29 Jan 2013: Randall installed replacement CoMo B sent by Gino

30 Jan 2013: (Jim & Jamie)

Ran calibration procedure using the profile dbbc_config_file_102_calibration.txt and found nulls as follows:

old values 100 75 80 100

new values 103 86 102 97

edited all _102_ config files on dbbcyg to these new values

Then a calibration run to peak up on pcal tones. Yg DBBC optimisation, Jan 2013

Dear Gino,

At Yarragadee today Randall swapped the replacement CoMo B you sent with the old one and I ran some tests. The conclusion is that it has made no difference and our phase cal amplitudes are about a factor of 3 lower in CoMo B than CoMo A. Here’s what we did:

1. Ran the usual calibration procedure with a 764 MHz tone and filter 1 in all CoMos, chose the first null as the phase.

2. DBBC config for a R1/R4:
	CoMO A : X-band with filter 1
	CoMo B: X-band with filter 4
	CoMo C: S-band, filter 2
	CoMo D : S-band, filter 2

3. Measured phase cal tone strength as a function of target power level. The peak strength (in Whitneys) for each CoMo was:

	CoMo	max ph cal		at this CoMo power level:
	A		~50				35000
	B		~15				25000
	C		~60				40000
	D		~70				35000

	Note that the maximum power level CoMo B could produce was ~35000. All the others can go past 50000

4. I then ran the above tests again but with the following changes:
	CoMo A: X-band with filter 4 (and corresponding high BBC frequencies)
	CoMo B: X-band with filter 1 (and corresponding low BBC frequencies)

	The results were:
		CoMo A was well behaved with a maximum phase can signal of ~40 Whitneys at a power level of 25000
		CoMo B was poorly behaved with maximum phase cal levels of ~15 Whitneys

	So we get low phase cal levels through CoMo B regardless of whether we use filter 1 or 4.

5. We also swapped the X-band IF inputs to CoMo’s A and B to see if the input signals were at fault, but there was no change.

6. Randall then changed the the SMA IF cable running from CoMo B to the core (?) board with a spare and checked all the SMA connections for tightness. Some of the S-band cables were a little loose but everything else was good. However, there was no improvement in performance through CoMo B.

So, given that the old and new CoMo B units have similar behaviour, my guess is that the CoMo hardware might be OK. It looks like the "IF out" cable of CoMo B is also OK. Given that the high end of X-band looks good through CoMo A, I don’t think the cause is our input signal strength or band shape.

One further check we could do would be to swap CoMo A and CoMo B. If the problems are then seen in the new "CoMo A”, then that unit (and the old one!) is faulty. I don’t think this is likely, but we can try. 

However, I guess the fault is further downstream. Can you suggest some other tests we can try? Please feel free to log in via Hamachi and try things out. I’ll leave the 764 MHz tone on, which is on input 3. The DBBC is currently set up for a calibration run.

If you want a R1/4 config, the file is dbbc_config_file_102_rapid.txt and IF settings are:

dbbcifa=1,agc,1
dbbcifb=4,agc,4
dbbcifc=2,agc,2
dbbcifd=2,agc,2

31 Jan 2013:

Randall replaced CoMo B IF cable and checked connections again. Nothing moved. Re-did phase calibration and new phases are:

103 86 80 105

Edited all cal files with these new values.

Re-ran phase cal amplitude peaking. Looks like S-band lower could be optimised a little by going to a target of 35000. All others seem OK