rjtjrt said:
"Penguin" in another thread mentioned the importance of EFIS latency in IFR use (see http://www.vansairforce.com/community/showthread.php?t=50144&page=6, post 54).
Can anyone give details of which EFIS' (certified and non certified) have what latencies, and opinions/information on the importance and minimal latency for IFR use?
John
Click to expand...

I agree with the figure quoted in the posts you mention: 100mS.

According to our own ergonomic testing anything less than 100mS total delay from start of measurement to full appearance of the result for an AHRS does not result in noticable advantage. More than 200mS starts to get bothersome if you have an aircraft that can react quite fast.

For numeric readouts, generally 250mS update rate between digits changing is ideal - slower can be OK depending on what you are looking at - faster can be problematic as the lower digits of the numeric may blur and not give your brain enough time to "lock on". There is a bit of "it depends" here - we find that the bigger the numerals the faster you can afford to make them change - but a maximum of perhaps 6-8 changes per second is the limit, beyond that it gets silly.

P.S: In my humble opinion this question should be put on the dusty shelves of history. I think any modern EFIS will be entirely capable of meeting any required minimum specs in this regard with ease, it is no longer relevant.

Rainier
CEO MGL Avionics
 
Last edited:
Rainier Lamers said:
P.S: In my humble opinion this question should be put on the dusty shelves of history. I think any modern EFIS will be entirely capable of meeting any required minimum specs in this regard with ease, it is no longer relevant.

Rainier
CEO MGL Avionics
Click to expand...

Amen to this statement. The modern systems will all be great for the end user. The effects of data latency are very subjective when you get into the sub 200mS range. As far as I know, all modern EFIS units are well below this value.
 
Rainier's is right, you have to be careful what you're talking about here. I also said that its possible to write a book on the subject, but the book would be out of date before it got to the printers!

Most of the data on an EFIS display is numerical and doesn't move around. Updating that 4 or 5 times a second (at the most) is fine. That means the delay between measurement and display of 200ms is also fine - and any system using modern hardware should be able to do that.

The dynamic data - such as a horizon bar - needs to be updated much more quickly (say 60 times a second, although 40 is OK), and the delay between aircraft movement and the display following should be less than 100ms to have no impact on your abilty to fly the aircraft well. Again, any well designed system using modern hardware should have no problem in achieving those figures.

I'm not convinced that all systems on the market do so - no I'm not prepared to name anyone.

Pete
 
penguin said:
Rainier's is right, you have to be careful what you're talking about here. I also said that its possible to write a book on the subject, but the book would be out of date before it got to the printers!

Most of the data on an EFIS display is numerical and doesn't move around. Updating that 4 or 5 times a second (at the most) is fine. That means the delay between measurement and display of 200ms is also fine - and any system using modern hardware should be able to do that.

The dynamic data - such as a horizon bar - needs to be updated much more quickly (say 60 times a second, although 40 is OK), and the delay between aircraft movement and the display following should be less than 100ms to have no impact on your abilty to fly the aircraft well. Again, any well designed system using modern hardware should have no problem in achieving those figures.

I'm not convinced that all systems on the market do so - no I'm not prepared to name anyone.

Pete
Click to expand...

Just so everyone is clear, a 100Ms latency between measurement and display will result in 1/10th of a second delay. If you are rolling moderately as you would in IFR (say, no more than 20 degrees/second) than the display will be NO MORE THAN 2 degrees off your actual bank angle. Since we all naturally slow our roll rate approaching the intended bank angle (level; 15; 20; 30 degrees) the difference between the actual and displayed values would reduce in proportion to the rate of roll - thus, by the time you "stop" rolling the error would be imperceptible.

This is why most folks aren't concerned with latency on any of the EFIS units - it won't be a problem unless you are performing violent aerobatics in the soup, and even then it provides enough of a trend to slow and then eliminate the unusual deviation.
 
breister said:
Just so everyone is clear, a 100Ms latency between measurement and display will result in 1/10th of a second delay. If you are rolling moderately as you would in IFR (say, no more than 20 degrees/second) than the display will be NO MORE THAN 2 degrees off your actual bank angle. Since we all naturally slow our roll rate approaching the intended bank angle (level; 15; 20; 30 degrees) the difference between the actual and displayed values would reduce in proportion to the rate of roll - thus, by the time you "stop" rolling the error would be imperceptible.

This is why most folks aren't concerned with latency on any of the EFIS units - it won't be a problem unless you are performing violent aerobatics in the soup, and even then it provides enough of a trend to slow and then eliminate the unusual deviation.
Click to expand...

This is a very sensible answer.

Rainier
CEO MGL Avionics

P.S: Yes, this is my shortest post ever.
 
Analog!

I want a full time unbuffered by software and "think for me designers" representation of my aircraft attitude.

Bob Axsom
 
Last edited:
Bob Axsom said:
I want a full time unbuffered by software and "think for me designers" representation of my aircraft attitude.

Bob Axsom
Click to expand...

Another example of latency would be on-line multi-player flight simulators. These often have even more latency than an EFIS, but players naturally "adapt" to the slight delay.

While in theory perfect real-time info is best, in reality after a short time you naturally adapt to the "lag" and "anticipate it unconsciously.

Knowing that in practice it will not be a safety issue, the only remaining question is, how much more are you willing to spend, and for what amount of lag reduction, just because it bugs you?

That is a personal question we all have to answer for ourselves.
 
Display Latency and Turbulence...

breister said:
If you are rolling moderately as you would in IFR (say, no more than 20 degrees/second) than the display will be NO MORE THAN 2 degrees off your actual bank angle. Since we all naturally slow our roll rate approaching the intended bank angle (level; 15; 20; 30 degrees) the difference between the actual and displayed values would reduce in proportion to the rate of roll - thus, by the time you "stop" rolling the error would be imperceptible.

This is why most folks aren't concerned with latency on any of the EFIS units - it won't be a problem unless you are performing violent aerobatics in the soup, and even then it provides enough of a trend to slow and then eliminate the unusual deviation.
Click to expand...

The scenario that you are missing here is that sometimes the plane doesn't always do what we plan. I'm thinking of turbulence here...often experienced in actual IFR conditions. If your flight attitude is rapidly and repeatedly upset and you are trying to fly the EFIS to keep the plane "shiny side up" excessive display latency could make your job quite difficult.

Skylor
 
skylor said:
The scenario that you are missing here is that sometimes the plane doesn't always do what we plan. I'm thinking of turbulence here...often experienced in actual IFR conditions. If your flight attitude is rapidly and repeatedly upset and you are trying to fly the EFIS to keep the plane "shiny side up" excessive display latency could make your job quite difficult.

Skylor
Click to expand...

With a typical latency of 100mS or less it will be impossible for any human pilot to fly the aircraft to take full advantage of this latency. Never mind the small fact that the aircraft has a significant mass and very defined reaction time. I honestly don't think that latency (within all but sensible range) will have any effect here...

I tightly coupled autopilot with very fast servos combined with equaly powerful control surfaces could take advantage of this - but with the consequence of possibly excessive airframe stress.

Rainier
CEO MGL Avionics
 
You must log in or register to reply here.