GPS week rollover issue

Jonathan Foster

unread,

Feb 6, 2019, 12:00:44 PM2/6/19

to

I just came across this interesting page: https://www.orolia.com/resources/blog/lisa-perdue/2018/gps-2019-week-rollover-what-you-need-know

On April 6th of this year the GPS week will roll over and on some older instruments it could mess up the date. How many of our soaring instruments might have this problem?

Jonathan

kinsell

unread,

Feb 6, 2019, 12:27:33 PM2/6/19

to

There's some potential for problems, but some of the firmware has been
hacked to delay the actual appearance of the problem for years in the
future. I doubt there's going to be a big ripple of problems in April.

Tim Newport-Peace

unread,

Feb 6, 2019, 12:45:04 PM2/6/19

to

Any of them, but the older instrument the more likely it is to have a
problem.

The date is transmitted by the satellite as a day oy the week and a week
number. there are 11 bits available for the week number (0-1023) and a
rollover occurs every 1024 weeks (an Epoch) or 19 years 8 months approx.

As the GPS engine in the instrument does not receive the Epoch number from
a satellite, it relies on a Battery Maintained Real Time Clock on the GPS
Engine module. If the battery fails or is allowed to become discharged,
Epoch 0 may be assumed, which it probably isn't. Epoch 0 started 6th Jan
1980 originally, but some recorders may have a different date set to fix
previous rollover problems.

The RTC battery should hold up for six months or so, but as the battery
becomes older, the capacity may decrease, reducing the figure. The best
advice seems to be to connect your instrument just before 6th April to
allow the GPS Engine battery to charge. And pray. If you Recorder has the
right date after 6th April 2019 all is well and good, but never allow the
RTC battery to become discharged.

If you have the wrong date after 6th April, it is a Return-to-Manufacturer
repair.

Good Luck!

Jonathan Foster

unread,

Feb 6, 2019, 12:56:06 PM2/6/19

to

Tim, thanks for that explanation. Is this what happened to the Cambridge instruments that needed an update to keep working several years ago?

Dan Marotta

unread,

Feb 6, 2019, 1:40:15 PM2/6/19

to

My old Cambridge gps had to be sent back to be reset. What about the
ClearNav or PowerFlarm portable?

--
Dan, 5J

Tim Newport-Peace

unread,

Feb 6, 2019, 5:15:04 PM2/6/19

to

A number of Legacy Cambridges had problems. They use the Garmin GPS25
Engine and lets face it, the batteries are very old now. I believe that CAI
replaced the GPS Engine, which includes the RTC Battery.

ClearNav are more recent by far, so are unlikely to be affected.

I'm not sure about Flarm, why not ask them?

Tim.

Tango Eight

unread,

Feb 6, 2019, 7:03:36 PM2/6/19

to

I'll post information pertinent to ClearNav products here and on the CN user forum as soon as I have the official story.

-Evan

Andy Blackburn

unread,

Feb 6, 2019, 9:13:00 PM2/6/19

to

Our very own Y2K bug. To be fair, they designed it before 2000 made us all worry about such things, but still, maybe they could've allocated an extra hex character.

9B

Bruce Hoult

unread,

Feb 6, 2019, 9:57:31 PM2/6/19

to

On Wednesday, February 6, 2019 at 9:00:44 AM UTC-8, Jonathan Foster wrote:
> I just came across this interesting page: https://www.orolia.com/resources/blog/lisa-perdue/2018/gps-2019-week-rollover-what-you-need-know
>
> On April 6th of this year the GPS week will roll over and on some older instruments it could mess up the date. How many of our soaring instruments might have this problem?

I'm guessing that will be the final end of my 1994 Cambridge Model 10. (It's already had the battery replaced, of course)

Tim Newport-Peace

unread,

Feb 7, 2019, 3:30:04 AM2/7/19

to

It is just transmissions on the L1 frequency that have this limitation.
Transmissions on L2 & L5 do not, but GPS Engines in Flight Recorders use
L1,
presumably on cost grounds.

Dan Daly

unread,

Feb 7, 2019, 10:54:08 AM2/7/19

to

I asked the Canadian dealer, Solaire Canada. He asked FLARM Technology. He just got back: "We have confirmed with the manufacturer of our GPS receiver modules (u-blox) that they handle the GPS week rollover event on April 6th 2019 correctly. We thus do not expect any issues on that date. This applies to all Classic and PowerFLARM devices, as well as OEM devices using a u-blox GPS receiver." So, no worries for PowerFLARMs and rollover.
Cheers

Tango Eight

unread,

Feb 7, 2019, 11:58:29 AM2/7/19

to

All ClearNav products are expected to handle the GPS Epoch rollover with no issues.

best regards,
Evan Ludeman for ClearNav Instruments

Dan Daly

unread,

Feb 7, 2019, 12:09:02 PM2/7/19

to

Thanks Evan!

Steve Leonard

unread,

Feb 8, 2019, 1:43:37 PM2/8/19

to

On Wednesday, February 6, 2019 at 8:57:31 PM UTC-6, Bruce Hoult wrote:

> I'm guessing that will be the final end of my 1994 Cambridge Model 10. (It's already had the battery replaced, of course)

Nah, Bruce. I bet if your GPS battery is still good, it will survive this one the same way it survived the first one. The problems didn't show up immediately on the first roll-over. They showed up when the old GPS batteries started to age out. At least, that is what I remember. May need to replace my memory battery. :-)

At least, that is my current gamble. Just mailed 4 Model 20s to Gary to get that GPS battery replaced!

Steve Leonard

kinsell

unread,

Feb 8, 2019, 10:16:48 PM2/8/19

to

I've got a 1996 era Garmin backpacker gps still holding the correct
date. Internal battery was replaced under warranty about 22 years ago.

I'd be surprised if it suddenly got the wrong date on April 6, but
really doesn't matter anyway. Lots of old Volksloggers got hit with the
date problem, maybe they'd be more subject to failure if any of them are
still in use.

Tim Newport-Peace

unread,

Feb 9, 2019, 10:45:05 AM2/9/19

to

Most of the Volksloggers were made with the Garmin GPS25 Engine, same as
legacy Cambridge.

Tim Newport-Peace

unread,

Feb 9, 2019, 10:45:06 AM2/9/19

to

At 03:16 09 February 2019, kinsell wrote:

JS

unread,

Feb 9, 2019, 11:31:21 AM2/9/19

to

One rollover problem the Volkslogger doesn't have is due to the robust chassis.
It makes a good wheel chock.
Jim

kinsell

unread,

Feb 9, 2019, 12:44:54 PM2/9/19

to

But not all GPS25's were the same, depends on what firmware was loaded.
Here's a more comprehensive article, at the very end it talks about some
1999 related failures that got kicked down the road many years based on
firmware.

https://www.gps.gov/governance/advisory/meetings/2017-11/powers.pdf

In the unlikely event you're able to update firmware in a GPS engine,
that should be done prior to April.

-Dave

kinsell

unread,

Feb 10, 2019, 9:34:28 AM2/10/19

to

At least the Volksloggers put out actual IGC files. The Models 10, 20,
and 25 should have been relegated to wheel chock duty long before the
Volksloggers.

Dan Marotta

unread,

Feb 10, 2019, 11:32:24 AM2/10/19

to

This raises a question with me. I've often heard that Cambridge started
the whole GPS flight logging thing. If that is true, why did FAI come
up with a different file structure? If it is true, why not simply use
the Cambridge file? Is it not "secure" enough?

--
Dan, 5J

Martin Gregorie

unread,

Feb 10, 2019, 12:14:34 PM2/10/19

to

On Sun, 10 Feb 2019 09:32:19 -0700, Dan Marotta wrote:

> This raises a question with me. I've often heard that Cambridge started
> the whole GPS flight logging thing. If that is true, why did FAI come
> up with a different file structure? If it is true, why not simply use
> the Cambridge file? Is it not "secure" enough?
>

Did the Cambridge file include baro height as well as GPS?
What about the task, pilot, glider and other stuff in the header, i.e.
A,C and H records?

FWIW, now that the use of finish rings and variable barrel size
turnpoints is becoming common in some countries the IGC 'C' records
(which list the start and end of a task as well as the turnpoints) really
need extending to allow the logger to record the type and size of all the
items that define a task, especially where variable barren sizes are used
in a handicapped event.

--
Martin | martin at
Gregorie | gregorie dot org

Dan Marotta

unread,

Feb 10, 2019, 12:59:33 PM2/10/19

to

Thanks. Way more complex than needed for my flying - completed my
diamonds with camera and barograph. Now I just fly for fun and use
OLC/Skylines to extract the data for my logs.

Dan, 5J

Tim Newport-Peace

unread,

Feb 10, 2019, 1:00:07 PM2/10/19

to

At 17:14 10 February 2019, Martin Gregorie wrote:
>On Sun, 10 Feb 2019 09:32:19 -0700, Dan Marotta wrote:
>
>> This raises a question with me. I've often heard that Cambridge
started
>> the whole GPS flight logging thing. If that is true, why did FAI come
>> up with a different file structure? If it is true, why not simply use
>> the Cambridge file? Is it not "secure" enough?

Bob Fletcher (General Manager of Cambridge Aero Instrument) and Hans
Trautenberg came up with it in 1993-4.

>>
>Did the Cambridge file include baro height as well as GPS?
>What about the task, pilot, glider and other stuff in the header, i.e.
>A,C and H records?

Of course.

>
>FWIW, now that the use of finish rings and variable barrel size
>turnpoints is becoming common in some countries the IGC 'C' records
>(which list the start and end of a task as well as the turnpoints) really

>need extending to allow the logger to record the type and size of all the

>items that define a task, especially where variable barren sizes are used

>in a handicapped event.
>

No, not really. The C-records are for defining Tasks in accordance with The
Sporting Code, which defines barrel radii and are not variable.

Charlie M. (UH & 002 owner/pilot)

unread,

Feb 10, 2019, 3:27:36 PM2/10/19

to

No real clue on GPS engine rollover, but I have dealt with other date issues.

Yes, I did most of my badges with a camera and a Replogle. Due to processing errors, I started doing my own film processing. I also developed other badge flights for our club until a club member (who was a pro photographer and could dump in a film load for almost nothing), sorta, "been there, done that".