* GadgetBuilder.com *
Last Modified:
The picture above shows the failure of my La Crosse WS-6002U-A radio controlled clock. The clock is on the left in front of my computer's monitor displaying the NIST time. The clock is slow by 1 hour and 18 seconds. The 18 second error is disturbing because the clock indicates that it updated recently - the tower icon is ON - so the time should be accurate within a second. In fact, my clock's crystal is quite accurate, it loses 1/2+ second per week, but the clock no longer synchronizes daily as it should - the 18 seconds is the cumulative error since I last changed batteries over 6 months ago!! This update failure also causes the hour error shown because it doesn't update to handle DST either. This failure to update is a firmware problem built into this and many other radio clocks - the failure wasn't apparent initially but it was always in there. The failure can pass as a problem due to the electronics deteriorating due to age or due to a change in modulation of the NIST radio - but a VERY simple check (see below) can eliminate these as the cause.
I received this clock as a present for Christmas 2005 so it is well beyond the 1 year warranty. The clock worked for a couple years on a set of batteries which were replaced over the years as needed. In 2013 I changed the batteries shortly before DST ended and it did not correct for DST so after a week I removed the batteries to reset the clock - and it synchronized to the correct time over night. However, in spring of 2014 it did not correct for the DST change and since the batteries were only a few months old I checked the voltage per La Crosse's site and found the batteries were fine.
At that point I decided to check the displayed time against NIST time on the net which resulted in the error shown in the picture above. Researching the problem on the net I found a site whose author had very carefully explored a similar problem with Skyscan radio clocks. Looking further, I found an Amazon review of La Crosse radio clocks reporting a similar problem.
The clock appears to be working properly and synchronizing daily via radio with NIST's atomic clock in Boulder, i.e. the tower icon is ON indicating recent sync. The most obvious - but infrequent - symptom is the failure to correct for DST when the time change occurs. The several second difference with NIST time from the net is an even clearer indication but who thinks to check this when the clock indicates it is sync'ed? I checked the batteries with a voltmeter since low batteries could cause a failure to sync - although this should also cause the tower icon to disappear. Batteries were above the minimum specified on the La Crosse site.
The next check was, with the tower icon displayed (and the time 18 second error), change the time via the buttons to make the tower disappear. I then waited 2 days for the tower to reappear but it did not: the clock did not sync. NIST has increased their transmit power such that their signal now works much of the day here in Connecticut so at 0845 I removed and reinstalled the batteries to reset the clock. All the digits and the tower icon immediately began flashing -- the clock synchronized by 0900. I checked the clock against NIST time via the net and they agreed within a second, i.e. it synchronized correctly.
Clearly the clock can receive the NIST signal, decode it and synchronize its display, verifying that the clock's electronics are working. After a couple weeks the clock had again drifted by over a second so it does not sync more than once after the batteries are installed, although the tower icon remains ON. That is, the symptoms agree with those from the site which investigated the Skyscan radio clock problem. My guess is that the La Crosse clocks of that era use the same type of integrated circuit and share the manufacturing defect: the firmware fault symptoms described on that site.
I wrote to Customer Service at La Crosse so they sent a PDF of the instruction manual (I have the manual that came with the clock). I wrote back and sent a picture of the clock and NIST time similar to the one above. Customer Service responded:
"Unfortunately this clock is past the one year warranty. The clock is nine years old and electronic components do fail in that time. If you would like another clock please look at option 2 below. The La Crosse Store will send you a 45% coupon for the next purchase on their website."
I believe this is a manufacturing defect, that is, a firmware bug. I made my living as a firmware engineer so I understand the problem - firmware bugs are forever because they're in the silicon. This means that the test plan for the firmware must be VERY rigorous because bugs that get through the test are expensive to fix, often requiring hardware replacement. Here, the engineering departments of both La Crosse and Skyscan failed to find this bug... but a talented amateur investigating the problem with a modest budget was able to pin it down quite nicely. In general, the engineers for chips like this have sophisticated debugging tools which would make locating this type of problem relatively easy - but only if you bother to look carefully at all the possibilities. Which is what firmware engineers are paid to do.
Having allowed their product to be sold with a bug in it, what should La Crosse do when it surfaces? The honest thing to do would be to describe the symptoms of the bug so people don't waste their time chasing a known problem. If the cost of replacing the units is prohibitive, suggest ways to work around or live with the problem. Offer replacement units at cost.
It is possible that La Crosse recognized the problem because the WS-6002U-B was released, likely a revision of the WS-6002U-A, perhaps fixing the failure described above. I wasn't able to find a description of the differences between these two models. (Apparently, the bug in my clock's firmware had no effect until after the warranty expired... and you must be alert to separate the bug from the reset that happens when batteries are replaced.)
I think the key point is La Crosse should be honest and up front. I spent considerable time chasing around figuring out what the problem was and how it affected my radio clock and whether I could live with it. The first place I looked was on the La Crosse site and there wasn't a hint that there might be an issue. I like the look of the clock I've got and don't like their current offerings. I could get a used B model via eBay but don't know whether it has the same old sync problem - because the La Crosse site doesn't admit there is a problem with their older units. Plus, it's hard to say whether their engineering department has improved their test plans so they catch similar issues in new designs - and if they haven't, how would I be able to tell given their reticence on noting problems even in older units which are well beyond warranty.
Looking around on the net, there are lots of discussions about radio controlled clocks not handling DST changes properly. How many of these are owners not recognizing this firmware problem and replacing the batteries which "solves" the problem... until the next DST change?
NIST recently improved the modulation method for the low frequency time sync transmission to incorporate phase modulation. Phase modulation allows reception at lower signal levels than the original amplitude modulation (which remains for compatability) but radio controlled clocks taking advantage of this will not be available until late 2014. The WS-6002U-A which I have works for my use now that I understand the need to re-sync manually so I'll wait for the clocks which take advantage of phase modulation to appear on the market.
Update: I gave up on finding a clock with phase modulation and got a Casio "Waveceptor Lineage" wrist watch in December 2015. It cost more than a clock but it syncs every night -- now I always have the right time within a second and it's solar powered so battery should last 10 years or so.
For most radio clocks the owners probably are happy if the clock syncs when the batteries are replaced. What's a few seconds a day? I'm more interested than most radio clock owners in those few seconds a day because I use this clock to calibrate and temperature compensate the gearless clock I built. I consider myself lucky that my radio clock's crystal has a very small error, considerably less than the error in my gearless clock. But that's luck of the draw, the radio clock could have a much larger error because it is assumed the radio will keep it accurate or if it can't will announce failure to sync by clearing the tower icon - neither of which work in my WS-6002U-A clock now.
If you have a comment on my site or its contents, click here scroll down and click again.