision performance (but maybe 
technology is not so tiny and so easy to repeat). 

   It remains to mention the multiplicity of current. She
almost equal to 8 - proving the number and independence of the 
triggers. But the import of chips really Ul1 8: 0.11mA * 8 = = 
0.88mA. And Ul0, taking into account the scatter, it so, too. 
Our mikruhi currents slightly higher than calculated ,8-fold, 
znacheniya.Vozmozhno, it is only because of the spread of 
triggers for the current and, possibly, on the border switching 
output voltage triggers significantly different from 0 and 
supply voltage, provoking a low current leakage through the 
gate. The latter circumstance is a small drawback (maybe a 
little reduces the load current gate on the verge of 
switching), and on every case should take this into account.




             The result of the comparison.


   Both mikruhi worthy of his appointment
and can be successfully used in the mouse.
Our chip is slightly more sensitive to input levels and more 
accurate on the Limits response, but voracious current. Import 
chip economical and more symmetrical the levels and current 
consumption, but with a noticeable spread in the level of 
switching logic 0.



                         By KSA-7G 11.2003