Jump to content

Garmin barometric altimeters - the evidence!


Recommended Posts

 

Thanks for the link. I have found these patent sites myself on previous searches to try to uncover the secret workings of the altimeter calibration routine, but unfortunately, I can only access plain text versions of the patent, which have all the words, but no equations, and no images, so sadly, we mathematicians can't do anything with this information. ;)

 

I would need a FULL copy of the patent (with figures and equations) to be able to "reverse engineer" the method. I am not prepared to pay for a copy of the full patent (and I suspect there are still some subtle bits of information which are not published in the patent - such as the actual values of the time increments between samples, etc) - I have no desire to go into business in opposition to Garmin, and I am simply happy that the auto-calibrating barometric altimeter on my Summit HC really works! :D

 

This link should take you to all of the images included in the patent: http://patimg2.uspto.gov/.piw?Docid=071421...View+first+page

 

Good luck pasting all of that into the browser...

Thanks - excellent! I have a copy of the full patent now. (I had to install the "AlternaTIFF" browser plug-in to be able to view all the images and formulae.)

 

I will take a look tonight, but it looks like the patented algorithm is pretty similar to what I postulated in my post #43 above:

 

http://forums.Groundspeak.com/GC/index.php...t&p=3517933

 

(Maybe I should contact Garmin, and offer my services for a job there?! :D )

 

Quoting from the patent:

 

... The method of the present invention accounts for the fact that changes in altitude by a user are reflected in both the barometric altitude reading and the GPS altitude reading, thereby allowing calibration to take place while the baro-altimeter and GPS are in motion. A user is not constrained to be motionless during "calibration mode". Furthermore, this method allows the barometric error to be continuously estimated and used to calibrate the system ...

 

This patent is dated November 28 2006, and describes a significant improvement in auto-calibration methods compared to previous technology - which I guess is why the latest Garmin models perform so much better than older units - the improved algorithm will only be found in very recent Garmin models and / or possibly recent software upgrades for older Garmin models. (My Summit HC is MUCH better at maintaining an accurate elevation reading than my previous B&W Vista.)

Edited by julianh
Link to comment

OK, I see what what #38 says in response to #36.

 

Now my #36 assumed that we did not know the proprietary algorithms, so I was expecting in-the-field, observed data at that time. But that is kind of too much to expect as it would require a recording barometer at the initial point (Sea Level).

 

Looking at #38, I notice that it was written in the future tense, ...will..., so I didn't see that as observed data.

 

If the 5 meters as quoted in #38 is derived from the algorithm, I accept that for what it is. ;)

Link to comment

OK, I see what what #38 says in response to #36.

 

Now my #36 assumed that we did not know the proprietary algorithms, so I was expecting in-the-field, observed data at that time. But that is kind of too much to expect as it would require a recording barometer at the initial point (Sea Level).

 

Looking at #38, I notice that it was written in the future tense, ...will..., so I didn't see that as observed data.

 

If the 5 meters as quoted in #38 is derived from the algorithm, I accept that for what it is. ;)

OK - let me explain this as simply as possible ....

 

While I have never had the opportunity to conduct the experiment exactly as you describe it (i.e. start at sea level, rise to 2,000 metres, while barometric pressure changes by 20 hPA, or whatever), I HAVE conducted many, many real-world tests with my Summit HC. You see, my primary use of my GPSr is for remote fieldwork for potential mining projects and the like, rather than geocaching (which is just an occasional distraction for me).

 

In the last 6 months, I have used my GPSr in the high Andes (> 3,500 m elevation to sea level in a single day), Indonesia (sea level to 500 metres elevation in a day), New Caledonia (sea level to 250 metres several times each day), and numerous trips around Australia (sea level to around 1,000 to 1,500 metres on a typical trip).

 

The nature of my work is that it is simply not possible for me to get quality topo mapping for the remote areas I go to on short notice, so I really rely on my GPSr to track my elevation. Because of the lack of topo mapping, I can't verify my elevation accuracy every step of the way, but I do get the opportunity to check my elevation from time to time - e.g. each time I return to "base camp", or whatever.

 

My experience is that my Summit HC almost always reports elevation to about 5 metre accuracy, regardless of weather conditions - and believe me, the weather can change extremely rapidly in the high Andes, or in the Indonesian jungle in the wet season! Occasonally, my elevation accuracy may drop to about 10 metres (especially if I have been travelling in a car or speed boat, or in extremely windy conditions), but once I get back on foot, and give the unit half an hour or so to settle down, my elevation accuracy will generally be back to about 5 metres accuracy - without any manual intervention or recalibration.

 

There are my actual observations, not theoretical predictions. Does that answer your question?

Link to comment

Well, I do have one more uncertainty to pose here. When you quote a + or - 5 meter accuracy, what is the basis for this result? What standard are you using for comparison? For example, is that observation that the calculated result is within 5 meters is taken a point of known altitude by other means such as survey?

TCP,

 

In my travels, I generally have SOME points of known elevation - sometimes, several such points each day; sometimes not at all for several days.

 

For example. whenever my travels take me to the coast, I can walk down to the water's edge, and be pretty sure I am at sea level (plus or minus a metre or two, allowing for tides, etc). For my remote site investigations, the project "base camp" may have been surveyed, so I will know the elevation of the camp office or similar. I sometimes have access to limited topo maps, which might pick up the tops of major hills or similar, but might be very light on for contours generally in the main areas of interest (or may even be non-existent).

 

That is, I often have no credible topo data for most of the areas I am inspecting. One of my tasks is often to do some preliminary terrain reconnaissance before we send in the surveyors and / or order aerial photography / LIDAR / etc to get quality topo data for the area in question. What I will typically do each day is as follows:

 

1) At the start of the day, turn on the GPS at a location with good sky visibility, let it "soak" for 15 minutes or so, and calibrate the altimeter. (Generally, I know the elevation of my start point, because it is where I finished up the previous day.)

 

2) A I travel through the day, I will frequently make a note of my auto-calibrated barometric elevation, and my GPS elevation. When I have good sky visibility, the two elevations will generally match pretty well (to about 3 to 5 metres typically). However, when my sky visibility is poor (under heavy rain-forest, in deep ravines, half the sky blocked because I am at the base of a big escarpment, etc), the two elevation readings can vary by significantly bigger margins - 10 to 20 metre difference is not unusual under poor sky visibility, 50 metres is not unknown. As explained previously, the GPS elevation can be expected to be in error whenever you don't have a good sky view, but the auto-calibrated barometric elevation will retain good accuracy through such circumstances.

 

3) Whenever I get to any location with reliable elevation data, I will ALWAYS make a note of the two elevation figures, to make sure that my current elevation reading is still well calibrated.

 

4) At the end of the day's travels, I will ALWAYS make a note of my reported auto-calibrated barometric elevation, even if I have no reliable figure for the "official" elevation of my location, as this will be the starting elevation I will use the next day.

 

On the basis of these practices, I am confident that my GPSr maintains auto-calibrated barometric elevation accuracy of about 5 metres or better almost all of the time, but may occasionally drift out to about 10 metres worst case, but the auto-calibration routine will self-correct such errors, so it will draft back into plus or minus 5 metre accuracy (approximately) pretty quickly. My GPS elevation is good to about 5 to 10 metres MOST of the time, but can easily drift out to 20 metre error (or worse) from time to time, especially in locations with a poor sky view.

 

Hope this helps!

Edited by julianh
Link to comment

UPDATE

 

I have received my replacement Colorado 300 from Garmin with a serial number up in the 55*** range and THE UNIT STILL DOES NOT RECORD PRESSURE DATA WHEN POWERED OFF.

 

In light of this development, I now do not believe anyone has a unit that records dynamic data when powered off and I am challenging EVERYONE who claims to have a working unit to post some screenshots of overnight trend data.

 

This issue is a potentially huge one for Garmin. The Colorado manual is rather vague in many respects but to this feature it is really quite specific (page 32 of the online PDF manual).

 

I am calling Garmin this morning about this issue, again (6th time). We wil see where this goes but as far as I'm concerned, I'm in the drivers seat. They clearly advertised a feature when they began selling the units back in January that does not work, has never worked. I believe there are legal protections against this type of thing.

Edited by yogazoo
Link to comment

UPDATE

 

I have received my replacement Colorado 300 from Garmin with a serial number up in the 55*** range and THE UNIT STILL DOES NOT RECORD PRESSURE DATA WHEN POWERED OFF.

 

yogazoo,

 

While I understand your frustration, please note that your specific issue is a bit off-topic for this thread. (The issue being debated on this thread is about how / if the Garmin auto-calibration algorithm really does work in keeping the barometric altimeter calibrated while the unit is turned ON.

 

For the record, my Garmin Summit HC does NOT claim to record barometer pressure when switched off, so the issue which is causing you grief is a non-issue for me - I have always had to re-calibrate my altimeter each morning, and I have never had (nor expected) the ability to track barometer pressure when the unit is off.

 

Cheers - and good luck! (I think you make a very valid point, when the capability IS claimed for the Colorado.)

Link to comment

 

On the basis of these practices, I am confident that my GPSr maintains auto-calibrated barometric elevation accuracy of about 5 metres or better almost all of the time, but may occasionally drift out to about 10 metres worst case, but the auto-calibration routine will self-correct such errors, so it will draft back into plus or minus 5 metre accuracy (approximately) pretty quickly. My GPS elevation is good to about 5 to 10 metres MOST of the time, but can easily drift out to 20 metre error (or worse) from time to time, especially in locations with a poor sky view.

 

Hope this helps!

 

It certainly does, Julian. I accept that explanation. You have provided sufficient detail to establish credibility regarding the auto-calibration. :D

 

Thanks,

 

Papa

Link to comment

 

On the basis of these practices, I am confident that my GPSr maintains auto-calibrated barometric elevation accuracy of about 5 metres or better almost all of the time, but may occasionally drift out to about 10 metres worst case, but the auto-calibration routine will self-correct such errors, so it will draft back into plus or minus 5 metre accuracy (approximately) pretty quickly. My GPS elevation is good to about 5 to 10 metres MOST of the time, but can easily drift out to 20 metre error (or worse) from time to time, especially in locations with a poor sky view.

 

Hope this helps!

 

It certainly does, Julian. I accept that explanation. You have provided sufficient detail to establish credibility regarding the auto-calibration. :D

 

Thanks,

 

Papa

Awww - geez! What are we going to argue about now?! Maybe you and I should head off to the Middle East, and see what we can do about organising global peace? :D

 

Seriously, I am glad that my explanations (and the contributions of others) have helped to clarify this issue.

 

Cheers!

Link to comment
Survey-quality and professional GPSr units will provide more information to an expert user to enable them to know more about their real-time accuracy (including VDOP etc), but this information is not provided on consumer GPSrs, and most users wouldn't know what to do with this information if it was.

Setup correctly, your older B&W Vista (not the HCx) should output it's DOP's via NMEA, just like my yellow eTrex does. Loggable to a PDA in the field. Here's a snippet from mine:

$GPGSA,A,3,,,,,11,15,17,,,28,,48,2.9,2.0,1.1*3F

As most users wouldn't know how to read this information, here's a translation:

$GPGSA = GPS DOP and Active Satellites (NMEA sentence code)

A = Mode: Automatic, 3D/2D

3 = Mode: 3D fix

,,,,,11,15,17,,,28,, = NMEA Satellite ID's of SV's used in position fix (null for unused fields)

48 = NMEA Satellite ID of PanAm Galaxy 15 (WAAS)

2.9 = Position Dilution of Precision (PDOP)

2.0 = Horizontal Dilution of Precision (HDOP)

1.1 = Vertical Dilution of Precision (VDOP)

*3F = checksum

 

Hope this helps!

Edited by coggins
Link to comment

Let me renew the topic.

 

I noticed (eTrex 30, FW 2.90) that pace of autocalibration is rapid after turning on. On the other hand - it takes more than hour if we calibrate it manually (by writing down altitude).

Also: if we turn off the unit during certain low-phase of autocalibration, it cannot be reverted after turing on (autocalibration starts again from the certain altitude which is remember by the unit).

 

See: Elevation plot

 

Practical note: it'd be better to turn off unit for a short while (some 5 minutes) if going outside from an airplane cabin or air-conditioned room (due to more or less stable ambient conditions).

Edited by wmarek74
Link to comment

Let me renew the topic.

 

I noticed (eTrex 30, FW 2.90) that pace of autocalibration is rapid after turning on. On the other hand - it takes more than hour if we calibrate it manually (by writing down altitude).

Also: if we turn off the unit during certain low-phase of autocalibration, it cannot be reverted after turing on (autocalibration starts again from the certain altitude which is remember by the unit).

 

Practical note: it'd be better to turn off unit for a short while (some 5 minutes) if going outside from an airplane cabin or air-conditioned room (due to more or less stable ambient conditions).

 

Or just recalibrate the altimeter. It should be understood that the barometer input can be turned off by choosing Settings->Altimeter->Fixed Elevation.

 

Not a fan of the Garmin algorithms in the 30 for combining GPS and barometer altitude myself. GPS only seems to give a better reading in most cases! Of course the barometer picks up quick and small altitude changes better, but enabling that input seems to introduce large constant errors which aren't magically just calibrated away even if reception is good.

 

Note: I see the thread is five years old. With old chipsets and poor reception - altitude readings sure weren't good. Back then there might have been a huge point with a barometric input, even with a poor algorithm behind it.

Edited by tr_s
Link to comment

It is hard to remember about manual calibration every time, therefore it is good to know that Garmin is able to quickly autocalibrate once turned on.

Off course if we use standard setting: Altimeter--->>>Variable Elevation

 

I find autocalibration extremely accurate and hardly trust GPS readings, especially when the GPS signal is weak (quite often in mountains)...

Therefore, I disagree and could not recommend Altitude--->>>Fixed Elevation at all.

Note that this setting also influences "Total ascents" and "Total descends" values of trip computer (which are more accurate with autocalibration, though still needs improvement).

Edited by wmarek74
Link to comment

Yup, I guess it depends on in what conditions the GPS is used. "Weird" constellations prone to huge altitude errors, blocking mountains on one side, extreme foliage etc. then one might be better off with the barometer activated. Also, if one is interested in quick altitude changes rather than absolute altitude.

Edited by tr_s
Link to comment

BTW, I've done some simple tests on my window sill:

1. Standard Setting-Altitude-->> Variable altitude

2. Altitude -->> Fixed Elevation

 

Results (just Trip computer):

Ad. 1. shows both "zeros" (true as it was lying on window sill) during 40 hours test (eTrex 30 in "economic mode" and with Eneloops XX: backlight, Glonass off, etc.).

Ad. 2. shows enormously huge "total descends" and "total ascents" - some 1000 metres both (after 40 hours non stop).

 

So, I think Fixed Elevation may cause quite huge innacuracy in trip computer when standing still (e.g. resting during a trip).

 

Also in field conditions autocalibration never lies to me.

Link to comment

I've never checked the total ascent and descent numbers. What I repeatedly found out was that when variable elevation was enabled, readouts would be 50 or 60 meters off the topo during entire 24-hour field trips close to sea level and never calibrated away. Going with GPS altitude gave repeatedly consistent readouts. It might have gotten better with the 2.80+ firmwares though, haven't checked into it much.

Link to comment

Glad i found this thread, i have garmin 62s, very new to bareometric,altimeter pressure so hope you do not mind can i ask hear!

 

1, barometer pressure and then ambient pressure? what is the def and which one should i be looking at if it starts going down that means the bad weather is coming in yes or no

 

2,i was winter camping should few days ago at 650 feet should i have set the altimeter for that hight!

 

3,what is the best setting to set on my 62s for a begginer?

 

4, is there any good links to this to help me understand what i am looking at i love to learn more

 

thanks for your time

 

Labrador Wild Man

Link to comment

I would be happy to answer your questions to the extent possible. However, in another thread regarding the same subject, my knowledge was severly discredited. You might want to review it first and then decide that my insight is correct or total trash as characterized here:

http://forums.Groundspeak.com/GC/index.php?showtopic=189245&view=findpost&p=3438512

 

Now:

1. For our purposes, the same thing. Specifically, ambient pressure is that outdoors. Now, using a barometer to measure pressure in a high pressure oxygen bottle (for welding), then that is "barometric" pressure, but not outdoors.

 

2. Yes, but if your GPS has the autocalibration function, that would have been fine, or just use the GPS derived readings.

 

3. Autocalibration or GPS derived. Raw, un-autocalibrated readings can vary due to weather by several hundred feet over several days time.

 

4. I am not aware of any other than this and the link that I posted above (if you accept my credibility).

Link to comment

 

Not a fan of the Garmin algorithms in the 30 for combining GPS and barometer altitude myself. GPS only seems to give a better reading in most cases! Of course the barometer picks up quick and small altitude changes better, but enabling that input seems to introduce large constant errors which aren't magically just calibrated away even if reception is good.

 

Note: I see the thread is five years old. With old chipsets and poor reception - altitude readings sure weren't good. Back then there might have been a huge point with a barometric input, even with a poor algorithm behind it.

Interesting opinion. That was essentially my opinion 4 years ago and that which Julian went into a rant over as "he (meaning me) just didn't get it."

 

Well that was then and this is now. At that time I did not have a GPSr with a barometric sensor, so I had no real, quantitative data, so I could only express my opinion. Which is essentially that as tr_s states above as "..GPS only seems to give a better reading in most cases!.."

 

Yes, I agree that to use the GPS data to correct raw barometric data to something more precise than the GPS data is like making a silk purse out of a sow's ear. Well, in the fullness of time, I now have a GPSr with a barometric sensor and I have gathered quantitative data which totally validates by prescient opinion of yore.

Link to comment

Well, in the fullness of time, I now have a GPSr with a barometric sensor and I have gathered quantitative data which totally validates by prescient opinion of yore.

This is more mind boggling than usual. Yesterday afternoon I pulled my raft out of its box and took it down to the seashore to inflate it. One chamber leaked. If I had two automatically calibrating gps devices, and if the plastic raft material was transparent instead of opaque, could I have placed one device inside and one device outside the leaking chamber and pumped it up again, and then determined the depth to which the boat would have sunk in the ocean if I did not eat lunch before paddling out to sea?

Edited by 39_Steps
Link to comment

What I repeatedly found out was that when variable elevation was enabled, readouts would be 50 or 60 meters off the topo during entire 24-hour field trips close to sea level and never calibrated away.

 

Wow, no wonder you aren't a fan of Garmin's autocalibration! If I had ever had results that bad, I wouldn't like it either; but I've never seen my 76CSx off by anything like that. I don't have need for great elevation accuracy, so I don't pay a lot of attention to it, but whenever I do have a chance to compare it with known elevations, it's usually at least close (typically within 10m, often much closer). I leave autocalibration on all the time, and I can't remember the last time I calibrated the altimeter manually.

 

Years ago, when I had more time to research the topic of autocalibration, I read somewhere that if the difference between the barometric altimeter elevation and the satellite-derived elevation is more than about 30m, autocalibration stops attempting to calibrate the barometric altimeter. I don't know why. This was discovered by someone who had experimented with deliberately calibrating the altimeter manually to several different incorrect elevations to see how autocalibration would respond.

 

This information is old, but I suspect it is still true for current units. So if there was 50 to 60 meters between the actual elevation and the barometric altimeter, that may be why the error was never auto-corrected.

Link to comment

This information is old, but I suspect it is still true for current units. So if there was 50 to 60 meters between the actual elevation and the barometric altimeter, that may be why the error was never auto-corrected.

 

Possibly.

 

Test today:

eTrex30 F/W 3.00 set to variable elevation

One hour with perfect sky view, GPS system only

Reported elevation at end of test 85m

True elevation as of topo map plus a couple of meters: 57m

 

Reported elevation eTrex H: 59m

Setting eTrex 30 to fixed elevation and rebooting: Values consistent to within two meters of eTrex H almost immediately

 

Not impressed, keeping mine on fixed

Link to comment

Any single data point is meaningless. The data set must be of statistically significant size. If one is trying to determine the effectiveness of the auto calibration function, the data set should span a wide range of ambient pressures. For consistency, all parameters must be recorded concurrently. (A misstep which negated the validity of the data expounded in the opening post.)

 

Suggested test procedure:

1. Place GPSr in a relativey unobstructed view outdoors about same time window each day.

2. Turn on and let stabilize for 10 - 20 minutes.

3. Record: ambient pressure, GPS derived elevation, autocalibrated elevation

4. Repeat 2 & 3 for 20 days (readings need not be on consecutive days)

5. Use Digital Dutch website to convert ambient pressure readings to equivalent elevation (raw) values.

6. Calculate standard deviations for GPS derived, autocalibrated and raw elevations.

7. Calculate averages for GPS derived and autocalibrated elevations.

Link to comment

Anybody who desires to learn beyond my capability to teach, take my posts to your local community college. Seek out professors of math chemistry, physics, and engineering. Get their comments pertaining to my posts and come back here with them.

 

Yes, where is the quantitave data which descredits my opinions? Or, is all that is to be offered just jabber-jabber and yappity-yap?

Link to comment

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...
×
×
  • Create New...