akschu

Well Known Member
Patron
I have a question that perhaps some of you can answer.

I'm working on designing my panel and I very much desire moving map. Not just the ability to display a an image, but real maps that have everything on them that I can scroll around like my garmin 296. I would also like to have synthetic vision since I do a tremendous amount of mountain flying.

Up until this point I was just going to get an MGL voyager since I saw the MGL stuff at oshkosh in 2007 and was very impressed with the system and absolutely love how open the company is, but playing with the demo is causing me to rethink since there aren't any maps for Alaska, the system seems to lack the ability to scroll the map around without moving to the dedicate map tool, and synthetic vision is still a little ways out.

Now I'm looking at other options but it doesn't seem like anyone has all of these options for a reasonable price. Is expecting your EFIS to be comparable to your garmin portable asking too much? Perhaps I should just buy a plain dynon EFIS and install a garmin 695 next to it.

Thoughts?

schu
 
Besides Garmin, the only one I know of right now who has an easily panning map is Advanced Flight (who incidentally has the same joystick switch/button/knob) on the Advanced Deck 4500's. Not shipping yet, but almost! Otherwise, it's pretty darned hard to beat Garmin at their own game.

Cheers,
Stein
 
there aren't any maps for Alaska, the system seems to lack the ability to scroll the map around without moving to the dedicate map tool, and synthetic vision is still a little ways out.
schu

Alaska is covered using our free North American vector map data set.
In addition to that you can scan paper maps or convert digital maps (if they are in an open format such as jpg or bmp) as secondary map source (you can install both map types at the same time and switch between them as you need).

We have been using sythetic vision for almost two years now and like other EFIS manufacturers have been threatend with patent infringement claims for this just recently (we don't infringe due to a completely different form of implementation and algorithms).
The synthetic vision includes 3D terrain and runways. Terrain depiction is using a terrain grid (similar to some other EFIS systems). This gives instant and clear depth of view (i.e. you can judge distances and orientation very nicely).

Our first hardware updrade path for existing and new Voyager and Odyssey EFIS systems consists of a plug in "brain" that replaces the current one. This one should be available at Osh. The new "brain" gives multi layered graphics with transparencies, texture mapping, video input and VGA output for a second monitor (and a couple of other things).
While the texture mapping engine is used to provide a typical "flight sim" style terrain display, the current terrain grid will remain an option as it works so well (in my opinion anyway).

You got me thinking on the "panning" BTW, should not be too difficult to add... HMMM....

Rainier
CEO MGL Avionics
 
Rainier, thanks for your reply,

I can't confirm that there are topo maps for Alaska. I downloaded the W180N90.DEM, W140N90.DEM, W100N90.DEM, W060N90.DEM files, put them in the correct directory and can confirm that they are working because I see the terrain data when I tell the simulator I'm in the lower 48, but when I change the GPS to Alaska I don't see any topo data. I can confirm that airspaces and the base maps are available so I can see a rough outline and airspace info, but it's not anywhere near as detailed as on my 296.

Even when I have the maps working in the lower 48 there are many things I can't do such as scroll over to an airspace and have it tell me the ceiling and floor. Another thing I do frequently when I fly is use the map pointer to scroll over to a land point where it meets the water, click on it and tell it to navigate me straight there. This is because much of the airspace in the Anchorage area is between a mouth of a river or the tip of some land. When I do cross country planning I'll frequently pick a peak or something to fly towards before making a turn though a pass so having the ability to put a waypoint anywhere is important, and being able to add one while flying by simply scrolling over and hitting enter makes it simple. Zooming in and out is also a big deal. Alaska is huge and only having 4 levels of zoom seems awkward.

At the end of the day I really want garmin cartography on a 8" display with everything else integrated (light weight is everything) but it looks like the cartography is the hard part. I suppose I can make my own maps, but I would want to see how that works on the sim before spending money. How do I make my own maps? I think I understand how the import system works. I do I just create new DEM files and import them? How will I be able to do this with vector maps? Raster maps will take up a huge amount of space before they have any real level of detail.

Thanks,
schu
 
Hi schu,
I'll check on the terrain, busy with some other stuff right now, give me a day or so.
You can eaily obtain relevant airspace information for the area you are in - simply press SHIFT and then IMS (information system) - button 6.
You get a list of airspaces you might be interested in with the word "IN" next to any airspace you are currently inside.
If you want more information on a particular airspace, simply press the numbered button next to it.
This will give you detailed information including ceilings, frequencies etc.
If you have a compatible radio connected, press the numeric next to the frequency and your radio is setup. It even tells you who to call.

You can also display ceilings and base as well as names of airspaces on the moving map - simply press SHIFT and then the MAP key (button 5). Then you can adjust map information as you need (declutter and detail). Note that this setting is independent for every zoom level, so you can remove a lot of stuff if you want to view a large area - it may get way to busy otherwise...
The system remembers your selections so you typically only do this once to suit your preferences.

Your comments regarding raster maps is quite valid, they can take around 1/2 gig to cover Alaska at a good detail level - but that hardly matters, SD card s are dirt cheap and even a 1 gig leaves you lots of space.

You use the Mapmaker utility to convert raster maps.

You can even edit our vector maps (you use the vector map editor). But before you do that, make sure that you have all map details checked on so you don't end up wanting to put in a road that is allready there.

The DEM files are standard terrain data files you can find from many places on the net - we don't do anything to them.
Raster maps once converted tend to result in several files, one for each source map and each of these normally spans several degrees horizontal and vertical. I tend to use the 1200 pixels per degree resolution option which is a nice compromise between resolution and file size and this also tends to load quite fast (around a second or so to zoom).
Once you converted your raster maps you copy the created files (with names like N70W123B.M43 (top left corner of map, map resolution, horizontal and vertical size of map in degrees). You also create and copy a file called Mapindex.MM0 which contains a list of which files contain which degree tiles at whatever resolution (it is possible to have the same area covered by several resolutions and these will then be selected depending on your zoom level).
To install them on the instrument is quite easy: Copy all the files to an SD card, insert in instrument - goto "Installation tasks" and select "install raster maps".
The mapmaker application contains a short text file that describes how to use it - effectively you load the map, then you set the number of degrees horizontal and vertical and desired resolution (number of pixels per degree). Then you go and click on the map image each intersection of a degree line so the image can be referenced. We do each degree rather than just two points so be can use any possible map projection - you could even use a hand drawn scetch !

Limitations that exist are as follows:
The map needs to have AT LEAST the number of source pixels per degree that you want to use as target resolution. More is good but less will not work.
You can only convert FULL degrees (this can be partly overcome if required by merging partial degrees from different maps using a suitable image processing program (such as Corel draw for example).

Hope this helps a bit to understand how the system works...

Your comments regarding waypoint creation have been noted and in fact some of them are on our wish list so it's only a matter of time...

Rainier
CEO MGL Avionics

Rainier, thanks for your reply,

I can't confirm that there are topo maps for Alaska. I downloaded the W180N90.DEM, W140N90.DEM, W100N90.DEM, W060N90.DEM files, put them in the correct directory and can confirm that they are working because I see the terrain data when I tell the simulator I'm in the lower 48, but when I change the GPS to Alaska I don't see any topo data. I can confirm that airspaces and the base maps are available so I can see a rough outline and airspace info, but it's not anywhere near as detailed as on my 296.

Even when I have the maps working in the lower 48 there are many things I can't do such as scroll over to an airspace and have it tell me the ceiling and floor. Another thing I do frequently when I fly is use the map pointer to scroll over to a land point where it meets the water, click on it and tell it to navigate me straight there. This is because much of the airspace in the Anchorage area is between a mouth of a river or the tip of some land. When I do cross country planning I'll frequently pick a peak or something to fly towards before making a turn though a pass so having the ability to put a waypoint anywhere is important, and being able to add one while flying by simply scrolling over and hitting enter makes it simple. Zooming in and out is also a big deal. Alaska is huge and only having 4 levels of zoom seems awkward.

At the end of the day I really want garmin cartography on a 8" display with everything else integrated (light weight is everything) but it looks like the cartography is the hard part. I suppose I can make my own maps, but I would want to see how that works on the sim before spending money. How do I make my own maps? I think I understand how the import system works. I do I just create new DEM files and import them? How will I be able to do this with vector maps? Raster maps will take up a huge amount of space before they have any real level of detail.

Thanks,
schu
 
Just my ignorance and curiosity: how useful is the 'panning map' thing? Any real-life experience out there? I haven't felt the lack but then I didn't feel anything was lacking before GPS and now it's hard to remember flying before it. So I'm willing to believe I'm missing out, but what am I missing?
 
Panning is great for moving ahead of your flight path to check airspace altitudes, metars, TAFs and obstructions. (on those MMs that have those features)

Also my wife is always askinh "whats that place?"

Pan over, press enter and voila.
 
Panning is great for moving ahead of your flight path to check airspace altitudes, metars, TAFs and obstructions. (on those MMs that have those features)

Also my wife is always askinh "whats that place?"

Pan over, press enter and voila.

I must admit, that the new Garmin 696 alleviates some of the panning required with the smaller models! :D

L.Adamson
 
Panning is great for moving ahead of your flight path to check airspace altitudes, metars, TAFs and obstructions. (on those MMs that have those features)

Also my wife is always askinh "whats that place?"

Pan over, press enter and voila.

I pan all over the place with the 696 - I probably leave it in "pan mode" more often than not when on the map. Just as Milt does, for checking weather and what is ahead or to the sides.

Paul
 
Rainier,

Your response is very helpful as usual. I'll take a look at the mapping and do some converting. Right now the airspaces aren't right, there isn't any topo information, and the map is so busy it's not useful or I declutter everything and it still isn't useful.

Give me a few days to import some maps, take some screen shots, and write some documentation and get back to you. I would like to lay out exactly what I want to do and why since Alaska can be so different to fly in. That way you will get good input from a user. This information may or may not be useful to you, but it will give you some ideas about how people use the product or want to use the product.

Thanks,
schu
 
I probably spend more time goofing around "panning" or moving the maps than I do on a static page. Like Milt/Paul said, once you have weather overlaid on the map, you can spend all kinds of time looking at TAF's, Weather, clouds, runway configurations, frequencies, etc.. I find it quite usefull to be able to easily scroll around with the joystick.

Cheers,
Stein
 
Agreed.

I scroll around on the maps for the following things:

1. Looking at airspaces beyond where my airplane is.
2. Setting random way points where I want to turn around a geographical obstacle.
3. Looking at a previous route or track that was logged on a prior flight.
4. Identifying lakes, roads, mountains, glaciers.
5. Looking for alternate places to land that aren't airports.
6. Looking for alternate routes around weather and obstacles.

Keep in mind that there are 6000-8000 foot mountains all around this area and I rarely fly about 4000 feet. This makes mapping a much bigger deal then simply which direction I need to be pointed.

I'm sure you guys figured out that I'm not building an RV :) However I find the knowledge on this site, especially when it comes to glass, very helpful.

schu
 
I'm sure you guys figured out that I'm not building an RV :) However I find the knowledge on this site, especially when it comes to glass, very helpful.
Your questions are valid and answers for them will help RV builders as well. Everyone wins so who cares? :D
 
I just tried it using a position around W150 / N61 using the Odyssey simulator with the W180N90.DEM terrain file.
Works like a charm !
Both synthetic vision as well as vector map topo backdrop...

For the map, please take note that the topo backdrop will only show if one of the two following conditions are met:

The map has been set to display vector maps (SHIFT+MAP).
The map setting permits the topo background to show.

Also note that with the lowest zoom level (widest area) topo background is automatically suppressed due to the large area that would need to be covered and consequent time it would take to process the map image.

Regarding your suggestion to add "panning", this can be done. There is one nasty issue, our system supports 100.000 waypoints and even with a fast processor it takes a little time to evaluate each one of them to see if it needs a icon drawn on the map or not. The map at the current location does not have that issue as we keep the complete list of waypoints sorted by distance from the current location - that way we know the icons to be shown are allways at the bottom of the list (closest to current location).
So, for panning - I would have to draw the map (and make that fast so there is immediate feedback) and then start going through the list of waypoints and add them to the map as it finds applicable ones - could take a few seconds for this to complete but that should not be too much of an issue I suppose...

Yes, certainly keep the suggestions coming - does not matter if you end up buying our stuff or not, that's not what we are here for.
None of our instruments are static (i.e. "finished") and I make no excuse for that - there is a long way to go and we have come a long way, but I understand, when it comes to development of an EFIS, there simply is no such thing as "finished".

Rainier
CEO MGL Avionics

Rainier,

Your response is very helpful as usual. I'll take a look at the mapping and do some converting. Right now the airspaces aren't right, there isn't any topo information, and the map is so busy it's not useful or I declutter everything and it still isn't useful.

Give me a few days to import some maps, take some screen shots, and write some documentation and get back to you. I would like to lay out exactly what I want to do and why since Alaska can be so different to fly in. That way you will get good input from a user. This information may or may not be useful to you, but it will give you some ideas about how people use the product or want to use the product.

Thanks,
schu
 
Ok, I see the problem, I followed the documentation on this page:

http://www.stratomaster.com/html/north_america.html

Which didn't show the W180N90.DEM file. For whatever reason I didn't even put it together that it had the lat/long in the file name. Perhaps all of these years using unix has caused me to go blind to non-cryptic file names. :)

At any rate it works fine now. I was also able to find the mapmaker program online so I'm playing with making my own maps as well. I have a fantastic Alaska atlas which would be really cool to have on screen.

As far as the maps and waypoints go, I'm not sure what language you are writing this in, but if it where me I would probably index the waypoint data by lat and long somehow then figure out what my min/max lat/long is onscreen then go find the waypoints using the index. Even if you only did longitude you could find your data really fast if the left side of the screen was 149W and right 151W so you only looked at waypoints between 149-151 using an index. Unless someone had 10000 waypoints directly north or south of each other it would perform well.

Anyway, not knowing how you do this stuff, it's just a suggestion, especially since I stick with database and email work anymore so I may be completely off base.

Later,
schu
 
I have another question real quick. When do you expect to have systems ready to ship with the 3d hardware? Also, do you expect the synthetic vision frame rate to increase when using the 3d hardware? I don't know how fast it is on the real thing, but the wire-frame stuff is pretty slow in the sim only giving me a frame every few seconds. Given decent computer hardware does the real thing perform faster or slower than the sim or is the sim really an emulator that clocks things so that they are similar?

Thanks,
schu
 
I have another question real quick. When do you expect to have systems ready to ship with the 3d hardware? Also, do you expect the synthetic vision frame rate to increase when using the 3d hardware? I don't know how fast it is on the real thing, but the wire-frame stuff is pretty slow in the sim only giving me a frame every few seconds. Given decent computer hardware does the real thing perform faster or slower than the sim or is the sim really an emulator that clocks things so that they are similar?

Thanks,
schu

The sythetic vision frame rate is determined largely by how fast GPS positions have GPS heading (track) become available. This varies a bit depending on satellite availability but can go up to 4 times per second (i.e. that is the rate at which the actual image transformation is done as only then do we know exactly where you are and where you are looking).
The image is coupled to the AHRS of course and the transformed image "moves" at the rate of the AHRS stabilised attitude data (which we fix at 10 times per second).
On the simulator it tends to be slower than in real life (we do the simulated GPS once per second).

Our MGL168 CPU card is a low cost upgrade to the Odyssey, Voyager and Explorer systems but the current MGL152 will still be used - in other words we may look at producing two versions of the systems, one low cost the other a little less low cost.

The MGL168 doubles the CPU speed but more importantly adds a dedicated graphics processor (a Fuitsu chip). With this a lot of the graphics work which is currently done on the main processor falls away so the doubling in speed is in reality much more than just double (Drawing the graphics takes perhaps as much as 90% of any cpu time currenty spent) - the actual EFIS by comparison would probably run on a much smaller processor just fine (if it would not be burdened with drawing screens).

We designed the systems in such a way that CPU and anything to do with memory or graphics would live on a small plug-in board - this is the one item that is most likely to need upgrades in the future as better silicon becomes available and the ever increasing software demands higher performance.
While we have a well established software upgrade path, we also wanted some form of hardware upgrade path and upgrading the "brains" makes the most sense.

Rainier
CEO MGL Avionics
 
Last edited:
At any rate it works fine now. I was also able to find the mapmaker program online so I'm playing with making my own maps as well. I have a fantastic Alaska atlas which would be really cool to have on screen.

That's exactly what we did when we started !
You will notice that there is an option to show vector airspaces drawn over the raster map.
Now you know why :)

As far as the maps and waypoints go, I'm not sure what language you are writing this in, but if it where me I would probably index the waypoint data by lat and long somehow then figure out what my min/max lat/long is onscreen then go find the waypoints using the index. Even if you only did longitude you could find your data really fast if the left side of the screen was 149W and right 151W so you only looked at waypoints between 149-151 using an index. Unless someone had 10000 waypoints directly north or south of each other it would perform well.

Anyway, not knowing how you do this stuff, it's just a suggestion, especially since I stick with database and email work anymore so I may be completely off base.

Later,
schu

No, definitely not off base.
I think one of our very first versions worked like this but we had to do the sort using both lat and long eventually - it just creates too much data otherwise (not for Alaska I am sure) but some areas in the U.S. are mind blowing when it comes to the number of waypoints, airfields, navaids, reporting points etc.

Rainier
CEO MGL Avionics
 
Regarding your suggestion to add "panning", this can be done. There is one nasty issue, our system supports 100.000 waypoints and even with a fast processor it takes a little time to evaluate each one of them to see if it needs a icon drawn on the map or not. The map at the current location does not have that issue as we keep the complete list of waypoints sorted by distance from the current location - that way we know the icons to be shown are allways at the bottom of the list (closest to current location).
So, for panning - I would have to draw the map (and make that fast so there is immediate feedback) and then start going through the list of waypoints and add them to the map as it finds applicable ones - could take a few seconds for this to complete but that should not be too much of an issue I suppose...

Rainier
CEO MGL Avionics

If you have control of the waypoint database, forget the sort. Give each waypoint links to other nearby waypoints. Then, when drawing a map, determine the waypoint nearest center and begin traversing links, breadth first. For each waypoint, mark 'visited', check if on map and, if so, draw it and add it's links, if unvisited, to the list of waypoints to check. Stop when the list is empty. Yes, it increases your storage for the link pointers, but the performance increase is crazy because you don't have to scan the entire list of waypoints. If you can thread the scan, it will draw from screen center out and you can kill the scan thread on redraw; so what if the periphery icons don't get shown while the map is panning?

Sorry, I can see everyone else's eyes rolling up into their foreheads. Computer Science major and retired IBM programmer; I just had to kibitz. I'm still not convinced about panning; I can zoom out and see everything I need inflight but, then again, I'm always using the map in Flight Simulator so I would probably enjoy the panning feature if I had it. Thanks for the responses to my question.
 
That is a good idea - but it does not solve the problem - I still need to find the waypoint nearest to the center - and if my list is not sorted I still need to go through the whole list.
Best to sort once when you load the list - and then keep it sorted as you move along (which can be done using a low priority bubble sort in the background).
This way when you need to draw the moving map (and that needs to be done real FAST) there is no need to search for waypoint icons.
Of course there is much more attached to this - nearest airport information (for 3D runways etc), obstacles etc etc. The system needs to be aware of it's suroundings even when you are not looking at the map.
Then when you want to do a quick "goto" using the GPS, you would want the waypoints/airports sorted by distance anyway...

I also prefer the system not to be dependent on preprocessed waypoints in any major way - one reason for that is that we want to be able to extend this to a lot of different waypoint sources in addition to the ones we allready have (Jeppesen is next BTW).

Rainier

If you have control of the waypoint database, forget the sort. Give each waypoint links to other nearby waypoints. Then, when drawing a map, determine the waypoint nearest center and begin traversing links, breadth first. For each waypoint, mark 'visited', check if on map and, if so, draw it and add it's links, if unvisited, to the list of waypoints to check. Stop when the list is empty. Yes, it increases your storage for the link pointers, but the performance increase is crazy because you don't have to scan the entire list of waypoints. If you can thread the scan, it will draw from screen center out and you can kill the scan thread on redraw; so what if the periphery icons don't get shown while the map is panning?

Sorry, I can see everyone else's eyes rolling up into their foreheads. Computer Science major and retired IBM programmer; I just had to kibitz. I'm still not convinced about panning; I can zoom out and see everything I need inflight but, then again, I'm always using the map in Flight Simulator so I would probably enjoy the panning feature if I had it. Thanks for the responses to my question.
 
That is a good idea - but it does not solve the problem - I still need to find the waypoint nearest to the center - and if my list is not sorted I still need to go through the whole list.

Just once at initialization; then keep pointers to current 'nearest-to-actual-location' and 'display-center' (the former lets you keep track of nearest waypoints even when the map is panned). You can add a check to ensure that the pointers are still correct and they only have to search their node lists.

I also prefer the system not to be dependent on preprocessed waypoints in any major way - one reason for that is that we want to be able to extend this to a lot of different waypoint sources in addition to the ones we allready have (Jeppesen is next BTW).

Rainier

Now, THAT is a good reason. I thought about that in regards to user-defined waypoints. However, you'd have a routine to do the pre-processing for you; it should be able to handle all your raw waypoints, no matter the source, and could be triggered when a user-defined waypoint is created as well as initialization. That way you maintain the raw data and the 'pre-processing' only occurs when needed.

Anyway, I'm sure there are more pitfalls; I was just struck by how similar the requirement was to some Virtual Worlds (simulation stuff) projects we had worked on here at Northern Arizona University. I'll try to stop kibitzing now. Really. I don't have a coding problem; I can quit anytime I want... :rolleyes: