Author Topic: The art of logic signal manipulation with analogue components (D/R/C)  (Read 6034 times)

0 Members and 1 Guest are viewing this topic.

Offline max.wwwangTopic starter

  • Frequent Contributor
  • **
  • Posts: 493
  • Country: nz
Re: The art of logic signal manipulation with analogue components (D/R/C)
« Reply #50 on: March 13, 2023, 04:35:11 am »
See the schematic https://www.alldatasheet.com/datasheet-pdf/pdf/50860/FAIRCHILD/CD4069.html


The input has protection diodes and there is that external series 220k resistor. The FETs experiences only small voltages beyond the rails and the currents in the diodes are acceptably low.
Appreciate your input as always.

Yes, its input may be protected, just like the one you are referring to (Fairchild). And I do know there is a 220k resistance. But if its input impedance is high enough (this is a reasonable expectation because the input current is very low), the 220k resistor will not make any difference in 'shielding' the gate from too low input voltage.

The problem is, this -0.5V is absolute maximum rating, not just a recommendation. And absolute is absolute (included is the Fairchild version; the same thing: -0.5V).

Perhaps I should read "absolute" in a different way? :palm:


Neutral | grounded
 

Offline pcprogrammer

  • Super Contributor
  • ***
  • Posts: 3631
  • Country: nl
Re: The art of logic signal manipulation with analogue components (D/R/C)
« Reply #51 on: March 13, 2023, 07:19:20 am »
It is not about the normal input impedance in this case. The protection diode has a very low impedance for negative voltages, but won't be able to sink a lot of current. Providing a negative voltage from a source with low impedance will allow to much current and this will kill the input protection diode and worst case the whole chip.

The 220K resistor gives an assurance the current won't be too high. Another limiting factor in this circuit is the 10K resistor with the diode in series, which will speed up the discharge rate of the negative pulse.

Also the datasheet states:

Quote
Note  1:  “Absolute  Maximum  Ratings”  are  those  values  beyond  which  the safety  of  the  device  cannot  be  guaranteed.  They  are  not  meant  to  imply that  the  devices  should  be  operated  at  these  limits.  The  table  of  “Recommended Operating Conditions” and Electrical Characteristics table provide conditions for actual device operation.

The safety cannot be guaranteed, but it does not mean that it will instantly die when these conditions are met.

Online wasedadoc

  • Super Contributor
  • ***
  • Posts: 1309
  • Country: gb
Re: The art of logic signal manipulation with analogue components (D/R/C)
« Reply #52 on: March 13, 2023, 10:42:07 am »
The gate input also an internal resistor before the diodes though that is likely to be relatively low value (for speed reason) compared to the external 220k.

If you have a 'scope look at that pin and see what negative voltages that pin experiences.
 
The following users thanked this post: max.wwwang

Offline max.wwwangTopic starter

  • Frequent Contributor
  • **
  • Posts: 493
  • Country: nz
Re: The art of logic signal manipulation with analogue components (D/R/C)
« Reply #53 on: March 18, 2023, 08:09:47 am »
Ah my assumption is only half correct. If the input voltage is positive, the input impedance is indeed high because of the MOS. However when the input is negative, there will be current through the highlighted path. In this case, although the voltage at the other side of the 220k resistor will be momentarily be as low as -11V, because of this current, the voltage on the input pin of the gate will be much higher than -11V.

I'm not interested in why the gate is not broken -- because of the variation of the specs, one component may well survive an input voltage that goes beyond the recommended or even absolute limits. What I'm interested is, I don't think the designer of this circuit will let it be exposed to an input voltage under normal working condition that goes beyond its absolute maximum based on wishful thinking "I think she'll be alright".

The remaining problem is -- with a correction of my previous assumption --- when the far side of the 220k resistor sees a -11V signal, the voltage on the input pin of the gate may still be higher than the absolute maximum -0.5m (because of the forward bias voltage of the protection diode, and the voltage drop over the internal resistor). This might be narrowly avoided if the voltage drop of the diode is 0.5V (I think there are some diodes the forward bias voltage drop of which is less than 0.6V) or even less than 0.5V, and the internal resistor is so small that its voltage drop is negligible (that's very likely).
Neutral | grounded
 

Online wasedadoc

  • Super Contributor
  • ***
  • Posts: 1309
  • Country: gb
Re: The art of logic signal manipulation with analogue components (D/R/C)
« Reply #54 on: March 18, 2023, 09:08:32 am »
11 volts and 220k is only 50uA. Don't need much forward voltage on a silicon diode to conduct that.
 
The following users thanked this post: max.wwwang

Offline pcprogrammer

  • Super Contributor
  • ***
  • Posts: 3631
  • Country: nl
Re: The art of logic signal manipulation with analogue components (D/R/C)
« Reply #55 on: March 18, 2023, 09:46:31 am »
Ever heard of schottky diodes?

These have lower forward voltages than a normal silicon diode. Germanium diodes also have lower forward voltages. (~0.3V)

Online wasedadoc

  • Super Contributor
  • ***
  • Posts: 1309
  • Country: gb
Re: The art of logic signal manipulation with analogue components (D/R/C)
« Reply #56 on: March 18, 2023, 10:36:47 am »
They are not drawn as Schottky diodes.

It isn't the voltage per se that would cause damage. The FET gates are not susceptible to modest excursions beyond the power rails. It is excessive current through the diodes that could damage them. The spec is saying that up to -0.5 volts at input will be low enough current. The equipment designer has used the external resistor to limit the current.
 

Offline pcprogrammer

  • Super Contributor
  • ***
  • Posts: 3631
  • Country: nl
Re: The art of logic signal manipulation with analogue components (D/R/C)
« Reply #57 on: March 18, 2023, 11:10:45 am »
They are not drawn as Schottky diodes.

I know. Was not suggesting that these are used in this IC, but merely providing information about there being diodes with lower forward voltages in a response to the below.

(I think there are some diodes the forward bias voltage drop of which is less than 0.6V)

Offline max.wwwangTopic starter

  • Frequent Contributor
  • **
  • Posts: 493
  • Country: nz
Re: The art of logic signal manipulation with analogue components (D/R/C)
« Reply #58 on: March 20, 2023, 05:23:10 am »
11 volts and 220k is only 50uA. Don't need much forward voltage on a silicon diode to conduct that.
That makes sense.
Neutral | grounded
 

Offline max.wwwangTopic starter

  • Frequent Contributor
  • **
  • Posts: 493
  • Country: nz
Re: The art of logic signal manipulation with analogue components (D/R/C)
« Reply #59 on: March 20, 2023, 08:43:51 am »
Yet another example.

From what I can see, the RHS of the circuit normally remains H, but will give a constant width pulse to L on the rising edge of the LHS (but not for falling edges). The problem is, on the falling edge of the LHS, the input pins of U29A will see a negative input as low as -11V, far below the absolute maximum of TC4071BP (which is, again, -0.5V). In this case, there is no buffering resistor between the capacitor and the input pins.

Again, my assumption is that the designer of the circuit should know with certainty that the voltage on the input pins of U29A will not go beyond the absolute maximum rating under normal working condition.
Neutral | grounded
 

Offline pcprogrammer

  • Super Contributor
  • ***
  • Posts: 3631
  • Country: nl
Re: The art of logic signal manipulation with analogue components (D/R/C)
« Reply #60 on: March 20, 2023, 09:29:52 am »
Did you measure it with a scope?

The input diode will still try to keep the voltage within limits. You can dive into the amount of charge that can flow through the diode in the short time it takes to discharge. It is the amount of energy stored in the capacitor. If the capacitor is small enough it won't harm the chip.

I myself would have added an external schottky diode to makes sure the input is protected.


Offline max.wwwangTopic starter

  • Frequent Contributor
  • **
  • Posts: 493
  • Country: nz
Re: The art of logic signal manipulation with analogue components (D/R/C)
« Reply #61 on: March 21, 2023, 07:09:52 am »
No I didn't. It's not in a condition that it can run the way it should. I've ordered faulty part and am waiting for it to arrive and hopefully it'll then be fixed, so I can probe all these things and verify all the analyses and theories.

I agree that probably there are also protection diodes inside, just like the case above, though I didn't spot them on the datasheet that I looked at. If that's the case, your theory of diode limiting voltage seems to make sense. The problem I have is, if this is the correct explanation, the datasheet does not need to state -0.5V as the absolute maximum rating for input voltage. It's a -11V voltage that has been (or will be) 'thrown' to the input directly without anything external in between as a buffer or whatever.

Don't get me wrong, by no means am I aiming at refuting your theory. I just find it difficult for all of these to reconcile. I've carefully checked the board; the schematic reflects the circuit correctly.

By the way, here attached a revised version with the value of the capacitor.

[Edit --- the value of the capacitor should read "470pF" not "471pF".]
« Last Edit: March 24, 2023, 08:20:06 am by max.wwwang »
Neutral | grounded
 

Online wasedadoc

  • Super Contributor
  • ***
  • Posts: 1309
  • Country: gb
Re: The art of logic signal manipulation with analogue components (D/R/C)
« Reply #62 on: March 21, 2023, 12:11:11 pm »
1.  Look at page 3 of https://www.ti.com/lit/ds/symlink/cd4071b.pdf to see 4 diodes in the input protection circuit of the TI version.  The max ratings section at the top of the page also shows "DC input current, any one input +/- 10 mA".

2.  On a capacitor the 3rd digit is usually a power of 10 multiplier.  So 471 denotes 470 pF.  In this case negligible difference from 471pF.  But 472 is 4700pf (4.7nF), 473 is 47000pF (47nF).
« Last Edit: March 21, 2023, 12:17:44 pm by wasedadoc »
 
The following users thanked this post: max.wwwang

Offline max.wwwangTopic starter

  • Frequent Contributor
  • **
  • Posts: 493
  • Country: nz
Re: The art of logic signal manipulation with analogue components (D/R/C)
« Reply #63 on: March 23, 2023, 06:39:34 am »
1.  Look at page 3 of https://www.ti.com/lit/ds/symlink/cd4071b.pdf to see 4 diodes in the input protection circuit of the TI version.  The max ratings section at the top of the page also shows "DC input current, any one input +/- 10 mA".

Good to have datasheet that does show this. But in this one, it's also stated as -0.5V for the absolute maximum rating for input voltage.

2.  On a capacitor the 3rd digit is usually a power of 10 multiplier.  So 471 denotes 470 pF.  In this case negligible difference from 471pF.  But 472 is 4700pf (4.7nF), 473 is 47000pF (47nF).
Thanks for mentioning this. I did wonder what the marking "471" means. It's almost identical to this guy ---

and I looked around before marking it as 471pF. There are many webpages taking this as 471pF, like this one. So this webpage is wrong?
I don't like to take it as 471pF if it means 470 despite the negligible difference. If 470 is correct, then 471 is simply wrong.
« Last Edit: March 23, 2023, 06:42:37 am by max.wwwang »
Neutral | grounded
 

Offline pcprogrammer

  • Super Contributor
  • ***
  • Posts: 3631
  • Country: nl
Re: The art of logic signal manipulation with analogue components (D/R/C)
« Reply #64 on: March 23, 2023, 10:18:11 am »
So this webpage is wrong?

Yep, they got it wrong, despite the fact that they got 102, 103 and 104 mentioned with the correct values on the main capacitor page.  :palm:

I use this site mostly to lookup the values of SMD resistors, but they do capacitors too.

Offline TimFox

  • Super Contributor
  • ***
  • Posts: 7934
  • Country: us
  • Retired, now restoring antique test equipment
Re: The art of logic signal manipulation with analogue components (D/R/C)
« Reply #65 on: March 23, 2023, 08:00:58 pm »
I was curious about that ignorant website that thought "471" meant 471 pF instead of 470 pF, but my malware filter blocked it.
 

Offline max.wwwangTopic starter

  • Frequent Contributor
  • **
  • Posts: 493
  • Country: nz
Re: The art of logic signal manipulation with analogue components (D/R/C)
« Reply #66 on: March 24, 2023, 08:16:34 am »
I was curious about that ignorant website that thought "471" meant 471 pF instead of 470 pF, but my malware filter blocked it.
That's a testimony of the intelligence of your malware filter.
Neutral | grounded
 

Offline max.wwwangTopic starter

  • Frequent Contributor
  • **
  • Posts: 493
  • Country: nz
Re: The art of logic signal manipulation with analogue components (D/R/C)
« Reply #67 on: March 24, 2023, 08:17:54 am »
Yep, they got it wrong, despite the fact that they got 102, 103 and 104 mentioned with the correct values on the main capacitor page.  :palm:
Thanks for confirmation. I didn't notice they got 102 and 103 right.  :palm:
« Last Edit: March 24, 2023, 08:20:31 am by max.wwwang »
Neutral | grounded
 

Offline max.wwwangTopic starter

  • Frequent Contributor
  • **
  • Posts: 493
  • Country: nz
Re: The art of logic signal manipulation with analogue components (D/R/C)
« Reply #68 on: March 25, 2023, 12:32:06 am »
Another interesting one, similar to the previous one but only a little different. This time I want to be more quantitative.

The same question of the input voltage momentarily being -11V, far beyond the maximum rating -0.5V, also exist, if from the perspective of the U24D gate (TC4071BP). This may well be 'clamped' down to safe voltage by the internal protection diodes. But even if this is the case, to me this is the internal business of the gate but does not change the fact that its input has been exposed to an unsafe input voltage (according to the datasheet). We can park this one (because it has been discussed, and I might be too pedantic here.)

This circuit, as far as I see, works this way. The output (RHS) remains L (i.e. active H), but rises to H whenever any of the inputs (LHS) rises to H. Dur to the existence of the R/C, any rising of the three highlighted inputs will be extended from its rising edge with a constant duration (determined by RC, which is given, the L/H threshold of the gate, and probably its propagation delay).

That is, any of these three input signals, whenever it comes and however short the pulse might be, will have at least the duration of this extension on the output side. [1]

To me this is meaningful only when the width of any of these three signals may be narrower than this extension. Because otherwise this extension will only be 'buried' (as overlap) in the input pulse.

Before figuring this out, here comes a question first about the input threshold voltage. According to the datasheet, say for Vdd 9V, these two voltages thresholds , which are different (highlighted as 5.5V and 4.5V, respectively). This looks like to me that there will be a voltage range where the logic state of the input is not determined. This will inevitably happen when the input is fed with the voltage that complies with the exponential curve of an RC charging/discharging circuit. What's going to happen when the input voltage is between 4.5V and 5.5V?

[Edit]
[1] With another look, this now does not seem to be correct.
« Last Edit: March 25, 2023, 04:56:53 am by max.wwwang »
Neutral | grounded
 

Online wasedadoc

  • Super Contributor
  • ***
  • Posts: 1309
  • Country: gb
Re: The art of logic signal manipulation with analogue components (D/R/C)
« Reply #69 on: March 25, 2023, 12:45:27 am »
You may be misinterpreting the datasheet but I need to see exactly how the table is labelled.  Please post a link to the original complete datasheet.
 

Offline max.wwwangTopic starter

  • Frequent Contributor
  • **
  • Posts: 493
  • Country: nz
Re: The art of logic signal manipulation with analogue components (D/R/C)
« Reply #70 on: March 25, 2023, 01:10:15 am »
Here it is—
https://pdf1.alldatasheet.com/datasheet-pdf/view/31643/TOSHIBA/TC4071BP.html

Please do let me know if it is a misinterpretation of the datasheet – that is very likely. Thanks.
« Last Edit: March 25, 2023, 04:43:34 am by max.wwwang »
Neutral | grounded
 

Offline pcprogrammer

  • Super Contributor
  • ***
  • Posts: 3631
  • Country: nl
Re: The art of logic signal manipulation with analogue components (D/R/C)
« Reply #71 on: March 25, 2023, 07:30:32 am »
The way I read it is that worst case the input voltage seen as a low is 3V but typical it will be 4.5V. This means the input has to go below this voltage to be treated as low. For it to be seen as high the worst case is as high as 7V, but typical will be 5.5V. So to be seen as high the input has to go above this voltage.

I interpret this as the gate having some hysteresis. The output will only change from low to high if one or both inputs rise above the given voltage, and will only go low again when both inputs drop below the other given voltage.

Offline max.wwwangTopic starter

  • Frequent Contributor
  • **
  • Posts: 493
  • Country: nz
Re: The art of logic signal manipulation with analogue components (D/R/C)
« Reply #72 on: March 25, 2023, 07:51:11 am »
The way I read it is that worst case the input voltage seen as a low is 3V but typical it will be 4.5V. This means the input has to go below this voltage to be treated as low. For it to be seen as high the worst case is as high as 7V, but typical will be 5.5V. So to be seen as high the input has to go above this voltage.
I read this differently.

It is fairly clear that there are three groups of columns for three different temperatures. Under any given temperature, say 25 degrees celsius, there is only one group (of columns) that applies.

In this particular group, there are two or three of the specs specified, which are "min", "typ", and "max". My interpretation is that these are for the variation between individual chips, not the range of drifting of any single chip. In other words, for any particular chip, the value for one parameter, such as input low voltage, is a single value, which is typically "typ", but may vary within the range between "min" and "max".

I interpret this as the gate having some hysteresis. The output will only change from low to high if one or both inputs rise above the given voltage, and will only go low again when both inputs drop below the other given voltage.
With the above being said, I do suspect that the discrepancy between the low and high input voltages for any particular chip, under any given temperature -- say 4.5V vs 5.5V -- indeed may mean some kind of hysteresis characteristic. But I'm not very sure.
« Last Edit: March 25, 2023, 07:55:31 am by max.wwwang »
Neutral | grounded
 

Offline pcprogrammer

  • Super Contributor
  • ***
  • Posts: 3631
  • Country: nl
Re: The art of logic signal manipulation with analogue components (D/R/C)
« Reply #73 on: March 25, 2023, 12:05:39 pm »
I read this differently.

It is fairly clear that there are three groups of columns for three different temperatures. Under any given temperature, say 25 degrees celsius, there is only one group (of columns) that applies.

In this particular group, there are two or three of the specs specified, which are "min", "typ", and "max". My interpretation is that these are for the variation between individual chips, not the range of drifting of any single chip. In other words, for any particular chip, the value for one parameter, such as input low voltage, is a single value, which is typically "typ", but may vary within the range between "min" and "max".

That is what worst case often means, the spread across a large set of the same components. When designing a circuit you have to take this into account to make sure all devices work the same coming of the production line.

But in this case only the extremes are specified for the two outer temperatures. Typical is only given for 25 degrees. This can mean that this typical value varies over temperature. Will it vary to the given extremes, probably not, but I would not count on them being fixed on the typical value over a wide temperature range.

With the above being said, I do suspect that the discrepancy between the low and high input voltages for any particular chip, under any given temperature -- say 4.5V vs 5.5V -- indeed may mean some kind of hysteresis characteristic. But I'm not very sure.

It is not specifically mentioned in the datasheet, but if it would not have some hysteresis, you can get very weird behavior due to signal noise, where the output could flip many times before getting stable.

Online wasedadoc

  • Super Contributor
  • ***
  • Posts: 1309
  • Country: gb
Re: The art of logic signal manipulation with analogue components (D/R/C)
« Reply #74 on: March 25, 2023, 06:51:18 pm »
With 10 volt supply and at any of the 3 temperatures an input of 7.0 volts or more is guaranteed to be treated as a HIGH by the input.  Additional info is that at 25 degrees C a typical chip will treat 5.5 or more as a HIGH input.

Similarly for any of the 3 temperatures an input of 3 volts or less is guaranteed to be treated as a LOW by the input.  Additional info is that at 25 degrees C a typical chip will treat 4.5 or below as a LOW input.

Nothing in the specs says what will certainly happen for input voltages between 3.0 and 7.0.  Any voltage up to but not including 7.0 could be treated as a LOW by an in spec chip.  Any voltage down to but not including 3.0 could be treated as a HIGH by an in spec chip.

Nothing in the specs says anything about hysteresis ie any difference in the changeover voltage when input going LOW to HIGH compared to input going HIGH to LOW.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf