Noise Injection into Cooling Pipe

 NEW RESULTS (25/04/02)

Ned Spencer's grounding scheme

Noise Injection over a transformer into the cooling pipe. In between the transformer and the cooling pipe is a resistor, to measure the voltage drop, from which the current conduction could be calculated.

Thermistor Temperature: 40 deg. C

Macros to analyse the data are provided by Joern.

1.) Measurements with sinus-waveformgenerator Wavetek 395 (as a loan from the barrel systemtest)

a.)     different Frequencies  4Mhz, 10Mhz, 20Mhz, 30Mhz and 40Mhz
         the current through the cooling pipe was for all measurements set to 50mA peak to peak.

        Difference in inputnoise

The highest increment for the inputnoise could be found for 10Mhz, but also for 20MHz and 30Mhz an increment could be observed. For 4MHz and 40Mhz the increment is small.

b.)    different currents:  12mA, 24mA, 36mA, 48mA and 60mA

        @ 10MHz                     @30MHz

        !!!! 1Vpp is equivalent to 6mA !!!
 
 
 
 

2.) Measurements with HP-Pulser 8012B

 

However, the current was not changed for the following measurements

Difference without noiseinjection / with noiseinjection

• ENC in absolute values for all four modules
• ENC difference for all four modules
• ENC difference for all four modules only Link 0 (link 1: preamp=0, shaper=0, dead) (Tthermistor=32 deg C!!)
• ENC difference for all four modules only Link 1 (link 0: preamp=0, shaper=0, dead) (Tthermistor=32 deg C!!)
• ENC difference for K5-302 single measurement
• ENC difference for K5-400 single measurement
• ENC difference for K5-402 single measurement
• ENC difference for K5-303 single measurement

Conclusions

• Noise was injected into the cooling pipe and could be observed as aditional noise in the modules.
• Structure of Noiseinjection different for the modules/positions/frequencies (10MHz seagull for 302 and 303)
• 10Mhz: the noiserise of the two modules on the deep cooling-blocks is in average the double of the noiserise of the modules in the upper positions. (low statistic!)

 

 
 
 
 
 

---


 

NEW RESULTS

Further Noise injection studies (P.Bernabeu, P.Dervan & A.Greenall)

• Noise injection procedure, as outlined above. 6MHz Sine Wave injected onto the cooling pipe.
• All modules have cooling block and Spine shunt shields implemented.
• Diagram showing module position and cooling pipe profile.
• Excess noise profile is shown for all 4 modules.

 

 

The procedure was then repeated but this time with the 2 'upper' modules K5-402 and K5-303 swapped - this was done to check whether the noise profile was position dependant and not module. The repeated noise profile plot can be seen here.

Module K5-402 was then modified so that the cooling block shunt shield was short-circuited to module AGND - this is not truely representative of Tony Smith's grounding scheme as PP0 (patch panel 0) DGND was connected to the radial foils and the power supply cables screen connected to DGND return. Noise injection was then repeated, with all 4 modules back into their original positions, this noise profile can be found here.
 

Observations

With the cooling block shunt shield:
 
• Geometry of the cooling blocks and pipes can be clearly seen.
• Excess noise can be seen at the edges of all modules, even where there is no overlapping module(s) present.

 
With the shunt shield bypassed, on a single module only - all the others still having their shunt shield(s):
 
• Module without shunt shield shows significant drop in excess noise - noise profile is flattened with also little evidence of excess noise at edges.
• There is also a global effect across all modules with their excess noise being reduced.
The action of swapping K5-402 and K5-303 highlights an apparent difference in module 'type' (Cicorel or Dyconex) performance and their suscesptibility to noise pick up!
 

Systemtest status

All 4 Outer modules have now been modified so that the cooling block shunt shield can either be implmented or short-circuited to AGND. This should allow further tests to continue in the evaluation of module performance with respect th the differing grounding/shileding schemes.