Why do digital multimeters not have zero-blanking?

[tooki]

I find the extraneous leading zeroes disturbing, especially on professional-grade equipment. They increase the cognitive load when reading a value on the display, which is not desirable on a technical instrument. Although there are many seemingly plausible hypotheses as to why the practice persists, I don't find them to be all that compelling and they've mostly been debunked in this thread.

Unnecessary leading zeroes seem to me to be more a case of The Pot Roast Principle. Keysight's response when queried about their inclusion of extraneous zeroes tends to support the notion.

Isn't that a sort of generalization of the concept of cargo-cult programming?

Where is a Keysight statement on this matter? I don't see anything in this thread.

[bitseeker]

Yeah, they do sound like related concepts.

Regarding Keysight, I was referring to the response High Voltage received (see Reply #26).

[retrolefty]

People can be so liberal with their hatred of things. I'll save my hatred for more important issues.

[cdev]

Perhaps they feel that zero shows the device is working properly and showing a null value- but no indication of anything indicates the equipment is malfunctioning or its battery is dead-

[PointyOintment]

- To make sure the most significant digits are working, so you won't read 30.000 Vac instead of 230.000 Vac because of a faulty digit.
- To easily approximate the expected range/resolution

On a quick glance 830.000 Vac often looks like 030.000 Vac.

Suggestion: underscores. For example: _30. That way there's no way you can mix up 30, 230, and 830 unless you have a partial digit failure.

This still doesn't solve, e.g., 1130 vs 130 on a 3.5-digit meter, because the half-digit can't display an underscore, but those are at least both somewhat dangerous voltages.

[tooki]

Again, what about my suggestion of dimming the leading zeros? That would show that the digits are working, but make it clearly different at a glance from the mantissa.

And on units without the ability to drive grayscale, how about using a small "o" for leading zeros, so that it looks like o30.000?

[madires]

IMHO underscores or 'o's are as bad as leading zeros. If you want users to see if the display is working add a lamp test. That could be done when powering on or via a lamp test button / menu function. Very simple and effective. UIs should be simple and intuitive, not overloaded with things that bewilder the user. Despite scientists researching UIs some vendors ignore very basic recommendations and totally screw up the UI of their devices.

[StillTrying]

A zero reading should show as 0, 0.0, 0.00, 0.000 etc.

Having to look though a row of leading 0s to check any aren't an '8' or if the DP has moved along them is a
fault/fwbug. 

[tooki]

IMHO underscores or 'o's are as bad as leading zeros. If you want users to see if the display is working add a lamp test. That could be done when powering on or via a lamp test button / menu function. Very simple and effective. UIs should be simple and intuitive, not overloaded with things that bewilder the user. Despite scientists researching UIs some vendors ignore very basic recommendations and totally screw up the UI of their devices.

As others explained, display failure isn't the primary reason for that. It's that leading zeros do give an indication of what range you're in. By leaving them, but reducing them to a visible placeholder, you get the advantages of leading zeros AND the advantages of zero blanking.

[ejeffrey]

You would think in this modern era, someone could put an option in a setup menu somewhere where you can select what you want..

That seems like the worst possible solution.  Useless settings are the worst UI/UX decision that any product can make.  They clutter the configuration menus making it more annoying to change settings that matter and they make the behavior inconsistent with almost no benefit.  Like if we have 3 of the same meter in the lab, and someone converted one of them to use zero blanking, that would be really annoying.  Or, if one of the people on this thread who thinks this is a *MAJOR ISSUE* (which it is categorically not), came by and helpfully turned on zero blanking for a meter I frequently use, that would just lead to extra confusion.

Zero blanking is a minor issue.  I can see some minor benefits to both options, and I might have a slight preference for blanking, but it is absolutely not worth making it a configuration option.  I think the term for what is going on here is not the pot roast principle or cargo cult engineering, but bike shedding.

[rsjsouza]

HPAK does not use leading zeros in all their gear.

You would think in this modern era, someone could put an option in a setup menu somewhere where you can select what you want..

That seems like the worst possible solution.  Useless settings are the worst UI/UX decision that any product can make.  They clutter the configuration menus making it more annoying to change settings that matter and they make the behavior inconsistent with almost no benefit.  Like if we have 3 of the same meter in the lab, and someone converted one of them to use zero blanking, that would be really annoying.  Or, if one of the people on this thread who thinks this is a *MAJOR ISSUE* (which it is categorically not), came by and helpfully turned on zero blanking for a meter I frequently use, that would just lead to extra confusion.

Zero blanking is a minor issue.  I can see some minor benefits to both options, and I might have a slight preference for blanking, but it is absolutely not worth making it a configuration option.  I think the term for what is going on here is not the pot roast principle or cargo cult engineering, but bike shedding.

I disagree as there are good reasons for both approaches. In labs where gear from many manufacturers are stacked, the lack of consistency causes confusion.

[floobydust]

It seems to be only about readability, as it has no effect on precision.
You would not find aircraft instrumentation doing this, although Gerber files can have them (leading zeros) lol.

I need to know ASAP if the voltage measurement is hazardous and after waiting for an autorange, I then have to rummage through extra useless characters, find the decimal dot and if the reading is moving around it's even harder to know WTF voltage is present.

It appears to be part of "good practice" and an omission from the SI. I am still working through them for an answer.
They define everything else, our V, A, ohms etc. electrical units, the spacing and the standards.

Underscores and the negative sign might be confusing/similar i.e. -02.34 vs -_2.34 ?

Since some multimeters are 4,000 or 6,000 or 10,000 count I think it's misleading in that 00.00 could go to 39.99 or 59.99 or 99.99 before moving the decimal place.

[bitseeker]

HPAK does not use leading zeros in all their gear.

Correct and they're not the only ones who refrain from displaying superflous leading zeroes. However, they're still inconsistent across product lines.

Speaking of HP, it's interesting that in some of their vintage gear they had custom versions of common logic ICs for zero blanking.

[rsjsouza]

It seems to be only about readability, as it has no effect on precision.
You would not find aircraft instrumentation doing this, although Gerber files can have them (leading zeros) lol.

Interesting you say that. This high profile aircraft accident was catalyzed by a problem with zeros.

[bitseeker]

Well, more specifically, it was a problem with decimals. Only aircraft equipped with an inertial navigation system used headings with decimals and the flight plan heading of 027.0 was incorrected assumed to be 0270 and, hence, misinterpreted as 270° (west) instead of 27° (north-northeast).

[rsjsouza]

Yes, but still a readability issue where zeros were involved...   

[David Hess]

Historic reasons:  Back in the days of dumb SSI and even discrete logic, leading zero blanking cost gates and thus money.   It was compounded by range switching that simply had an auxiliary contact to select one of the decimal points to feed power to to light it. The logic typically didn't even 'know' which d.p. was lit, so it would have required a lot of extra circuitry to blank all except the zero in front of the active d.p.

The Fairchild 3814 4.5 digit digital voltmeter IC from 1975 (earlier?) supported optional leading zero blanking but the contemporary Siliconix converters did not and I do not see it included in the modern Intersil 7106 implementations.  Maxim's single chip voltmeters include optional leading zero suppression though.

If Fairchild implemented it in the 3814, the cost is trivial except in requiring another pin to select it.  Since Intersil did not implement it, there must have been no demand.

Huh, I just found my Fairchild 4000 series CMOS databook.  I remember ignoring it as exotic.

[madires]

As others explained, display failure isn't the primary reason for that. It's that leading zeros do give an indication of what range you're in. By leaving them, but reducing them to a visible placeholder, you get the advantages of leading zeros AND the advantages of zero blanking.

I think the main goal should be to display the measurement in a unambiguous way without any redundant leading zeros or strange symbols. Those cause more harm than good, because the measurement value can be misread easily. If a range indication is needed then add a dedicated one. Presumbaly this whole mess was created by minimizing BOM cost, since misusing the digits and drivers already available as a range indication is cheap, no additional segments and drivers are needed. But there's no excuse for graphical displays besides ignorance.

[Ian.M]

Another reason would be segmented displays with + and - annunciators to the left of the first digit (and units annunciators to the right of the last digit).   Displaying:
   +0001.6^mV
would be vastly preferable to displaying:
   +   1.6^mV
However there is no excuse on modern MCU controlled bench meters with dot matrix displays and persistent user settings not to make leading zero blanking user configurable.

[StillTrying]

Another reason would be segmented displays with + and - annunciators to the left of the first digit (and units annunciators to the right of the last digit).   Displaying:
   +0001.6^mV
would be vastly preferable to displaying:
   +   1.6^mV

You must be kidding.

[Neomys Sapiens]

"Computerized" displays lead to all kinds of stupidity in terms of displaying measured values. Worst to see are the ones that treat everything as a float and display e.g. a room temperature reading with 5 digits after the decimal point - no joke, you can see such stupidity quite often in PC based software written by coders that do not know the usage of resolution and accuracy.

Especially stultefying when you try to explain this and the designer explains to you that those microkelvins do really exist, because it is analogue!

BTW.
Though beeing an European and beeing taught "," as the decimal separator, I do consider "." as _the only valid_ decimal seperator in science and technical issues, so any scientific or technical application displaying a "," gets cursed by me.

Concur totally. As far as I know, most German engineers do. This does not hold true to our austrian neighbours. Austrian company took ours over and changed all document formats to use a comma instead of a DP.

[tooki]

As others explained, display failure isn't the primary reason for that. It's that leading zeros do give an indication of what range you're in. By leaving them, but reducing them to a visible placeholder, you get the advantages of leading zeros AND the advantages of zero blanking.

I think the main goal should be to display the measurement in a unambiguous way without any redundant leading zeros or strange symbols. Those cause more harm than good, because the measurement value can be misread easily. If a range indication is needed then add a dedicated one. Presumbaly this whole mess was created by minimizing BOM cost, since misusing the digits and drivers already available as a range indication is cheap, no additional segments and drivers are needed. But there's no excuse for graphical displays besides ignorance.

I don't think you understand the value of visual placeholders. They reduce cognitive load, not increase it, because they let you know exactly what you're not seeing, which is often easier than seeing. (Do you drive a car? If so, then you'll know how the purpose of glancing in your mirrors isn't actually to see what's there, but rather to verify what isn't there: another car.)

Another example of how placeholders reduce cognitive load: dimmed commands in software menus. Dimming ("graying out") commands that are unavailable, as opposed to actually removing them, helps users a) learn what commands are available, even when not available right now, b) help you understand the relationships between the circumstances that make a command available or not, (and you don't go around looking for commands that are unavailable but invisible because they were removed) and c) arguably most importantly, they keep commands from moving around. That last one reduces cognitive load, because you know that for example, that copy, cut, and paste always come in that order, so if you need paste, it'll always be third, even if cut and copy are currently dimmed.

Regardless, my entire point is that using something that's neither a space nor a zero has massive advantages.

[Naguissa]

"Computerized" displays lead to all kinds of stupidity in terms of displaying measured values. Worst to see are the ones that treat everything as a float and display e.g. a room temperature reading with 5 digits after the decimal point - no joke, you can see such stupidity quite often in PC based software written by coders that do not know the usage of resolution and accuracy.

Especially stultefying when you try to explain this and the designer explains to you that those microkelvins do really exist, because it is analogue!

BTW.
Though beeing an European and beeing taught "," as the decimal separator, I do consider "." as _the only valid_ decimal seperator in science and technical issues, so any scientific or technical application displaying a "," gets cursed by me.

Concur totally. As far as I know, most German engineers do. This does not hold true to our austrian neighbours. Austrian company took ours over and changed all document formats to use a comma instead of a DP.

I prefer ",", as it's 'bigger' and can be seen easier even for people as short-sighted as me.... :-)

Enviado desde mi Jolla mediante Tapatalk

[tooki]

Another reason would be segmented displays with + and - annunciators to the left of the first digit (and units annunciators to the right of the last digit).   Displaying:
   +0001.6^mV
would be vastly preferable to displaying:
   +   1.6^mV

Wut? The second is exactly what I'd want to see. That or +0001.6 mV.

Or rather still, 0001.6 mV for positive values and –0001.6 mV or –1.6 mV for negative ones. No meter I have ever used shows an annunciatior for positive values, only for negative. No need to clutter the screen with them for positives.

For example, a Fluke 87V would display those as 1.6mV and –1.6mV (it floats the minus sign along with the value, one leading zero only, e.g. 0.5mV).

[David Hess]

No meter I have ever used shows an annunciation for positive values, only for negative. No need to clutter the screen with them for positives.

Displaying both + and - for DC measurements distinguishes them from AC measurements which have no polarity.  If I see 1.23 volts, is that positive DC or AC?