Re: FDC cableing

["Fernando J. Barbosa" <barbosa@jlab.org>]

Hi Gerard and Joe,

I just have two comments:

1.    I did a few calculations on the relative radiation lengths on the 
cable components and I got the following:

    50 Copper wires - 1.75%
    Copper Braid - 5.3%
    PVC insulation on 50 wires - 0.4%
    PVC on cable wrap - 1.3%.

    I used 1.43 cm and 30 cm radiation lengths for Cu and PVC, 
respectively. These are just approximations and perhaps David and Simon 
can get a better set of numbers. In any case, Joe was correct in 
pointing out the contributions from the shield. But how does this 
compare to the rest of the material inside the solenoid?

2.   The shield should only be connected at one end as Gerard points 
out. However, to maintain the shielding effectiveness over 100 feet, the 
impedance of the shield must be low and this is the function of the 
braid/foil combination. Note that in order to take full advantage of the 
ADC resolution (12-bit), the CMRR must be kept to better than 70 db up 
to 50 MHz (that's about 240 micro volts for a 1V differential dynamic 
range from the preamp). This is a function of the differential receiver 
(front end) and the shielding effectiveness of the cable.

    I like Gerard's suggestion of removing the shield inside the 
solenoid to minimize multiple scattering, but keep the foil. Maybe use a 
heat shrink wrap just to keep the integrity of the cable bundle. A good 
shield should be maintained for the rest of the cable, though. I believe 
the foil/braid combination provide a low impedance and about 95% coverage.

Regards,
Fernando



Gerard Visser wrote:
> Hall D Electronics:
>
> Hi guys,
>     Ok it seems this is generating enough discussions or opinions that 
> we (I, since I have the similar cable sample) should just make some 
> measurements.
>     But in the meanwhile I will toss in more opinions, they are nearly 
> free (unlike measurements :) )
>     The DC resistance of the shield is irrelevant. It has no DC 
> function. In fact, we specifically do not want to carry currents of 
> any kind through the shield, DC or AC. Equipment ground on both ends 
> is connected to the experiment clean ground. The shield is connected 
> on one end only (probably the ADC/TDC board end). The ADC boards (and 
> TDC boards) must have input receivers that are capable of extracting 
> the differential signals despite a reasonable amount (maybe 1 Vpp) of 
> common mode ground difference noise. Any current the shield carries 
> will (obviously) be carried on the preamp-to-detector ground pins and 
> therefore impose a voltage on the preamp inputs, aka an interfering 
> signal which the preamps will respond to, which we do not want.
>     Neither braid nor the foil will have much impact on 60,180, etc. 
> line frequency pickup. Because of the random lay loose pair 
> construction, this pickup (in differential mode) should be small 
> anyway. But any trouble we have, the shield is not going to help on 
> this. Obviously should not run the cable parallel to power trays 
> carrying AC line (unless that's in rigid steel conduit).
>     Very low frequency interference is not a problem, as Joe says it 
> is easily taken out by any pedestal subtraction algorithm. There will 
> be some pedestal subtraction no doubt.
>     We want to cut down on interference pickup in the range 100kHz up 
> to 50 MHz or so. We want to reduce radiated fields from the cable in 
> the same range. We want to have the cable propagation be independent 
> of the location of other nearby cables and metal objects. We want to 
> not have a signal dependent current injected into the experiment 
> ground (as Elton points out). These are the goals of the shield. I'm 
> rather sure the foil shield suffices. I'll try to justify this with 
> some measurements.
>     I certainly agree with Joe, it is a lot of copper in the shield so 
> it is probably good/necessary to eliminate it. At least, it would seem 
> crazy to jump through many other hoops to minimize mass but ignore 
> this item. It is possible however (if safety people allow it) to strip 
> back the entire jacket and shields from the last 1m or so of cable at 
> the detector end. Possibly would have to add a woven plastic braid to 
> keep the pairs bundled nicely. Once it is (nearly) outside the magnet 
> bore the jacket and shield must be present. I think this would 
> probably work ok from an electrical performance point of view. Just 
> another option to consider, perhaps.
>
>     Gerard
>
>
> Elton Smith wrote:
>> Hall D Electronics:
>>
>>
>> Hi Joe,
>>
>> If at all possible we want to avoid capacitively couple all signals to
>> ground. Therefore, I do not think you want to eliminate the copper. 
>> Also,
>> realize that the DC resistance of the thin Al may be quite high.
>>
>> Elton.
>>
>>
>> Elton Smith
>> Jefferson Lab MS 12H5
>> 12000 Jefferson Ave
>> Suite # 16
>> Newport News, VA 23606
>> elton@jlab.org
>> (757) 269-7625
>>
>> On Thu, 7 Jun 2007, Joe Beaufait wrote:
>>
>>> Hall D Electronics:
>>>
>>>  To Elton:  I am more worried about the high frequency than the low. 
>>> Any
>>> low frequencies can be filtered out without affecting data, so removing
>>> the braid still sounds good to me.
>>>
>>>  To Fernando: The large amount of copper in the braid is much denser
>>> than any of the plastic and is a radiator. If we can get half to 3/4 of
>>> it out of the cable, that has to be a huge improvement on the multiple
>>> scattering front. With hundreds of cables there is over a thousand 
>>> ft of
>>> cable in the bore, it would probably remove > 25lbs of copper from the
>>> active area. It will also decrease the diameter over all.
>>> Joe
>>>
>>>
>>>
>>>
>>> Fernando J. Barbosa wrote:
>>>> Hi Joe, Gerard, Dave and Elton,
>>>>
>>>> One more suggestion?
>>>>
>>>> I suggest we get a full picture from the manufacturer regarding their
>>>> design criteria for the electrical and non-electrical specs before we
>>>> start designing a new version of the cable. I am not convinced that
>>>> removing the braid will improve things. After all, we have space and
>>>> the metal in the braid is a small fraction of the total material, 
>>>> right?.
>>>>
>>>> I also concur that whatever we order must be in the final design
>>>> configuration.
>>>>
>>>> Regards,
>>>> Fernando
>>>>
>>>> Joe Beaufait wrote:
>>>>> I am glad we have decided to go ahead and purchase a sample of the 25
>>>>> pr. cable. This cable is going to be the standard for much of the
>>>>> data acquisition system and it would be very nice to see how hard it
>>>>> is to work with.
>>>>>
>>>>> On that note, there were some comments made at the Wednesday FDC
>>>>> meeting  concerning the possible redesign of the cable. Daniel
>>>>> mentioned that we might ask to make the outer jacket thiner and I
>>>>> promised to look into it but, while looking at the cable diagram I
>>>>> also noticed that this cable is double shielded. It has an aluminized
>>>>> mylar layer with drain wire, and a copper braid. I believe we could
>>>>> do away with the copper braid and reduce the copper in the cable by
>>>>> at least half.
>>>>>
>>>>> I would like comments from the group on whether or not you would like
>>>>> me  to pursue this.
>>>>>
>>>>> Thanks
>>>>> Joe
>>>>>
>>>
>>>
>>>
begin:vcard
fn:Fernando J. Barbosa
n:Barbosa;Fernando J.
org:Jefferson Lab
adr;dom:;;12000 Jefferson Ave.;Newport News;VA;23606
tel;work:757-269-7433
version:2.1
end:vcard