Featherweight Tilt checker/timer

The Rocketry Forum

Help Support The Rocketry Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

Adrian A

Well-Known Member
TRF Sponsor
TRF Supporter
Joined
Jan 21, 2009
Messages
3,157
Reaction score
2,806
Location
Lakewood, CO
After some requests on the TRA email list, I decided to try my hand at making a simple, small, inexpensive, gyro-based device that measures the rocket flight angle and inhibits an airstart ignition if the flight angle exceeds a user-selected angle threshold. This would be like Frank Hermes' Rocket Tiltometer, but without as many extra features. I have picked out the major parts and completed part of the layout, so I think it's time to finalize the features and interface, and for that I'd like any feedback people are interested in providing.

Here are the specs:
  • 3-axis gyro, and 3-axis accelerometer, and a bunch of math in a microcontroller determine the rocket's orientation before and during the flight.
  • 1.5" long by 0.75" wide
  • One switch connected to the + half of the circuit will prevent current to the ignitor by opening a solid-state switch when the flight angle exceeds a user-selected threshold.
  • You can select the no-go angle from 2-34 degrees in increments of 2 degrees by turning a rotary DIP switch with a screwdriver.
  • The board also has a solid-state switch that acts like a simple rocketry timer. You can connect the switches together to make a stand-alone tilt-safe timer.
  • The timer values are selected by another rotary dip switch, that can select from 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 15, 20, or 30 seconds from liftoff, or an apogee setting that lets this switch be used by itself for a backup apogee charge.
  • Single battery needed, 4-16V, up to 30 Amps current capability.
  • LEDs to show when the unit has power, when the timer output has continuity, and when the tilt threshold is exceeded. I may also provide an LED and a pair of terminals to bring out a pad health check status remotely.
  • No beeper, but the tilt meter is designed to work with the continuity check for most altimeters so there's just one device beeping status at the pad.
  • No data recording
  • Cost: Hopefully under $70

This is still in the early design stages, and wouldn't be available for a couple of months, at least.
 
Last edited:
[*]1.5" long by 0.75" wide
[*]Cost: Hopefully under $70

Those two, I think, are huge.

The Tiltometer is a cool device. Its downsides are size and cost, with the latter being significant enough, I think, to prevent most people from considering one.

-Kevin
 
Adrian- oh brudder! If you were a girl, I'd propose! HELLYAH!! Do it and sign me up~! (this will put the final nail in the coffin of Mercury switches-and I don't mean the ones John Glenn flipped!). I've been doing this a looonngg time and seen my share of powered augers doing core samples or smearing payloads all over the landscape. Build it and they will come!
 
Just in time for my next build. Airstarts on an XL Gizmo.




----Yes please----
 
After some requests on the TRA email list, I decided to try my hand at making a simple, small, inexpensive, gyro-based device that measures the rocket flight angle...
Very cool. Do you actually use the accelerometer for anything but launch detection? I don't want to encourage feature creep, but being able to set a delay from burnout detection instead of launch detection would be nice.
 
Sign me up for one from the first batch.

I just happen to be building an timer airstarted rocket right now. It's tricky because the first stage is a 3-way cluster, with the second stage being a single motor. So....lots to go wrong.

And quite honestly I'm a bit nervous about the safety aspect of this rocket. I've already decided that I won't be launching this around crowds - only out in the open desert by myself - to minimize the risks. I simply can't afford to trick it out with current tilt/failsafe options.

But this sounds like EXACTLY what I need. Sweet.

So, again, sign me up.

s6
 
I would get one, since I'm planning an airstarted rocket. It might influence my design, especially if it has a second output for an apogee charge. If it were shorter, that would be nice...

It would be capable of detecting apogee by integration even when the rocket arcs over, right?
 
Very cool. Do you actually use the accelerometer for anything but launch detection? I don't want to encourage feature creep, but being able to set a delay from burnout detection instead of launch detection would be nice.

I plan to use it to also measure the initial angle of the launch rail, for the cases where the thrust doesn't exceed the accel's range. Burnout detection would be feasible. I would like to have the timer start counting at burnout, but I also want it to be able to trigger right at liftoff detection. I think I would need to add a Liftoff/MECO switch in order to make it selectable. If it's selectable, inevitably someone will forget to set it the way they want and be disappointed when the sustainer cooks their booster. Or maybe I'll make the times in general based on MECO, and then reserve a special DIP slot for liftoff the way I plan to do for apogee. Actually, I like that idea best so far. I don't think many people would mind not being able to select a time between liftoff and MECO.

Sign me up for one from the first batch.

I just happen to be building an timer airstarted rocket right now. It's tricky because the first stage is a 3-way cluster, with the second stage being a single motor. So....lots to go wrong.

And quite honestly I'm a bit nervous about the safety aspect of this rocket. I've already decided that I won't be launching this around crowds - only out in the open desert by myself - to minimize the risks. I simply can't afford to trick it out with current tilt/failsafe options.

But this sounds like EXACTLY what I need. Sweet.

So, again, sign me up.

s6
Thanks. I don't know how many people fly airstarts, but maybe this product could help grow that market safely.

I would get one, since I'm planning an airstarted rocket. It might influence my design, especially if it has a second output for an apogee charge. If it were shorter, that would be nice...

It would be capable of detecting apogee by integration even when the rocket arcs over, right?

Yes. My current plan is to make the hookups so that the tilt check switch and the timer switch can be hooked up independently or together. The independent case would be useful for a flight like Jim Jarvis's , where the tilt check can be used together with a separate altimeter that ignites the upper stage at a particular velocity. Then the timer function is freed up to be used for apogee detection, which is when the rocket tilts over 90 degrees. This has the potential to be the most accurate apogee detection for flights over 120kft.

But if you're using the timer switch as a timer, it would not also be available for apogee detection.

Adrian glad you are taking up this challenge. It is a real need and will keep me from relying on my crude solution to this problem.

Thanks. These things always take longer than you expect, though.
 
Last edited:
Wow.. Im so far away from air starting a rocket that I feel like a freshman on the first day of school....

With that said, the concept is dead on.. and the unit will be a must have for me when the time comes, anything you can do to make a rocket safer and aid in its recovery is awesome.

I have a two questions though... If the fail safe trips, and the angle is off, Will that tilt meter fire the ejection charge? and the second question is, Do you sell an "AV" bay completed ?

Example, Dual deployment, with altimiter, tracking and tilt options say in a 4" body tube? or maybe a 3" tube ?

One stop shop for high quality recovery !!!

Thanks and again, Awesome Job... I look forward to giving you some of my money....

Tom
 
Wow.. Im so far away from air starting a rocket that I feel like a freshman on the first day of school....

With that said, the concept is dead on.. and the unit will be a must have for me when the time comes, anything you can do to make a rocket safer and aid in its recovery is awesome.

I have a two questions though... If the fail safe trips, and the angle is off, Will that tilt meter fire the ejection charge? and the second question is, Do you sell an "AV" bay completed ?

Example, Dual deployment, with altimiter, tracking and tilt options say in a 4" body tube? or maybe a 3" tube ?
Tom

Thanks, Tom. Once the flight angle gets large enough after launch for the tilt check switch to open up, it won't close again during the flight. If the trigger switch is set up to fire at apogee, it will still do that.

I sell complete av-bay kits that cover just the altimeter, battery, and arm switch functions. In some of them, you can squeeze a Beeline tracker along for the ride. The Telemetrum is the most complete unit so far, for combining GPS tracking and basic deployments.
 
Out of curiosity, what would it do to the cost to include some basic recording?

Specifically, for those going for high altitudes, being able to know angle over time might be useful.

-Kevin
 
Although recording would be nice, my vote is to make this as simple as possible since the device is in series with mission-critical functions, i.e., a "Missileworks PET2+" with a tilt detect switch.
 
Although recording would be nice, my vote is to make this as simple as possible since the device is in series with mission-critical functions, i.e., a "Missileworks PET2+" with a tilt detect switch.

I was about to agree and to suggest that someone probably makes a small altimeter with recording capabilities that you could use along with the tilt detector.

But, on second thought, the recording capabilities of Adrian's altimeters have proven themselves to be very useful for determining what happened when things haven't gone exactly right.

Having capability in the tilt checker to record some data and to log events might be helpful in the same kind of situations.

If anything, it might allow Adrian to say, "Hey, dummy. You set the timer to fire two-seconds after lift-off, not burnout," or whatever. :)

-- Roger
 
Out of curiosity, what would it do to the cost to include some basic recording?

Specifically, for those going for high altitudes, being able to know angle over time might be useful.

-Kevin

Yes. And it would be especially useful, if not necessary, during product development, so I can find out how well the attitude determination is on its test flights. For free (other than development time) I can store some flight data in leftover volatile RAM in the microcontroller, and bring out a serial interface to some holes on the board. Then use hyperterminal to download it using a serial/USB converter that I would have to make separately. That's how I'm leaning at the moment.

I'm not sure if I would even advertise that capability, though. I have found that just by making it known that you can hook up a computer to something, it turns off a lot of people who think that means that they have to.

Although recording would be nice, my vote is to make this as simple as possible since the device is in series with mission-critical functions, i.e., a "Missileworks PET2+" with a tilt detect switch.
It's funny that you mention that, because I just read the PET2+ timer manual last night, and that's a pretty capable and complicated device. This would definitely be a PET2-- as far as flexibility and options go. I'm trying hard to include only the options that can fit on a 16-position selector switch.

I was about to agree and to suggest that someone probably makes a small altimeter with recording capabilities that you could use along with the tilt detector.

But, on second thought, the recording capabilities of Adrian's altimeters have proven themselves to be very useful for determining what happened when things haven't gone exactly right.

Having capability in the tilt checker to record some data and to log events might be helpful in the same kind of situations.

If anything, it might allow Adrian to say, "Hey, dummy. You set the timer to fire two-seconds after lift-off, not burnout," or whatever. :)

-- Roger

If I were going to do a good job on the recording, I would need a separate microcontroller, because keeping up with the rocket's rotations will consume almost all of the CPU throughput. On the Raven, recording all the data consumes almost all of the CPU throughput, so I'm not planning to make a combined product anytime soon.

By the way, thanks for the feedback. This is the kind of discussion I was hoping for.
 
You say "single battery needed"; would that mean that I could fly a Raven and this on a single battery? I interpreted that the tilt checker would have a backup capacitor just like the Raven does, so that when the battery voltage drops under load, each would keep plugging away on stored power.

Also, what sort of added impedance would the tilt-check switch add to the circuit? Depending on the supply battery's internal resistance, its voltage, and the igniter(s) used, could this cause airstarts to fail in cases where they wouldn't have?
 
Good info, Adrian.

Recording the orientation data would be cool, but if it ads bulk or significantly adds to the cost, it's not worth it. Size and cost are two major plusses to this endeavor!

Having a separate document available on how to extra data via the serial port, for folks who would want, would be cool. Possibly even offer a small header wire that can be attached and left on the board.

-Kevin
 
You say "single battery needed"; would that mean that I could fly a Raven and this on a single battery? I interpreted that the tilt checker would have a backup capacitor just like the Raven does, so that when the battery voltage drops under load, each would keep plugging away on stored power.

Also, what sort of added impedance would the tilt-check switch add to the circuit? Depending on the supply battery's internal resistance, its voltage, and the igniter(s) used, could this cause airstarts to fail in cases where they wouldn't have?

Yes, you could run a Raven and the tilt checker on a single, shared battery. I still need to size the hold-up capacitor for this device, but I'm hoping I don't need the size or expense of the one used on the Raven.

The tilt-check will add very little resistance to the circuit. The FET resistance is around 13 mOhms, IIRC, and it's right next to the terminal. For comparison, a 130mAhr cell has about 500 mOhms internal resistance, and a 9V battery has about 2000.
 
After some requests on the TRA email list, I decided to try my hand at making a simple, small, inexpensive, gyro-based device that measures the rocket flight angle and inhibits an airstart ignition if the flight angle exceeds a user-selected angle threshold. This would be like Frank Hermes' Rocket Tiltometer, but without as many extra features. I have picked out the major parts and completed part of the layout, so I think it's time to finalize the features and interface, and for that I'd like any feedback people are interested in providing.

Here are the specs:
  • 3-axis gyro, and 3-axis accelerometer, and a bunch of math in a microcontroller determine the rocket's orientation before and during the flight.
  • 1.5" long by 0.75" wide
  • One switch connected to the + half of the circuit will prevent current to the ignitor by opening a solid-state switch when the flight angle exceeds a user-selected threshold.
  • You can select the no-go angle from 2-34 degrees in increments of 2 degrees by turning a rotary DIP switch with a screwdriver.
  • The board also has a solid-state switch that acts like a simple rocketry timer. You can connect the switches together to make a stand-alone tilt-safe timer.
  • The timer values are selected by another rotary dip switch, that can select from 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 15, 20, or 30 seconds from liftoff, or an apogee setting that lets this switch be used by itself for a backup apogee charge.
  • Single battery needed, 4-16V, up to 30 Amps current capability.
  • LEDs to show when the unit has power, when the timer output has continuity, and when the tilt threshold is exceeded. I may also provide an LED and a pair of terminals to bring out a pad health check status remotely.
  • No beeper, but the tilt meter is designed to work with the continuity check for most altimeters so there's just one device beeping status at the pad.
  • No data recording
  • Cost: Hopefully under $70

This is still in the early design stages, and wouldn't be available for a couple of months, at least.

Looks like with a few additions (like recording and a Kalman filter) you would have a full 6-dof flight computer. Any plans to extend the design in this direction?
 
Looks like with a few additions (like recording and a Kalman filter) you would have a full 6-dof flight computer. Any plans to extend the design in this direction?

The CPU burden associated with the attitude propagation and full high-rate recording combined are more than I would be able to handle with the hardware I'm familiar with. I am interested in getting proficient with Kalman filters for several longer-term projects I have in mind, though.
 
Frank Hermes has done the hobby a great service by bringing the RocketTiltometer and Rocket Tiltometer2 to the market. He has some well-deserved awards to that effect, and he has been very generous to me with his knowledge and encouragement. We have talked a little about this project a few days ago, and he was surprised and dismayed to see this discussion thread, because it's so soon in my development process. I would keep this under my hat, but I have found that a collaborative design effort with my customers, early in the process, leads to better products, and that's my top priority.

I want to emphasize that a fully-tested, working product that does this tilt checking function is available right now, from Rocket Electronics. My design is only on the drawing board, and it's quite possible that with the hardware I have, and the algorithms I plan to use, that the tilt checking function I'm going to implement will be crude, at best. That's what Frank believes, and he should know. I can't guarantee that my development will be as accurate as I want, or even successful at all, so all I can do is promise to be open about the performance and accuracy that I do get as this development goes along. If you have an airstart project that you're planning to launch this spring, and you want a gyro-based tilt check, don't wait for me to get this done, get a Rocket Tiltometer2 instead.
 
Though I didn't realize that the RocketTiltometer 2 was available, it doesn't suit my purposes. The size of the RocketTiltometer 2 makes it too large to fit in my planned cluster project.
 
Frank Hermes has done the hobby a great service by bringing the RocketTiltometer and Rocket Tiltometer2 to the market. He has some well-deserved awards to that effect, and he has been very generous to me with his knowledge and encouragement. We have talked a little about this project a few days ago, and he was surprised and dismayed to see this discussion thread, because it's so soon in my development process. I would keep this under my hat, but I have found that a collaborative design effort with my customers, early in the process, leads to better products, and that's my top priority.

I want to emphasize that a fully-tested, working product that does this tilt checking function is available right now, from Rocket Electronics. My design is only on the drawing board, and it's quite possible that with the hardware I have, and the algorithms I plan to use, that the tilt checking function I'm going to implement will be crude, at best. That's what Frank believes, and he should know. I can't guarantee that my development will be as accurate as I want, or even successful at all, so all I can do is promise to be open about the performance and accuracy that I do get as this development goes along. If you have an airstart project that you're planning to launch this spring, and you want a gyro-based tilt check, don't wait for me to get this done, get a Rocket Tiltometer2 instead.

There's a lot of validity in what you say, and by all accounts, Frank does indeed have a fantastic product.

The issue that keeps me from being interested? Cost.

Yep, it slices, it dices, it juliennes. But the idea of adding yet another $250 in electronics to a project makes it that much more difficult to do.

It's kinda like comparing the cost of an RDAS to most of the altimeters out there -- the $400 price tag is what keeps most people I know from even considering it.

-Kevin
 
There's a lot of validity in what you say, and by all accounts, Frank does indeed have a fantastic product.

The issue that keeps me from being interested? Cost.

Yep, it slices, it dices, it juliennes. But the idea of adding yet another $250 in electronics to a project makes it that much more difficult to do.

It's kinda like comparing the cost of an RDAS to most of the altimeters out there -- the $400 price tag is what keeps most people I know from even considering it.

-Kevin
Agreed. In a world where life was fair, the people who would buy the Rocket Tiltometer despite its size and cost would keep buying them, and I would just be selling to everyone else. I don't have that attitude toward many rocketry electronics manufacturers, but Frank is one of them. I'll probably buy one to use as a testing reference.
 
I know the cost of RocketTiltometer (RTOM) up there, but are you balancing it with the waste and financial loss of a sustainer motor that ignites when you don’t want it to? For example, in one of Frank’s test flights he used a CTI M2020 for his sustainer. Retail on that motor is $370. The flight went awry, but the COS feature aborted ignition and saved the motor for a subsequent flight.

I may be missing something, but if one is using expensive sustainer motors it seems like RTOM has the potential to save you some bucks.

- Rich
 
You said that you weren't sure of the accuracy of your algorithms, given the hardware limitations of the microcontroller. But accuracy isn't necessarily the end goal: it is safety. What were you planning to use?

There are plenty of simplifications that you can make if you automatically abort all airstarts once the rocket leaves the region where the small-angle approximation holds to within 15% (30 degrees). That way, you only have to do a 2D transform of the horizontal-axis rate gyroscopes, simplifying the calculations greatly. Additionally, in that case it would always overestimate the rotation angle, erring on the side of safety.

I'm not absolutely sure, but basically any case where this algorithm would abort an airstart when the full algorithm would keep going would be far from an ideal flight. And it would be far simpler to implement, I bet.
 
You said that you weren't sure of the accuracy of your algorithms, given the hardware limitations of the microcontroller. But accuracy isn't necessarily the end goal: it is safety. What were you planning to use?

There are plenty of simplifications that you can make if you automatically abort all airstarts once the rocket leaves the region where the small-angle approximation holds to within 15% (30 degrees). That way, you only have to do a 2D transform of the horizontal-axis rate gyroscopes, simplifying the calculations greatly. Additionally, in that case it would always overestimate the rotation angle, erring on the side of safety.

I'm not absolutely sure, but basically any case where this algorithm would abort an airstart when the full algorithm would keep going would be far from an ideal flight. And it would be far simpler to implement, I bet.

I wasn't planning to make any simplifications, other than skipping the use of a magnetometer in flight. The full DCM transformations needed for the full-up propagation aren't that hard, just a 3x3 matrix multiplication each timestep. If the gyro were perfect, this method can compute the full 3D propagation, even with roll rates up to 2000 dps, without significant errors. But the gyro errors can be significant, even for a 30 second flight. When UAV aircraft do this, they use a magnetometer to help identify and correct gyro bias errors as you go along, which is necessary for a multi-minute flight. That's what the Rocket Tiltometer does, too.

But with the specs of the part I'm planning to use, the error budget should come out with just a few degrees of error on a 30-second flight, if the gyro temperature only changes a few degrees during the flight. The noise and temperature sensitivity are much better than the best gyros available a couple of years ago. I think I need to calculate the worst-case error more rigorously though. There aren't any published figures on the gyro response to constant acceleration, however, so mostly this is just going to require prototyping and testing.
 
Back
Top