DLP2000: Using DLP2000 or DLP650LE with a pulsed laser

Hello User,

Please see the responses from the team in red. 

1. Mirror Surface to Bulk Mirror Delta
   95. q=95.49 W/cm^2
       1. focused on a 2mm diameter area at the center of the DMD
   96. q_adjusted = q*(1-MR) = 5.73 W/cm² (this assumes MR = 0.94, at 520 nm MR = 0.89 and q_adjusted = 10.5 W/cm^2)
       1. MR=0.89 at 520nm (from DLPA027B)
   97. deltaT_Mirror Surface to Bulk Mirror = 2*q_adjusted*(1/k) sqrt((alpha*t_pulse)/π)+T_i = 20.3 C deltaT_Mirror Surface to Bulk Mirror = 2*10.5 W/cm^2 *(100^2 cm^2 / 1 m^2)*(1/160 W/m-K)* sqrt((6.47e-5 m^2/s*[2.5 ms * (1s / 1000 ms)]/ π)+ 0 = 0.3 ºC

*The mirror surface to bulk temperature rise is quite small in this case

1. k=160 W/(m-C) (from DLPA027B)
2. alpha=6.47E-5 m^2/s (from DLPA027B)
3. t_pulse = 2.5ms
4. T_i=20 ºC (assumed starting ambient temperature) The mirror temperature rise above the bulk is just 0.3 ºC.  If the bulk mirror is 20ºC, then the surface is 20.3 ºC

2. Bulk Mirror to Silicon Delta
   5. Q_incidident mirror = q * pitch² = 5.46E-5 W 
   6. Q_mirror = Q_incident mirror * (FF_on*(1-MR)) = 5.59E-6 W
      1. FF_on = 0.931 (from DLPA027B)
      2. MR=0.89 at 520nm (from DLPA027B)
   7. T_f  =  T_i + (Q_mirror * R_mirror to silicon) = 22.5 ºC
      1. T_i=20 C (assumed starting ambient temperature for initial pulse)
      2. R_mirror to silicon = 4.47E5 ºC/W (from DLPA027B not in DLPS140B)
   8. T_off = 0.1225s >> 5τ [fully cools]
      11. τ=11.49μs (from DLPA027B)
   9. deltaT_Bulk Mirror to Silicon = T_f + (T_i + T_f) e^(-t_pulse/τ) = 22.5 ºC
      1. T_i=20 ºC (assumed starting ambient temperature for initial pulse)

Calculation is correct.  The bulk mirror temperature rise above the silicon is just 2.5 ºC. 

3. Silicon to Ceramic Delta
   1. Q_electrical = 0.045 W (from DLPS140B example)
   2. α_DMD = (1 - Overfill) *((FF_off * (1 - MR)) + (1 - FF_off)) + (2 * α_window) + Overfill = -4.99 <= This seems wrong

This calculation assumes the entire array is illuminated or it possibly even has overfill outside of the active array.  If the illumination spot is entirely within the active array, simply assume overfill = 0.  Then α_DMD = 0.36964.  This is the absorptivity of the mirrored array plus bulk window absorption.

1. FF_off = 0.724 (from DLPA027B)
2. MR=0.89 at 520nm (from DLPA027B)
3. α_window = 0.007 (from DLPA027B and cross-referenced with DLPA031E)
4. overfill = 1 - (Active Array Area/Incident area) = -17.6 <= This seems wrong (just assume overfill = 0 for the underfilled array case)
   7. Active Array Area = (7.57μm * 1280) + (7.57μm * 800) = 5.85E-5 m² (DLP2000 is a 640 x 360 array with 7.56 um pixel, DLP650LE is a 1280 x 800 array with a 10.8 um pixel. Just choose one or the other for this calculation.)
   8. Incident Area = 0.25 π d^2 =3.142 d²  = 3.14E-6 m²
      1. d = 2mm (assumed projection area)
      2. I do not intend to project onto the entire DMD, as the pendulum will oscillate along the DMD width. 

• I believe that the overfill is intended to reflect how much light is incident onto the ceramic. 

1. duty cycle = t_pulse/(t_pulse+t_off) = 0.02
   2. t_pulse = 2.5ms
   3. t_off = 0.1225s
2. q_average optical power density = q * duty cycle = 1.91 W/cm²
3. q_average optical power = active area area * q_average optical power density = 1.12 W (This should be average power density * illumination area for your case, so 1.91 W/cm^2 * 3.14E-6 m² = 0.06W)
4. Q_illimunation = α_window / q_average optical power = -12.3 W <= This seems wrong (This should be α_DMD * Q_incident = 0.36964 * 0.06W = 0.022W)
5. deltaT_Silicon to Ceramic = (Q_electrical + Q_illumination) * R_Silicon to Ceramic = -6.12 ºC <= This seems wrong (delta_T_Silicon to Ceramic = (0.045W + 0.022W) * 8ºC/W = 0.54ºC
   1. R_Silicon to Ceramic = 0.5 ºC/W (from DLPA027B, not in DLPS140B) (Technically we should generate a new R_Silicon to Ceramic from a package thermal model since the value will be higher because we have all the heat load in a small spot size. Note: DLP2000 is 8ºC/W with full array illumination.  The 2 mm spot covers about 24% of the array.  Even if the package resistance increased to 16 ºC/W, the temperature rise would be ~1ºC.  Electrical load is spread over the entire area so the effect would be less than this, so 1ºC is probably worst case.)
6. Mirror Surface to Ceramic
   56. T_mirror surface = T_ceramic + deltaT_Mirror Surface to Bulk Mirror + deltaT_Bulk Mirror to Silicon + deltaT_Silicon to Ceramic = 56.68 ºC

(This should be T_mirror_surface = 20ºC  + 0.3ºC + 2.5ºC + 1ºC = 23.8ºC.)

A few related questions: 

1. In Section 6.5, the DLP2000 and DLP160CP each provide only the Thermal Resistance as measured at TP1. Why not also provide the R_mirror to silicon and R_Silicon to Ceramic? (Mirror to silicon temperature rise is only significant with pulsed sources. Most of our applications use continuous sources.  R_silicon-to-ceramic is the same as thermal resistance as measured at TP1 since TP1 is on the ceramic)
2. The DLP160CP provides a small area for a thermal interface, but the DLP2000 does not. Is there any documentation on the thermal interface for the DLP2000 ribbon? (The DLP2000 is intended for low power applications and therefore does not have a dedicated thermal interface area)
3. Above 800 nm, both DMDs are limited to 10 mW/cm^2; however, does this not depend on the duty cycle? I am asking this to understand if the 5W 980nm laser I have available will be a useful substitute. (These DMDs were intended for visible applications and therefore were not tested above 800 nm. TI offers several DMDs intended for IR applications such as the DLP4500NIR and DLP650LNIR)

Regards,

Alex Chan