Path: EDN Asia >> Design Centre >> IC/Board/Systems Design >> The puzzling disappearance of duo-binary signalling
IC/Board/Systems Design Share print

The puzzling disappearance of duo-binary signalling

02 Apr 2015  | David Banas

Share this page with your friends

Instead, I'll explore two other cases: mandating equal peak-to-peak signal amplitude at the Rx input, and mandating equal eye heights. In order to accomplish this, I rerun the duo-binary simulation, changing nothing but the master Tx amplitude control, in order to produce equal peak-to-peak amplitudes at the Rx input for both cases. The new duo-binary result is shown in figure 6 and the results are re-tabulated in table 2.

Table 1: Eye diagram for duo-binary, after enforcing equal Rx input amplitude.

Table 2: Eye heights and widths, after enforcing equal Rx input amplitude.

In this case, duo-binary modulation offers better than a 2x improvement in eye height, with widths, again, essentially equal. I had to increase the Tx master output voltage to 4 V, to achieve equal Rx input amplitudes, and this represents a relative increase in Tx power dissipation of 16x. These results suggest another comparison constraint: equal eye heights. In this case, the Tx master amplitude is adjusted to produce an eye height, in the duo-binary case, which is equal to that of the PAM-4 case. Then, the Tx power dissipations are compared and this comparison is weighed against the increased Tx and Rx circuit complexity, in the PAM-4 case, before deciding which modulation scheme makes sense for a particular design. To produce an eye height of 43 mV, the master Tx amplitude must be set to 2 V for the duo-binary case. This represents an increase in Tx power dissipation of 4x, relative to the PAM-4 case.
Possible explanations
Given that duo-binary modulation seems to be able to produce eye openings on par with PAM-4, albeit with increased Tx power dissipation, why does it seem to have been dropped by the experts as a potential solution to 56Gbit/s design challenges? I can think of four potential reasons:

 • Momentum of an arbitrary decision. It could be that one of the major telecommunications equipment manufacturers simply decided, for better or worse, that PAM-4 was the way to go, and that they had enough clout to sway the industry.
 • Dependency of the Tx tap weights on channel. In the duo-binary case, the Tx tap weights need to be somewhat matched to the channel, since the two work together, in order to produce the proper system impulse response. Perhaps, designers don't like the idea of having to accommodate this dependency.
 • Power is more important than circuit complexity. More power is required, in the duo-binary case, to produce an eye opening equivalent to the PAM-4 case. And, perhaps, designers have decided that, in the "green" age, power is more important than circuit complexity.

Actual differences in bandwidth required. Despite the reporting of equal bandwidth requirements between duo-binary and PAM-4, consideration of the ideal system impulse responses for the two suggests that PAM-4 requires only 3/4 the bandwidth of duo-binary, in order to transmit optimally. This may have been considered by forward looking engineers concerned about the scalability of their 56Gbit/s design choices to higher data rates.

Perhaps duo-binary just missed its opportunity to win favour with serial communication link design experts before other alternatives captured a critical mass of their attention. (As engineers practicing our trade in the post-VHS/Beta era, we've had to learn to live with these market-driven realities.) Or, it could be that PAM-4 offers a very real technical advantage over duo-binary, which the experts know about, but aren't speaking about publicly. In either case, don't be surprised if some enterprising upstarts do some deliciously disruptive things with duo-binary along the way. After all, it's very difficult to objectively argue against its three main advantages:

 • Lack of need for any Rx equalisation.
 • Utilisation of existing Rx architectures for final data detection.
 • Simpler circuit design for both Tx and Rx.
Duo-binary signalling appears to be an extremely attractive solution to challenges faced by designers of 28 and 56Gbit/s serial communication links today. Given this, the apparent abandonment of duo-binary modulation by the various industry working groups and task forces charged with next generation modulation scheme selection is puzzling. In particular, the recent achievements of the consortium formed by FCI, Ghent University, and Alcatel-Lucent would appear to demand a reconsideration of the duo-binary approach as a serious contender for 28 and 56Gbit/s transmission over copper media.

About the author
David Banas is a principal SERDES applications engineer. He is a 23-year veteran of the semiconductor industry, currently, providing design, modelling, and characterisation support for high speed serial transceivers in the programmable logic space.

 First Page Previous Page 1 • 2 • 3

Want to more of this to be delivered to you for FREE?

Subscribe to EDN Asia alerts and receive the latest design ideas and product news in your inbox.

Got to make sure you're not a robot. Please enter the code displayed on the right.

Time to activate your subscription - it's easy!

We have sent an activate request to your registerd e-email. Simply click on the link to activate your subscription.

We're doing this to protect your privacy and ensure you successfully receive your e-mail alerts.

Add New Comment
Visitor (To avoid code verification, simply login or register with us. It is fast and free!)
*Verify code:
Tech Impact

Regional Roundup
Control this smart glass with the blink of an eye
K-Glass 2 detects users' eye movements to point the cursor to recognise computer icons or objects in the Internet, and uses winks for commands. The researchers call this interface the "i-Mouse."

GlobalFoundries extends grants to Singapore students
ARM, Tencent Games team up to improve mobile gaming

News | Products | Design Features | Regional Roundup | Tech Impact