Wrong Elevation & position not always recorded in Auto

I don't see anything in the f/w release notes that impacts elevation as recorded in the activity. 

Offhand, I don't know if the activity records barometric altitude or GPS-derived altitude. I rather suspect that it uses barometric, but I am just guessing. The accuracy of the barometric altitude will be affected by how and when you last calibrated the altimeter, as well as the auto-calibration setting (which would affect timing of automatic re-calibration). At least for relatively short hikes in stable weather conditions, you might get better results if you manually calibrate to a known altitude at the start and disable auto-calibration. This would, of course, be a seriously bad idea in changing weather conditions, regardless of the length of the outing.

If you are curious, it is possible to display both the elevation and GPS-derived elevation in different fields on any screen which allows you to change field content. The "trip computer" is the typical place for this kind of thing.

I am not aware of any issues with auto position recording. However, it IS a fuzzy technique for determining when to record a point. You might get better results with time- or distance-based recording, or by setting the auto interval to more or most often, rather than the default normal. Note that both auto and distance-based will be affected by your ability to obtain a good-quality fix.

Thank you for the feedback twolpert.

I think you are correct about the activity being recorded using barometric pressure to derive altitude. I have gone through the settings for the altimeter to attempt to figure out what is happening.

* Auto Calibration: This was set to "Once" by default.

Choices were Off / Once / Continuous.

I set it to "Continuous". I assume with "Once", if the unit does not calibrate properly right away when you turn the unit on, then you will be off for the rest of the day. Hopefully "Continuous" will provide a better accuracy.

* Barometer Mode: "Variable Elevation" by default.

I left it at this default since the other choice is "Fixed Elevation" which does not match my type of activity.

What I infer from this setting is that it seems that barometric calibration is used as a default on an on-going basis to calculate elevation and we can only choose how it calibrates not "if" it calibrates using barometric pressure.

* Pressure Trending: "Save When Power On" by default.

I left it with this setting. The other option is "Save Always". It seems to record all the time even when the unit is off? I guess it can help to calibrate faster but is probably not kind to batteries overall.

* Plot Type: "Elevation / Distance" by default. Left it at that since it is what I am looking for in a GPS unit.

The available choices are: Elevation / Time - Elevation / Distance - Barometric Pressure - Ambient Pressure

* Calibrate Altimeter: It seems on top of the default Barometric calibration, we can do one-off calibrations using 1 of 3 methods.

+ Current Elevation

+ Sea Level Pressure

+ GPS altitude

I have not used any of these yet.

Since I have found that the positions extracted by the satellites are very accurate to within 3m or less on this unit, I am wondering why the 66i does not allow for GPS altitude on a continuous basis or even an hybrid mode barometric/GPS. I can understand that in areas where sat coverage is not great the altitude may be iffy but I doubt it can be worse than the up to 130m error using the default barometric pressure.

Where I live there is great satellite coverage and I use a good 1/25k map with 10m contour lines on my unit so that could generate very good accuracy. I will test this at some stage with the One-off calibration feature.

twolpert said:

If you are curious, it is possible to display both the elevation and GPS-derived elevation in different fields on any screen which allows you to change field content. The "trip computer" is the typical place for this kind of thing.

I do not know how to do this. I am not sure what you mean by "trip computer" but I will look into it. If this allows me to compare the elevation generated by both methods then it would be great.

Thanks for your input for position recordings. Given the good sat coverage in my region, getting a fix should not be an issue but I will play around with the settings to try to determine the best one. The weird thing is that the last time I was in the same zone skiing, the auto feature worked flawlessly. Very odd. I shall monitor.

Going through the GPSMAP 66i user manual, I noticed that a lot of points are not clearly defined or explained at all in some cases. It feels like they quickly slapped something together based on other variations of the GPSMAP series other than the 66i. I find this disappointing and not very professional but we'll just have to make do with it.

My understanding of auto-calibration is based on something I once read about a different Garmin unit. However, they all tend to behave the same way. As I understand it, "once" means that auto-calibration occurs once when you start an activity. The calibration method isn't well defined, but involves some combination of prior manual calibration, current GPS location, DEM (digital elevation) data available in any of the on-board maps, and GPS-derived altitude. "Continuous" means that auto-calibration occurs frequently during the activity. In that case, it's based on the combination of position, DEM data, and GPS-derived altitude.

If the unit is set to auto-calibrate at all, it does auto-calibrate at power-up. But "once" really refers to the start of an activity.

Manual calibration is still the best way to start. "Known altitude" provides the best combination of ease of use and accuracy. It requires you to know the altitude at your current location. Sea-level pressure requires you to know the equivalent sea-level pressure at your current location. This is usually the pressure reported by your local weather service/station. However, unless you're sitting at the location where the measurement was taken, at the time it was taken, it's not likely to be entirely precise. GPS-derived altitude would seem to be easy to use. The problem is that it's not as accurate as you might like. Typically, the error in the altitude reading is about 3 times the error in the X-Y location reading. This is due to geometric considerations in the measurement technique, not shortcomings in the unit.

On the subject of position error - the approximate error reported by the GPS unit is really stated in terms of probability. Suppose that the satellite page on the 66i says the GPS accuracy is 10 feet. To oversimplify a bit, what this really means is that the reported position is within 10 feet of the real position with a probability of x%. No manufacturer, including Garmin, discloses just what x% is. It could easily be as low as 50%. Some people say that, as a rule of thumb, you should double the number displayed. So 10 feet is better treated as 20 feet. Which would mean a vertical error of as much as 60 feet.

With regard to the trip computer: From the main menu, press Page repeatedly until Trip Computer is the currently selected shortcut. Wait briefly for the trip computer page to appear. Press Menu, rocker down to select Change Data Fields, Enter. Rocker arrows to select the field you want to change, Enter. Rocker down arrow to select Elevation category, Enter. Rocker up/down arrow to select Elevation (which is, I think, barometric) or GPS Elevation. Enter to select. This returns you to the trip computer page, still in field selection mode. When you've changed all the fields you want to change, press Quit to leave field selection mode.

Edited to add: Yes, you do want Variable Elevation. Fixed Elevation is only useful if you are stationary and wish to watch pressure trends (for example, mountain camp site watching the weather). 

The plot type is only going to affect the graph, not the measurements. For variable elevation, you will want elevation vs. time or elevation vs. distance. The other two are for stationary (fixed elevation) use. Ambient means the raw pressure recorded by the sensor. Barometric means the pressure adjusted to sea level (which depends on the device having an accurate idea of elevation).

With regard to auto working differently during different trips to the same location: Even with a clear sky view, the quality of the fix can vary by time of day. The GPS constellation moves. In a given location, the geometry repeats approximately every 12 hours. Readings take at different times can vary quite a bit based on satellite geometry. In addition, things like where you carry the GPS unit (which side of your body, how high up on your body, etc.) can effect the quality of the fix.

I agree with tw, I have added GPS Elevation to my Trip Computer screen along with Barometer and Ambient Pressure.  I might suggest the following:  

• Main Menu - System - Recording - Auto Start:  Off
• Main Menu - System - Altitude - Auto Calibration:  Turn Off
• Start a walk watching the two elevation boxes.
• If the values are not identical and when they change, they change at different times.
• If so, Auto Calibration is Off.
• If the both change at the same time and two the elevations are identical, Auto Calibration is ON.
• After you have determined the Auto Cal status, 
• Main Menu - Setup - Recording Controls - Auto Start:
• Return to Trip Computer Screen and watch both elevation boxes to determine Off/On statue of Auto Cal function.  

Thank you for the detailed answer twolpert.

Thanks to your detailed explanations I have changed the fields in the trip computer screen. I have added GPS elevation and also barometer and ambient pressure as suggested by CowboySlim.

It seems the autocalibration for elevation is a bit of dark art and people are not entirely sure how it works. I was also guessing that a 3m accuracy on a GPS really meant "around" 3m but a 50% probability seems a bit low. I would have thought that GPS makers would use a 95% - 2 standard deviations when making accuracy claims but maybe not after all if it allows to look good with customers. That being said I have noticed really good accuracy when the GPS is actually recording my positions which is very nice.

I understand the concept of fixed elevation and why it is accurate. It is a bit strange though to have to enter the elevation yourself in order to get an accurate elevation :-) (I know it is to help further calibration down the road with the initial anchor point) Also, having to calibrate manually at the start of each trip is a bit of a pain, I would have liked for an ON and go unit without too much manual fiddling to get it to work. Between my GPS, my avalanche beacon, my radio and my phone each with their own idiosyncrasies I am starting to feel that it is making it easier to make mistakes. At least my new 66i does not generate RFI incapacitating my avalanche beacon like my old 60csx used to which is great.

Understood also with regards to the number of "birds" in the sky at a given time and the potential fluctuation of the accuracy of the GPS position.

So far I have tested from home surrounded by tall buildings not on an actual hike outdoors and it seems that the Elevation and the GPS elevation are different by about 25m. The Elevation is within 3m of my actual elevation and the GPS Elevation is off by 20+ meters. This is what I expected given that a lot of sats are blocked by buildings, thereby affecting accuracy. I will keep testing once I go on an actual hike again.

Given what I am seeing so far I think that using "continuous" calibration will fix the problem I had for the elevation.

For tracks not being recorded on some sections of my trip I will test again and may have to manually set the accuracy/number of points being plotted and see how it goes.