Calculation of PTR and the relationship between PTR and obscuration

Hello . I have two questions .

1. calculate  PTR

Refering to

PTR = Ipd/Iledx_pk/Rpd/ηled


      Code = output of FIFO ; Q=1.15  (when R = 100k) ; PULSE_COUNT = 2 ;

(2)ILED_PK = Icourse*Ifine*Iscale

(3)Rpd=0.26(for blue);0.42(Ifor IR)

(4)ηled =  ηN*K*C

          for blue ηN=0.38 ; for IR ηN =0.22

          for IR : k=1

          for Blue :  k=A0+A1*ILED_PK+A2*(ILED_PK)^2+A3* (ILED_PK)^3;

      for Blue: C = (Gain*ILED_PK +INTER)/ NORMINAL_SCALAR


            Gain= 17*( Gain_COEFF-112)/256-17

            INTER = 8*( Int_COEFF-128);

            NORMINAL_SCALAR = 17*ILED_PK+622;

So I will calculate Blue PTR as a example .

     (a) IPD=Code*Q*(1/PULSE_COUNT)= Code*1.15/2=0.575000*FIFO

     (b) Icourse = 50.3+19.8*6;    //6  0x23bit[3:0]

          Ifine = 0.74+0.022*12; //12 0x25bit[4:0]

          Iscale = 0.1+0.9*1;     //1  0x23bit13

          ILED_PK = Icourse*Ifine*Iscale=169.776398mA

(c) Rpd=0.26

(d) blue ηN=0.38

(e) k=9.8976e-01-5.1448e-03*ILED_PK+2.0287e-05*(ILED_PK)^2-2.9645e-08* (ILED_PK)^3

     =0.555975  ( for ILED_PK , I should use mA or A here?)

(f) As I read from efuse register ,for Blue,



So :

       Gain= 17*( Gain_COEFF-112)/256-17 = -22;

       INTER= 8*( Int_COEFF-128) = 552;

       NORMINAL_SCALAR = 17*ILED_PK+622 = 3508.198761;     ( for ILED_PK here , I should use mA or A?)

       C=(Gain*ILED_PK + INTER)/ NORMINAL_SCALAR= -0.907326

So for Blue PTR = IPD / ILED_PK * (1/Rpd) *( 1 / (ηN*k*C) )  = -0.067954*FIFO

Similarly,for IR, Gain_COEFF=0xC5,Int_COEFF=0xAA, C=0.841255 , So IR PTR = 0.044633*FIFO

Obviously Blue PTR is less than 0 ,It‘s wrong . And I don't know if  IR PTR is  right or wrong .

If FIFO>1000,Then PTR is over 40 . As I see in the files you provide ,Is it to high?

  the max PTR is 4.4

I think I have followed you instruction . Maybe I confused what unit  of ILED_PK should be used ?

2. convert PTR to obscuration 

As CN0537 says, PTR =γ*β . So I think If I know PTR ,I can calculate β , which  is the obscuration level in ft-1.

But in

You say "our device does not measure obscuration per foot. We measure what is essentially the inverse of obscuration per foot"

I dont quite understand it .I think that the higher FIFO is, the PTR is higher . It means the the smoke is thicker .

So the obscuration is higher .  Is it wrong ? So If you can expain it to me , or is there any  files I missed ?


  • +1
    •  Analog Employees 
    on Jan 29, 2021 6:03 PM


    For question 1, you have to use LED current in mA so use 169.77 in the calculations. Gain calculation is incorrect, you subtracted 17 instead of adding. Gain_cal_x should always be positive. The intercept calculation for blue looks correct. I suggest checking the gain calculation for both LEDs. If IR LED current is ~230mA (I'm assuming you are using the default config described in CN0537 or the datasheet) then PTR should be <10

    For question 2, your understanding is correct. Obscuration measurement is a measure of how much the light signal got attenuated when smoke was present. In an obscuration measurement light from an source is shining directly on a detector so with no smoke, light signal is maximum. When smoke is greater obscuration (light attenuation) is higher. ADPD188BI measures the light scattered by smoke particles, so when more smoke is present more light is scattered and PTR is higher. So when smoke increases both obscuration and PTR increase.

    The statement  "our device does not measure obscuration per foot. We measure what is essentially the inverse of obscuration per foot" is referring to the mechanism. ADPD188BI receives more light with more smoke, obscuration measurement receives less light. But mathematically as explained above obscuration measurement and PTR are proportional to each other.



  • Hi ,

          For my second question,there is one thing I forgot to mention ."As CN0537 says, PTR =γ*β ."  I think if I can get FIFO ,I can get PTR directly. 

          But in

    Kevin says

          "As I mentioned in the above email, our device does not measure obscuration per foot. We measure what is essentially the inverse of obscuration per foot. In order to understand how the output of the ADPD188BI correlates to obscuration per foot, this can only be done by empirical testing. You need to build your smoke detector and test it in the standards facility, e.g. UL smoke labs, so you can empirically measure what the output of the ADPD188BI yields at the obscuration per foot thresholds dictated by the standards."

          As I konw, for diferent kinds of smoke ,the IR_PTR/Blue_PTR should be a constant . 

    So If I know PTR , why should I " build your smoke detector and test it in the standards facility"?

    If  I want to distinguish different kinds of smoke , knowing PTR isn't enough?


  • +1
    •  Analog Employees 
    on Mar 2, 2021 7:51 PM in reply to hitlhd


    If you're building a smoke detector that is going to be deployed in a residential or commercial setting it needs to be certified. For certification your smoke detector design has to meet the standard that's applicable to the geographical region you're targeting.

    Standards specify smoke measurements in obscuration units because the measurement is performed using an obscuration technique. Standards specify alarm thresholds based on obscuration units. With ADPD188BI, you get PTR units. If your smoke detector design uses ADPD188BI that means your alarm thresholds will have to set in PTR units. By going to an approved facility, you can see how PTR units correlate to obscuration units. Without that you don't know what to set the alarm thresholds to.

    The ratio of blue PTR to IR PTR shown in the bar graph are for a specific hardware design, and are not based on a large population. Your hardware design (smoke chamber, mechanical housing etc.) and preferred ADPD188BI device configuration will affect the PTR measurements. The PTR measurements may also vary within a large sample of smoke detectors. The sample to sample variation has to be well understood when setting alarm thresholds. When you make your smoke detector design, it's possible that the PTR measurements will be different from the numbers shown in the bar graph.

    You can see from the bar graph, the ratio of blue PTR/IR PTR is similar for some of the smoke sources. The bar graph doesn't say how big the variance was in the underlying dataset, only average numbers are shown. The ratio of blue/IR can be used to distinguish steam from smoke, because there's a big difference. 

  • Hi,

         Thank you for your answer . You have solved my question !