This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

DM6467T EVM printed circuit board impedance specifications (aka Spectrum Digital HD1080P)

Why does page 20 of the Gerber pdf for the DM6467T EVM (aka Spectrum Digital HD1080P) specify the following controlled impedances?

Note 16: "Controlled Impedance of 50 ohms to be matched on all signal layers, on 5 mil lines."  This does make some sense to me, but I've never had to specified controlled impedance before.  Any insight?

Note 17: "5.1 mil traces on layer 12 to be 45 ohm, single ended and 90ohm differential." What are these signals and why are they special? 

Note 18: "5.1 mil traces on layers 1 and 3 to be 50 ohm, single ended and 100 ohm differential".  What are these signals and why are they special?  

Most especially, why the specified difference here in Note 18 versus Note 17? I find DDR routing on all three of these layers but also on others.  So Notes 17 and 18 must refer to either special DDR lines or something going elsewhere, like ATA (hard disk), PCI, or VLYNQ perhaps.

ref: http://support.spectrumdigital.com/boards/evmdm6467t/revc/files/Davincihd1080p_EVM_Gerber_revc.pdf

Thanks,

Helmut

  • Helmut,

    Following is the response from Spectrum Digital Inc.

    Instead of giving the exact line width and distance of separation we tend to specify a specific width( for each impedance rule )  so the board fabrication shop can tweak the separation and distance to their process for the impedance rules specified

     

     

    Note 17 is USB lines 90 ohm

     

    Note 18 is ether-net or similar 100 ohm

    Regards, Srirami.

  • EXCELLENT.  Thank you VERY much.  My board was already part way through fab before I had the mental bandwidth to pay attention to this item.  At least these specs were not for the DDR.  From my USB experience, I should be able to operate the USB without careful attention to note 17.  (At least at 1.1 speeds if not 2.0 speeds.) And I guess I'll just have to cross my fingers about the ethernet connection.

    Regardless, this is very GOOD TO KNOW.  It will inform me while diagnosing my new board, as well as on the next board rev, as well as inform other readers before they get to this point.

    If you have time for a purely academic discussion...  I can see why you want the impedance of those lines controlled.  But, frankly, I don't see why there's any difference between the two.  That is, make them all 100 ohm, including the USB.  Or maybe 95.  I don't *believe* the controls are that accurate anyway.  And I can't imagine the circuitry is that sensitive, either.  I've had decades of mixed signal experience, although never high speed.  Still, my gut tells me not to worry too much about these, and so my eyes and my gut are at odds with each other when I read such manufacturing notes!  Of course, such gut/eye conflict is how one learns.  (Another tortured metaphor.)

    -Helmut