AMC6821: I2C issues with Intelligent Temperature Monitor and PWM Fan Controller

Chris,

Do you have pull resistors for the I2C bus?
Have you debugged your I2C code?

-Kelvin

Hi Kelvin,

Thanks for the prompt reply.

Yes I have pull-up resistors on the I2C bus and I know that they are working, because I have other I2C devices on the same bus working correctly and I can also see data on the SCLK and SDA lines from a CRO.

I am using the same basic i2c code I used on a MAX7318 i2c chip on the same board (which works), which is written in LabVIEW (same language you use for your AMC6821EVM Evaluation board - is this available for download?).

The Master (microcontroller) sends the 8-bit register command for the Device ID Register (0x3D) and I then try to read one byte from this register (it should be the default value of 0x21), but it says there are no bytes available.

Hi Chris,

Another item that would benefit the debug of you communication is to share an oscilloscope capture of the I2C transaction. Can you provide the readable I2C frame?

Best Regards,
Matt

Hi Matt,

The attached CRO images is of me sending the Slave Address 0x18 and then the Command Register Address 0x3D (Device ID Register).  In my software I then also send the Slave Address again and the command to read this Register Byte.  However you don't see that on the CRO, because the first send receives a NACK indicating (I assume) that the AMC6821 didn't respond to this command.

After replacing the IC and before I connected up the 24Vdc 4-wire Sanyo 97BMB33_E fan, I checked the voltage on the fan PWM Input line to the fan and it measured 4V.
Could this have damaged the IC (when connected to the PWM-OUT pin 1 on the IC), because VDD on the IC was +3.3V?
From the data sheet the absolute maximum ratings is +6.5V so we are well within that, but I'm just wondering if this voltage, because it is higher than VDD will cause damage?

The PWM-Out is an open drain pin so I don't think a 4-V pull up voltage is a problem as long as the max sink current is less than 10 mA.

Thanks Kelvin,

Yes it's all going well now (mostly anyway).  Software-DCY mode works fine, but it is open loop and there are no alarms if the fan stops working.

Software-RPM mode is more complicated as it is closed loop, but it does give us alarms if the RPM level falls below a set level.

However when we are reading the TACH-DATA registers (0x08 and 0x09) for some PWM settings we are getting considerable noise on the TACH input line.

eg. for PWM = 75% we will get an RPM reading of say 7220RPM (TACH-DATA=831)  then all of a sudden it will jump to say 14,634 RPM - it appears very random.

Funnily enough at PWM=100% there is very little noise on the TACH input line and the RPM readings are rock solid at around 8185RPM (TACH-DATA=733).

How can we reduce the noise on the TACH input line for all PWM settings (ie. 0% to 100%) to obtain stable RPM readings?

Chris

OK don't worry about the TACH input I've sorted that issue out.

The last issue I have (hopefully) is simply how to turn the fan off in the event of an alarm (ie. fan stalls or something)

On start-up I have set the registers as follows.  All other registers are left at their default.

Config Register 1  -> 0x19 = Software-DCY mode
Config Register 2 -> 0x07 = TACH/PWM enabled
Config Register 3 -> 0x83 = leave as default
Config Register 4 -> 0x88 = Set MODE to 1
DCY Duty Cycle   -> 0x40 = PWM% is 25%
TACH Low Limit (Low Byte) ->0xA0 = TACH Low Limit (Low Byte)
TACH Low Limit (High Byte) -> 0x0F = TACH Low Limit (High Byte)

I then read the FANS (bit 1) value from the Status Register 1 and when it is set to "1" (0x42) the fan has gone to slow and so I want to shut it down.

To do this I then clear the PWM-EN (bit 0) value in the Configuration Register 2 and it does shut down the fan.

However when I re-start the microcontroller (and after setting the registers as above) the fan immediately has a Status of (0x42) and shuts down - why is this happening?

Chris

It seems I have an issue with writing to the Config Register 2.
To turn the fan off I write 0x20 to this register to disable the PWM-EN bit.
However when I repeat the above sequence of writing to the 7 registers - the Config Register 2 stays at 0x20 (it should have changed to 0x07).
Even stranger if I write 0x82 to this register (ie. to set RST bit to "1" to reset the fan) it resets/crashes my whole microcontroller.

Is there a particular order I should be writing the above registers in?
Chris

Chris,

Did you figure out the issue? Sorry for the late response.

-Kelvin

More or less yes - thanks Kelvin.

It turns out the IC was faulty and when we replaced it we could control it, albeit at a basic level.

If I had more time I would look into the closed loop control of the fan, but for now we are just setting a PWM value and that does what we need for now.

Chris