• TI Thinks Resolved

Want to know about transformer etc

Guru 14355 points

Replies: 3

Views: 80

Part Number: DP83822I


My cusotmer are planning to use DP83822I and
they have a few questions related to this PHY IC .

They are thinking to use YDK RJ-45 connector(Part# 46F-1305GYDNWZ2NL).
It has internal magnetics and it includes capacitors for each center tap(CT) of both PHY side and connector side.
In DP83822I datasheet page.101, recommended transformers are listed in Table 138.
These transformers don't have internal capacitors.
Is transformers with capacitor not recommended?
Should they change to other RJ-45 connector which don't have internal capacitor?

This question is related to above Q1.
If they need to change the above RJ-45 connector,
they are thinking to use J0026D01BNL(Manufacture:Pulse Electronics) and they want to know
if this transformer could be use with DP83822.
From J0026D01BNL datasheet, transformer electrical specification are close to
the Table 139 of DP83822I datasheet page.101. So, I guess this transformer could be used with DP83822I.
Do you think this transformer could be used with DP83822I?
I understand that this transformer haven't tested at TI so that the customer need to
test themselves and decide if it could use with DP83822I.
But if possible could you please give us an advise?

From DP83822I datasheet page.74 Table 69. Line Driver Control Register,
it seems that we can select the swing of the driver.
What are the cases of setting values other than default values?
Is it a case such as long pattern wiring, noise concerns, suppressing heat generation etc?

Related to Q3.
When setting the swing to 2.0V at 100Base-TX, does it mean that the differential peaks of
TXD_P and TXD_N or RXD_P and RXD_N become 2.0V?

Related to Q4.
Why there is no mention of "peak to peak" in the description of 10Base in Table 69?

best regards,

  • Hi g.f.,

    A1) Transformers with internal caps are fine, but it depends on how they are connected. The transformer you sent the link for has a series cap. This will block the DC. The DP83822I is a current mode driver and requires the center tap line to be tied to VDD. That is not possible with the transformer you sent.

    A2) The pulse transformer they have will work with the DP83822I.

    A3) The line driver control gives you options to change the swing depending on the insertion loss of your channel (PCB, external components, etc).

    A4) When setting swing to 2V, it means the pk-pk swing (V+ to V-).

    A5) 10BASE-T is manchester (i.e. 1010) while 100BASE-TX is MLT3 (+1, 0, -1). We could put pk-pk for both, but it was just chosen to have it how it currently is.

    Best regards,

    Ross Pimentel

    Ethernet - Applications Engineer

  • Guru 14355 points

    In reply to Ross Pimentel:

    Hi Ross,

    Thank you for the reply and we understood.
    But my customer have additional questions.

    It's about Table 69 of DP83822I datasheet.
    The default value of 10Base-Te is 3.7V . Is this 3.7V the peak voltage?

    We want to know what type of termination resistor(1/2W?1/4W?1/16W? etc..) should be used.
    If 3.3V are supplied to AVD, VDDIO and Center Tap of Transform are connected to AVD,
    100Base-TX voltage swing is 2.0Vpp(default) so that required power consumption of resistor would be 1[V]^2 / 49.9[ohm] = 0.02[Watt],
    10Base-Te voltage swing is 3.7V(default) so that it would be 1.85[V]^2 / 49.9[ohm] = 0.068[Watt].
    So, more than 0.02[watt] resistor(49.9ohm) for 100Base-TX,
    and more than 0.068[watt] resistor(49.9ohm) for 10Base-Te are needed, am I correct?

    Does TI have tested DP83822 with RJ-45 which have internal transformer?

    best regards,
  • In reply to g.f.:

    Hi g.f.,

    A6) Yes, 10BASE-Te has a requirement of peak differential output between 1.54V - 1.96V.
    This a pkpk swing of 3.08V - 3.92V. The DP83822 default setting is 3.7V pkpk.

    A7) Please see this post: e2e.ti.com/.../302618

    A8) Yes, please see the DP83TC811EVM.

    Best regards,

    Ross Pimentel

    Ethernet - Applications Engineer