Tom Sheppard's GPS article is wrong.
- Thread starter Robthebrit
- Start date
Scott Brady
Founder
Robthebrit said:
While that could be (missed opportunity), your above statement is a long way from your original post, and certainly the title of your post, which stated, quite directly that "Tom Sheppard's GPS article is wrong".
Lacking detail- Maybe, but again, a one page article intended for those with limited knowledge of GPS
4th satellites function could have been explained in more detail- Again, maybe, but I believe the description was appropriate to the audience.
Tom Sheppard was "wrong"- It appears that he was most certainly not.
Saying that one of Overland travels greatest ambassadors was wrong "in lots of ways", and "an error is an error", when it appears that the only "error" was not going into enough detail to satisfy your expectations.
Given your personal knowledge of our team, and the efforts we go to producing the publication, I have to say that I am more than disappointed...
edgear
aventurero, Overland Certified OC0012
Well said, Lance.
I myself am a graduate student in GIS (Geographic Information Systems), and would love to jump into discussing the technical aspects of GPS technologies. Do I claim to be an expert on the topic? By no means! But I do hope to further my knowledge of the subject. And I think this thread might just encourage others out there to do some homework on how GPS really works.
I had always heard it said that the 4th satellite provided altitude. When I heard my professor in a lecture state that it provided 'time,' I thought he was giving us false information. That then prompted me to do some more research myself. Since altitude is a by-product of the accurate time the 4th satellite provides, Tom isn't "wrong," nor are the other sources.
Can we expect a 1-page article with no diagrams to delve into the 0's and 1's of how a technology works? I should hope not. I'll find that out in a 500 page text book and some monotonous lectures by PhD's. In my opinion, the article put forth by Tom is a wonderful mix of storytelling and technical writing. I think it should appeal to a wide audience of adventurers, both technical and non-technical.
I myself am a graduate student in GIS (Geographic Information Systems), and would love to jump into discussing the technical aspects of GPS technologies. Do I claim to be an expert on the topic? By no means! But I do hope to further my knowledge of the subject. And I think this thread might just encourage others out there to do some homework on how GPS really works.
I had always heard it said that the 4th satellite provided altitude. When I heard my professor in a lecture state that it provided 'time,' I thought he was giving us false information. That then prompted me to do some more research myself. Since altitude is a by-product of the accurate time the 4th satellite provides, Tom isn't "wrong," nor are the other sources.
Can we expect a 1-page article with no diagrams to delve into the 0's and 1's of how a technology works? I should hope not. I'll find that out in a 500 page text book and some monotonous lectures by PhD's. In my opinion, the article put forth by Tom is a wonderful mix of storytelling and technical writing. I think it should appeal to a wide audience of adventurers, both technical and non-technical.
Robthebrit
Explorer
What I meant by that is if every 'authority' web site can get it wrong than so can the OJ. If you printed the correct information you would probably get complaints of people telling you it doesn't agree with what Garmin say. It would be nice to have a single page article that shows the simple elegance in how it works.
I'm sorry you are disappointed but I still holding out that the article is wrong in at least a couple of places. I have great respect for Tom's and his contributions but everybody can be wrong and just because he is an experienced traveller and used GPS doesn't mean I will take his overview of GPS at face value. Afterall, I am probably more qualified than he is to state how it works. Just because I haven't written an expedition book or drove across africa doesn't mean I don't understand it. I have a degree in microelectronics and masters degree in Math. I have actually wrote the software for a GPS receiver and I am willing to find my source code and post it. In addition I work in 3D graphics so am working with geometry every day.
I have great respect for you and your magazine and I am glad I have helped out where I can. I am sorry you have to be so defensive, I laid out all my facts with no hand waiving. I am not trying to cause a problem or upset anybody.
Lets agree to disagree.
Rob
I'm sorry you are disappointed but I still holding out that the article is wrong in at least a couple of places. I have great respect for Tom's and his contributions but everybody can be wrong and just because he is an experienced traveller and used GPS doesn't mean I will take his overview of GPS at face value. Afterall, I am probably more qualified than he is to state how it works. Just because I haven't written an expedition book or drove across africa doesn't mean I don't understand it. I have a degree in microelectronics and masters degree in Math. I have actually wrote the software for a GPS receiver and I am willing to find my source code and post it. In addition I work in 3D graphics so am working with geometry every day.
I have great respect for you and your magazine and I am glad I have helped out where I can. I am sorry you have to be so defensive, I laid out all my facts with no hand waiving. I am not trying to cause a problem or upset anybody.
Lets agree to disagree.
Rob
Robthebrit
Explorer
tacozona said:
Your professor is correct. The solution you get from the first 3 satellites includes the altitude, the error on the altitude component is the same as the error on log/lat. The 4th satellite corrects time to give the most accurate measurment of distance to satellite which gives the best geometric solution. Altitude and accurate time don't just appear when the 4th satellite shows up.
Rob
Scott Brady
Founder
Robthebrit said:
Followed by
Robthebrit said:
Tom's explanation is all most people need, yet the title of your thread say "Tom's GPS Article is Wrong".
That sends a very confusing message to people who will casually read this thread, and all you needed to do was send an email to Jonathan or I (since you have both our emails and phone numbers) to determine if we had fact-checked the article before making a very disparaging set of comments in a public forum.
The above situation is what required a response defending (i.e. defensive) myself/the journal and a person I respect.
That is what was disappointing to me...
Tucson T4R
Expedition Leader
From what I have read in this thread, I have to agree with Lance. If you take it to a scientific, mathematical level, there are many details that can be discussed. The majority of OJ readers are not interested in that level of detail to be covered by the OJ writers. We read the OJ for insight into areas that are new to us or for detailed descriptions of adventures we wish we could have gone on.
I expect unbiased reviews by knowledgeable people but I don't expect in-depth details that support the theories behind the design. I read Tom's article on the GPS technology and I am still satisfied that it is accurate, even though it may not have covered the mathematical foundation to the level some would have liked. I believe the majority of OJ readers enjoy the level that the Tom's article was written. As Lance said, if you want more in depth details behind the technology, specific professional trade journals and publications would be more appropriate. I personally would fall asleep if the article attempted to cover the details at that level.
I expect unbiased reviews by knowledgeable people but I don't expect in-depth details that support the theories behind the design. I read Tom's article on the GPS technology and I am still satisfied that it is accurate, even though it may not have covered the mathematical foundation to the level some would have liked. I believe the majority of OJ readers enjoy the level that the Tom's article was written. As Lance said, if you want more in depth details behind the technology, specific professional trade journals and publications would be more appropriate. I personally would fall asleep if the article attempted to cover the details at that level.
Scott Brady
Founder
DaveInDenver said:
Dave,
Are you suggesting that we did either?
I can promise you that we would never intentionally present information incorrectly, and in this case, our version was audited against three scientists in technical volumes, a GPS manufacturer AND one of the people that originally designed and deployed the GPS system to begin with. All of which validated Tom's description.
And, I will guarantee that Overland Journal will never "dumb down" a subject. As I am sure you are aware, technical subjects all have various levels of instruction and learning and documentation; including executive summaries, functional reviews, technical specifications and then the 1 and 0. A functional overview of a subject, which was confined to the space available on one page certainly does not constitute "dumbing it down" in my opinion.
I can fully appreciate that this subject is of significant technical interest to you and other members of this board (and Overland Journal readership), but I also feel it is important to temper ones criticism against the filter of the intended audience and editorial space available. There is no other consumer publication in North America that will do a 14 page review of GPS units AND tell you how the technology works from the function level.
ntsqd
Heretic Car Camper
FWIW I've run across this sort of issue in the past in my own ramblings. The best solution I've come up with is to say "there's some complicated things that happen behind the scenes, but here's what the end user sees."
In other words acknowledge that there is more depth to the topic than available space can cover, note that the 'common' explanation is incorrect though easily misconstrued as correct and explain why in very simple terms, and then move on. This is where wordsmithing ability makes or breaks the article. Which is why I don't write magazine articles........
IMHO most readers won't care about the subtle distinctions, but that is no excuse to ignore them or propagate the incorrect. Though dwelling on them does no service either. It is a fine line to walk. Of that I have no doubt. From what I've read of OJ (waiting for my first issue) I believe that they've done a good job of this.
In other words acknowledge that there is more depth to the topic than available space can cover, note that the 'common' explanation is incorrect though easily misconstrued as correct and explain why in very simple terms, and then move on. This is where wordsmithing ability makes or breaks the article. Which is why I don't write magazine articles........
IMHO most readers won't care about the subtle distinctions, but that is no excuse to ignore them or propagate the incorrect. Though dwelling on them does no service either. It is a fine line to walk. Of that I have no doubt. From what I've read of OJ (waiting for my first issue) I believe that they've done a good job of this.
Last edited:
mmtoy
Adventurer
I find this discussion fascinating.
In my field, I consistently come across articles targeted at the "layman" that incorrectly describe how something works. Usually the justification for the incorrect explanation is that the target audience "doesn't need to know the details." This annoys me to no end because usually the correct explanation doesn't take up any more space and would give the audience a true understanding of the topic.
With the knowledge that I may be getting myself into trouble, I'd like to point out that Rob's first post is also "wrong." The clocks of each of the GPS satellites are not in perfect sync, and are especially not in sync with any clocks on the earth.
Because the satellites are moving within different frames of reference and within different gravitational fields, General and Special Relativity come into play. Because they're in weaker gravitational fields, the satellite's clocks tick faster than a clock on the ground. Because they're moving at higher velocities, they tick more slowly than a clock on the ground. These affects have to be taken into account. If they aren't, the satellite clocks will gain time faster than an identical clock on the ground, at a rate of about 38 microseconds/day. The GPS system works in nanoseconds, and 38,000 nanoseconds/day is a lot of accumulating error.
My understanding is that relativity is taken into account when the clocks are built; the satellite's planned orbital altitude, velocity, etc. are used to calculate the difference in clock rates on the ground and in orbit. The satellite clock is then adjusted so it ticks about 38 microseconds/day slower while on the ground. When it gets to its orbit, the clock should now be ticking at about the right rate. The GPS receivers still have to do the relativistic calculations when they receive the signal.
I guess my point is that there are details that shouldn't be left out and then there are even more details that are terribly interesting but may not be necessary to further the basic knowledge of how a system works. Since I'm a math geek, I find I would prefer Rob's mathematically correct solution (with the relativistic explanation included) to Garmin's hand-wavy explanation.
Just my $0.02 to add to the fire.
--Moses
In my field, I consistently come across articles targeted at the "layman" that incorrectly describe how something works. Usually the justification for the incorrect explanation is that the target audience "doesn't need to know the details." This annoys me to no end because usually the correct explanation doesn't take up any more space and would give the audience a true understanding of the topic.
With the knowledge that I may be getting myself into trouble, I'd like to point out that Rob's first post is also "wrong." The clocks of each of the GPS satellites are not in perfect sync, and are especially not in sync with any clocks on the earth.
Because the satellites are moving within different frames of reference and within different gravitational fields, General and Special Relativity come into play. Because they're in weaker gravitational fields, the satellite's clocks tick faster than a clock on the ground. Because they're moving at higher velocities, they tick more slowly than a clock on the ground. These affects have to be taken into account. If they aren't, the satellite clocks will gain time faster than an identical clock on the ground, at a rate of about 38 microseconds/day. The GPS system works in nanoseconds, and 38,000 nanoseconds/day is a lot of accumulating error.
My understanding is that relativity is taken into account when the clocks are built; the satellite's planned orbital altitude, velocity, etc. are used to calculate the difference in clock rates on the ground and in orbit. The satellite clock is then adjusted so it ticks about 38 microseconds/day slower while on the ground. When it gets to its orbit, the clock should now be ticking at about the right rate. The GPS receivers still have to do the relativistic calculations when they receive the signal.
I guess my point is that there are details that shouldn't be left out and then there are even more details that are terribly interesting but may not be necessary to further the basic knowledge of how a system works. Since I'm a math geek, I find I would prefer Rob's mathematically correct solution (with the relativistic explanation included) to Garmin's hand-wavy explanation.
Just my $0.02 to add to the fire.
--Moses
Skylinerider
Adventurer
So I read the article last night and thought it was great. If I wanted to know the inner (and outer) intimate workings of a GPS I would pick up a technical manual. I read the OJ for information pertinent to overland travel.
You all can split hairs all day long on the how's and why's, but what I'm concerned with is will it work. Cudo's to OJ for a great article.
You all can split hairs all day long on the how's and why's, but what I'm concerned with is will it work. Cudo's to OJ for a great article.
Jonathan Hanson
Well-known member
Rob, I have to agree with Scott about your presentation of the subject. You posted a thread in the General section of the Expedition Portal forum - the most-read section, rather than the more appropriate Navigation section - titled it in the most inflammatory way possible, and accused the article of being wrong "in lots of ways" in your first sentence. It was apparent to me you were trying to attract as much attention as possible. You have now backed well away from that to "wrong in at least a couple of ways," but what most Expo readers are going to remember is your first headline and your first sentence. I don't think that's fair to Tom or to us.
If you had emailed me, the editor, first, you would have learned that Tom and I went back and forth for some time trying to edit a complex subject down to the one page for which we had room. (Incidentally, your post was longer than Tom's article.) Thus, for example, the perfectly accurate description of how the portion of the sphere that a satellite "sees" as it measures distance to the GPS unit appears to a user on the surface of the earth as a circle. To describe the entire sphere and then describe its intersection with the planet would have alone taken half the page. What we described is accurate from the perspective of the user.
I'm willing to bet that for the vast majority of Overland Journal readers, Tom's short article gave them a far better understanding of how GPS works.
I am always open to criticism and suggestions regarding the content of Overland Journal. Part of our strength is the pool of talented and experienced people to which we have access - including yourself. I just think there are better ways to contribute to that than hastily posted, broad-stroked dismissives void of perspective.
If you had emailed me, the editor, first, you would have learned that Tom and I went back and forth for some time trying to edit a complex subject down to the one page for which we had room. (Incidentally, your post was longer than Tom's article.) Thus, for example, the perfectly accurate description of how the portion of the sphere that a satellite "sees" as it measures distance to the GPS unit appears to a user on the surface of the earth as a circle. To describe the entire sphere and then describe its intersection with the planet would have alone taken half the page. What we described is accurate from the perspective of the user.
I'm willing to bet that for the vast majority of Overland Journal readers, Tom's short article gave them a far better understanding of how GPS works.
I am always open to criticism and suggestions regarding the content of Overland Journal. Part of our strength is the pool of talented and experienced people to which we have access - including yourself. I just think there are better ways to contribute to that than hastily posted, broad-stroked dismissives void of perspective.
Robthebrit
Explorer
Direct Errors:
"If there is a 4th satellite overhead you will get your elevation as well - although with little else to intersect it, height accuracy won't be as good". Alltitude is not computed seperately, its does not just appear when the 4th satellite appears. Altitude is just as accurate as position but its measured in a strange coordinate system based on the center of the earth. The biggest error in "effective altitude" is not knowing where the ground is (due to landmass, shape of earth etc) and not an inherent problem with how the position is calculated.
"The first satellite draws an accurate circle; all it knows is you are somewhere on the line of that circle. The second satellite draws its own circle which intersects the first at 2 points. When the third draws its circle it will intersect only one of those points" Rubbish... its using spheres not circles and it never geometrically picks a point, at the final stage of geometry calculations there are still two points and one of them is selected via logic. If as you beleive and the article states, 4 satellites are requried for altitude, this explanation has no room for a 4th satellite.
"...and the time is spot on to the nano second time the satellite ceasium clocks are using". Wrong. Using standard C/A L1 signals, your receiver is only accurate to 1us, the bit period of the sync signal. Its actually slightly under 1us due to having 1024bits in 1ms.
"If you remember that the 1.5GHz frequency equals a wavelength of 200mm, that, incredibly, is the order of accuracy you are dealing with". Complete rubbish, that signal is never used in computing position, it is demodulated to give the data payload and is no more impressive than a radio receiver. That statement to me is a complete missunderstanding of what the singals are used for, there is never a 200mm accurate position. The sync signal gives time within 1us which gives the sphere raduis to within 300m. The accuracy does not come from the absolute size of the spheres, the final position is the intersection of these spheres.
In addition to the above there is alots of handwaving. From my point of view, these errors are so fundamentally wrong I can only beleive all the people involved with editing it don't get it.
Edit: "A satellite direclty overhead yeilds a small circle, one close to the horizon draws a circle nearly the dimension of the earth". I don't know what that means, it sounds like somebody is projecting the spheres so it can be drawn on a piece of paper. In the modelling the earth is not used, in fact its not even there. The only places the earth is used is in converting ECC back into log/lat/altitude and a sphere on the center of the earth can be used in lieu of a 4th satellite.
Rob
"If there is a 4th satellite overhead you will get your elevation as well - although with little else to intersect it, height accuracy won't be as good". Alltitude is not computed seperately, its does not just appear when the 4th satellite appears. Altitude is just as accurate as position but its measured in a strange coordinate system based on the center of the earth. The biggest error in "effective altitude" is not knowing where the ground is (due to landmass, shape of earth etc) and not an inherent problem with how the position is calculated.
"The first satellite draws an accurate circle; all it knows is you are somewhere on the line of that circle. The second satellite draws its own circle which intersects the first at 2 points. When the third draws its circle it will intersect only one of those points" Rubbish... its using spheres not circles and it never geometrically picks a point, at the final stage of geometry calculations there are still two points and one of them is selected via logic. If as you beleive and the article states, 4 satellites are requried for altitude, this explanation has no room for a 4th satellite.
"...and the time is spot on to the nano second time the satellite ceasium clocks are using". Wrong. Using standard C/A L1 signals, your receiver is only accurate to 1us, the bit period of the sync signal. Its actually slightly under 1us due to having 1024bits in 1ms.
"If you remember that the 1.5GHz frequency equals a wavelength of 200mm, that, incredibly, is the order of accuracy you are dealing with". Complete rubbish, that signal is never used in computing position, it is demodulated to give the data payload and is no more impressive than a radio receiver. That statement to me is a complete missunderstanding of what the singals are used for, there is never a 200mm accurate position. The sync signal gives time within 1us which gives the sphere raduis to within 300m. The accuracy does not come from the absolute size of the spheres, the final position is the intersection of these spheres.
In addition to the above there is alots of handwaving. From my point of view, these errors are so fundamentally wrong I can only beleive all the people involved with editing it don't get it.
Edit: "A satellite direclty overhead yeilds a small circle, one close to the horizon draws a circle nearly the dimension of the earth". I don't know what that means, it sounds like somebody is projecting the spheres so it can be drawn on a piece of paper. In the modelling the earth is not used, in fact its not even there. The only places the earth is used is in converting ECC back into log/lat/altitude and a sphere on the center of the earth can be used in lieu of a 4th satellite.
Rob
Last edited:
Robthebrit
Explorer
Johnathan,
You published it, you have to stand behind it and you have a right to defend it - I think you got egg on your face and you don't like it. Thats as personal as I am going to make it. You are not arguing the facts you are throwing accusations at me about about my wording. I put all my cards on the table, there is no hand waving on my part.
The internet is a dangerous place to vet information. You know there are poeple on these forums who how how it works. Why didn't you run it past Dave or myself.
I am not going back and forth either. I said its probably is ok for the magazine because I'm the only one who bought it up and people in this post have said they found it acceptable. There is no need to explain the details, Fixing just the errors I posted above and keep the handwaiving would make an accurate article.
You have to expect that using a big name to make a article which is incorrect is going to get some flak.
Rob
You published it, you have to stand behind it and you have a right to defend it - I think you got egg on your face and you don't like it. Thats as personal as I am going to make it. You are not arguing the facts you are throwing accusations at me about about my wording. I put all my cards on the table, there is no hand waving on my part.
The internet is a dangerous place to vet information. You know there are poeple on these forums who how how it works. Why didn't you run it past Dave or myself.
I am not going back and forth either. I said its probably is ok for the magazine because I'm the only one who bought it up and people in this post have said they found it acceptable. There is no need to explain the details, Fixing just the errors I posted above and keep the handwaiving would make an accurate article.
You have to expect that using a big name to make a article which is incorrect is going to get some flak.
Rob
Last edited:
articulate
Expedition Leader
I saw this thread yesterday when there were only 2 responses.
My immediate reaction was that this was 100% the wrong place and manner in which to approach the problem . . .
My immediate reaction was that this was 100% the wrong place and manner in which to approach the problem . . .