Skip to main content

iPhone 11 and above have a U1 chip which is supposed to be much more accurate than phones without that chip.  It uses UltraWideBand technology and (I read somewhere) it will get down to 30 cm.  Here shown is a plot done with geo data (Details tab) from an iPhone 6 (no U1) photo.    Has anyone tested a U1?

Attachments

Images (1)
  • mceclip0
Original Post

Replies sorted oldest to newest

I am copying the posts from the previous thread (which had a title not descriptive of the current discussion) below. Please continue the discussion on this thread.

Admin wrote:
Oscar, I don't understand what you are showing in your photo.

oscarwagner wrote:
I took my iPhone 6 to a manport about 600 feet from my house with the phone OFF. That manport can easily be seen on GE so it could be located later on. I stood on the manport, turned on the phone and started taking pictures about every 4 seconds. Turning on at the port is only important to learn how long it takes to get to GPS lock after travel. I loaded each of the photos onto my Desktop, right clicked, selected properties and clicked the Details tab and scrolled down to lattitude and longtitude. This time all but the first of the photo coordinates were about 15' SE of the manport. To locate them, I opened a "placemark" and maneuvered it till the coords matched the photo. Quick check says 1/100th of a " (second) is about one foot. The U1 chip in the iPhone 11 and 12 is supposed to do much better than a nominal 15'. I want to know how much better.

Admin wrote:
Thanks for the explanation.

Another good test to do would be to go out a couple months after the first test and repeat the process to see how much the points have moved. If they haven't moved much then your device is still useful for locating waypoints you established earlier, even if it has large absolute error.

duanerussell wrote:
It appears that precision is not gained after about 15 seconds of the phone acquiring the satellites. The variation in locations also tells me that 1) there is something nearby causing satellite-signal bounce, or 2) the algorithm doesn't get good accuracy. Close, but not repeatable to great precision. I would have thought it would have taken a bit longer for the phone to have that close of a cluster.

Still, not bad for a phone. It will be interesting to see if someone with a newer model can do the same, but get a tighter cluster of locations. I will have to try it with my Galaxy S7 Android.

In response to your second question:

"Somewhat on same subject: What is your favorite iPhone app for plotting points while riding or hiking?"

I use Solocator to take photos of my split points. The app stamps the photo with GPS data and any other labels I attach to the photo (split name, course name, my business name). I then supply these photos to my clients.

Example below.

Cowtown 10K 2020 Mile 5

Attachments

Images (1)
  • Cowtown 10K 2020 Mile 5
Last edited by garybrumley

Went out and bought an iPhone 12 Pro to test.  It does better than the 6 BUT my buddy was just lucky that day.  I do get about 20% closer with the 12.  It will come in handy mapping out new trails not yet on Google Earth.  Using the Solocator app shows accuracy before the shot. Best for the 6 was +-16 feet and the 12 says 9 some of the time but almost always does 13 feet.  The 12 is a lot faster.  I am guessing it throws off fewer wild readings because it locks onto satellites faster.  Solocator takes about 3 seconds to finish a picture and the iPhone 6 takes 3 or 4 times that long.   So between faster satellite lock and knowing guessed at accuracy, I have had no wild readings with Solocator saying +-13 feet.  Taking 2 shots of important points helps verify accuracy. 

Now, fess up time: the U1 chip is likely not involved in any accuracy improvements.  It is supposed to let you put an active tag (Tile) on something and get within a foot of it over a WiFi like range, say 150 feet.  Tiles are $25+.  Remember the the old shoe chips and the threat of a $35 "fine" if you did not return it?

The LIDAR function works and really gets close...at distances < 16 feet.  Still, nothing you would use to mark a 2 x 4.  Quite disappointed at how little the sales people knew about these functions.

Keep realizing more things Gary Brumley's discovered Solocator can do faster and better.  This is a shot of a reference bench leg that has the radial to the S F point right there and included.  No more, "Oops, I forgot to record that."

Not sure how but it could make triangulation easy for faster, more accurate descriptions.  I have asked the author if his app could generate actual links to a map location.  Fewer and fewer people keep a regular computer and printer around.  At $10,000 / gallon (my accountant cousin's calculation) HP inkject ink is a tougher and tougher sell.  It is so expensive that you are afraid to use it and when you do, it has dried up!  The world is going mobile. A big part of service to our customers is making it easy for them to find stuff the morning of the race.

Attachments

Images (1)
  • mceclip0
Last edited by oscarwagner

Thanks Oscar for bringing this topic up. We're all using GPS for multiple purposes these days, so whatever we can learn about its accuracy is very important.

For single point location, what we are call "accuracy" depends on your purpose. If your purpose is to drop a waypoint with your device at a mile mark and then 2 years later use that same device to find that location, then repeatability of your device is all that matters. As long as it sends you to that same location it doesn't matter if the GPS coordinates it assigns to that point are accurate or not.

If however, your purpose is to determine the GPS coordinates of a mile mark, and then give those coordinates to a race director so he can use them with his device to determine the location of the mile mark, then absolute accuracy of BOTH devices is important. In that case we need to consider the accuracy not just of the latest and greatest devices, but also the less expensive and older devices that race directors might have in their pockets.

With all that in mind, and now that it's warmer, I'm going to do an experiment similar to Oscar's. Except that I'm going to take multiple readings over time, probably about once a week. I'm curious how the accuracy of different devices will be different, and also how repeatable the reading of each device will be over time. If you're interested you can follow the experiment at the link below as I add new readings each week.

http://www.rrtc.net/driveway_GPS.html

Last edited by Admin

For the benefit of an "old guy", how many of your race directors can use GPS coordinates at all?  I can't think of any that I've measured for.  Additionally, I am sure that mile splits are established on race day by pacing from a permanent landmark.

If I understand Oscar's very well done work, I could use my phone to provide a GPS Coordinate set that I could record by opening Maps or Google Maps (iPhone 11) and marking the location where the phone is.  I could send these coordinates to a race director who could go to the same spot using Maps or Google Maps.  How close to the original spot is the subject of Oscar's work.  I agree that if the process was accurate within say 1-ft and repeatable also within 1-ft, it would indeed be very valuable.  It could still, even now help locate a paint mark on a country road with no convenient landmarks.  The caveat here is that the GPS coordinates must be used to find a permanent (paint included) landmark.

On a trail, or in the woods, maybe not so much.

"Close enough to find the landmark used to locate the split" is another measure of accuracy. But I would say any device that does GPS is accurate enough for that. It doesn't really matter if the device gets you within 1 meter or 15 meters.

There are some courses though, like ones that are primarily on bike paths through parks or on rails to trails, that don't have many landmarks. For those GPS is really the only thing you can use to define the split location. And therefore, accuracy matters.

Over the past 3 weeks I've taken GPS readings each week from 3 different devices and put them on the map I shared before. So far the  newer and much more expensive iPhone does appear to do better than my cheap, old Android device.

http://www.rrtc.net/driveway_GPS.html

My learning style is confabulation: make something semi-logical up and learn the truth when people crawl out of the woodwork to shoot you down: "The LIDAR function works and really gets close...at distances < 16 feet.  Still, nothing you would use to mark a 2 x 4.  Quite disappointed at how little the sales people knew about these functions." 

That is BAD wrong!  About 50 yards fom my truck yesterday morning with no tape and needed to describe a mile mark from a point about 50' away.  Wondered if the 16' LIDAR limit could be used in multiples.  Tried it and worked fine BUT also noticed that with the right app, you could keep going.  Just now went out into street to my steel taped 200' horseback course and tried it.  Fixed one point and started walking.  Was admonished to move slower once but at the end, without trying for precision, it said 200' 2".  No idea what limit it is but this should work fine for description of miles at least that far.

Attachments

Images (1)
  • mceclip0

It gets even more astonishing.  Next obvious question WAS: Will it go around corners or through objects?  Ran a 25'  tape under center line of my truck.  Set one end at zero point of the tape, walked around vehicle and the set other point at 22' and took picture.  A Riegel Hub at 23.xxx counts per rev on a 27" tire resolves down to about 3".

Attachments

Images (1)
  • mceclip0

Add Reply

Post
×
×
×
×
Link copied to your clipboard.
×