This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

LM3648: pin TEMP function doesn't work

Part Number: LM3648

circuit details:
device is LM3648TT.
pin IN=STROBE=HWEN=2,8V; TX=not connected;TEMP connected to NTC 100k.
L=1uH; Cout=22uF;LED=white UF=2.7 to 3.4V I_LED=0 to 400mA.
using torch mode only.

initialization sequence (reg 0x03=0x07=0x08=reset default) :
reg 0x02=0x00 //IVFM disable
reg 0x05=0x80 //lowest current
reg 0x09=0x01 //select TEMP, threshold 0,2V
reg 0x01=0x00 //standby, LED off
LED on sequence:
reg0x01=0x18 //awake from standby
reg 0x05=0x80 to 0xD9 //set torch current
reg 0x01=0x1B //LED on

action: LED doesn't light up. current through NTC, but insane values instead of specified 50uA, dependent from led current, see some listet measurements listet below:
U_NTC=131mV @ I_LED=9,5mA;
U_NTC=13mV @ I_LED=31,9mA;
U_NTC=23mV @ I_LED=54,3mA;
U_NTC=33mV @ I_LED=76,6mA
NTC voltage measured 20ms after turning on the LED current.

with last command reg 0x01=0x0B instead of 0x1B the LED works fine, but there is no current out of TEMP pin and FLAG2.TempTripFault is set to high always.

I ask for support activating torch function together with NTC temperature monitoring.

thanks in advance for helping.
best regards.

  • Hello,
    With the torch command 0x0B in reg 0x01, you are using the I2C controlled torch and the NTC function is disabled so the LED should turn on.
    You can remove the write of 0x18 to register 0x01. This puts the device in a somewhat unneeded state where the part is somewhat on with the LED off. Not sure this is the issue but it could be.

    At room temp, with a 100k NTC, with a 50uA bias, you will get 5V which is a bit high. The voltage will not hit this value because it is powered off of the input voltage. What temperature are you wanting this to turn off at? What is the Beta value? Also, if you can try putting a 10kohm resistor in there to see if everything powers up right it would be interesting data.

    Greg
  • Hi Greg, after posting and receiving the incomming reply I realized, that the text looks really terrible! All line feeds and new lines are gone. I am so sorry! What am I doing wrong? How to send a spreadsheet file?
  • Richard,

    There is a small paperclip icon in the toolbar you can use to add an attachment.  Once you have it attached, I will give it a look and try out the settings myself in our lab on Monday.

    Greg

  • Greg, sorry, I do not find any toolbar! i use Firefox. Do I need IE?
  • To get the tool bar, you need to click the Use Rich Formatting link in the bottom right corner of the reply window.  I believe it is browser independent. 

  • Hi Greg, I did as you told me, but there is no toolbar visible on the site. Only two buttons: 'Select Tags...' and 'Post'. Subsequent I repeat the post in rich text format. Hopefully this works! Tell me, if not. I have to write the spreadsheet informations this way manually then. REPEATED POST: removing write 0x01=0x18 doesn' t change anything. The NTC data are R_25=100kOhm; 1%; B_25/50=4700K; B_25/85=4750K. Due to data sheet table R_NTC @ 100°C = 3939Ohm. U_NTC= 50uA*4000Ohm=0,2V. That's the temperature threshold I want the driver shut down automatically. I tested some logically and also some illogical driver settings, because I learned, that data sheets not always are error free or incomplete. The results are listed in the document attached. I testet the TEMP function at torch mode only. Maybe it is available at flash mode only? Please have a look at the attached document. Cripes! I don't find anywhere the 'Paste from Word'-button at the site! As quick alternative I attach the spreadsheet file in csv format. Hope, this makes not too much discomfort, sorry. LM3648TT TEMP function,,,,,,,,,,,,,,,,,,,,,, ,,,,,,,,,,,,,,,,,,,,,, ,,,,,,,,,,,,,,,,,,,,,, NTC resistor:,R_25=100kOhm; 1%; B(25/50)=4700K; B(25/85)=4750K,,,,,,,,,,,,,,,,,,,,, "temperature threshold using TEMP detect voltage threshold = 0,2V:",,,,,,,,,,,,,,,,,,,,,, "R=0,2V / 50uA = 4000 Ohm",,from data sheet table: R_NTC @ 100°C = 3939Ohm,,,,,,,,,,,,,,,,,,,, measured LED temperature: less than 84°C @ I_LED=402mA.,,,,,,,,,,,,,,,,,,,,,, ,,,,,,,,,,,,,,,,,,,,,, ,,,,,,,,,,,,,,,,,,,,,, uniform initializing sequence,,,,,,,,,,,,,,,,,,,,,, register name,register address,data direction,value,comment,,,,,,,,,,,,,,,,,, Enable,0x01,wr,0x00,"mode=standby, TX pin and Strobe and TORCH/TEMP pin disabled",,,,,,,,,,,,,,,,,, IVFM,0x02,wr,0x00,IVFM disabled,,,,,,,,,,,,,,,,,, Flash Brightness,0x03,wr,0x80,set lowest current,,,,,,,,,,,,,,,,,, Boost Config,0x07,no access,0x09,we use the reset value,,,,,,,,,,,,,,,,,, Timing Config,0x08,no access,0x1A,we use the reset value,,,,,,,,,,,,,,,,,, ,,,,,,,,,,,,,,,,,,,,,, ,,,,,,,,,,,,,,,,,,,,,, testet settings following the initializing with parameter I_LED = 4/99/402mA,,,,,,,,,,,,,,,,,,,,,, register name,register address,data direction,value(s),,,,,,,,,,,,,,,,,,,comment Torch Brightness,0x05,wr,0x80,0x91,0xC7,,0x80,0x91,0xC7,,0x80,0x91,0xC7,,0x80,0x91,0xC7,,0x80,0x91,0xC7,set current TEMP,0x09,wr,0x01,0x01,0x01,,0x01,0x01,0x01,,0x41,0x41,0x41,,0x00,0x00,0x00,,0x40,0x40,0x40,set TEMP function Enable,0x01,wr,0x0B,0x0B,0x0B,,0x1B,0x1B,0x1B,,0x1B,0x1B,0x1B,,0x1B,0x1B,0x1B,,0x1B,0x1B,0x1B,turn LED on FLAG1,0x0A,rd,0x00,0x00,0x00,,0x00,0x00,0x00,,0x00,0x00,0x00,,0x00,0x00,0x00,,0x00,0x00,0x00, FLAG2,0x0B,rd,0x00,0x00,0x00,,0x01,0x01,0x01,,0x01,0x01,0x01,,0x00,0x00,0x00,,0x00,0x00,0x00, LED illuminate,,,on,on,on,,off,off,off,,off,off,off,,on,on,on,,on,on,on, U_NTC [mV],,,0,44>,208>,,0,0,0,,0,0,0,,0,44>,177>,,0,10,157>, U_LED [V],,,"2,55","2,74","2,9",,0,0,0,,0,0,0,,"2,55","2,75","2,94",,"2,54","2,71","2,94", ,,,,,,,,,,,,,,,,,,,,,, Annotation:,'>' means voltage decreases with increasing LED temperature,,,,,,,,,,,,,,,,,,,,, ,,,,,,,,,,,,,,,,,,,,,, As could be seen: In case U_NTC not zero the U_NTC initial voltage at ambient temperature (about 28°C) differs with LED current!,,,,,,,,,,,,,,,,,,,,,,
  • Richard,

    It looks about the same as before but that is ok.  I will be able to follow it.  I will get back you Monday with some more information from my side.

    Greg

  • Richard,
    I tried your sequence you initially provided and I don't see any issues. I see the thermistor voltage come up and does not appear to cause the LED current to move. Have you ordered a LM3648 EVM to try out your sequence? While the torch current will be different as the LM3648 max torch current is lower, but in general it should behave the same with regards to the control.

    Greg