First thermal measurements on modules on stave

1
First thermal measurements on modules on stave
Ge 31/10/03
M.Dameri, F.Vernocchi et al. (R. Cereseto,
G.Gariano, C.Gemme, M. Parodi, C. Pizzorno,
L.Rossi, A.Rovani, E.Ruscino)
3 FEI1 modules 20210020420058 in position 3,
power 3.3W (chip7 off, chip8 malfunctioning) 202
10020420026 in position 4, power 2.4W (chip14
off) 20210020420023 in position 11, power 2.5W
(chip 813 off) have been mounted on a prototype
stave with the production machine and
procedure. Thermal images have been taken with
the module fully configured and with a power
consumption as detailed above. Cooling was done
circulating fluid at 10C in a room temperature
environment.
2
chip7
chip0
Tntc120C
20210020420023 no power
High digital current on pigtail
Tntc17.50C
20210020420023 power on
3
Tntc130C
20210020420026 no power
Tntc16.50C
20210020420026 power on
4
Tntc12.60C
20210020420058 no power
Tntc16.10C
20210020420058 power on
5

• Measurements were done switching on one module
  at the time.
• Comparison between power-off and power-on
  cancels most of the systematic effects.
• There is good agreement between Tntc and thermal
  images
• The temperature over the modules is quite
  uniform with the exception of chip8 of module
  20210020420023 that has too much digital
  consumption.
• It is not possible to see individual chips which
  are switched off
• The DTntc between the fluid temperature
  (assumed to be the same as the module when power
  is off) and the module with power on is 1.67
  (mod. 20210020420058), 1.46 (mod. 20210020420026)
  and 1.4 (mod. 20210020420023) degrees per W
• Assuming DT1.50C/W implies that the maximum
  allowed temperature drop happens with 10W modules
  (i.e. we are very much on the safe side).
• ()The NTC measurement should be redone at fluid
  temperature closer to room temperature to see if
  the above assumption holds.

6
Glue thickness has then been measured with a
microscope, looking at the module cross section,
as shown
chip0
chip7
Glue thickness
Results are
7
Glue layer is not uniform and the temperature on
NTC (on chip 14 toward the module center) feels
this. The glue layer thickness does not appear to
be very critical (1 C per 10W), unless we go
above 200mm. We do not observe local hot spots
indicating that the silicon and C-C heat
transmission coefficient smears out local effects.
8
Geometrical position has been checked, measuring
the edge of the pads on the FE chips (reference
crosses are not visible in those modules) in
positions A and B
A
B
The lateral displacement relative to the stave
reference system are 20210020420058 A-20mm,
B 0 mm 20210020420026 A 0mm, B 10
mm 20210020420023 A-40mm, B-100 mm All
measurements indicate a problem in the placement
of module 20210020420023