Hacker News Reader

Comment by tyho | Hacker News Reader

How does this affect timekeeping? Normally, if there is a drop in grid frequency, a small increase will be planned later to maintain a long term average frequency of 50Hz so that clocks that keep time by grid frequency stay accurate [0].

The network split seems to have made this impossible, during the split, the cycle counts for the two regions diverged, and the split ended before this was reconciled. Will people in one of the regions have to adjust their clocks?

[0]: https://en.wikipedia.org/wiki/Wide_area_synchronous_grid#Tim...

View on HN

My suspicion is that NTP (and GPS) have generally superseded the synchronous grid for timekeeping. It has for every application that I know of.

Simultaneously, I suspect some systems still do rely on grid time.

Computers (or anything containing a proper computer) don't care about grid frequency for time keeping.

On the consumer side, your microwave, coffee machine and even bedside clock will.

Not sure on the industrial side.

> On the consumer side, your microwave, coffee machine and even bedside clock will

I'd be very surprised. It's cheaper to build in a 32.768kHz crystal for timekeeping than try to access the grid frequency from the isolated low-voltage circuit.

It's cheaper to build in a 32.768kHz crystal for timekeeping than try to access the grid frequency from the isolated low-voltage circuit.

Those products don't use isolated power supplies --- normally they use a capacitive dropper. A high-value resistor to an IC pin is sufficient to drive the clock counter, and a resistor is definitely cheaper than a crystal.

It is very true.

A similar event happened in Europe a year ago (also due to problems in former Yugoslavia) and all electric clocks in my house (microwave, oven, alarm clocks) went out of sync temporarily. I was just as surprised as you.

Basically, it's exactly the devices that lose track of time in case of a power outage, and that you need to manually adjust for daylight savings time, that are synchronized to the grid. Devices that use a battery (such as laptops, mobile phones, and CMOS clocks in PCs) must necessarily use some other means.

It doesn't seem to be much correlated with price either: my alarm clock cost over $100 and it was still affected. Surely at that price point they could have afforded a crystal oscillator in the design, if they wanted to, but it seems this just isn't typically done.

At a $100 price point, I'd expected a DCF77 receiver tbh, not a crystal oscillator.

As an industry, it requires a special customs declaration + fire safety measures on boats and planes when you transport an oven with a battery inside. That may be the reason for not using battery-backed ovens, and relying on the power supply.

Just curious what kind of alarm clock costs $100

One that plays soothing sounds, like dogs barking or seagulls fighting over trash, probably.

It might even project the moon phase on the ceiling.

Something like this: https://www.amazon.co.uk/dp/B008LR3KD8/

The idea is that it wakes you up gently by flooding the room with light resembling a natural sunrise. One reason that it might be expensive is that it needs a lot of high luminance hue shifting LEDs, and possibly I overpaid a bit (there do seem to be much cheaper competitors) but it was still a good investment because it does make my mornings a lot more pleasant.

I have a microwave and oven from the same brand, and their digital clocks diverge about 2minutes in 24h. I just don't understand how it's possible to build a device that poorly in 2020 (they are new) and I also wonder which tech they use for timekeeping. Is it possible that one of them uses a crystal and the other uses the 50Hz? That would explain a small difference.

It's easy enough to mistune a 32768Hz clock crystal by shitty electronics engineering, unfortunately :/

50Hz might have temporary variations, but it's controlled so that over a longer period of time, you always get the correct number of cycles, e.g. 180000 in an hour.

This also means that going off the 50Hz power grid theoretically has better (or even perfect) long-term accuracy. Also means that if you adjusted your clock for this separation event, you'll have to adjust it back :D

> I have a microwave and oven from the same brand, and their digital clocks diverge about 2minutes in 24h.

That's much, much worse than a cheap digital watch from 20 years ago.

I know. It's worse than the pendulum clock I have to crank at my grandma's once a week! Progress!

Much worse than a digital watch from the 80s. (Maybe 70s?)

It's either a poorly made bottom of the barrel crystal or they cheaped out even further and used a ceramic resonator. That is the price of race to the bottom globalization. Measuring line frequency is a no go because the safety compliance inceases costs.

US current is 60Hz. EU current is 50Hz. Maybe your over maker did some rounding when adjusting by 5/6 and here you go? Btw, does your oven have a 50-60Hz switch?

>> On the consumer side, your microwave, coffee machine and even bedside clock will

> I'd be very surprised. It's cheaper to build in a 32.768kHz crystal for timekeeping than try to access the grid frequency from the isolated low-voltage circuit.

IIRC, the grid frequency is usually more accurate. Probably because it's carefully monitored and managed by the power authorities like the OP describes. If you build a clock with its own frequency reference, than any error will accumulate and have to be monitored and managed by the user.

Maybe, but I bet those designs are very traditional and date from the time where it was cheaper to use grid frequency.

If not, why don't coffee machines that can start on a timer have a backup battery for their clock? That might be cheap enough too. And vital for those moments when there was a 10 second brownout which leads to your coffee not being ready in the morning...

> why don't coffee machines that can start on a timer have a backup battery for their clock?

As a European the thought wouldn't have come to me. Brownouts just don't happen (see this 0.5% frequency dip making headlines), and a blackout is a once-a-decade event for any given house (probably even less frequent than that).

So your answer is probably that demand for that feature is far from universal, and a coffee machine is much lower stakes than for example an alarm clock.

> Depends. I'm in a rural area, and I get 5-second power cuts every couple of months. It's the main reason I have a UPS under my computer desk.

This. Can't put the coffee maker on an UPS though. Or not on a cost effective UPS.

> So your answer is probably that demand for that feature is far from universal, and a coffee machine is much lower stakes than for example an alarm clock.

Actually for the alarm clock i'm just going to be late. Not having coffee hurts much more!

Depends. I'm in a rural area, and I get 5-second power cuts every couple of months. It's the main reason I have a UPS under my computer desk.

I don't think price is significant factor. It just needs optocoupler, diode and few passives.

You don’t need an optocoupler to measure grid frequency. You can just use a resistor. Dirt cheap, cheaper than a crystal.

Yes, but OP talked about isolated circuits. It's likely that most appliances don't need to be isolated though.

OP was talking about consumer circuits, I think they were mistaken about how likely they are to be isolated.

> your microwave, coffee machine and even bedside clock will.

I guess a ~0.5% deviation for a couple minutes is going to be tolerable here.

The grid is never exactly 50Hz, it varies quite significantly. On the long run, the grid operators try to average to 50Hz as close as possible, just so that clocks continue to operate somewhat accurate.

In fact, it is even in the FAQ of the article:

"The transmission grids of the countries of Continental Europe are electrically tied together to operate synchronously at the frequency of approximately 50 Hz"

Based on the article the control limits are 49.8 to 50.2 Hz (+-0.2hz). That's less than 1%, not very significant.

One of the places I lived in Austin, TX had a range with a clock that used grid frequencies (60hz in the US). It was dead accurate. I only changed it for daylight savings, it never need adjustment.

The US used to guarantee 5,184,000 cycles per day. That’s not, like, 1% or anything comparable, that’s zero (with temporary deviations). Any clock which referenced line frequency would be dead accurate over the long-term. There’s a system called Time Error Correction (TEC) which would do this (refer to WEQ-006 and the like). Obviously this error is measured relative to some reference clock.

I’m not sure what the current status is, but my understanding is that there are efforts to retire this system and allow the speed to drift more.

Oh found it. Looks like this went on for more than a couple minutes:

https://www.theguardian.com/world/2018/mar/08/european-clock...

I was under the impression semaphores use grid power as a clock source. And the case of missing payer in southern Balkan got the timing shifted by a few minutes.

Clocks on our ovens and some microwaves got out of sync during that period. I haven't seen any other device rely on the network frequency for timekeeping - everything time sensitive is already on NTP and GPS time.

OT: Several commenters have mentioned both their oven clocks and microwave clocks getting out of sync. I'm curious as to why people set both of these clocks in the first place.

The microwaves and ovens I've seen do not actually require you to set the clock. None of their cooking features and functionality depends on it. It's just a convenience feature. I doubt that most people actually need two clocks in the kitchen, so why not just set one of them (or neither if you don't need a kitchen clock at all)?

Lots of ovens and microwaves don't allow you to disable the clock, so the alternative to setting them is to have them run out-of-sync, which is much more irritating than having to set them.

Ah...I'd forgotten about the ones that always show the time regardless of whether or not it has been set. I can see how that would get annoying.

I guess I've been lucky in that all the microwaves I've had over the last couple of decades, and the ones we had at work, do not show the time unless you've set it. My current one, which has 7 segment LEDs for the time digits, blanks all the digits, leaving just the colons. My prior one showed "--:--" if time was not set. The one before that had a graphical display and just showed blank where the time would normally go.

Mine has two alternatives: Clock or blinking red light. I choose the clock...

I can give you a personal counter-example: my oven requires the time to be set to run - without it, neither the fan nor any of the two heating elements turns on.

It's an old Bosch HBN202S

I have in my kitchen + living-room one clock based on NTP on a dashboard, one on a WiFi radio and one on the oven.

Depending on where I am I see one of them so all must be synchronized to the millisecond so that children are not late to school (school is nearby, which exponentially raises the risk of them being late because it is just "4 minutes" away).

This is one of the scientific uses of such multiple clocks.

There was this great design for a microwave that a UX designer proposed. One button for +30 seconds, no clock, and a door handle. That's it. Everything else is superfluous.

On most microwaves I encountered one can just press the start button and it will do just that: Start with a 30s countdown and pressing it again will add 30s.

As a person who uses a microwave a lot this sounds terrible, I could live with 30s resolution, but not having adjustable power level would not be acceptable. Cooking anything on 100% power is just a recipe for uneven heating.

Based on the other commenter maybe two buttons: start/+30 seconds and power level in 25% increments.

I just got rid of a GE microwave which required me to set both the time and date before heating food after a power outage. I never did figure out what the date was for since it was never displayed and the clock did not automatically adjust for daylight savings time.

On top of that it had a dedicated button for toggling between AM and PM when setting the clock, which served no purpose whatsoever after that point. I wish I knew what the designer was thinking when they came up with that.

Literally because it annoys me more having it not set properly

In Europe, many countries have radio signals that many clocks use to sync their time that is based on some atomic clocks. For example, in Germany these are called Funkuhr and the signal comes from an atomic clock in Braunschweig if I'm not mistaken.

We have this in North America as well. The station is called WWVB.

https://www.nist.gov/time-distribution/radio-station-wwvb

Reply on news.ycombinator.com