Fast clock- ferrite ring

Status
Not open for further replies.

Hv&Lv

Senior Member
Location
-
Occupation
Engineer/Technician
Anyone ever heard of clocks running fast in an installation? We have put a PQM on the service and can’t see anything.
This clock is on a stove.
we are thinking it’s something distorting the wave and causing extra counts on the clock timer.

Kind of wondering if a ferrite ring would help..
Thoughts?
 
Are the clocks even line-sync'd? I'd check with the manufacturer (if possible) to see if it's crystal-timed or line-timed before doing any wiring changes. Or get a replacement clock module?

It's common for cheap crystal clocks to be off (the radio clock in one car runs fast, the other slow; by maybe 3-4 seconds a week each).
 

winnie

Senior Member
Location
Springfield, MA, USA
Occupation
Electric motor research
I'd check for a defective clock first.

First you will need to confirm that the clock actually uses the mains as a timebase; while a mechanical clock with a synchronous motor is clearly using the grid as a timebase, a digital clock might use the grid or might use an internal crystal.

If the clock is using the mains as a timebase, then there is probably chip with a pin sampling the AC at reduced voltage, going through a simple gate to square the AC off to a digital square wave. Noise on that pin could cause false transitions and excess counts, but there should be filtering so that you don't see extra counts at the transisiton.

-Jon
 

Hv&Lv

Senior Member
Location
-
Occupation
Engineer/Technician
Are the clocks even line-sync'd? I'd check with the manufacturer (if possible) to see if it's crystal-timed or line-timed before doing any wiring changes. Or get a replacement clock module?

It's common for cheap crystal clocks to be off (the radio clock in one car runs fast, the other slow; by maybe 3-4 seconds a week each).
I just assumed since it’s in a new stove (digital) that it keeps time through line frequency.
I know voltage arcing and PLC meter reading that causes enough of a blip on the wave it COULD be counted as a crossing, just not positive
 

Hv&Lv

Senior Member
Location
-
Occupation
Engineer/Technician
I'd check for a defective clock first.

First you will need to confirm that the clock actually uses the mains as a timebase; while a mechanical clock with a synchronous motor is clearly using the grid as a timebase, a digital clock might use the grid or might use an internal crystal.

If the clock is using the mains as a timebase, then there is probably chip with a pin sampling the AC at reduced voltage, going through a simple gate to square the AC off to a digital square wave. Noise on that pin could cause false transitions and excess counts, but there should be filtering so that you don't see extra counts at the transisiton.

-Jon
That’s where I’m wondering if a ferrite ring on the power cord would help.
Would the led dimmers cause enough line noise to cause these types of problems?

This isn’t a one off problem, nor is in a specific area. The last two complaints are 40 miles apart..
 

synchro

Senior Member
Location
Chicago, IL
Occupation
EE
I'd check for a defective clock first.

First you will need to confirm that the clock actually uses the mains as a timebase; while a mechanical clock with a synchronous motor is clearly using the grid as a timebase, a digital clock might use the grid or might use an internal crystal.

If the clock is using the mains as a timebase, then there is probably chip with a pin sampling the AC at reduced voltage, going through a simple gate to square the AC off to a digital square wave. Noise on that pin could cause false transitions and excess counts, but there should be filtering so that you don't see extra counts at the transition.

-Jon
I concur. If the clock is running off of the 60 Hz line waveform then there should be a Schmitt trigger on the input that will prevent false triggering on noise near the zero crossings (unless the noise is very large). It was common practice to include this function but perhaps they're cheapening designs and not including it.


Are all of the problem clocks running on meters that are read using PLC signalling? I remember you posted something relating to this several months ago. The signalling you mentioned then was quite low in frequency and so I don't think a ferrite choke would remove it. However, the PLC signalling might reduce the slope of the voltage waveform (i.e, flatten it) at the zero crossings, and that could increase the susceptibility to noise from other sources (including other devices in the home). If that was happening then a good line filter before the stove's power cord might help reduce this other noise and therefore prevent false triggering that increments clocking events and makes the clock speed up.

I agree with zbang that a manufacturer should be contacted to see if they can provide info on whether the timebase is a crystal oscillator or the power line. If crystal controlled timebases are having the problem then it's not likely the problem is being caused by a POCO power quality issue.
 

Hv&Lv

Senior Member
Location
-
Occupation
Engineer/Technician
I concur. If the clock is running off of the 60 Hz line waveform then there should be a Schmitt trigger on the input that will prevent false triggering on noise near the zero crossings (unless the noise is very large). It was common practice to include this function but perhaps they're cheapening designs and not including it.


Are all of the problem clocks running on meters that are read using PLC signalling? I remember you posted something relating to this several months ago. The signalling you mentioned then was quite low in frequency and so I don't think a ferrite choke would remove it. However, the PLC signalling might reduce the slope of the voltage waveform (i.e, flatten it) at the zero crossings, and that could increase the susceptibility to noise from other sources (including other devices in the home). If that was happening then a good line filter before the stove's power cord might help reduce this other noise and therefore false triggering that increments clocking events.

I agree with zbang that a manufacturer should be contacted to see if they can provide info on whether the timebase is a crystal oscillator or the power line. If crystal controlled timebases are having the problem then it's not likely the problem is being caused by a POCO power quality issue.
No, we read about 38,000 meters this way and only occasionally run into something like this. It’s really aggravating but we hate to tell a customer it is what it is, deal with it..

I found one with a radio that turned out to be an arcing fuse assembly. This caused two clocks on this line to run fast.
We have looked all over for this cause..
 

winnie

Senior Member
Location
Springfield, MA, USA
Occupation
Electric motor research
That’s where I’m wondering if a ferrite ring on the power cord would help.
Would the led dimmers cause enough line noise to cause these types of problems?

This isn’t a one off problem, nor is in a specific area. The last two complaints are 40 miles apart..

It is plausible to me that a poorly designed clock circuit could be extra susceptible to line noise. 60Hz is very slow by logic standards, so for the many microseconds when the 60Hz is near zero noise could jump you above and below zero and cause extra counts.

A proper design would include significant filtering to eliminate frequencies above 60Hz, and then as synchro describes a schmidt trigger would be used to further reduce noise sensitivity. If these clocks are sensitive to line noise, I really would complain to the manufacturer

Ferrites might help, but they pretty much only reduce common mode noise. You might consider a plug in EMI filter, but you mention this is on a stove, which would get a tad pricy. You might try something like this in the power feed to the clock itself, but now you are talking about modifications inside an appliance, with all the headache and liability that entails: https://www.digikey.com/en/products/detail/ohmite/AH-01/11689194

Jon
 

Hv&Lv

Senior Member
Location
-
Occupation
Engineer/Technician
Would it be possible for a Schmidt trigger circuit to be “bad” and cause this?
can it be that simple?
 

drcampbell

Senior Member
Location
The Motor City, Michigan USA
Occupation
Registered Professional Engineer
I saw this happen on a microwave oven in a boat. They switched from shore power to the generator and back so frequently that they didn't know or care whether the clock was accurate, but the generator apparently supplied enough third harmonics that the timer ran three times as fast as expected and undercooked everything.

I say "apparently" because nobody ever put a power-quality monitor or spectrum analyzer on it, but it seemed to run exactly 3x as fast as expected; no more, no less.

If it is a low-order harmonic of a 60-Hz powerline, a ferrite bead (or ferrite pair) will not be effective. They're effective against high-frequency and radio-frequency intrusions, and Dirac spikes. (which can get past a Schmitt trigger)
 

Speedskater

Senior Member
Location
Cleveland, Ohio
Occupation
retired broadcast, audio and industrial R&D engineering
Three or four decades ago, lots of clocks used the power line for a reference, but in 2020? I don't think so.
Almost all modern devices use a SMPS (Switch Mode Power Supply) that chops the AC up into tiny higher frequency bits, so the 60Hz reference is lost.
Any info on the stove?
 

Speedskater

Senior Member
Location
Cleveland, Ohio
Occupation
retired broadcast, audio and industrial R&D engineering
Getting back to old clocks and ferrites. Ferrites filter radio frequency noise. Most work above 1MHz but a few work above 100kHz. So they won't have any impact on low frequency zero crossing noise pulses.
 

Hv&Lv

Senior Member
Location
-
Occupation
Engineer/Technician
Getting back to old clocks and ferrites. Ferrites filter radio frequency noise. Most work above 1MHz but a few work above 100kHz. So they won't have any impact on low frequency zero crossing noise pulses.
Thanks for that..
Ill look into it.
 

Hv&Lv

Senior Member
Location
-
Occupation
Engineer/Technician
What I’m getting is “minutes a day”..

after the suggestions received this far I’m going to have them give me a weeks worth of times.
I want to see the minutes gained in an exact 24 hour period.
I’m going to have the customer check the clocks at the exact same time every day for a week and give me the times.

After thinking about this awhile, it would be real simple if they gain exactly 10 minutes a day. That would make the diagnosis simple (I think)
 

hbiss

EC, Westchester, New York NEC: 2014
Location
Hawthorne, New York NEC: 2014
Occupation
EC
Would it be possible for a Schmidt trigger circuit to be “bad” and cause this?
can it be that simple?

To answer that question, no. The whole clock/timer is one semiconductor chip with the display and buttons connected to it.

You say you have two such clocks showing this problem. Are the stoves from the same manufacturer (keeping in mind that several "brands" are actually made by the same manufacturer).

-Hal
 

winnie

Senior Member
Location
Springfield, MA, USA
Occupation
Electric motor research
Would it be possible for a Schmidt trigger circuit to be “bad” and cause this?
can it be that simple?

As hbiss says, the schmidt trigger would be part of an ic clock chip, not a stand alone component.

I think it could be 'bad', in the sense of a poor design with inadequate filtering, or with insufficient hysteresis, or with defective filter components. Something that is functional in most situations but noise sensitive.

And to echo Speedskater: we don't know that these clocks are mains synchronized. Because of available chips many modern clocks do use mains as their frequency standard but many do not.

Jon
 
Status
Not open for further replies.
Top