Re: Preamp discussion

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

Hi Mitch,

Thanks for your reply.

1. Package - It's a good idea to use another packager for the 
flexibility in choosing the package and also if it is cheaper. Let's 
plan on using a QFN64 package as it makes thermal management on the 
detector cards easier and cheaper also.

2. Costs - You mentioned in a previous email an NRE of $1K plus $0.75/ 
package for packaging the die. Therefore the total cost would be:
        $81,250 for 3K parts - $27.08/part - $3.39/channel
        $84,000 for 4K parts - $21.00/part - $2.63/channel.
Is this correct?

3. Testing & Yields - Do the die costs include tests? Maybe just wafer 
tests? What's your plan for testing the devices? It would be easier 
perhaps to test the packaged devices either through a third party or we 
would need to plan for a "testing station" that some group in the 
collaboration could take up as a project. Yields are usually high, but I 
suggest we should plan for 4K parts, or the largest number we can get, 
as the cost differential is not that significant between the 3K and the 
4K runs. What are your thoughts?

4. Part Number - I will create a tentative schematic symbol and PCB part 
soon so we can arrive at a device pinout. Do you have a part number for 
this device?

Thanks and best regards,

Fernando

mitch@hep.upenn.edu wrote:
> Budgetary Quote from MOSIS for TSMC 0.25um process with 3X3.2mm die.
>        3K die $26/die
>        4K die $20/die
> No charge for MIM caps and thick metal that we might want to use.
>                                        Mitch
>
>
>
> Quoting "Fernando J. Barbosa" <barbosa@jlab.org>:
>
>   
>> Hi Mitch and Gerard,
>>
>> I have a few items that I would like to bring to your attention and
>> which are based on previous discussions:
>>
>> 1. Preamp Costs - For the prototype run of 40 Die, the Mosis cost is
>> $21.5K for the assembled parts. What would the cost be for 2500-3000 Die
>> and assembled parts?
>>
>> 2. Package - I agree with using a 64 pin package as it provides for
>> additional power, ground and additional pins not yet defined (if
>> needed). However, I would like to suggest the use of the QFN64 (64 pin,
>> ground-paddle, 0.5 mm pitch). This package provides improved thermal
>> characteristics compared to the LQFP64. The preamp cards on the FDC (and
>> most likely the CDC) will be installed in regions with very limited
>> space and most of the heat will be removed by conduction to some sort of
>> detector cooling assembly (not yet defined). I2A carries this package as
>> standard.
>>
>> 3. Double termination - Given the thermal limitations, I agree that
>> there should be no termination resistor on the package. I also see no
>> need for double termination on the card as we already have differential
>> lines and the SNR may actually be adversely effected with the double
>> termination given the long cable runs.
>>
>> 4. I should have a preliminary diagram for the preamp card soon. I
>> should also review an old FDC electronics diagram that I distributed a
>> long time ago.
>>
>> 5. I would also like to request a number of the prototype packaged
>> preamps for assembly on this new preamp card as soon as they are
>> available. We can quickly test them on the prototype FDC on site. Maybe
>> 10 would be more than enough.
>>
>> 6. A 24-channel board is fine per Gerard's calculations. I still need to
>> check the spacing required on the FDC (connectors and such) due to the
>> plane-to-plane pitch.
>>
>> 7. A detector test pulse system is a must and we need to discuss this
>> soon. There are pros and cons to having the pulser on-board the preamp
>> card or external: power, compactness, flexibility (amplitude
>> adjustment?), cabling, etc. Capacitive coupling to the preamp inputs via
>> orthogonal traces is a simple and effective way of injecting some fixed
>> charge.
>>
>> Best regards,
>>
>> Fernando
>>
>>
>>     
>
>
>   
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