Index Home About Blog
From: Henry Spencer <henry@zoo.toronto.edu>
Newsgroups: sci.space.shuttle,sci.space.tech
Subject: Re: Why don't big launchers fall at lift-off time ?? (serious question 
	:-& )
Date: Wed, 31 Jul 1996 19:42:15 GMT

>Would it be possible to have a ring or a disc or even better the whole fuel
>turning very fast at the top/middle of the rocket???
>The purpose of such a solution would be to simulate a gyroscope which keeps
>anything in a very stable position.

A gyroscope large enough to keep a large rocket stabilized, simply by its
inertia, would be impossibly heavy.  Small upper stages are sometimes
stabilized by spinning the entire rocket, including the payload, but this
is an inferior approach which is used because of its simplicity rather than
its high effectiveness.

A rocket's control system typically does use gyroscopes to *detect* changes
in attitude, but the correction is done by altering the direction of rocket
thrust, not by use of gyroscopic forces.
-- 
 ...the truly fundamental discoveries seldom       |       Henry Spencer
occur where we have decided to look.  --B. Forman  |   henry@zoo.toronto.edu



From: Henry Spencer <henry@zoo.toronto.edu>
Newsgroups: sci.space.tech
Subject: Re: Inertial compass
Date: Sat, 24 Aug 1996 02:58:01 GMT

In article <n9630117.840698049@statler> n9630117@cc.wwu.edu (Bjorn Townsend) writes:
>I'm wondering... is there a way to make a compass (perhaps 
>inertial/gyroscopic?) that always points in the direction of the Earth's 
>rotation?

Yes and no.  There's no way (that I know of) to sense Earth's rotation
instantaneously... but any half-decent inertial-quality gyro system
sitting still on the ground will fairly quickly notice that the ground is
rotating, and given ten minutes or so of data can measure the position of
Earth's axis quite accurately.  This is called "gyrocompassing" and it's a
standard way of aligning an inertial navigation system. 

>And does the vector of the rotation vary widely, or is the 
>variance low enough to make navigation practicable? (this would be from a 
>ground-based platform, of course)...

The geographical position of the axis does vary, but only very slightly.
-- 
 ...the truly fundamental discoveries seldom       |       Henry Spencer
occur where we have decided to look.  --B. Forman  |   henry@zoo.toronto.edu

From: henry@spsystems.net (Henry Spencer)
Newsgroups: sci.space.tech
Subject: Re: Short Lifespan of Gyroscopes in Space?
Date: Sat, 13 Mar 1999 22:22:05 GMT

In article <7cc027$2bf$1@nnrp1.dejanews.com>,
 <wjlservo@my-dejanews.com> wrote:
>> 	3) Are there no alternatives to a physical gyroscope?
>
>Obvious choice of inertial sensor for measuring spacecraft
>rotation would be ring laser gyro. Sensor itself is all
>optics; big bale of fiber optic cable in older units, chunk
>of low TCE glass with precision beveled prism faces
>in newer ones. ... Nothing moves except electrons
>and photons; things are rugged as hell.

Unfortunately, the standard RLGs *do* have moving parts.  There's a little
bit of cross-coupling between the beams going in opposite directions,
which makes the gyro's output tend to stick a bit at the zero point.  This
is defeated by having the gyro mechanically wobbled back and forth --
"dithering" -- so that it's never sitting at zero rotation rate.  There
are piezoelectric dithering schemes which don't involve moving the whole
gyro, but they are a more recent development and I'm not sure how widely
used they are yet.

My impression is that the fiber gyros don't have this problem and don't
need dithering, but I don't know that for sure.  They're quite new.

>...Suspect that they are
>not used in spacecraft because they are too new, don't yet have
>enough statistical failure rate data logged to keep NASA
>reliability engineers happy.

They're showing up in new spacecraft, but remember that Hubble is old, and
often it costs less to plug in a replacement part of the same type than to
re-engineer for a new design.
--
The good old days                   |  Henry Spencer   henry@spsystems.net
weren't.                            |      (aka henry@zoo.toronto.edu)


From: henry@spsystems.net (Henry Spencer)
Newsgroups: sci.space.tech
Subject: Re: Short Lifespan of Gyroscopes in Space?
Date: Sun, 14 Mar 1999 07:20:33 GMT

In article <7cf4im$lco@lace.colorado.edu>,
Frank Crary <fcrary@rintintin.Colorado.EDU> wrote:
>>You have one undetermined rotation axis...
>
>No, this isn't a problem. A conventional compass wouldn't do, but
>that's an instrument problem. A set of flux gate magnetometers
>can measure all three components of the magnetic field relative
>to the spacecraft.

Unless I've missed something, Frank, that gives you the strength and
direction of the field... but rotation around the field axis is still
undetermined.  To sense that instantaneously, you'd have to be able to
measure the field *gradient*, which is pretty damn small.

If you are not in a hurry, mind you, you can fix this by watching the time
variation of the field (after compensating for a fixed -- well, you hope
it's fixed -- spacecraft rotation rate).  In fact, sufficiently determined
monitoring of the field over many minutes can even give you enough data
for fairly precise orbit determination.

>I believe Hubble has both RSUs and momentum wheels (i.e. gyros
>for sensing and gyros for turning), but I'm not sure about that.

Correct.  The ones for turning are called reaction wheels in some
descriptions, but I have the impression that they are actually gyrodynes
(aka control moment gyros) -- the terminology here is confused and
inconsistent.

(One consistent set of terms is:  A reaction wheel turns to turn the
spacecraft and stops to stop it.  A momentum wheel is similar except that
it's biased toward rotation in one direction, with a non-zero spin rate
when the spacecraft is stopped, and higher or lower rate for turning; this
gives you some of the same benefits, and some of the same problems, as a
spin-stabilized spacecraft.  A gyrodyne, aka control moment gyro, holds a
fixed speed and is rotated around another axis to turn the spacecraft by
precession -- "pushing against the gyro", roughly speaking.)
--
The good old days                   |  Henry Spencer   henry@spsystems.net
weren't.                            |      (aka henry@zoo.toronto.edu)


From: henry@spsystems.net (Henry Spencer)
Newsgroups: sci.space.tech
Subject: Re: Short Lifespan of Gyroscopes in Space?
Date: Tue, 16 Mar 1999 04:30:31 GMT

In article <7ch8iv$or5$1@samba.rahul.net>,  <jeff@computer.com> wrote:
>	   ...I bet the Hubble operators would be quite
>	   happy right now, if they had 2 or 3 more gyros on board
>	   that they could just fire up right now.

They'd be even less happy than they are right now if Hubble hadn't
*already* had three spares.  There is a limit to how many spares you can
reasonably carry.

>	2) Since some of the use of the Gyroscopes is for small
>	   repositioning of the craft, could you use a flywheel
>	   instead?

The gyroscopes which are failing are sensors only, and their mass is
irrelevant.

>	3) Could you mkae a gyroscope or flywheel that had no
>	   bearings - using a magnetic suspension? - so you had
>	   no wear-and-tear of any mechanical nature?

I don't think anyone has done this for sensor gyros, presumably because
it's too easy to perturb the motion of the rotor.  For reaction wheels and
such, magnetic suspensions are now quite common.
--
The good old days                   |  Henry Spencer   henry@spsystems.net
weren't.                            |      (aka henry@zoo.toronto.edu)


From: henry@spsystems.net (Henry Spencer)
Newsgroups: sci.space.tech
Subject: Re: Short Lifespan of Gyroscopes in Space?
Date: Sat, 13 Mar 1999 22:13:30 GMT

In article <7c6lk9$nse$1@samba.rahul.net>,  <jeff@computer.com> wrote:
>...now we hear that all 3 of Hubbles backup
>gyroscopes have failed.  Hubble started with 6, but they need a minimum of
>3 to work, so they are down to just 3 and one of those is flaky.

As I recall, it's not quite that bad:  they're down to 4 with one flaky,
so if the flaky one goes they have no spares left.  This is making them
nervous but it's not a disaster yet.  (They faced this exact decision a
while before the first servicing flight, by the way -- two gyros dead and
one ailing -- and decided to leave the flight on its original schedule.)

>	1) Why do these fail so often?  Is this just chance?

Traditional gyros are historically a trouble area -- including spares is
common -- at least partly because the behavior of bearings and lubricants
in free fall is a tricky and poorly-understood area.

>	2) If in fact they are unreliable, then why didn't
>	   NASA replace them, on Hubble, during the last
>	   maintenance mission...

Lots of work to do and this didn't seem urgent.  The first servicing
flight replaced three of them, and I believe they were all in good shape
at the time of the second visit.

> why don't craft like SOHO have more backups; etc

Backups are heavy.

>	3) Are there no alternatives to a physical gyroscope?
>	   There ought to be a circuit that could do this that
>	   had no moving parts.

There are now several alternatives to spinning-rotor gyros, and some new
spacecraft are flying with them -- for example, both NEAR and Cassini use
hemispherical-resonator gyros, which still technically have moving parts,
but parts which vibrate rather than spinning.  Ring laser gyros often *do*
have moving parts because of some complications, but I believe fiber-optic
gyros don't, and Clementine flew them.

>	   ...If a spacecraft is near a planet like
>	   Earth that has a magnetic field, then the rotation
>	   of the craft is a movement of this coil through
>	   the Earth's magnetic field; this ought to provide
>	   something measurable to determine the relative motion.

Magnetometers can be, and are, used for rough attitude determination, but
the usual types aren't as precise as gyros, and also they don't give a
full three-axis orientation (you can't sense rotation on an axis parallel
to the field).

There are several kinds of sensor which are okay either for low-precision
work, or in favorable orientations or at favorable times (for example, Sun
sensors do not work in eclipse), but nothing that quite equals gyros for
precision and the ability to work at all times in all positions.
--
The good old days                   |  Henry Spencer   henry@spsystems.net
weren't.                            |      (aka henry@zoo.toronto.edu)


From: henry@spsystems.net (Henry Spencer)
Newsgroups: sci.space.science
Subject: Re: gyroscopes
Date: Thu, 18 Nov 1999 02:35:30 GMT

In article <3831C54E.3EF9DD67@vtacks.com.net>,
Dan  <stevans@vtacks.com.net> wrote:
>I was just curious why gyroscopes seem to fail so often in spacecraft.

They're very precise machinery with rapidly-moving parts.  Moving parts in
general are trouble in space -- tape recorders are another historically
troublesome subsystem -- and the very rapid rotation of gyros makes it
worse.  The long-term free-fall behavior of lubricants, in particular, is
not as well understood as one might like.

Solutions are now available, fortunately.  NEAR and Cassini are using
tuning-fork gyros with no bearings; some of the latest optical gyros have
no moving parts at all.  (Older laser gyros typically *do* have moving
parts, because of some complications in their design.)
--
The space program reminds me        |  Henry Spencer   henry@spsystems.net
of a government agency.  -Jim Baen  |      (aka henry@zoo.toronto.edu)


From: henry@spsystems.net (Henry Spencer)
Newsgroups: sci.space.science
Subject: Re: gyroscopes
Date: Fri, 19 Nov 1999 04:35:37 GMT

In article <942961015.345.0.nnrp-03.9e98d142@news.demon.co.uk>,
Ian Stirling  <root@mauve.demon.co.uk> wrote:
>>They're very precise machinery with rapidly-moving parts.  Moving parts in
>>general are trouble in space...
>
>Is there a reason why momentum wheels at least have not been made with
>magnetic bearings?

They are being made that way, now.  Magnetic bearings add complexity and
some mass -- ball bearings are simple and relatively light -- but they do
confer some advantages, and they're slowly taking over.  (Remember, this
is a very conservative business, and getting something flown for the first
time is particularly difficult because nobody wants to be the guinea pig.)

Unfortunately, I don't believe they've been made to work for sensing gyros,
as opposed to reaction wheels / momentum wheels / gyrodynes.

(Incidentally, Hubble's problems are with sensing gyros, not with the
gyrodynes -- reaction wheels?  I forget -- that it uses as actuators.)

>A solution with a ball, with a pattern of poles on it's surface, suspended
>inside several drivers would seem to offer a 3 axis positioning solution,
>though would need a bit of development.

The usual preference is to use three wheels instead.  Simpler, on the
whole, and by adding a fourth wheel set at an angle to the rest, you can
lose any one wheel and still have full control.

Mind you, some of the fancier magnetic-bearing wheels can play cute games
that actually come pretty close to this.  With a well-designed magnetic
bearing and good controls, you can *move* the spin axis a small but useful
amount on command; unlike a ball bearing, it's not constrained to a single
axis position.
--
The space program reminds me        |  Henry Spencer   henry@spsystems.net
of a government agency.  -Jim Baen  |      (aka henry@zoo.toronto.edu)


From: Brett Buck <brett.buck@lmco.com>
Subject: Re: Gyros - Why required for Guidance?
Date: 14 Oct 1998
Newsgroups: rec.models.rockets

Jeff Taylor wrote:
>
>    Mark Kempisty <xyzMKEMPISTY@GI.COM> writes:
> Several non-mechanical systems
> > have been devised including ring laser gyros where laser light goes into
> > a loop of fiber optics.  As the body twists the light takes a little
> > less of more time to go around the loop.

    Sort of. A ring laser gyro sends a coherent laser beam around a (typically
triangular) path with mirrors at the corners. It sends it in both directions
and it start out in-phase.  If there is a rate one path gets slightly shorter
that the other, and the phase shift of the light is proportional to the rate.
The phase shift is determined by interferometry. The gyro tends to "lock-in"
at low rates, so the entire trinagular block is usually dithered (jiggled back
and forth) so that it never sees 0 rate.

   I think you are referring to a fiber-optic gyro, which is similar except
the triangular path is replaced by a huge length of fiber-optic (maybe 1/2
mile) lopped around. They are easier and cheaper to make, but they have
reduced performance.

   The cutting edge is the hemispherical resonator gyro. The gyro uses a
ceramic "wine glass" sensor that is excited to oscillation just like a wine
glass "sings" if you rub you  finger around the rim. The rate of change of
orientation of the standing waves is a measure of the external rate. These
have been used for the first time in space (to my knowledge) on the Cassini
mission. The mission duration is such that no one thought that the bearings on
 little wheel spinning at 14,400 rpm would last.

>
> Cool, never heard of these before.  Will they fit in my rocket?
>

    Depends on the rocket, and your budget.

   Brett


From: jtkare@ibm.net (Jordin Kare)
Newsgroups: sci.space.tech
Subject: Re: Inertial Navigation Systems: how cheap?
Date: Fri, 19 May 2000 20:35:12 GMT

In article <3923f017.0@cfanews.harvard.edu>, Gregg Germain
<gregg@elway.harvard.edu> wrote:

> Zubrin is holding meetings every week or so in Colorado with about 12
> local Mars Exploration Zealots.  They are planning a field trip into
> the rocky mountains to simulate exploring the surface of Mars. They
> are introducing communications times delays etc to try and make it a
> little more realistic.
>
>  A good friend of mine atends these meetings and tells me what they
> are planning. That led us into an email discussion about how to
> navigate around the martian surface. There are problems such as no
> North Star, weakand variable magnetic field (i.e. no compass).
>
>  We were wondering what the state of the art of Inertial Navigation
> Systems were. How cheap are the ring laser gyro based systems? How
> accurate? Can they be easily carried by a man in a space suit?
>
>
>  One person in sci.military.naval suggests that they can build a RLG
> system on one PC board for around $1000. Does anyone know if this is
> true? Does anyone know how accurate they would be?  I surmise that the
> technology hasn't been pursued with utmost vigor once GPS arrived, and
> wonder how accurate and lightweight they could be made.
>
>  Thanks

State of the art a few years ago (1993, when Clementine flew) was RLG's or
fiber gyros about the size of half a soda can, and around 1 kg.  MEMS
accelerometers are even smaller.  However, the cost was considerably more
than $1K (more like $50K, IIRC) even for the gyro alone, and I'd be
surprised if it had come down below $10K for a commercial package.  (In
principle the parts for an RLG or FOG are quite cheap these days, but
building one from scratch would take considerable work).

The difficulty with such small systems is that they have comparatively
high drift rates, measurable in degrees/day (and in some cases
degrees/hour) for gyro drift, so they need frequent calibration.
Depending on your exploration scenario, you could recalibrate using star
sightings (as long as you didn't plan to travel during sandstorms), or by
using RF links to one or two satellites (range/range rate data over a
horizon-to-horizon pass can provide pretty good location accuracy).

For navigation within a limited area, there are various ways to use one or
more beacon transmitters, either alone or in combination with satellite
data. You probably wouldn't want to use GPS pseudolite technology directly
(since it depends on the satellite constellation for timekeeping) but
time-difference-of-arrival with 3 stations will give a 2-D position with
good accuracy using only a passive receiver.  Even a single beacon (at
your base station) with an active-transponder portable unit and a suitable
direction-finding antenna system at one end (either end) of the link can
give accurate range and useful bearing information.   However, for best
accuracy you're talking about fairly sophisticated semi-custom hardware,
so the cost is back upwards of $10K for a terrestrial system.  "On the
cheap"  you might be able to cobble something together out of ham gear and
PC's for well under $10K.

For the sake of amusement, you might want to figure out how accurately you
can locate a suitable RF source on Mars using Long Baseline Interferometry
on Earth.    A rough handwave:  at X band (3 cm wavelength), resolving
1/100 fringe (0.3 mm) on a 3000 km baseline gives an angular resolution of
10^-10 radians.  At 100 Gm, that's a position error of 10 meters; even at
300 Gm it's only 30 m.  Pretty good.  But it only works on the
Earth-facing hemisphere, and it breaks when Earth is behind the Sun.

Jordin Kare




Newsgroups: sci.space.policy
From: henry@spsystems.net (Henry Spencer)
Subject: Re: State of the Art of Inertial navigation Systems
Date: Sat, 20 May 2000 18:13:59 GMT

In article <3925629c.0@cfanews.harvard.edu>,
Gregg Germain  <gregg@elway.harvard.edu> wrote:
> In thinking over navigating on the Martian surface (where you have no
>North Star and you cannot use a magnetic compass and no GPS), you have have to
>have a way of locating the lat long of interesting things you run
>across, and figure out the way back to base camp.

If you are using a rover -- which you'd want for any substantial excursion
anyway -- the obvious thing to do is what the Apollo lunar rover did:  a
gyrocompass for direction, and a wheel-revolution counter for distance.
The gyrocompass can realign itself any time the vehicle is stopped for a
while, by measuring Mars's rotation.  (That method is routinely used to
align inertial navigation systems on Earth.)

There would be some buildup of error due to wheel slippage etc., but it
should be manageable.  The navigation system could recalibrate itself a
few times a day using satellite passes.  (An expedition might not want to
take along a whole GPS constellation, but one or two small satellites
would probably suffice for doing occasional corrections.)

> - be easily man portable without too much weight and not requiring too
>much power.

I don't think they're quite at that point, not if you want high accuracy
over long operating times.  Missile INSs can be very small and light, but
they are designed for flight times measured in seconds.  The limiting
factor, in general, is gyro drift, which is proportional to operating time
and inversely proportional to the cost and weight of the gyros. :-)

> - advanced enough to not need constant re-alignment

INSs generally do need frequent re-alignment, unless you are prepared to
go to very heavy and expensive systems designed for things like submarines.
Nobody frets much if you're half a kilometer off at the end of flying the
Atlantic.
--
Microsoft shouldn't be broken up.       |  Henry Spencer   henry@spsystems.net
It should be shut down.  -- Phil Agre   |      (aka henry@zoo.toronto.edu)


Newsgroups: sci.space.policy
From: henry@spsystems.net (Henry Spencer)
Subject: Re: State of the Art of Inertial navigation Systems
Date: Mon, 22 May 2000 18:08:38 GMT

In article <3929412c.0@cfanews.harvard.edu>,
Gregg Germain  <gregg@elway.harvard.edu> wrote:
>: I don't think they're quite at that point, not if you want high accuracy
>: over long operating times.  Missile INSs can be very small and light, but
>: they are designed for flight times measured in seconds.  The limiting
>: factor, in general, is gyro drift, which is proportional to operating time
>: and inversely proportional to the cost and weight of the gyros. :-)
>
>	What about ring laser gyros? They are fairly light ar ethey not?

Yes. :-)  As with mechanical gyros, it depends on how much you spend and
how much you want them to weigh.  The early ones were big heavy boxes.
Some of the modern ones are smaller and lighter, but they're still
substantial pieces of machinery.  (And they usually have moving parts, by
the way -- there are some little problems which are tricky to solve with
all-solid-state solutions.)

>How does their accuracy compare to spinning chunks of metal?

If you're willing to spend enough money and mass, equal or better, at
least in normal applications.  I'm not sure whether they're yet up to
equalling the *very best* mechanical gyros (which are complex, very heavy,
and very very expensive), but they're pushing mechanical gyros out of the
usual applications like aircraft and space-launcher navigation.

>: INSs generally do need frequent re-alignment, unless you are prepared to
>: go to very heavy and expensive systems designed for things like submarines.
>
>   However Apollo fretted over more stringent accuracy requirements
>did they not? I wonder how often they re-aligned?

A couple of times a day, I believe, including just before any substantial
maneuver.  (That's why the navigation system included optics as well as
inertial hardware.)  Spacecraft inertial systems usually are not required
to be especially low-drift, because they don't have to hold high accuracy
*without* *realignment* for great lengths of time -- generally no more
than the duration of a maneuver.

> Read the Apollo 1 fire report recently where they detected crew
>movement within the capsule because the INS system was very sensitive.

Note that short-term sensitivity and long-term drift rate are different
things.
--
Microsoft shouldn't be broken up.       |  Henry Spencer   henry@spsystems.net
It should be shut down.  -- Phil Agre   |      (aka henry@zoo.toronto.edu)

Index Home About Blog