Menu Sign In Contact FAQ
Banner
Welcome to our forums

How to make an electronic AI

Thus it was necessary to realign the inertial platform periodically by sighting on stars.

How cool isn’t that

The elephant is the circulation
ENVA ENOP ENMO, Norway

This is an interesting thread. Found THE web site about anything “gyro”: Vectornav

They have loads of downloadable stuff and a dedicated page for Educational Material

Different “applications” are usually defined as follows:

The interesting part is the accuracy of different types of sensors: accelerometer and gyros

The most accurate “solid state” rate gyros are FOG and RLG (Fiber Optical Gyro and Ring Laser Gyro). But even those can be several orders of magnitude less accurate than a mechanical gyro. A mechanical gyro may be old fashion, but it’s the most accurate of them all. The largest problem with a mechanical AI is probably that the earth is round and that it rotates around it’s axis. It needs something to point the axis of spin to the center of the earth.

Then GNSS aided INS.

The elephant is the circulation
ENVA ENOP ENMO, Norway

I saw that site earlier. Amazing how good solid state components can be if you pay enough. Presumably they are doing some very good temperature compensation because I see the 10 year old sensors I have here drifting from 0.98G to 1.02G from +20C to +65C and that (4%) is of the same order as G variations when flying a coordinated turn of a few degrees roll angle.

More amazing is that mechanical gyros outperform the laser ones.

Administrator
Shoreham EGKA, United Kingdom

Peter wrote:

More amazing is that mechanical gyros outperform the laser ones.

Indeed. And who would have thought. I mean, nothing is more accurate than a laser – NOTHING They are completely different principles though. Both FOG and RLG measures rate, while a mechanical gyro doesn’t measure anything in the correct sense of the word. It just is. In a practical application with vibrations, temperature variations and maintenance requirements and so on I would still think a FOG/RLG would be a better solution.

The elephant is the circulation
ENVA ENOP ENMO, Norway

Well, I can see that a mechanical rate gyro would have zero temperature drift. Expansion would just change the moment of inertia of the flywheel, which will do precisely nothing. And if you make the bearings frictionless (which is almost possible, using magnetics) then it will have “zero” drift.

What I don’t get is how you make a really precise mechanical accelerometer. This is because you can’t let the reference mass move relative to its mounting (it would “escape” ). It must be kept centred, and you use a magnetic field to hold it in the central position, and the current required for this magnetic field gives you the G. Now, this isn’t going to be linear… any magnetics employing anything other than vacuum have a nonlinear B-H curve, but perhaps this comes out in the wash, mathematically. This problem must have been solved in weighing machines, however, which are pretty accurate and cheap. Some use the obvious strain gauge method while others use electromagnetic force restoration e.g. these and those must be linear enough too. This suggests that nonlinearity is compensated in software. But none of this will be good for double integration…

Administrator
Shoreham EGKA, United Kingdom

wleferrand wrote:

Checkout https://westphae.github.io/stratux/ – he has an open-source ahrs that plugs back on the stratux

What’s the difference compared to the built in AHRS for Stratux?

always learning
LO__, Austria

Am I right that CS25 aircraft use RLG hardware whereas the Garmin stuff is more consumer level using chips? Where can I read up more on this?

I reckon a certified G5 technically functions the same as the ahrs chip in a Stratux, but has fancier code?

always learning
LO__, Austria

I don’t think laser gyro INS is mandatory in CS25. It certainly wasn’t some years ago; for example airliners flying only intra-US routes didn’t have to install it. Whether you can buy a B737 without it, I doubt. @Antonio may know.

AFAICT all solid state “gyro” products use the same principle as was used since for ever, in Polaris, Gemini, Apollo, etc: three yaw rate gyros at 90 deg to each other, and three accelerometers at 90 deg to each other. There must be some dead common set of equations for processing this into a “navigation solution”, or – as a subset of that, and all you can do if the gyros are solid state i.e. relatively crap – into an attitude reference solution.

The solid state devices range from crappy ones for maybe $1, all the way to $100s for a complete AHRS module. I last looked at this some years ago and Honeywell were a big player in the latter end. But now you get some impressive stuff like Vectornav which seamlessly integrates GPS with inertial, but is still no good for “navigation” without GPS. Well, not for more than minutes. Then for some $20k+ you get fibre based products like the KVH CNS-5000 which has a G antenna one end and outputs NMEA (or some such) at the other end, and if the GPS fix is lost the output just carries on… Airliner INS is another level above that; whether it is more accurate I don’t know.

Administrator
Shoreham EGKA, United Kingdom

Snoopy wrote:

What’s the difference compared to the built in AHRS for Stratux?

I believe this has been pulled back into Stratux and is the basis of the native AHRS solution, where all active development is happening.

tmo
EPKP - Kraków, Poland

Yes absolutely, sorry, it’s been years since I checked on that project. At some point (2016) I was trying to build a portable autopilot using their dev branch.

Sign in to add your message

Back to Top