Re: Preamp tests

[Gerard Visser <gvisser@indiana.edu>]

Hall D Electronics:

Hi Fernando,
	Yes these are good points... I have already asked it too, why are the 
wires connected in this fasion? There was some problem I believe, that 
lead them to disconnect the shields. Yves/Curtis, can you comment on that?
	I have already bought some 330 pF 2.5kV capacitors for AC coupling, 
noting as you did that 33 pF is just too low. Even 330 pF is a bit low ( 
1/(2pi*10MHz*330pF)=48Ohms compared with the preamp input impedance ~100 
Ohms ). But it is probably as good as it can be in practice. I'll be 
retrofitting the HV board I was sent with the new capacitors.
	The intended connection of the shields, as I discussed long ago with 
Mike McCracken, is as you note below. There is a trace, bypassed to 
ground, at the HV potential, that is intended to be connected to the 
shields of the cable. Besides the low voltage rating of the cable, the 
other reason to bias the shields in this way is to not have a large 
stored charge in the cable, which could easily produce extra noise if 
there was any capacitance fluctuation (vibration).
	The connecting wires in the present CDC setup are obviously much longer 
than we would want in the real design, this will have to be addressed, 
and probably contributes to the oscillation problem. Especially it is 
needed to make short direct connections to the chamber endplate, 
otherwise it is perhaps the whole detector acting as an antenna not 
merely the signal connecting wires. The signal connecting wires should 
also not be routed right near the output signal cable (i.e. preamp board 
has wrong orientation on the HV board in the present design - see 
http://www.jlab.org/Hall-D/software/wiki/images/8/84/Upstream_endplate.jpg 
  - we can surely hack this up to be better).

	Gerard

Fernando J. Barbosa wrote:
> Hall D Electronics:
> 
> Hi Yves,
> 
>     Did you say the shields are not connected at all? That's it! The
> cause of the oscillations is due to the fact that the cables connecting
> the sense wires to the HV board are just antennas - their shields, to be
> effective, need to be connected to some low impedance reference point.
> Because these cables are RG174, I believe from a previous picture, the
> maximum potential they can tolerate is about 900 V. Therefore, their
> shields cannot be connected to GND.
> 
>     Looking at the layout of the HV board, I notice there is a resistor
> (R50) connecting a trace to the HV and with a decoupling capacitor (at
> the bottom of the board). Was this trace meant to connect the shields of
> the cables to the HV reference? In this case, the potential between the
> inner conductors and the shields is about zero. The shield has then a
> low AC impedance to GND through the capacitor. You have to be careful
> how you route the cables for safety but try connecting the shields to
> this trace. You may want to the decrease R50 to 10 kOhm or so.
> 
>     Alternatives? Use thicker cables that can handle 3 KV (RG58?) - this
> may not be an option; bring the sense wire pins out of the gas volume
> and design a HV board with short traces and good shielding. There may be
> others that we need to look at carefully.
> 
>     Also, I noticed that the decoupling capacitors are 33 pF - a bit on
> the low side. I would change the capacitors to 220 pF, 3KV. This value
> is not critical - just a commonly available capacitor.
> 
> Best regards,
> Fernando
> 
> Yves Van Haarlem wrote:
>> Hi Fernando,
>>
>> I made a report of all the tests I did today, you can find it on the
>> wiki:
>>
>> http://www.jlab.org/Hall-D/software/wiki/index.php/CDC_19_12_2007
>>
>>
>> It should include answers of your questions and measurements you
>> suggested.
>>
>>
>> Cheers,
>>     Yves
>>                           /--------------------------------
>>                          /   Yves Van Haarlem
>>        _--~~--_         /
>>      /~/_|  |_\~\      /      Carnegie Mellon University
>>     |____________|    /        Department of Physics
>>     |[][][][][][]|:= /          Wean Hall room 8404
>>   __| __         |__ \            Pittsburgh, PA 15213
>>  |  ||. |   ==   |  | \           USA
>> (|  ||__|   ==   |  |) \
>>  |  |[] []  ==   |  |   \           Tel.:   +1 412 268-6949
>>  |  |____________|  |    \                  +1 412 641-9252
>>  /__\            /__\     \          Fax.: +1 412 681-0648
>>   ~~              ~~       \-----------------------------------
>>
>> On Wed, 12 Dec 2007, Fernando J. Barbosa wrote:
>>
>>> Hi Yves,
>>>
>>>   I have attached the results I obtained from the tests with the
>>> interposer + Preamp Card + cable + termination. As you can see, the
>>> pulse response is quite nice and the noise is also quite low.
>>>
>>>   As I mentioned before, the oscillations you observe with your set
>>> up are definitely a result of the connections to the preamp card. I
>>> believe your tests over the last few days showed that ground loops
>>> are not the cause.
>>>
>>> Best regards,
>>> Fernando
>>>
>