# Length of time needed to measure digital clock accuracy

On a piece of equipment, I purchased new, the digital clock ran slow. The second digital clock that was installed ran slow---it lost approximately 4.1 minutes in one month. Below is what I sent the Service Manager who I understand thinks it is possible to test the newly installed clock immediately after installing it with a cell phone (the NH Rep thinks it's possible also, I understand. (I did not measure the time initially with the precision indicated below but just included all the digits displayed by the Windows computer calculator and actually, if you round up, the clock loses approximately 27.8 ms per hour.) Even if a person had a super precise stopwatch, there would be a time lag between seeing the digit or digits change on the clock being timed and pushing the stopwatch button. So I suggest timing a newly installed clock for a period of preferably approximately two weeks using a digital wristwatch to get an accurate assessment of the clock accuracy. What are your comments?

Sent to New Holland: "You mentioned something about possibly checking the accuracy of the clock with a stopwatch while up here. It’s not possible to check the accuracy of the clock right away after installing the instrument panel, like I mentioned since the current clock loses a small but cumulative amount and it is impossible to accurately time such a small interval of time with a stopwatch. In my email dated February 7, 2018, I write, “Difference between clock and watch in three weeks was 14 seconds or 14 / 21 = 0.6666666666666667 loss in seconds per day.” So, in one hour, it loses 0.6666666666666667 / 24 = 0.0277777777777778 seconds. That is 27.7 milliseconds in an hour. I don’t know how you could time it so accurately, since just pushing the stopwatch button has some variability and you would have to wait one hour at least and if you were off just a few thousandths of a second, that would greatly affect the accuracy of your results."

Thanks

davenn
Gold Member
On a piece of equipment, I purchased new, the digital clock ran slow

Hi, Welcome to PF

1) make and Model ?
2) links to info/datasheets for the unit
3) what are you using this for ?
4) what accuracy do you need or expect ?

If the circuitry of the unit can be accessed and it isn't to complex, then it may be possible to use a higher accuracy
oscillator that is say GPS or Rubidium Standard controlled/locked.
This is what I do for my seismic recording system and some of my higher microwave transceiver gear

Dave

russ_watters
Mentor
Simply put, the accuracy of the test is a function of the length of the measurement -- as long as the accuracy of the clock is consistent.

jim hardy
Gold Member
Dearly Missed
I wouldn't trust a digital watch. It might run fast. Get some sort of standard that's connected to NBS.
In my day we calibrated our plant computer's time of day using the WWV 10 mhz time standard radio broadcast from Colorado. We included the ~10msec propagation from Boulder OOPS Ft Collins to Miami.
Nowadays there are wall clocks that keep themselves calibrated to WWV - about twenty bucks at Walmart.
https://www.lacrossetechnology.com/media/catalog/product/w/t/wt-8005u.pdf

Last edited:
Simply put, the accuracy of the test is a function of the length of the measurement -- as long as the accuracy of the clock is consistent.

You are correct. I am requesting a comment to the information supplied as to how long would you suggest to wait in this case using the timing method as described and timing checking devices such as a wristwatch or cell phone which does not display time in increments smaller than one second? So far, no one has commented on my suggestion such as waiting two weeks and that the New Holland people thought they could immediately check to see if the newly installed clock was accurate by waiting a minute or two using the timing equipment mentioned.

jim hardy
Gold Member
Dearly Missed
With 86,400 seconds in a day , and once per second update,
that's the best you can resolve your drift in a day.
To what precision do you want to measure ? One part in 86,400 will take a day .

If you could reach inside that clock and measure its oscillator you could get a faster measurement using a sophisticated frequency counter.

The New Holland people either weren't thinking, or they know of some trick like using a frequency counter on an internal test point.
Chances are there's a test point inside the clock for just that purpose. The factory had to check it when they built it and i doubt they waited two weeks.

If you could measure the time between one second updates with something electronic and precise you could get a quick answer by eliminating the variable "time to trigger by hand"

old jim

davenn
anorlunda
Staff Emeritus
In your OP, you told what you said to New Holland, but not what they said.

The accuracy of a digital clock depends on the frequency of a crystal oscillator. Using laboratory instruments, one can measure that frequency very accurately without waiting a lot of time. At the New Holland factory, I presume that's how they do it.

But a wristwatch is merely another clock with unknown accuracy. You can never establish accuracy by comparing clocks, you need an external source such as WWV for GPS. We used to tell the story, "When I had a watch, I always knew what time is is. But now I own two watches and I never know what time it is."

By the way, @jim hardy , I always though that my watch bowed toward Ft. Collins, CO for WWV, not Boulder. I also remember being out at sea, totally out of communication with everyone (not really ). I wondered if the Apocalypse happened, how would I know? The answer was to tune the SSB radio to WWV and I could hear that Colorado was still there.

jim hardy
Gold Member
Dearly Missed
By the way, @jim hardy , I always though that my watch bowed toward Ft. Collins, CO for WWV,
You're exactly right... Sorry !

btw i've always found them pretty darn close.

russ_watters
Mentor
You are correct. I am requesting a comment to the information supplied as to how long would you suggest to wait in this case using the timing method as described and timing checking devices such as a wristwatch or cell phone which does not display time in increments smaller than one second? So far, no one has commented on my suggestion such as waiting two weeks and that the New Holland people thought they could immediately check to see if the newly installed clock was accurate by waiting a minute or two using the timing equipment mentioned.
Well, the second part is an easy answer: you are correct that you can't tell anything in a minute or two.

For the first part, I think you haven't gotten a clear-cut answer because you haven't given clear-cut acceptance criteira or precision in the comparison test. But I'll give an example:

If your clock rate is consistent and you require an accuracy of 1 minute per month and you are capable of discerning a 0.1 second deviation with your test, then a failure will show up in 1/600th of a month or about 1.2 hours.

For a better signal to noise ratio you may want to multiply that by 10 (12 hours) to get a full digit of precision and if you aren't certain the deviation rate is consistent you may want to gather, say, 10 data points.

You may also consider putting this back on the manufacturer: if they they they can measure the error in a few minutes, let them -- if they are willing to provide a written guarantee.

anorlunda
Staff Emeritus
Don't forget that crystal based digital clocks are also sensitive to temperature and atmospheric pressure, and possibly other factors.

Do you have a specified accuracy requirement for your clocks? Has the vendor agreed to meet that requirement?

russ_watters
jim hardy
Gold Member
Dearly Missed
We used to tell the story, "When I had a watch, I always knew what time is is. But now I own two watches and I never know what time it is."

You know of the old timekeeper at the factory in a small Midwest town . He blew the factory's lunch whistle every day precisely at noon by his 23 jewel Hamilton pocket watch.
Every morning on his walk to work he set that fine watch by an equally fine German made grandfather clock in the watchmaker's store window.
One morning he stopped and exchanged pleasantries with the shopkeeper. He remarked on the beauty of the grandfather clock. The shopkeeper replied "Yes and it keeps perfect time, too. I check it every day against the factory's noon whistle."

old jim

Asymptotic, NTL2009, russ_watters and 5 others
f95toli
Gold Member
The accuracy of a digital clock depends on the frequency of a crystal oscillator. Using laboratory instruments, one can measure that frequency very accurately without waiting a lot of time. At the New Holland factory, I presume that's how they do it.
True, but that is somewhat irrelevant. A real clock does not behave in a way that makes it possible to characterize it in a short amount of time; the reason is that clocks (or oscillators in general) drift due to external factors and also suffer from e.g. slow 1/f noise which does not average out.

If you want to fully characterize a given oscillator you need to plot what is known as the Allan Deviation (or if you prefer Allan Variance); just google "ADEV oscillator" for some examples. You will see that there is -typically- a white noise region for short time but then the ADEV becomes flat (1/f) and can then go up again (e.g drift). Measuring for a very short time is equivalent to just extrapolating the slope of the white noise to long times; i.e. it is simply incorrect.

In "real" timekeeping we use multiple systems: typically a system with good short term stability (rubidium standards or even better a hydrogen maser) is being steered by the GPS signal (or your own CS fountain if you have one).

Hence, "clock accuracy" isn't a single number; it is a graph. You need to decide over what timescale you need a certain accuracy and how often you are happy to calibrate the clock.

(I've spent a LOT of time over the past few years potting ADEVs....)

With 86,400 seconds in a day , and once per second update,
that's the best you can resolve your drift in a day.
To what precision do you want to measure ? One part in 86,400 will take a day .

If you could reach inside that clock and measure its oscillator you could get a faster measurement using a sophisticated frequency counter.

The New Holland people either weren't thinking, or they know of some trick like using a frequency counter on an internal test point.
Chances are there's a test point inside the clock for just that purpose. The factory had to check it when they built it and i doubt they waited two weeks.

If you could measure the time between one second updates with something electronic and precise you could get a quick answer by eliminating the variable "time to trigger by hand"

old jim

The New Holland people were not using any test point or a frequency counter in the field on the tractor after the instrument panel was installed which has the clock built into it but only looking at the display. These people do not appear to have a science background and I am probably quite correct in saying that they have never used a frequency counter or an oscilloscope. I am sure that the "clock" frequency produced by the quartz crystal could be measured in the factory after its assembled but manufacturers don't test everything they make---that would cost too much money. They only test a few units out of a batch and if they test okay, then all of the rest are supposedly okay. Instead they let the end user be part of their Quality Control department, which I don't think is very nice!

russ_watters
anorlunda
Staff Emeritus
In "real" timekeeping we use multiple systems: typically a system with good short term stability (rubidium standards or even better a hydrogen maser) is being steered by the GPS signal (or your own CS fountain if you have one).

Hence, "clock accuracy" isn't a single number; it is a graph. You need to decide over what timescale you need a certain accuracy and how often you are happy to calibrate the clock.

(I've spent a LOT of time over the past few years potting ADEVs....)

They only test a few units out of a batch and if they test okay, then all of the rest are supposedly okay. Instead they let the end user be part of their Quality Control department, which I don't think is very nice!

We are loosing sight of the context here regarding what's real and not real. It sounds like the clock in question is a panel clock on a tractor. It probably cost only $1 from the manufacturer. How much quality and quality-control do you really expect for that price? How much of @f96toli's expertise do you expect from a tractor mechanic? The pragmatic remedy is to buy a digital watch (or a GPS) and glue it to the tractor panel. NTL2009, davenn and Averagesupernova We are loosing sight of the context here regarding what's real and not real. It sounds like the clock in question is a panel clock on a tractor. It probably cost only$1 from the manufacturer. How much quality and quality-control do you really expect for that price? How much of @f96toli's expertise do you expect from a tractor mechanic?

The pragmatic remedy is to buy a digital watch (or a GPS) and glue it to the tractor panel.

This is getting off course and doesn't answer my issue, including purchasing something, which I should not have to do. The clock is defiantly worth more than $1 and is part of a costly instrument panel so your assumption is incorrect and is on an expensive tractor and I don't know about you but I expect everything to work properly. The clock on the tractor does not perform nearly as well as an old, maybe 30 year old Seth Thomas clock battery powered wall clock that I have. As for a tractor mechanics and the Reps expertise, they should be properly trained to resolve issues on the equipment they service but manufacturers obviously don't do that, like in this case. russ_watters Mentor I'm a little surprised a modern tractor doesn't have a GPS linked clock like a car. This would annoy the hell out of me too. I really hate simple, stupid problems that I could work-around easily but shouldn't have to. anorlunda Staff Emeritus Below is a New Holland dashboard with a clock showing 4:16. Does your's look like that? #### Attachments • slask.png 73.2 KB · Views: 741 I understand more expensive tractors have GPS systems so then there hopefully wouldn't be a clock accuracy issue. It is annoying since I like things to work as expected, especially on an expensive tractor. My very inexpensive battery operated radio controlled clock on my nightstand keeps excellent time---as you might expect. They could have put a radio controlled clock in the tractor. Anyway, I have started timing it with my digital wristwatch which is very reasonably accurate for me in this application and will wait at least one week, maybe two weeks and then will do a comparison check and plan on posting the results. (New Holland just changed out the second instrument panel a few days ago.) Something which I haven't mentioned because it's not directly relevant is that New Holland made it difficult to set the clock. With my 2011 Ford Ranger, all I need to do is push the "Clock" button and I'm immediately in the clock setting mode. Not so with the New Holland tractor. They have two rocker switches you have to operate which is not intuitive at all. In the owner's manual, it takes approximately four full sized around 8 1/2 pages which includes diagrams to set the clock! Also, I timed the ranger's clock and the previous tractor's clock ran approximately 12 times slower. russ_watters Below is a New Holland dashboard with a clock showing 4:16. Does your's look like that? View attachment 227196 No, I attached a photo, (a reflection shows my hand). The clock is displayed in the right hand LCD (the longer rectangular one). The tractor has a CVT transmission and after the new instrument panel is installed, it has to be set up with New Holland's proprietary software using a laptop computer. That brings up another issue about serviceability. Some service on the tractor can only be done with New Holland's proprietary software using a laptop computer. The result of this is that cost of repair is greatly increased, especially if you are hours away from dealer and something you might be able to fix yourself is made impossible! That's why there is a movement for tractor manufacturers to allow a person to acquire equipment to service their own tractor. This is a very important thing to those who want to do their own work as much as possible. See info here, if you're interested: https://www.economist.com/business/2017/09/30/a-right-to-repair-movement-tools-up https://modernfarmer.com/2016/07/right-to-repair/ https://motherboard.vice.com/en_us/...acking-their-tractors-with-ukrainian-firmware #### Attachments • NH Boomer 3050 Instrument Panel.jpg 155.2 KB · Views: 268 256bits anorlunda Staff Emeritus I sympathize. I've heard before about the plight of farmers wanting "the right to repair." But respectfully, I think your issues do not belong on an engineering forum, they should go to a customer service forum (we don't have one on PF). Because it is your dealer who should do what's necessary to make a customer of an expensive item happy, no matter whether the complaint is an inaccurate clock, or a stain on the leather seat. The "right to repair" issue should go to New Holland management or to your congressman. But is not reasonable for a tractor company to employ clock engineers or to apply military/scientific grade accuracy requirements on a non-essential accessory like a clock, or to ask an engineering forum to solve non-engineering complaints. I also stand by what I said. The clock in that dashboard probably came from China and probably cost$1 or less. For example, one like this, which costs $0.10 in quantities of 1000. NTL2009 and Averagesupernova I sympathize. I've heard before about the plight of farmers wanting "the right to repair." But respectfully, I think your issues do not belong on an engineering forum, they should go to a customer service forum (we don't have one on PF). Because it is your dealer who should do what's necessary to make a customer of an expensive item happy, no matter whether the complaint is an inaccurate clock, or a stain on the leather seat. The "right to repair" issue should go to New Holland management or to your congressman. But is not reasonable for a tractor company to employ clock engineers or to apply military/scientific grade accuracy requirements on a non-essential accessory like a clock, or to ask an engineering forum to solve non-engineering complaints. I also stand by what I said. The clock in that dashboard probably came from China and probably cost$1 or less. For example, one like this, which costs \$0.10 in quantities of 1000.

Concerning your statement about my issues not belonging on an engineering forum, perhaps you should read what I posted first. For some reason, you think I wrote that a tractor company should employ clock engineers and etc. but I did not, nor did I ask that you actually solve a non-engineering complaint. Actually, the problem may be caused by improper design.

I'm not sure why you are getting into the actual cost of the clock, since it has nothing to do with the reason I started the post.

anorlunda
Staff Emeritus
You had answers in #2, #3, and #6. Are those not enough?

sophiecentaur
Gold Member
2020 Award
If you could reach inside that clock and measure its oscillator you could get a faster measurement using a sophisticated frequency counter.
If you can do the comparison at the high frequency oscillator stage inside the watch then you can watch the rate of phase drift relative to a high quality standard. This doesn't require a long measurement time. A mate of mine did just this with his cheapo watch and a small probe coil and tweaked to small variable C inside. For a few weeks it was spot on but then it started to drift off.

Instead they let the end user be part of their Quality Control department, which I don't think is very nice!
Annoying but it does mean that you and I can afford things which otherwise we couldn't. It's the Rolls Royce and Model T Ford effect. (And, of course, the personal computer etc. etc.)

Averagesupernova