Tim Wescott wrote:>> Unfortunately, the best laser gyros are orders of magnitude less >> accurate then the best mechanical ones. The mode locking and the noise >> are the problems.> Someone should take the time to check -- my understanding is that ring > laser gyros that work in free space are pretty darn good. > > I have experience with fiber-optic laser gyros, or at least their > specifications. For a given volume & weight I've always seen better > specifications from mechanical gyros, so when they can be found I have > used mechanical ones.I have dealt with the laser ring gyros somewhat 16 years ago. Although the technology does improve with time, they do have the fundamental limits of the performance by the nature of their operation. There is no doubt that the laser gyros are very convenient, however they can't get close to the performance of the best mechanical gyro.> For a good long time the defense department was actively pushing FOGs > because they were going to be ever so cool compared to those nasty > greasy mechanical things.As far as I know, if the goal is the ultimate accuracy, the INSes are still built using the mechanical gyros. To date, that promise hasn't materialized,> but because of their pushing there are a lot of systems using FOGs, and > not very many vendors selling good mechanical gyros.Perhaps, the accuracy offered by the FOGs is good enough for most of the practical applications. Vladimir Vassilevsky DSP and Mixed Signal Design Consultant http://www.abvolt.com
Rate Gyros and Accelerometers
Started by ●April 22, 2007
Reply by ●April 23, 20072007-04-23
Reply by ●April 23, 20072007-04-23
Tim Wescott <tim@seemywebsite.com> wrote in news:RYKdnbEDwNKefbHbnZ2dnUVZ_qbinZ2d@web-ster.com:> http://www.analog.com/UploadedFiles/Data_Sheets/ADIS16120.pdfThis IS a mechanical device! The tuning fork is actually vibrating at resonance, and coriolis is being measured. Micromachining has come a long way in a decade. I've used Watson rate sensors, which use the tuning fork technology, and they drifted big time. This chip doesn't drift nearly so badly. The spec sheet I posted lists 0.04% "typical" nonlinearity over full scale, and 0.02 deg/sec/sqrt(Hz) noise. The Null drifts 0.05 volts over a 100 deg Celcius temperature range, with about a 2% sensitivity change over that range. At constant temp, the Null one-sigma seems to be 0.005 deg/sec "typical", given a good Vcc. Also, this technology is very insesitive to linear acceleration, 0.05 deg/sec/g. Given an ample warmup time, this looks pretty good. In fact, it looks better than most people could align and calibrate a device like this. Do the mechanical devices really go to 0.004% nonlinearity, with an acceleration sensitivy of 0.005 deg/sec/g?? -- Scott Reverse name to reply
Reply by ●April 23, 20072007-04-23
"Tim Wescott" <tim@seemywebsite.com> wrote in message news:RYKdnbEDwNKefbHbnZ2dnUVZ_qbinZ2d@web-ster.com...> Scott Seidman wrote: > > Tim Wescott <tim@seemywebsite.com> wrote in news:rdSdnR- > > 5h58iR7HbnZ2dnUVZ_rGinZ2d@web-ster.com: > > > >> Someone should take the time to check -- my understanding is that ring > >> laser gyros that work in free space are pretty darn good. > > > > http://www.analog.com/UploadedFiles/Data_Sheets/ADIS16120.pdf > > > > This is an Analog Devices MEMS gyro. It uses an orthogonal tuning fork. > > > Right. And its performance is several orders of magnitude worse than > the mechanical gyros I work with. Its value comes about because it is > small and cheap. > > -- > > Tim Wescott > Wescott Design Services > http://www.wescottdesign.com >Here are some hard numbers: For a "decent" quality inertial navigation system (INS), your gyro should be better than around 0.01 deg/hour (drift). This will yield an INS with mile-an-hour accuracy, meaning that after an hour of purely unaided flight (i.e., unaided by GPS or other position updates), the INS will have a position error of around one nautical mile. This is decent enough for most commercial airlines. The cited MEMS gyro lists its drift as 0.005 deg/sec, which when multiplied out is 18 deg/hour. Clearly, such a gyro cannot be used for long-term inertial navigation. Why the difference? Inertial-grade gyros do not actually measure angular rate; rather, they measure delta-angle. The best gyros are spinning-wheel gyros on a 2D hinge, and the output is a measure of the work needed to maintain the wheel at a null position. Hence, the output is (in a sense) the integral of angular rate since the last measurement, or delta-angle. The rate gyros, on the other hand, output pure angular rate. There's a lot less information inside of a rate measurement, as opposed to the delta-angle information from high-quality gyros. For one example, suppose you integrate the output of a rate gyro in order to estimate angle. Since each individual measurement has noise, and since there are Nyquist issues with the bandwidth of the angular rate actually experienced by the vehicle as opposed to the bandwidth of the device, the integrand will have square-root growth in error-over-time on the estimate of angle. On the other hand, since the output of high-quality gyros is delta-angle, a simple sum suffices to obtain an estimate of angle. Moreover, since the device is essentially a mechanical integrator, there are not the same Nyquist issues as in a rate gyro. Furthermore, the error caused by measurement noise in estimation of the angle is constant, and does not grow over time. Ring laser gyros can be designed to achieve INS-quality 0.01 deg/hour drift, and also to output delta-angle. Such gyros will count the number of interference fringes, created by counter-rotating laser beams, that pass in front of a sensor. Fiber-optic ring lasers do not do this. They merely output the current number of interference fringes visible on a sensor. They are therefore lower-grade rate gyros. The gyro from the cited article is a tuning-fork design that in principle of operation is similar to so-called Watson devices. These are pure rate gyros. Each type of gyro has its purposes/uses. Mostly, rate gyros are used for feedback stabilization and control of a vehicle's angular rotation. For example, when you are trying to zero-out the angular rotation of a vehicle, you wouldn't care that there might be a residual rate of 0.005 deg/sec (or 18 deg/hour). In the short-term, such low angular rotation is meaningless. Likewise, the delta-angle devices in INSs are not usable for feedback control and stabilization of a vehicle, precisely because of the low bandwidth caused by their integrating effect. Rate gyros can sometimes be used for very-short-term inertial guidance, where position error is measured in feet (not miles), like in a GPS-aided navigation system. The mile-an-hour curve in position error initially grows like cosine(time), so initial growth in the error is slow. A rate gyro with 2.0 deg/hour accuracy (which is 10 times better than the cited gyro, but still 200 times worse than an INS-quality gyro) will have a position error of 200 feet after 200 seconds of flight. If this is acceptable, then such a rate gyro can be used for both control and for navigation. Mike
Reply by ●April 23, 20072007-04-23
"Michael K. O'Neill" <MikeAThon2000@nospam.hotmail.com> wrote in news:QI7Xh.7474$H_5.1218@newssvr23.news.prodigy.net:> Here are some hard numbers:Very nice summary. -- Scott Reverse name to reply
Reply by ●April 23, 20072007-04-23
Michael K. O'Neill wrote: [very nice summary snipped] It might interesting to know that the first trial of an airborne INS took place from Bedford Air Force Base around 1953. It was conducted by a team of the MIT Instrumentation Lab under Charles Draper. At the time, the words "inertial" and "navigation" in the same sentence were classified. The equipment was mounted in a bomber which flew a pretty direct route out past Nahant and back. When they landed about half an hour later, the readout indicated that they were over Nova Scotia, going 700 knots straight up. The entire system used AC servos which required a stable AC voltage reference. The reverence in that first flight had mechanical elements that couldn't withstand the vibration. The second run used an entirely electronic reference that was much less accurate on the ground, but more satisfactory in flight. I debugged some of that equipment even though I didn't know what it was for. One of my housemates designed many of the amplifiers. He was a whiz at theory, but useless at fixing things. (Both he and his technician were color blind.) Al would bring home a redacted schematic, with waveforms copied from a scope or DC voltages at various points -- sometimes I would send him back for more -- and I would tell him what to fix. For one fix, I marked the grid of a pentode and told him that it was somehow shorted. It turned out that the grid-leak resister which should have been a megohm was in fact ten ohms. (Black brown brown) instead of (black brown green). Close enough! Jerry P.S. The best technician I ever had was colorblind. We got an early (rack-mount size) digital ohmmeter on the strength of that disability. -- Engineering is the art of making what you want from things you can get. ¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
Reply by ●April 24, 20072007-04-24
On Apr 24, 8:33 am, Jerry Avins <j...@ieee.org> wrote:> Michael K. O'Neill wrote: > > [very nice summary snipped] > > It might interesting to know that the first trial of an airborne INS > took place from Bedford Air Force Base around 1953. It was conducted by > a team of the MIT Instrumentation Lab under Charles Draper. At the time, > the words "inertial" and "navigation" in the same sentence were > classified. The equipment was mounted in a bomber which flew a pretty > direct route out past Nahant and back. When they landed about half an > hour later, the readout indicated that they were over Nova Scotia, going > 700 knots straight up. > > The entire system used AC servos which required a stable AC voltage > reference. The reverence in that first flight had mechanical elements > that couldn't withstand the vibration. The second run used an entirely > electronic reference that was much less accurate on the ground, but more > satisfactory in flight. > > I debugged some of that equipment even though I didn't know what it was > for. One of my housemates designed many of the amplifiers. He was a whiz > at theory, but useless at fixing things. (Both he and his technician > were color blind.) Al would bring home a redacted schematic, with > waveforms copied from a scope or DC voltages at various points -- > sometimes I would send him back for more -- and I would tell him what to > fix. For one fix, I marked the grid of a pentode and told him that it > was somehow shorted. It turned out that the grid-leak resister which > should have been a megohm was in fact ten ohms. (Black brown brown) > instead of (black brown green). Close enough! > > Jerry > > P.S. The best technician I ever had was colorblind. We got an early > (rack-mount size) digital ohmmeter on the strength of that disability. > -- > Engineering is the art of making what you want from things you can get. > =AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF==AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF= =AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF surely the V2 German rockets pre-date this?
Reply by ●April 24, 20072007-04-24
gyansorova@gmail.com wrote:> On Apr 24, 8:33 am, Jerry Avins <j...@ieee.org> wrote: >> Michael K. O'Neill wrote: >> >> [very nice summary snipped] >> >> It might interesting to know that the first trial of an airborne INS >> took place from Bedford Air Force Base around 1953. I...> surely the V2 German rockets pre-date this?I don't know what they had on board. Marine torpedoes were steered by gyroscopes, but as far as I know, neither they nor the V2s had stable platforms. Jerry -- Engineering is the art of making what you want from things you can get. ¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
Reply by ●April 24, 20072007-04-24
Jerry Avins wrote:>> surely the V2 German rockets pre-date this? > > I don't know what they had on board.V2 had a simple appliance-like timer which controlled the gas rudders and the other elements. As far as I know, there was no stabilization or feedback at all. VLV
Reply by ●April 24, 20072007-04-24
"Jerry Avins" <jya@ieee.org> wrote in message news:4v-dnegFuraW5LPbnZ2dnUVZ_judnZ2d@rcn.net...> gyansorova@gmail.com wrote: > > On Apr 24, 8:33 am, Jerry Avins <j...@ieee.org> wrote: > >> Michael K. O'Neill wrote: > >> > >> [very nice summary snipped] > >> > >> It might interesting to know that the first trial of an airborne INS > >> took place from Bedford Air Force Base around 1953. I > > ... > > > surely the V2 German rockets pre-date this? > > I don't know what they had on board. Marine torpedoes were steered by > gyroscopes, but as far as I know, neither they nor the V2s had stable > platforms. > > JerryAgree. V2's were *stabilized* by gyros, but they were not *guided* by gyros. The V2 was basically point-and-shoot, with a purely ballistic trajectory. PS to Jerry: Draper Labs still makes the finest gyros on the planet.
Reply by ●April 24, 20072007-04-24
Jerry Avins <jya@ieee.org> writes:> gyansorova@gmail.com wrote: >> On Apr 24, 8:33 am, Jerry Avins <j...@ieee.org> wrote: >>> Michael K. O'Neill wrote: >>> >>> [very nice summary snipped] >>> >>> It might interesting to know that the first trial of an airborne INS >>> took place from Bedford Air Force Base around 1953. I > > ... > >> surely the V2 German rockets pre-date this? > > I don't know what they had on board. Marine torpedoes were steered by > gyroscopes, but as far as I know, neither they nor the V2s had stable > platforms.What's a stable platform? A type of woman's shoe? -- % Randy Yates % "So now it's getting late, %% Fuquay-Varina, NC % and those who hesitate %%% 919-577-9882 % got no one..." %%%% <yates@ieee.org> % 'Waterfall', *Face The Music*, ELO http://home.earthlink.net/~yatescr
