Continue to Site

Welcome to

Welcome to our site! is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

SATA interconnect total insertion loss

Not open for further replies.


Advanced Member level 3
Oct 24, 2005
Reaction score
Trophy points
Florida, USA
Activity points
sata 50 insertions


what is the limit of the total insertion loss in a sata interconnect chip-to-chip?
All what i found is about cable testing/measurement, and cable loss can be max 6dB.

What i want to know, is what is the max allowable loss in a chip to chip interconnect? i am not a consumer, but a designer. I have to design 2 boards, connected by cables, and have a control over the total loss in the system. So what is allowable on a transmitter board, on a receiver board (the max cable loss is known 6dB)?
The SATA specs doesnt tell me too much, and very confusing about this. For example they are stating that V_in_min=325mV and V_out_min=400mV, which produces an impossible 1.8dB loss limit.

sata insertion loss caused by

Generally, the SATA spec (and also respective chip specifications) contain signal level margins as quoted. However, I didn't find any cable transmission loss specification (neither 6 dB nor below), I only see return loss specifications. Thus I think, a 6 dB attenuation (for the signal fundamental) is most likely beyond the specification. Just my impression, I'm not an expert in this field, I did some SATA hardware wiring within a system design, involving PCB differential microstrip and several connectors, operational so far, that's all.

The other point is, that specifications and standards aren't necessarily consistent in all details.

sata insertion limit

ok. i just like to know what i am doing, and what is happening on my boards.

**broken link removed**
it says cable insertion loss.
this too: (on page 42)
**broken link removed**

sata return loss

O.K., I see. The said insertion loss specification (-6 db, 10 - 4500 MHz) can't be related to the basic signal level specification, cause the 1.5GBps fundamental frequency is 750 MHz only (the domain of your said impossible 1.8dB spec), both specifications exist independantly (and are almost consistent, I think).

Not open for further replies.

Part and Inventory Search

Welcome to