SparkFun Forums

[asilloo]

Hej,

I'm still working on this sensor;

But it dose not work , I send a email to sparkfun about this problem , this is the replay;

Hello,I'm very sorry to hear about the trouble this sensor is giving you. The example code we've posted on the product page (http://www.sparkfun.com/commerce/produc ... ts_id=9371) should work perfectly, and you can load it onto an Arduino with the need of an AVR programmer. This tutorial: http://www.sparkfun.com/commerce/tutori ... als_id=142, in the 'Serial bootloading an Arduino board' section describes how you can load the code using WinAVR.

There have been discussions of an incorrect capacitor on the board, but our tests show that the board does work with the current setup.

Hope that helps. Let me know if you have any further questions or concerns.

is there any one can help me

I want just get X,Y,Z of the magnetic filed. How!!!!

My items;
1.Arduino ATMEGA 328
2.logic level converte
3.HMC5843 sensor
Back to top

[tz]

Sparkfun is fudging the truth. SOME boards may work with the current capacitor, but it is very unstable. I could not get mine to work before I change the capacitor. If you check the product page (or this thread) you find many people with the same problem.

[zybr]

i have tremendous problems with this sensor when trying to interface it to a ARMexpress. I cant even get an ACK. i get nothing, zip, nada. I tried putting eletrolytes(ESR unknown) in parallel with the black cap, i also tried removing it all together and have only eletrolytes. Nothing. i dont really hae access to a whole lot of compoenents due to facility constraints, thats why i buy stuff from SF. premade so i dont have to bother with it....

in pure desperation i even tried putting a ceramic cap paralel with the tantal and eletrolyt: nothing....

Im at the verge of giving up on thise sensor. I just cant get it to work and it really annoys me because i have now spent 3 weeks on something that should have taken 2 days, and an IMU without a magnetometer is like a car with no wheels

[noether]

As I posted at the website.

Be careful operating this compass with your extra electrolytic capacitors. Aluminum tends to increase its ESR dramatically as the temperature goes down.

I actually have problems with the default capacitor, I attached a electrolytic one and solved.... but if the operation temperature is around 0ºC, the compass doesn't work properly, I tried other capacitors and the problem remains.

Check it out here: http://www.illinoiscapacitor.com/pdf/Pa ... iction.pdf

I'm going to test the board with a polymer one during the next week, I'll post the results.

[VIKI]

Hi all!

I seem to be having the same issue as asilloo, comparing the data, so basicly i do get data from the HMC5843 but it makes no sense.
Running on a Arduino mega, tried one of the HMC libs, wrote something myself, tried several caps parallel to C1 (datasheet), added external 10K pull-up resistors for the SDA and SCL lines and i still don't get any normal data.
Although recently bought it still looks exactly like the pics on the honeywell page.

Anything more i can try? (apart from just eating it )

Thanks

[seiji8503]

Hello!!

I am having a different problem interfacing HMC5843. I can't configure it to Measurement Mode. When I send the 0x3C Write Command the sensor answers ACK. But in the following byte, when I send 0x02, the sensor answers NACK. The bits are being correctly sent for I've acquired them with a scope.

For what I've read here, the main actions people are taking are:
1 - Replacing C1 with a low ESR (ceramic - http://my.execpc.com/~endlr/esr.html) higher capacitance capacitor (>4.7uF, higher capacitance lower ESR);
2 - Clock stretching I2C (http://www.i2c-bus.org/clock-stretching/) and adding delays, since sensor cannot answer too fast;
3 - In Single Measurement Mode, sensor provides one measurement and enters Idle state, so new measurements can only be obtained switching mode to Continous Measurement.

I have taken all actions possible but I can't figure out what the problem is. Has anyone gotten similar problem?

Thanks for the help!!!

[tz]

[quote="seiji8503"]Hello!!

I am having a different problem interfacing HMC5843. I can't configure it to Measurement Mode. When I send the 0x3C Write Command the sensor answers ACK. But in the following byte, when I send 0x02, the sensor answers NACK. The bits are being correctly sent for I've acquired them with a scope.

[/quote]

Have you tried any of the sample programs (or which one are you using that does this)?

If there is a power problem, it can miss bytes, so check your power supply that it isn't drooping.

Also, I think there are termination resistors on the I2C. You might want to use some lower values in parallel or use a current mirror.

[seiji8503]

Thank you tz for the suggestions!

I checked the supply and it is ok, no drops or significant noise.

When you say termination resistors do you mean pull-up? I have used them, 2.2k in each line.

However, I am thinking that the connections I made with DVDD are causing the trouble. Here I post some answers I got from DigiKey and Honeywell support, hope it may help!

- "The main purpose of C1 cap is the reservoir for storing the energy for the set/reset pulses. The second function is to act as a decoupling cap for DVDD in single supply mode so the answer is yes".
- The internal regulator can operate with 3.3V I2C.
- "DVDD pin should tie to C1 in single supply mode (VREN enabled). It should not be tied to another source. This may damage the internal digital ASIC. The maximum voltage for DVDD is 2V."
- "The exact ESR of C1 is not critical but we recommend 200 milli-ohm or less. What’s the capacitance of C2 cap? Make sure the ratio of C1/C2 is at least 10."

If tests run OK I will post a follow-up!

Thanks!!

[pizza]

Hi people!!!!

i am using an imu breakout board tt some other student designed. apparently the student also used a tantalum cap on C1 causing it to fail to work. so changed the tantalum cap in the circuit to a 10uf ceramic and it has been working for hours. however the ceramic cap is not low esr. will it cause noisy readings? because my readings are very noisy.

this is the ceramic cap that im using
http://sg.farnell.com/kemet/c1206c106z8 ... dp/9227903
found it in the lab. well i'm not sure how to measure esr.

can't get my hands on low esr ceramic caps since the minimum quantity order is like 65 bucks worth of low esr caps..zzz

THANKS

[tz]

[quote="pizza"]Hi people!!!!
i am using an imu breakout board tt some other student designed. apparently the student also used a tantalum cap on C1 causing it to fail to work. so changed the tantalum cap in the circuit to a 10uf ceramic and it has been working for hours. however the ceramic cap is not low esr. will it cause noisy readings? because my readings are very noisy.
THANKS[/quote]

A pic of the board and a definition of what you mean by "noisy" would help - are the readings varying 1%, 10%, .1%?

Generally any cermic cap should have a low enough ESR. Many sensors are going to be noisy unless it is locked down in a faraday cage.

That said, I would look to how you are decoupling, isolating, or otherwise insuring the analog side of the circuit is avoiding noise from the various possible sources. The supply voltage might have ripple or spikes (when something else turns on), there may be things carrying current nearby inducing spikes, etc. An inductor (or simple ferrite bead) can sometimes help especially RF noise. Is it shielded?

[thebecwar]

can't get my hands on low esr ceramic caps since the minimum quantity order is like 65 bucks worth of low esr caps..zzz

Did you check Mouser? A quick search found this little guy 10uF 10V Low ESR SMD MLCC @ $0.14 a pop:
http://www.mouser.com/ProductDetail/TDK ... Kma%2fo%3d

Not sure if that will work in your specific application, but there are a ton of low ESR caps there < $3.

Oh and if my parts vendor had a minimum order quantity, I wouldn't shop there. Mouser, for example, is perfectly happy to send out one $.05 component, as long as you're willing to pay about $6 for shipping (USPS Priority <1lb rate to the US).

[pizza]

A pic of the board and a definition of what you mean by "noisy" would help - are the readings varying 1%, 10%, .1%?

Generally any cermic cap should have a low enough ESR. Many sensors are going to be noisy unless it is locked down in a faraday cage.

That said, I would look to how you are decoupling, isolating, or otherwise insuring the analog side of the circuit is avoiding noise from the various possible sources. The supply voltage might have ripple or spikes (when something else turns on), there may be things carrying current nearby inducing spikes, etc. An inductor (or simple ferrite bead) can sometimes help especially RF noise. Is it shielded?

[/quote]

please see attached logged data(10Hz from hyperterminal) and schematic and advise. thanks!
from eye observation the worst case of noisy data varies for about x=-21 fo x=20 so it's about a difference of 40 in decimal.
i did not design the board. there are four mounting holes at the corner with diameter of m2 size screws, so i mounted it to my platform with m2 screws, pillars and nuts.

[tz]

The noise isn't that unusual for a typical setup, but I would really need the layout, not the schematic for a reason. The schematic only shows there is a capacitor somewhere, not where it is located - near to or far from the chip, using thin or wide traces. Reducing noise is somewhat of an art, but the rules are simple. Also remember that if you are around DC or AC current, it will have a small magnetic field. You are trying to measure under 1% of the earth's magnetic field (that is about where the noise is), so how do you know it isn't just picking up the surrounding environment? You could try shielding it in an iron or steel box or go outside away from things and see if it reduces the problem. Beyond that it might be switching or induced noise.