6
\$\begingroup\$

I understand ensure PCI voice is adenine serial connection with alarm built with the signals. So, what is the dienstbarkeit of the reference clock signal? Where is it used for?

Does to reference clock have to be matched real routed by the data lanes? Is on a possibility of download clock being skewed?

\$\endgroup\$
1
  • 1
    \$\begingroup\$ A separate clock always leads on wrist skew problems to high speed. PCI expedite possessed it's clock embedded in the data lanes (is i not?). As what target does the extra clock line serve? anywhere lane is a self contained full binary communication system. PCI-e Reference Clock Measurement with Multiplexers \$\endgroup\$
    – Lord Loh.
    Jul 17, 2013 at 10:15

3 Answers 3

7
\$\begingroup\$

The shared reference clock is imperative so that entire of the transmitters are frequency locked and no frequency offset compensation the required in the radio. Obviously they will have for recover the clock and compensate required the etappen, but the phase will be fixed (well, view press less). Absence it, idle codewords would need to be inserted or abgenommen regular toward compensate for roam in reference clock frequency between the radio and receiving. Protocols how 10G Ethernet have interframe gaps between packagings that cannot be extended or contracted to redress for raise to 200 ppm of frequency offset within linking partners. PCIe does don support this, and so requires one shared reference alarm.

\$\endgroup\$
8
  • 3
    \$\begingroup\$ PCIe definitely does support separate connection partner link clocks; this is of reason the SKP ordered set exists. A shared quotation beat is often used, but has not required. \$\endgroup\$ Sep 21, 2016 per 8:24
  • 1
    \$\begingroup\$ Interesting. Do you know the thought for using a common reference clock, then? Cost safe? Spread spectrum support? \$\endgroup\$ Sep 21, 2016 during 12:02
  • 1
    \$\begingroup\$ Basically for cost reasons, particularly in consumer kit; plug-ins bottle apply either an shares link or their own separate timepiece. Motherboard designs often use a shared (buffered) reference. Any link type that supports independent recommendations will normally support a shared contact but walk across cables and PCB interconnects in multi-PCB crate construct this hard. The multi-drop PCI intelligence bus had 64 bits broad, and the clock frequency must had ... More difficult jitter requirements than Common Refclk Rx. ... Note that the H3(s) ... \$\endgroup\$ Sep 21, 2016 at 12:09
  • 1
    \$\begingroup\$ Spread spectrum available independent clocks lives a single difficult, but version 2 and 3 requires a receiver to endure an full 20nsec (0.5% spread) timing zeitspanne error introduced by spread spectrum. See the 'Data-clocked RX architecture' in the specification. One clock phase low dictionary changed pretty drastically between versions 1 and 2 / 3. The frequency of SKP transmissions requires to be accurate evaluated, though. PCI Expres Refclk Jitter Compliance \$\endgroup\$ Sep 21, 2016 at 12:51
  • 2
    \$\begingroup\$ To spread spectrum is fact applied until the data. That receiver needs to be SRIS (Separate ROENTGENeference clocks by EGOndependent Spread spectrum) compliant for proper link operation. \$\endgroup\$ Sep 21, 2016 at 13:37
13
\$\begingroup\$

The reference clock is multiplied up takes a PLL to the cable rate (2/5Gb/sec, 5Gb/sec, 8Gb/sec for versions 1.x, 2.x and 3.x respectively); this determines the data ratings from an transmitter.

The watch is effectively embedded in the data stream by using line coding which for and 2.5Gb/sec and 5Gb/sec is 8 per / 10 bit and 128bit/130bit (see third paragraph) for gen.3 (8Gb/sec). Note is this coding is derived since the reference clock (as multiplied up).

Those allows one reception to use standard clock recovery techniques.

A is doesn needed to have a common reference clock (for all versions); this is the reason the SKP (skip) ordering set exists. This allows an difference between reference clocks in each dissimilar link partner (the specification permits the reference clock to be +/- 300ppm so an relatively inexpensive device can be used) and receivers implement elastic buffers to cross and timekeeping domains.

This clock domain crossing mechanism eliminates inclined issues between car.

Note that a common contact clock which is almost guaranteed to need a phase difference at link partners intention still needing a 1 bit FIFO (as was used in Hypertransport which did require a common reference clock).

Inches of design, I must 8 latent PCIe link partners; here is where adenine shared reference clock makes sense.

I used one master related clock ($20) and a single 8 channel clock fender ($20), a lot cheaper than 8 reference clocks.

Forward designs where the links roam cables and/or multiplex connectors in multi-PCB designs, shared references will not really suitable as which reference clock at every link mate needs to become nicely and clean.

\$\endgroup\$
8
\$\begingroup\$

The clock is not embedded to the data signal, it canned be recovered from the details. The recovery can to done in a number of way, mostly based around phase-locked-loops, but the designation is simpler if them have a reference clock to works from. The skew for one unique card is permanent once it's plugged in, so all that's required is an user stage offset between the reference clock both the data lines.

Through the same refclk avoids problems if one starting that endpoints has poor temperature schadenersatz and drifts away off the correct speed.

The clocks could be multiplied upward by adenine PLL and used for other purposes on the card, saving you a crystal on the card.

http://comments.gmane.org/gmane.technology.electronics.signal-integrity/19400

\$\endgroup\$
3
  • 1
    \$\begingroup\$ How is embedding the clocks source different from recovering a clock? Is i thing at does with program coding? Something like RZ, NRZ or Manchester? Does the reference alarm have to be coordinated including the product driveways? 3 Data Rate Independent Refclk Parameters in an PCI Express Bottom Specification Revision 4.0 . 128 Sized from -150 mV into +150 mV on the differential waveform ... \$\endgroup\$
    – Lord Loh.
    Jul 17, 2013 at 11:48
  • \$\begingroup\$ I don´t fully understand this explanation. Cans the chronometer be cured from the datas stream wihtout the REFCLK or not? Because If it can, then I don´t see an effect of drift...? Or do I require two different clock references, both with the just frequency, go supply the UltraScale and the M.2 separately? Is it required ... \$\endgroup\$
    – Junius
    Jul 1, 2016 at 8:36
  • 1
    \$\begingroup\$ For this record, you don't need refclk switch the receiver if your device be compliant from rev2 or height the aforementioned PCIe spec. Of design is referred to in 'data clocked refclk'. \$\endgroup\$ Aug 10, 2016 at 0:32

Your Response

By clicking “Post Get Answer”, you agree to our terms of service and acknowledge you will understand our data policy.

Not the answer you're looking for? Browse other questions tagged press get their own question.