Introducing 'AltiLogger', a new high accuracy altimeter

The Rocketry Forum

Help Support The Rocketry Forum:

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

stefel

Active Member
Joined
Oct 22, 2013
Messages
29
Reaction score
0
Hi Everyone!

I've been working on this for the last 6 months so I thought I'd share it with other rocketeers as well to see what people think. I wanted a logging Altimeter that is very accurate and at the same time very lightweight, so I can use it in smaller rockets as well without the weight of the altimeter having too much of an influence on the rocket's flight. I read an article about how temperature can make a huge difference to the measured altitude so the device also logs temperature and altitude calculations derived from pressure will be temperature-compensated. Also included is a 3-Axis accelerometer and magnetometer on the board, which also allows the acceleration and flight direction of the rocket to be logged in detail. There's an integrated USB interface so there's no separate download kit needed, just plug it straight into a laptop with a microUSB cable. All this is squeezed into a board just 20x14mm weighing 2 grams with battery. Below is a picture of the first prototype.

AltiLoggerRev1.jpg

I've done one test flight with it and it worked great. I'm working on an improved version now which I'll be prototyping next month. I'm calling the device 'AltiLogger'.
 
First prototype needed M1 screws for mounting. I'm upgrading to M1.2 for the V2 board as it's much easier to find (commonly used on RC helicopters etc.)
 
Cool!

Any idea on what the software functionality will be? Any interest in developing a way to record motor performance? Any price point targets?
 
I have an early functional version of the data analysis software. It can download data from the device, you can load/save data to/from files and it can do various plots as well. Eventually I'll be developing various algorithms for filtering and deriving different things from the data. One particular area I'm interested in is to actually reconstruct the flight path from the data by integrating the accelerometer data and using magnetometer data as a direction vector, so that I can see how the rocket spins and how the angle of it changed during the flight.

I think that evaluating motor performance is more difficult because the weight of the rocket changes throughout the burn and you also have drag to take into account. With such dynamic conditions it would be difficult to get an accurate thrust curve. If you have a look at the image below from a test flight you can see that the acceleration data closely resembles the shape of the thrust curve of the B6 motor I used and I could estimate the force in Newtons if I know the weight of the rocket before and after the burn, but the weight is not reduced linearly making it very difficult. We can also get a reasonably good indication of air drag from the acceleration graph after burnout. If there was no drag, the rocket would decelerate at a constnat 9.8 m/s^2. If you look at the acceleration curve, deceleration is about 17m/s^2 which reduces to about 9.8 by the time velocity reaches 0 at apogee.

Accelerometer.PNG

The velocity graph (red) on the plot is calculated by integrating acceleration and the distance (green) is derived by integrating velocity. The purple line (height) is the altitude derived from the barometer data.

With regards to price, I would like to make it as cheap as possible. I think that with all these features packed in a sale price of around $60 would be possible, but it depends on the quantity of boards produced (higher quantity = lower price).
 
Last edited:
I have an early functional version of the data analysis software. It can download data from the device, you can load/save data to/from files and it can do various plots as well. Eventually I'll be developing various algorithms for filtering and deriving different things from the data. One particular area I'm interested in is to actually reconstruct the flight path from the data by integrating the accelerometer data and using magnetometer data as a direction vector, so that I can see how the rocket spins and how the angle of it changed during the flight.

I think that evaluating motor performance is more difficult because the weight of the rocket changes throughout the burn and you also have drag to take into account. With such dynamic conditions it would be difficult to get an accurate thrust curve. If you have a look at the image below from a test flight you can see that the acceleration data closely resembles the shape of the thrust curve of the B6 motor I used and I could estimate the force in Newtons if I know the weight of the rocket before and after the burn, but the weight is not reduced linearly making it very difficult. We can also get a reasonably good indication of air drag from the acceleration graph after burnout. If there was no drag, the rocket would decelerate at a constnat 9.8 m/s^2. If you look at the acceleration curve, deceleration is about 17m/s^2 which reduces to about 9.8 by the time velocity reaches 0 at apogee.

View attachment 150508

The velocity graph (red) on the plot is calculated by integrating acceleration and the distance (green) is derived by integrating velocity. The purple line (height) is the altitude derived from the barometer data.

With regards to price, I would like to make it as cheap as possible. I think that with all these features packed in a sale price of around $60 would be possible, but it depends on the quantity of boards produced (higher quantity = lower price).

Motor performance can be done: there are fairly safe assumptions that can be made about mass consumption (rate of mass consumption is proportional to thrust), and you can back out the drag from the coast.


What's the range of the accelerometer? Most 3-axis accels are pretty wimpy (+- 8 G) for rocketry applications. I like having accels onboard, but at the moment the cheapest device with a wide range is the Raven.
 
The accelerometer I'm using is +/-16g on each axis. If you look at the photos you'll notice that one of the chips is oriented at 45 degrees. This "trick" allows G forces up to 22.6g's (sqrt[16^2+16^2]) to be detected along the direction of travel as the force is evenly distributed along the X and Y axes. Some rockets will generate higher G's so this board will not be for every rocket, although you will still get accurate altitude data if you saturate the accelerometer.
 
I would love a small recording board (or for that matter, a deployment altimeter that also records) with an Analog Devices ADXL375, a digital 3-axis 200g accel.

Its sensitivity is 20.5 LSB per G, easily enough resolution for a rocket, so there'd be no need for low- and high-range models.

Its price is not unreasonable: Digikey has it at eleven dollars and change for a single chip.

https://www.analog.com/static/imported-files/data_sheets/ADXL375.PDF
 
I would love a small recording board (or for that matter, a deployment altimeter that also records) with an Analog Devices ADXL375, a digital 3-axis 200g accel.

Its sensitivity is 20.5 LSB per G, easily enough resolution for a rocket, so there'd be no need for low- and high-range models.

Its price is not unreasonable: Digikey has it at eleven dollars and change for a single chip.

https://www.analog.com/static/imported-files/data_sheets/ADXL375.PDF

It looks like the perfect chip for a rocket altimeter. It would only fit onto my board however if I drop the magnetometer as I'm using the LSM303DLHC which is the same package size as the ADXL375 but contains a combination of 3-axis Accelerometer and 3-axis Magnetometer. I do have plans for a more advanced and larger board with dual deployment and the ADXL375 is a great candidate for that in combination with a stand-alone magnetometer and/or gyroscope.
 
I do have plans for a more advanced and larger board with dual deployment and the ADXL375 is a great candidate for that in combination with a stand-alone magnetometer and/or gyroscope.

That I would very likely buy.
 
By the way I created a Facebook page and will be posting updates on there so those interested can keep track of the project: https://www.facebook.com/flytrak

I will also post some updates here in the forum. Layout of the AltiLogger v2 board is completed. Initial V2 prototypes will be produced in a couple of weeks.
 
That's a nice little logger
icon14.png
 
This is a really nice design. I am also working on micro altimeters based on my kits but they are a lot less complex.
What about the software ? Is it multi platform?
I am currently working on a Java altimeter interface if you are interrested
https://rocket.payload.free.fr/index.php?option=com_content&view=article&id=13&Itemid=11&lang=en

Hi there, nice altimeter page you have there! I've been working on an app in C# but the plan is to make a Java app so that it's multi platform, I just found it easier to put something together quickly using C# to get the firmware/hardware tested. As far as the interface is concerned I'm identifying the USB device directly (based on Product ID and Vendor ID) rather than selecting a virtual COM port as this allows the application to automatically detect when a device is inserted so it's easier to use.
 
Hi Everyone!

I've been working on this for the last 6 months so I thought I'd share it with other rocketeers as well to see what people think. I wanted a logging Altimeter that is very accurate and at the same time very lightweight, so I can use it in smaller rockets as well without the weight of the altimeter having too much of an influence on the rocket's flight. I read an article about how temperature can make a huge difference to the measured altitude so the device also logs temperature and altitude calculations derived from pressure will be temperature-compensated. Also included is a 3-Axis accelerometer and magnetometer on the board, which also allows the acceleration and flight direction of the rocket to be logged in detail. There's an integrated USB interface so there's no separate download kit needed, just plug it straight into a laptop with a microUSB cable. All this is squeezed into a board just 20x14mm weighing 2 grams with battery. Below is a picture of the first prototype.

View attachment 150469

I've done one test flight with it and it worked great. I'm working on an improved version now which I'll be prototyping next month. I'm calling the device 'AltiLogger'.
How much did it cost you to build?
 
How much did it cost you to build?

As I am also doing that kind of project let me try to answer that one.

The cost cannot be only the price of the components + board that you see on the picture
To achieve such board you need to do several prototype and test everything separatly on breadboards, you end up buying componants that you don't use etc ... You also need to have the right tools (ie: reflow oven for smt)
You need to buy motors to test it
Plus you have to take into account the number of hours spent designing etc....
So it is very difficult to estimate how much the firt working prototype cost
 
This looks interesting.

When you say "high accuracy", what makes it accurate and what benchmark testing have you done to confirm that it is accurate?

Greg
 
Cool!

Any idea on what the software functionality will be? Any interest in developing a way to record motor performance? Any price point targets?

I agree that research motor development/flight testing would be a prime application if the precision was sufficient.
 
I agree that research motor development/flight testing would be a prime application if the precision was sufficient.

That would be better with a high-range accel, hence my interest in the ADXL375.
 
Even though I'm a LPR/low MPR guy, I'm always interested in new altimeter devices. I'd be interested in at least one of these just for that reason.
 
It looks nice and compact. One issue I see it the battery holder. Under the G-loading of a rocket launch, it is possible for the battery to loose contact with the holder and dropout the power. Have you considered using a small LiPo battery and using the USB port as a charger? The PerfectFlite Pnut does it this way.

Bob
 
How much did it cost you to build?

That's difficult to say. The PCB itself cost about $40 for 5 prototype boards. The components cost about $150 for 5 prototypes so that's about $38 material cost per board. I assembled a couple of prototype boards by hand using solder paste, a hot air rework station and a microscope. It takes many hours and very steady hands to get it made in this way so for anything more than a couple of prototypes I will use a PCB assembly service. After assembly, every board will need to be programmed and tested on a test bed, etc.
 
That's difficult to say. The PCB itself cost about $40 for 5 prototype boards. The components cost about $150 for 5 prototypes so that's about $38 material cost per board. I assembled a couple of prototype boards by hand using solder paste, a hot air rework station and a microscope. It takes many hours and very steady hands to get it made in this way so for anything more than a couple of prototypes I will use a PCB assembly service. After assembly, every board will need to be programmed and tested on a test bed, etc.

Definitely a "labor of love." I applaud your efforts!
 
This looks interesting.

When you say "high accuracy", what makes it accurate and what benchmark testing have you done to confirm that it is accurate?

Greg

It's high accuracy because I'm using a barometer that is very high resolution and is calibrated to a high accuracy by the manufacturer and also has built-in temperature compensation. On top of that the pressure-derived altitude also depends on temperature, therefore using actual temperature measurements rather than a static temperature value will give you more accurate altitude calculations. This was inspired by the article on page 9-10 in the following Apogeerockets newsletter: https://www.apogeerockets.com/education/downloads/Newsletter257.pdf

As you can see the difference due to temperature can be very significant and the error becomes greater with altitude. The actual altitude that you launch from also has an effect on the accuracy. The device actually logs pressure, not altitude, therefore if you know the elevation of the launch site, you will be able to enter this in the data analysis software to get an additional level of accuracy.
 
It looks nice and compact. One issue I see it the battery holder. Under the G-loading of a rocket launch, it is possible for the battery to loose contact with the holder and dropout the power. Have you considered using a small LiPo battery and using the USB port as a charger? The PerfectFlite Pnut does it this way.

Bob

I don't think there will be contact issues with the battery, it sits in there very firmly, I can drop the device on the floor giving it hundreds of g's and it keeps running just fine and I also had no issues during a real launch. If you're really worried you can put a tape over it but it's not necessary.

I did consider using a LiPo but it increases cost and board size as you need a power management IC for charging and the LiPo battery also costs more and is difficult to find such a small LiPo. For this design the coin cell is the best solution.
 
Last edited:
I did consider using a LiPo but it increases cost and board size as you need a power management IC for charging and the LiPo battery also costs more and is difficult to find such a small LiPo. For this design the coin cell is the best solution.

Adding a small JST connector and a pin header jumper might allow users to use LiPo's if they desire... ones such as these: https://www.motionrc.com/admiral-250mah-3s-11-1v-30c-lipo-battery/?gclid=CPr4tofKrroCFYdQ7AodBB8Aaw

Also, nice altimeter! :D I've worked in electronics assembly, so I know how much work those little SOT packages can be. The company I worked for might be a good one if you want to do small production runs. They have the right equipment for manufacture and testing, but are a mom & pop shop so I don't think they're particularly expensive. I'm definitely interested in the chip, in particular in the larger one with dual deploy capability. If possible, please incorporate igniter short detection :)
 
Last edited:
Adding a small JST connector and a pin header jumper might allow users to use LiPo's if they desire... ones such as these: https://www.motionrc.com/admiral-250mah-3s-11-1v-30c-lipo-battery/?gclid=CPr4tofKrroCFYdQ7AodBB8Aaw

Also, nice altimeter! :D I've worked in electronics assembly, so I know how much work those little SOT packages can be. The company I worked for might be a good one if you want to do small production runs. They have the right equipment for manufacture and testing, but are a mom & pop shop so I don't think they're particularly expensive. I'm definitely interested in the chip, in particular in the larger one with dual deploy capability. If possible, please incorporate igniter short detection :)

Hi, Thanks :)

I'm located in New Zealand. I found several places locally where they do electronics assembly for small batches at reasonable prices. One of the things I'll have to figure out is how many boards to produce. They can place 180 boards on a panel. If I get 180 made, that's thousands of dollars for the parts and manufacturing. I can finance it as long as I know that people will buy most of them. Alternatively I'll get them to make a panel and I only populate a certain number of boards and throw the rest of them away. It is for these reasons that the per unit cost will depend greatly on quantity. Also the cost of components get significantly cheaper in the hundreds. Not sure how I'll figure out how much demand there is for such a board. I think it's primarily for smaller rockets, but there's no reason why you couldn't include it in high power rockets as a secondary payload even if this board itself will not deploy your parachutes.
 
Hi, Thanks :)

I'm located in New Zealand. I found several places locally where they do electronics assembly for small batches at reasonable prices. One of the things I'll have to figure out is how many boards to produce. They can place 180 boards on a panel. If I get 180 made, that's thousands of dollars for the parts and manufacturing. I can finance it as long as I know that people will buy most of them. Alternatively I'll get them to make a panel and I only populate a certain number of boards and throw the rest of them away. It is for these reasons that the per unit cost will depend greatly on quantity. Also the cost of components get significantly cheaper in the hundreds. Not sure how I'll figure out how much demand there is for such a board. I think it's primarily for smaller rockets, but there's no reason why you couldn't include it in high power rockets as a secondary payload even if this board itself will not deploy your parachutes.

I really would want a slightly larger one with 9-axis IMU capabilities. Deployment not necessary; I just want very accurate, reliable data logging. Add a few millimeters in length, if you don't want to sacrifice the 14mm dimension?
 
I really would want a slightly larger one with 9-axis IMU capabilities. Deployment not necessary; I just want very accurate, reliable data logging. Add a few millimeters in length, if you don't want to sacrifice the 14mm dimension?

I would purchase one with these capabilities. But if you keep it as is I may purchase one just to play with.
 
Back
Top