by

|   Sep 10, 2009 at  5:51pm

What is the actual bit resolution on this bugger?

by

|   Sep 10, 2009 at  6:02pm

it's got a 12-bit ADC built-in according to the datasheet and also stores each axis's values as a pair of bytes.

by

|   Sep 11, 2009 at  4:43am

Thanks.
I have read the 2bytes storage part, but I missed the ADC.

by

|   Sep 11, 2009 at  2:47pm

Sweet! I was the one who suggested this (multiple times a while ago). However, on digikey the chip itself is only $20? Am I supposed to believe that the various support components and tiny board account for $30? Maybe the price will go down.

by

|   Sep 15, 2009 at  8:25pm

You have to look at manufacturing and assembly cost. Plus there are inevitable human and machine errors that lead to unusable product (notice the ding + dent products section).

These BOBs are meant for people like me that can't/won't smd solder or don't have access to reflow ovens etc.

by

|   Sep 16, 2009 at  9:48pm

Nagmier is right, these are evil little buggers to solder. And remember, the honeywell breakout board is $100. taktoa - email us, point us to where you suggested it, we'll send you a free one for suggesting we carry it.

by

|   Sep 19, 2009 at  4:51pm

http://forum.sparkfun.com/viewtopic.php?t=15563

Here is a forum post by me (Faraday's Cage == taktoa@gmail.com == taktoa), but I also sent in a request via the "feedback" form.

by

|   Sep 19, 2009 at  4:53pm

Oops, forgot to say that I wasn't sure who to email, so I commented instead.

by

|   Sep 21, 2009 at  8:46pm

I have bought it. Seems the I2C cannot work! I am using the AVR MCU. After the SLA+R or SLA+W is trasmitted, and NACK is always received. Any suggestions?

BTW, in the HMC5843 datasheet, it says the C1 and C2 should have low ESR. It seems C2 is 0402 which may not have a low ESR.

by

|   Sep 22, 2009 at  3:25pm

Hey fenghua9801,

Sorry to hear you're having trouble with this little guy. If you could send us an email, with as much information as possible about your setup, we'd be happy to help you get this figured out.

-techsupport at sparkfun dot com

by

|   Sep 22, 2009 at  8:01pm

Thanks! I will first try to fix it. If it does not work. I will contact you.

by

|   Sep 22, 2009 at  5:51pm

I also had problems with the I2C not responding probably. I added another capacitor in parallel with C3, I used a 100uF because I had it to hand. I have had no problems since. I tried this because the same fix was reported on a couple of forums - diydrones and Surveyor robotics.

by

|   Sep 22, 2009 at  7:51pm

I also see what you said at AVRfreaks, but I only have 10uf Tan cap at hand, and I tried to parallel one to C3. But it still not work! Do you use a 100uf Tan cap? Thanks!

by

|   Sep 22, 2009 at  9:38pm

It wasn't a tan, it was a big electrolytic.

by

|   Sep 23, 2009 at  1:31pm

It works! Thank you so much! I also used a 100uF electrolytic cap now.

by

|   Sep 29, 2009 at  4:45pm

This might shed some light on the issue.

From the datasheet:

The two external capacitors should be ceramic type construction with low ESR characteristics. Reservoir capacitor C1 is
nominally 4.7 micro-farads in capacitance,
...
Low ESR characteristics may not be in many smallest size SMT ceramic capacitors (0402), so be prepared to up-size the capacitors to gain Low ESR characteristics.

I have a custom 4-layer board that uses this chip that I should get in the mail today. Assuming I can solder it without ruining it, I provide more feedback on the cap size I used and it's ESR.

by

|   Oct  5, 2009 at 12:14am

My custom board works fine with a 10uf ceramic chip cap. I wonder what the ESR of that tantalum cap is on this board? My layout is about the same except I have a 4 layer board. My guess is that this chip prefers a ceramic cap.

by

|   Oct  5, 2009 at  2:21pm

Yes, the problem may be the tan cap, which is 3 ohm from the datasheet. I also ordered ceramic cap to replace that tan cap. I will see if it works or not.

by

|   Oct 20, 2009 at  6:07am

The sensor did not operate properly until I put 100 uF capacitor in parallel with capacitor C3 as suggested in this thread. Does SparkFun have plans to replace capcitor C3 with the one that works?

by

|   Sep 24, 2009 at 10:19am

Although I can clock the data out, I had another problem. The X axis data is most of time negative, while the Y axis data is usually positive. But this should not be the case since I put it on the horizontal plane and rotate the board, which means the data should half time positive and half time negative. Have you experienced the same problem? Thanks!

by

|   Sep 24, 2009 at 12:32pm

I dont believe I have that problem since I can calculate the bearing correctly - at least a 2axis bearing, still having problems calculating a 3axis bearing.
Are you holding the board flat? That can affect the reading.
I init the chip by reading the id registers for H43 and then setting reg 2 to 0
To read I read 1 byte from reg 9 and send a i2c stop.
If the value back has bits 0 and 2 set, I then do a read of 6 bytes (without sending a reg), i.e. i2c start, write 3D followed by 6 reads. Make sure the reg 9 read and the last read of the 6 reads are naked rather than acked.

by

|   Sep 25, 2009 at  5:52pm

Done as what you said, but same problem exists. Thanks anyway. Maybe I will try another board to see if similar problem exists.

by

|   Oct  2, 2009 at  3:29am

For those having problems reading data in continuous mode read page 18 in the spec... the the 5 milli-second power-up time is important!

from page 18:
To change the measurement mode to continuous measurement mode, after the 5 milli-second power-up time send the
three bytes:

0x3C 0x02 0x00

This writes the 00 into the second register or mode register to switch from single to continuous measurement mode
setting. With the data rate at the factory default of 10Hz updates (100 milli-seconds interval), a 100 milli-second delay
should be made by the I2C master before querying the HMC5843 data registers for new measurements. To clock out the
new data, send:

0x3D, and clock out DXRA, DXRB, DYRA, DYRB, DZRA, DZRB located in registers 3 through 8. The HMC5843 will
automatically re-point back to register 3 for the next 0x3D query, expected 100 milli-seconds or later.

by

|   Oct 23, 2009 at  9:15am

The datasheet, if I read it correctly, says it loops back at register 9, the status register, not 8 (i.e. read the status and if data is ready it will wrap and read the data).

by

|   Nov 12, 2009 at  8:05am

The current revision of this board has a potentially incorrect capacitor.

I was unable to get it to work with either of my Arduinos and there are a few forum posts detailing the problem.

http://forum.sparkfun.com/viewtopic.php?t=17475&start=18

A newer revision is in the works but there isn't an ETA on it (as per a support email I received from SparkFun on 2009-11-11). The workaround is to replace the capacitor with one like this:

http://www.sparkfun.com/commerce/product_info.php?products_id=96

chris.

by

|   Nov 16, 2009 at 12:23am

It seems the capacitor problem may be fixed in new ones. I got one yesterday that works without modification.

I wrote an Arduino library to interface it: http://eclecti.cc/hardware/hmc5843-magnetometer-library-for-arduino

by

|   Nov 29, 2009 at  2:28pm

Hi,

The datasheet says that the I2C lines of this device are nominal 3.3V... I would appreciate an advise or share comments regarding interfacing this magnetometer with 5V dsPIC MCU.

Thank you!

by

|   Dec  6, 2009 at  3:43am

I was able to talk to HMC5843 chip using an arduino mega, with built-in pullups to 5V. Communication seems fine, but I cannot get the right measurements ( seems like the problem with capacitor or something - i have to investigate it, though I tried putting in a 460uF electrolytic, with no luck. Will try ceramic)

by

|   Dec 25, 2009 at 11:26am

One slight problem - it will only read 1 byte, ie. if I send 0x3d then try to read 4 bytes, the first one will be correct (from the corresponding register), but the rest of the bytes read as 0xff. I'm doing 400khz clocking.

I saw the test code can just read sequential bytes.

I don't know about the commands - I might need to send them sequentially too.

Is anyone else seeing this?

by

|   Dec 25, 2009 at  2:59pm

Does almost the same - compiled the test code and I get x=32, y=-1 x=-1 with a few glitches.

I'm using a 3.3v arduino pro and changed the program to 8Mhz (that was the only change).

There are two caps stacked, so I assume that has been fixed, but I'm not getting any magnetic readings though it is communicating.

Adding in something, the ID bytes work. The status register is always 4 (regulator enabled, zero conversions ready). The command registers (even after being written to) read the defaults of 0x10, 0x20, but the mode register reads 0 or 1 depending on how quickly I read.them.

by

|   Dec 29, 2009 at  8:11am

It works! Of course there is now a huge normal electrolytic cap (47uF) larger than the chip hanging off the existing cap/stack.

My recommendation to anyone getting this board is to have a cap ready to solder on and if it acts at all strange, add the cap.

(Come on SparkFun - don't you test these things?)

by

|   Dec 30, 2009 at 11:09am

One more note on I2C - I have a compliant high speed assembly driver for ATtinyX5 chips. It does NOT work with this chip unless you add a small delay for a larger setup time before the clock line is pulsed. And it isn't just the ACK on reads (which will return 0xff after the first) - putting delays only there (to insure it would be acked) didn't help, it needs the delay for every clock pulse. (On the scope, the waveforms are compliant, and the arduino version seems to just have a long risetime v.s. the very square ATtiny waves on my setup).

by

|   Jan  1, 2010 at 10:50am

Another interesting thing - probably too slow for commercial use, but it does pick up credit card magstripes.

by

|   Jan 13, 2010 at  5:44pm

Just got two-- work fine after adding an 0805 10 uF ceramic in parallel with the present cap.

by

|   Jan 14, 2010 at  9:42pm

Digikey PN PCC2395CT-ND / ECJ-1VB0J106M

Panasonic ECG Ceramic 10uF 6.3V X5R 0603.

Small, lowESR, and makes the thing actually work well.

by

|   Jan 14, 2010 at  9:55pm

My 47uF electrolytic can didn't work as well - it would work for a few minutes then stop. This one from Digikey just runs and runs.

by

|   Mar  9, 2010 at  5:05am

Another input.

Somebody has attempt to use it at low temperature? (~0ºC).

I have to solder the extra capacitor, but at low temperature, I tried several capacitors and the magneto doesn't work properly, it has the same behavior without the extra cap.

by

|   Jul  6, 2010 at  2:23pm

Does anyone know if the capacitor problem has been fixed with the latest boards?

by

|   Jul  7, 2010 at  4:21pm

Yep. The PCB was updated in February, 2010. The tantalum capacitor was replaced towards the end of March. All boards are fully tested before they leave production.

by

|   Jul 13, 2010 at  9:13pm

I received a copy of this part just yesterday, and was able to get this part talking to my Logomatic. It works fine for me, spinning it on a turntable generates the sine and cosine signal I expect.

I guess that new cap works fine.

by

|   Jul 14, 2010 at  2:11pm

Good to know, I have heard people say this is a tough compass to use even without hardware issues.

by

|   Jul 23, 2010 at 10:16pm

Sparkfun Guys,

You would make a much, much better board if you bring our the DRDY pin to the header - there is a good reason that Honeywell decided to bring out the RDY bit (SR0 of the Status register) to a separate pin and it is pity that you did not do the same!

Having the DRDY pin available would make asynchronous (interrupt-driven) reading data from this magnetometer much, much easier; for now it is quite cumbersome...

by

|   Aug  5, 2010 at  8:38am

Doing a rev that has the DRDY pin broken out to a header would be nice. If you have any other suggestions for a rev send them to techsupport@sparkfun.com

by

|   Aug 10, 2010 at  7:43pm

Toward the end of these comments Nate says that the capacitor problem has been fixed. However the schematic linked to here still shows the 4.7 capacitor. Is this just an out-of-date schematic?

thanks

Joe

by

|   Aug 11, 2010 at 12:07pm

The v11 breakout had a 10uF tantalum cap that was causing problems for some but not all users:
http://www.sparkfun.com/datasheets/Sensors/Magneto/HMC5843-v11-schematic.pdf
This problem has been fixed in v12 by replacing the 10uF tantalum cap with a 4.7uF ceramic cap.