[SI-LIST] Re: proper Rref for return loss sim
- From: agathon <hreidmarkailen@xxxxxxxxx>
- To: si-list <si-list@xxxxxxxxxxxxx>
- Date: Thu, 24 Feb 2011 19:52:42 -0500
Steve! Thanks for the input.
Yes, I expect the optimal mstrip Z0 to end up being < = pkgZ0 but not
necessarily since the RL spec profile overrules all.
The question I had, though, is different -- because my RL results do a
180deg change DEPENDING ON Rref of the measurement (in simulation).
The goal is to optimize RL, mainly doing a mstrip Z0 sweep.
--->> But I need the "proper" Rref first because the results depend on
it.
The sweep results completely REVERSE their order of worst to best at the RL
peaks @ 11 and 18GHz. The data rate is 17Gbps. The mstrip is 0.35 to 0.55
in. The pkg diffpair is ~18mm.
Some of the dependence on Rref arises from reflections and SWR at the
measurement node, no doubt, so I think it's best to use Rref=mstrip_Z0.
Right???
So, I need to determine the proper measurement/sim Rref before I sweep
mstrip Z0 and make a judgement on the RL results!!
Again, the results reverse their worst-to-best order. The cases of
measurement Rref:
Rref=50 then I have mstrip Z0=50 is optimal and Rref < 50 is worse.
Rref=Z0 of mstrip, I get the OPPOSITE. Then, Z0 is optimal, even when it's
40 ohms. THAT'S 10 OHMS DIFFERENCE. I'm sure I could continue the sweeps
and get even worse contradiction.
These are given / cannot change:
Die & pkg sparam models use Rref=50.
Pkg Z0 = 42.5 nom. and pkg Zeven ~ Zodd.
Thanks for revisiting this.
On Thu, Feb 24, 2011 at 5:08 PM, steve weir <weirsi@xxxxxxxxxx> wrote:
> If by chance the die is using 50 Ohms while the package is 42.5, then
> unless the package Tx line is very short, there is little you can do about
> that discontinuity. Your best bet is to keep the remainder of the channel
> at a constant impedance matched to that of the package. The ustrip and
> termination should both match the package at 42.5 Ohms.
>
> Steve.
>
> agathon wrote:
>
>> Hello All,
>> It's real simple.
>>
>> ................ BACKGROUND ....................
>>
>> I'm asked to optimize the RL to be measured on a test board (tb) that will
>> evaluate an asic serdes io cell.
>> In general, RL depends on the measurement Rref, since some signal at
>> measurement node will include that node's own reflections.
>>
>> asic -- pkg --|--- microstrip -- cap --- microstrip (taper if
>> necessary) -- SMP
>>
>> The asic diffpair cell model is a 2-port (p/n) with Rref=50. The pkg is
>> s-param with Rref=50 and the diffpair in the pkg has Zdiff=85 loosely
>> coupled (Z12 ~ 1ohm). First, I'm just looking at the die/pkg/mstrip
>> without
>> any cap or rightside mstrip or SMP. I really things will change w/ those
>> in
>> the sim.
>>
>>
>> ............... THE QUESTION .....................
>>
>> In simulation, just with mstrip to left of cap + pkg + asic, the RL is
>> found
>> with Rref termination at mstrip end = 50 and various mstrip Z0
>> ([Zodd+Zeven]/2).
>> The best to worst, against a template, is for mstrip Z0= 50, 45, 42.5, 40.
>> Ie: max RL.
>>
>> asic -- pkg --- mstrip <--- Rref---gnd
>>
>> If I change the mstrip Rref to = mstrip Z0 then the best to worst reverses
>> order. So, there's an issue in my mind... WTF is the "proper" Rref for
>> the
>> sim? The term "50 ohm system" doesn't really apply, it seems.
>>
>> The die Zdd11 smoothly goes: 75.6.. ~54.7 ..~94.6 ohms
>> (@200Mhz..10GHz..18GHz) and Zcc11 (18.9.. 15.. 23.8ohms). So, for the die
>> the coupling is loose and the "Z0 equivalent" ranges approx. 38 .. 29 ..
>> 47
>> ohms.
>>
>> Since I'm optimizing mstrip for asic+pkg, shouldn't I be using Rref =
>> Z0(mstrip), instead of Rref=50? Obviously, we're lowering mstrip Z0 below
>> 50 to better match the die & pkg and will recommend that to customers.
>> This
>> is for eval and for customer demo. We're not trying to make RL
>> unrealistically good, just trying to support accurate measurement.
>>
>> Thanks for the input.
>>
>> ------------
>>
>>
>> ------------------------------------------------------------------
>> To unsubscribe from si-list:
>> si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
>>
>> or to administer your membership from a web page, go to:
>> //www.freelists.org/webpage/si-list
>>
>> For help:
>> si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
>>
>>
>> List technical documents are available at:
>> http://www.si-list.net
>>
>> List archives are viewable at:
>> //www.freelists.org/archives/si-list
>> Old (prior to June 6, 2001) list archives are viewable at:
>> http://www.qsl.net/wb6tpu
>>
>>
>>
>>
>
>
> --
> Steve Weir
> IPBLOX, LLC 150 N. Center St. #211
> Reno, NV 89501 www.ipblox.com
>
> (775) 299-4236 Business
> (866) 675-4630 Toll-free
> (707) 780-1951 Fax
>
>
>
------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
or to administer your membership from a web page, go to:
//www.freelists.org/webpage/si-list
For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
List technical documents are available at:
http://www.si-list.net
List archives are viewable at:
//www.freelists.org/archives/si-list
Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu
Other related posts:
