HMC5983 EOL Suggested replacements

[Philip Rowse]

Hi All,

   I am working to find a replacement for the HMC5983 compass as it is EOL.

Please have a look here https://docs.google.com/spreadsheets/d/1gVlKZBvRNTXldoxTXwipGaaHmtF9DNPaftDrzKA47mM/edit?usp=sharing

at the comparison so far of sensors that are available.  The HMC5983 is on the left, data to the right is the new options.  

the only requirements are that they be digitally interfaced, and a couple of $ max.

May the best sensor win...

Thank you

Philip

[Nikolay Arsov]

There are just a few alternatives, each with its pro and cons:
1. LSM303D is not reliable, not temperature compensated. Almost 10-15% of the sensors soldered show problems in time;
2. FXOS8700CQ - maybe better alternative to LSM303D, but it hasn't temperature compensation too.......a compromise alternative;
3. RM3100 - maybe the best mag ever ( http://www.pnicorp.com/products/geomagnetic-sensor/ ), but huge and expensive ....could be an alternative if 10x price and space are acceptable;
4. To make our best and make MPU9250 mag work.....this could be a solution too.

[Philip]

Kartik from 3DR suggested 

http://www.memsic.com/userfiles/files/Datasheets/Magnetic-Sensors-Datasheets/MMC3416xPJ_Rev_C_2013_10_30.pdf

It has an interesting method of cancelling out error... May also be worth a look.

[Andy Little]

The LIS3MDL does have temperature compensation according to http://www.st.com/web/en/resource/technical/document/application_note/DM00136626.pdf section 6

Have to question if HMC5883 is best then why is it EOL ?

regards

Andy

[Jonathan Challinger]

Probably internal politics

--
You received this message because you are subscribed to the Google Groups "drones-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to drones-discus...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Andy Little]

The Datasheet also says that the zero gauss level is pre-calibrated at the factory. If that means what  think then surely that makes it a killer application, way better than HMC5983. so I have ordered one to see what it does actually mean

[Phillip Kocmoud]

I ordered several of these to test. They arrive today. https://www.pololu.com/product/2737 . My plan is to compare with the HMC5983 during flight tests.

Phil

--

[Philip]

Thanks All!

[pop eye]

Hi All,
As per my tests with data captures and heading logs, the invensense MPU9250 and FXOS8700 are not able to provide stable 1 degree readings, I have been testing these 2 devices and they show signal to noise ratio not as good as HMC5983. It is a pity Honeywell stops this IC, I am also still looking for a suitable replacement...
Rgds
Florent

On Friday, February 19, 2016 at 5:11:11 AM UTC+8, Philip Rowse wrote:

Thanks All!

[Nikolay Arsov]

Hi Florent,
You may stop searching as there is no suitable replacement. The only set which can give better alternative is RM3100, but the best price obtained is $12.75 / 1000
BR
Nick

[pop eye]

Hi Nick,
As I expected :(... I wanted to order a demo board with ST LIS3MDL 3 axis magnetometer for evaluation, any feedback on this device?
Thanks again,
Florent

[Nikolay Arsov]

Hi Florin,
LIS3 is under tests at the moment, but I wouldn't expect too much from it.
Next week we'll have some test results and I hope we can compare to 5983.
BR
Nick

[Phillip Kocmoud]

We have also had acceptable flights using the AK8963 that is integrated inside the MPU9250. With the prevalence of external MAGs, I am not sure the necessity of a standalone Magmetometer.

--

[pop eye]

Hello,
As per my testings the AK8963 is the worst in term of signal/noise ratio vs freescale and honeywell, it can t provide a stable output within a degree but this can be acceptable for numerous applications.
I have not tested any other magnetometers, hope to find one to replace the HMC5983.

[Phillip Kocmoud]

Please send me your test results.

[Craig Elder]

Please post the test results to this group

[pop eye]

From the captures one can see that the invensense has nearly 50% less signal for a more noisy output compared to freescale FXOS. I used all the setup available to reduce the noise as much as I could for my application (digital filters bandwidth/etc...). From the angle output we can see the jitter especially when it comes to 359 degrees/roll over. For my platform honeywell has been providing state of the art response with 0.3 degrees accuracy (9 axis solution gyro gimbaled), 4-8 times better than FXOS/AK solution. For smartphones and tablet I guess 1 degree accuracy is not needed.
Florent

[Phillip Kocmoud]

Here is where we are tracking the performance of the AK8963 on PX4. https://github.com/PX4/Firmware/issues/3953#issuecomment-197538985

[Philip]

I would prefer we kept all data in this chain or on the original Google doc. It's important to keep this effort unified.

[Phillip Kocmoud]

Hi Philip,

We are currently flying with the 9250 and will be flying with the LIS3MDL by Friday. Shortly there after with the RM3100. I think specs are only a one part of the equation, we actually want to fly them. Has there been a consensus from your perspective? 

On Wednesday, March 16, 2016, Philip <phi...@proficnc.com> wrote:

I would prefer we kept all data in this chain or on the original Google doc.  It's important to keep this effort unified.

[philip]

The next one to try is ST’s latest one…  LSM303AGR

So is there a Px4 driver for the LIS3MDL already?  Can you please point me to it? We are just assembling the test hardware now.

 

Philip Rowse

--
You received this message because you are subscribed to a topic in the Google Groups "drones-discuss" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/drones-discuss/XhTl5hmko04/unsubscribe.
To unsubscribe from this group and all its topics, send an email to drones-discus...@googlegroups.com.

[pop eye]

Hi Philip,
Is not the LSM303 a LIS3MDL with an accelerometer? (I may ask my ex work collegues from ST)
Rgds
Florent

On Thursday, March 17, 2016 at 7:01:23 AM UTC+8, Philip Rowse wrote:

The next one to try is ST’s latest one…  LSM303AGR

So is there a Px4 driver for the LIS3MDL already?  Can you please point me to it? We are just assembling the test hardware now.

 

Philip Rowse

 

From: drones-...@googlegroups.com [mailto:drones-...@googlegroups.com] On Behalf Of Phillip Kocmoud
Sent: Thursday, 17 March 2016 9:23 AM
To: drones-...@googlegroups.com
Subject: Re: [drones-discuss] HMC5983 EOL Suggested replacements

 

Hi Philip,

 

We are currently flying with the 9250 and will be flying with the LIS3MDL by Friday. Shortly there after with the RM3100. I think specs are only a one part of the equation, we actually want to fly them. Has there been a consensus from your perspective? 

On Wednesday, March 16, 2016, Philip <phi...@proficnc.com> wrote:

I would prefer we kept all data in this chain or on the original Google doc.  It's important to keep this effort unified.


--
You received this message because you are subscribed to the Google Groups "drones-discuss" group.

To unsubscribe from this group and stop receiving emails from it, send an email to drones-discuss+unsubscribe@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

[Philip]

Quite possibly... 
I would be interested in your enquiry results...

[Nikolay Arsov]

Hi Florent,
LSM303C integrates the same mag core as LIS3MDL.
LSM303AGR is the newer one but is still in preview stage.

[Robert Dickenson]

You can find the beta version of the LIS3MDL driver here:

https://github.com/Inspirati/Firmware/tree/lis3mdl/src/drivers/lis3mdl

--

You received this message because you are subscribed to the Google Groups "drones-discuss" group.

To unsubscribe from this group and stop receiving emails from it, send an email to drones-discus...@googlegroups.com.

[Philip]

Thanks Robert!

[pop eye]

Hello Guys,
the LIS3MDL is not too bad, a bit more noise in the data compared to the HMC5983, but needs to be calibrated in temperature (big heading deviation if not compensated).
Florent

[Andy Little]

FYI

Recently saw this

http://www.infineon.com/cms/en/product/sensor/magnetic-position-sensor/3d-magnetic-sensor/channel.html?channel=5546d4624c9e0f0e014c9e105a8a001c

£0.90 at mosuer uk

http://www.mouser.co.uk/ProductDetail/Infineon/TLV493DA1B6HTSA2/?qs=%2fha2pyFadujqKQdZUFp8%252bDauYIEloESnfv15ezRoifm9G2YqjD%2fU3Q%3d%3d

[Baruch Kantor]

HI,

I've tested the 9250 (AK8963) Vs HMC5983 and found it noisier, in contrast to others findings.
Am I missing something?

bar

[jackyxu.s...@gmail.com]

Try this :)

https://groups.google.com/forum/#!topic/drones-discuss/m7kc0ZF7qYQ

[Nikolay Arsov]

IST8310 is not full 5983 equivalent as it has just I2C bus, while we use 5983 over SPI bus.

[Jacky Suffice]

Oh, yes.. if spi bus is certainly required~

--
You received this message because you are subscribed to the Google Groups "drones-discuss" group.

To unsubscribe from this group and stop receiving emails from it, send an email to drones-discuss+unsubscribe@googlegroups.com.