LM3492: It can NOT be read COMM pin Bit Pattern

Hello,

From 8.3.9 in the datasheet, DIM1 needs to be used:

"Except for the power good indication and the overtemperature alerts, all data flow through the COMM pin is
serial and is latched by the falling edge of the signal applying to the DIM1 pin, even when channel 1 of the
current regulator is disabled. If the DIM1 pin remains only low or only high, either by an external circuit or by
allowing it to open and pull low internally, data does not flow. Figure 24 and Figure 25 show timing diagrams of
reading and writing a bit from and to the device through the COMM pin.
Pull up the COMM pin by an MCU I/O pin, which has pullup capability, or an external resistor RCOMM connected
to the VCC pin. Without this capability, the voltage of the COMM pin remains at zero. The rise time of the output
signal of the COMM pin depends on the pullup power. If the rise time is long (RCOMM is too large or pullup power
from the connecting MCU I/O pin is too weak), data may be ready after a longer duration after the falling edge. In
this case, the design requires a longer delay between the falling edge latching and the (input or output) bit."

Please read this section of the datasheet.

Best Regards,

Hello Mr. Irwin,

Thank you for your response.

However we read the datasheet and tried to read COMM pin Bit Pattern, we still have problems what we can NOT get the Bit Pattern.

Even the first bit "0" can NOT be detected. 

Would you have some more detail information a kind of timing-chart of DM1 pin clock input and COMM pin Bit Pattern output?

Best Regards,

Hello,

Do you have a pull-up on the COMM pin?

Are you cycling DIM1 low and waiting the appropriate time to read the first bit, then the second, third and fourth?

"Except for the power good indication and the overtemperature alerts, all data flow through the COMM pin is
serial and is latched by the falling edge of the signal applying to the DIM1 pin, even when channel 1 of the
current regulator is disabled."

Data will not flow if DIM1 remains high or low.  The data latches on the falling edge of DIM1.

Best Regards,

Hello,

1) Yes, we pull-up the COMM pin by 52.3kOhm to Vcc line.

2) Yes, we are cycling DIM1 low by 4 times to read the COMM pin Bit Pattern.

 But COMM pin remains high every DIM1 pin's falling edge.

Would you guess what is the reason we can NOT read the Bit Pattern? 

Regards,

Hello,

1)  Have you created faults to read?

2)  How fast and for how long are you cycling DIM1?

Is the drive enabled when you are reading?

Best Regards,

Hello,

1) Yes, we have created open-status. One of 2 LED-strings has been made open.

2) Cycling frequency is about 300kHz.

We just input 600 cycles to DIM1 pin by MCU.

After 600-cycling DIM1, we pulled EN pin low about 190-nsec.  

Then, in order to read the Bit Pattern, we input 4-cycling DMI1.

Is there any miss understanding of us?

Best Regards,

Hello,

Are you waiting after the EN pin pulled low?  The device has to go through a test to see if there is a fault:

"8.3.9.3 Output Current Undervoltage Indication
The LM3492/-Q1 device gives an IOUTn (n = 1, 2) undervoltage indication if the voltage of the IOUTn pin when
DIMn is high is lower than its minimum required voltage which can regulate ILED, and the voltage of the CDHC
pin reaches its maximum. These conditions remain while the device applies 508 consecutive dimming signals on
the DIMn pin. This means that the current of the LED string n does not reach the regulation value. In most cases,
the IOUT undervoltage indication can be regarded as an open fault of the LED string n. A bit pattern (see
Table 1) can be read from the COMM pin. The device does not latch off and continues to operate in the
presence of the IOUT undervoltage indication."

"Apply the clock signal on both DIM1 and DIM2 pins when the COMM bit pattern is read by an external MCU.
Before reading the COMM bit pattern, pull the EN pin low for approximately 200 ns to reset the COMM bit
pattern. This situation does not affect the operation of the boost converter and the current regulator. After EN is
reset, if the IOUT overvoltage or undervoltage condition lasts for 508 consecutive clock cycles, the COMM pin
sends the COMM bit pattern for the MCU to read."

I am just trying to interpret the datasheet commands on this part.  I would try waiting for the 508 consecutive clock cycles before trying to read the COMM bits, this wait is after the EN cycle.

Best Regards,

Hello,

Thank you for your reply, I could read the COMM Bit Pattern after applying the clock signal of about 510 cycles on both DIM1 and DIM2. 

In order to read the Bit Pattern, it required to apply the clock on not only DIM1 but also DIM2 at the same time.

And the 508 consecutive clock cycles were not enough to read in case that it's dimming function is NOT used.

Is it correct to use this product?

Best Regards,

Hello,

It should only require cycling DIM1.  I asked a colleague more about this.  There is limited information on this part.  He states:  "the limitation of LM3492 communication link is the need of a clock signal appeared on DIM1 pin (it has to get 99% duty cycle if customer needs a full brightness and communication at the same time)."

Best Regards,

Hello,

Sorry, I haven't responded on this issue and late reply.

Actually, We don't supply the continuous clock to DIM1 that you mentioned 99% duty clock to achieve full brightness.

Is the following usage wrong? Is there any problem that LM3492 shall be damaged otherwise be malfunctioned?

 - DIM1 and DIM2 are wired together and connected to an output port of external MCU. 

 - To keep DIM1 and DIM2 high generally to achieve full brightness operation.

 -  In order to read some error status or power-good indication on the LM3492 operating,    

  At the periodic, ex. every 500msec,  the MCU apply about 510 clocks to both DIM1 and DIM2 then read the COMM pin Bit pattern.

I'm waiting for your reply.

Best Regards,

Hello,

It should not get damaged.  Have you been successful in reading the COMM bit pattern reliably?

Best Regards,

Hello, 

Thank you for the good information for us.

Yes, we have been successful in the reading process.

We have 2 LED strings and each string connects to IOUT1 and IOUT2. 

Then we have tried to read the COMM bit pattern in any case that one or both of LED string was/were open or short.

We could read the bit pattern correctly. 

I have another question regarding this issue.

Our targeting LED is specified 5.7V as typical Vf value. 

Therefore we designed LM3492 board that LED driving voltage should be boosted about 10% and more to light LED strings on full brightness.

However it doesn't boost the LED voltage, because the LED's Vf value will be varied +/- 20% as individual differences.

Finally, we found that it doesn't boost the LED voltage in most cases.

Is there any problem that some internal circuit block of LM3492 would be stressed or be damaged?

I'm waiting for your reply.

Best Regards,

Hello,

What is your input voltage, including range?  Also what is your input voltage during your testing where it is not boosting.  The LM3492 runs down to 4.5V input.  80% of 5.7V is 4.56V.  It probably won't boost since the DHC is trying to minimize the headroom, the DHC.  To test this you can try increasing your LED voltage.

Best Regards,

Hello,

At first, I would like to make sure what is our concerning point.

It is NOT matter for us that it doesn't boost the input voltage to light LED strings.

Our targeting LED's Vf figure is including range in individual differences and we know it cause the case that it is not boosting the input voltage.  

Therefore I would like to know whether LM3492 should be damaged and/or be degraded or not if it couldn't be boosting.

Could you teach us any problem what you could confirm and /or guess internal circuit point of view?

Next, let us inform the figures of testing board for answering to your request, as followings;

 - The input voltage including rage is 12V +/-5%.

 - The input voltage during testing is about 12V which is supplied from AC adapter.

 - A LED string has 2 LEDs and two LED strings are connected to LM3492.

Best Regards,

Hello,

There are a lot of possibilities.  Are you saying that your LED loads are 5.7V times two or 11.4V?  If the input is 12V it cannot boost to 11.4V.  If this is the case I would try lowering the input voltage or increasing the LED stack voltage.  The LM3492 shouldn't be damaged however it is a boost so if the load is a lower voltage than the input the boost diode and load could be damaged.  If you look at a boost the input goes through the inductor and a diode to the load.  A boost output needs to be higher than the input.  Two channel current regulator should prevent damage but it is not intended to run this way.

Best Regards,