Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

Re: [8023-10GEPON] PIN-PD D/S power budget proposal



Dear Hamano-san, 
I am not wasting my time by no means. It is an interesting discussion for me as well  even though I do not have first hand data to work on. 
My comment about the DS and US links being separate in the ONU makes sense from one point: the selection of the ONU Tx is not related with the selection of the ONU Rx. Thus it should not go into cost comparison between an APD and PIN based ONU. Otherwise we are not going to compare apples with apples. Let's assume that we have an ONU with a Tx that meets the given power budget, say PR20. The question then is what is the relative cost benefit per ONU in a system which uses PIN at ONU and higher power Tx at the OLT (whatever the structure is i.e. EML/DFB/DML + SOA/EDFA, just decide how You want to reach the launch power) versus APD at the ONU and lower power Tx at the OLT (again, decide what the launch power is etc.). When comparing the scenarios, disregard uplink for the time being, account for all the required protection mechanisms for the OLT (if higher power is used) and ONU (if APD needs protection against overload), sum them up and divide by the number of the ONUs in the!
  system. Only then we can do the realistic comparison. 
If we take in additionally the cost of the uplink, that just should give us a constant factor in the calculations which does not impact the final outcome. The maths of this is relatively simple ... 
If what You're saying is true i.e. "ROSA has relatively a small impact to the total optics cost." then why are the ONU cost differences so high ? Because of the complete optical subassembly ? 
Thank You for Your time 

Marek Hajduczenia (141238)
NOKIA SIEMENS Networks S.A., Portugal - R
Rua Irmãos Siemens, 1
Ed. 1, Piso 1
Alfragide
2720-093 Amadora
Portugal
* Marek.Hajduczenia@siemens.com
http://www.marekhajduczenia.info/index.php
(+351.21.416.7472  4+351.21.424.2082
"C makes it easy to shoot yourself in the foot; C++ makes it harder, but when you do, it blows away your whole leg." - Bjarne Stroustrup 

-----Original Message-----
From: Hiroshi HAMANO [mailto:hamano.hiroshi@JP.FUJITSU.COM] 
Sent: quinta-feira, 6 de Setembro de 2007 4:31
To: STDS-802-3-10GEPON@LISTSERV.IEEE.ORG
Subject: Re: [8023-10GEPON] PIN-PD D/S power budget proposal

Dear Dr. Hajduczenia,

Thank you for your comments.
What I only want to say is if you focus on D/S, cost definitely 
ends up to ROSA cost only at high splits.  I admit again that 
current APD-ROSA costs 2-times, and I cannot change that reality.
But the system does not work without U/S, and ROSA has relatively 
a small impact to the total optics cost.  I only want to confirm 
that APD-ROSA costs twice, but ONU or system cost does not.
I apologize for making you spend your time.

Best regards,
Hiroshi Hamano

%% "Hajduczenia, Marek" <marek.hajduczenia@SIEMENS.COM>
%% Re: [8023-10GEPON] PIN-PD D/S power budget proposal
%% Wed, 5 Sep 2007 13:51:35 +0100

> Dear Hamano-san, 
> 
> Please forgive my ignorance but what has ONU Tx have to do with the ONU Rx ? Apart from the fact that they are located in the same unit, I believe there is little relation between both. I agree that both Rx and Tx impact the resulting ONU cost and we can safely assume (I think) that we can use the same laser at the ONU regardless of whether we use a PIN or an APD at the ONU. 
> I am saying that since the Tx at the ONU depends only on the OLT Rx and I suspect that nobody is assuming we can do the PR20 with a PIN based OLT. We will need every fraction of the dB we can squeeze out of the OLT Rx for PR20 and especially PR30 systems. 
> 
> Form where I stand, the PIN/APD discussion can be viewed as follows:
> 
> APD: higher ONU cost (at least for now and probably in the future, though the APD/PIN cost ratio will tend to 1 with the ramp up in the production volume as everyone claims), slightly more complicated Rx electronics, much easier power budgets for PR10 and PR20 systems (perhaps we could go even without a FEC as some people suggested), PR30 with no problems, concerns about APD overload values (may need additional attenuators to be plcaed in the signal path)
> 
> PIN: lower ONU cost, lower sensitivity than APD but likely to improve 1-2 dB in the next few years, simpler Rx electronics, PR20 budget may require SOA/EDFA, PR30 will require SOA/EDFA amplifier, concerns about eye safety, additional safety mechanisms required for the OLT box. FEC utilization likely to boost ONU Rx sensitivity to cope with PR20 and PR30 budgets. 
> 
> I do not quite understand Your comment on the "misleadning discussion of the DS channel". The ONU Rx operates in the downstream. Unless the ONU Tx depends directly on the type of the Rx we use there, I do not see any point in including the upstream channel in this discussion. What purpose would that serve? 
> 
> Thank You for Your time 
> 
> Marek Hajduczenia (141238)
> NOKIA SIEMENS Networks S.A., Portugal - R
> Rua Irm達os Siemens, 1
> Ed. 1, Piso 1
> Alfragide
> 2720-093 Amadora
> Portugal
> * Marek.Hajduczenia@siemens.com
> http://www.marekhajduczenia.info/index.php
> (+351.21.416.7472  4+351.21.424.2082
> "C makes it easy to shoot yourself in the foot; C++ makes it harder, but when you do, it blows away your whole leg." - Bjarne Stroustrup 
> 
> -----Original Message-----
> From: Hiroshi HAMANO [mailto:hamano.hiroshi@JP.FUJITSU.COM] 
> Sent: quarta-feira, 5 de Setembro de 2007 12:01
> To: STDS-802-3-10GEPON@LISTSERV.IEEE.ORG
> Subject: Re: [8023-10GEPON] PIN-PD D/S power budget proposal
> 
> Dear Dr.Li and Dr.Hajduczenia,
> 
> Thank you for discussing APD-ROSA cost assumptions.  I admit that 
> current APD-ROSA costs 2 times of PIN-ROSA, and it is important 
> to see how it will be reduced in the future volume production.
> But I think people are too much micro-focusing on ROSA cost.  
> ROSA is only one of the component in the ONU, and it does not 
> indicate the ONU cost or total optics cost.
> U/S TX is the major cost contributor for ONU and also the total optics.  
> And considering the total cost, APD@ONU is always cost effective 
> up to 16 splits, as I have shown in my presentation 3av_0707_hamano_2.pdf..  
> Even at 32 splits the cost difference is only 10-15%, and if you assume 
> the overall system, it will be much smaller. 
> Some PIN group people always stress D/S cost only to give the impression 
> that APD@ONU is 2-times expensive, but it is not true.  
> APD-ROSA costs 2-times, but APD@ONU does not.
> D/S cost discussion only is not fair and misleading.
> 
> Best regards,
> Hiroshi Hamano
> 
> %% "Hajduczenia, Marek" <marek.hajduczenia@SIEMENS.COM>
> %% Re: [8023-10GEPON] PIN-PD D/S power budget proposal
> %% Wed, 5 Sep 2007 07:12:53 +0100
> 
> > Dear David, 
> > Thank Yout for the answer. 
> > We have one vote in favour of the APD price reduction after the beginning of the 10GEPON deployment. In Your opinion, the price difference can be as low/high as 210 - 220% (APD vs. PIN). Are there any other cost components which need to ba taken into account when considering APD based ONU ? 
> > Regards
> > 
> > 
> > Marek Hajduczenia (141238)
> > NOKIA SIEMENS Networks S.A., Portugal - R
> > Rua Irm腴‰ Siemens, 1
> > Ed. 1, Piso 1
> > Alfragide
> > 2720-093 Amadora
> > Portugal
> > * Marek.Hajduczenia@siemens.com
> > http://www.marekhajduczenia.info/index.php
> > (+351.21.416.7472  4+351.21.424.2082
> > "C makes it easy to shoot yourself in the foot; C++ makes it harder, but when you do, it blows away your whole leg." - Bjarne Stroustrup 
> > 
> > -----Original Message-----
> > From: David Li [mailto:dli@LIGENTPHOTONICS.COM] 
> > Sent: ter艸┃feira, 4 de Setembro de 2007 17:14
> > To: STDS-802-3-10GEPON@LISTSERV.IEEE.ORG
> > Subject: Re: [8023-10GEPON] PIN-PD D/S power budget proposal
> > 
> > Marek,
> > 
> > In the past, the APD/TIA price is about 5X to 10X of the PIN/TIA price for
> > the 1.25G and 2.5G  applications.
> > 
> > With the fast deployment of the GPON systems, the price of the APD/TIA price
> > has dropped to about 2X of the PIN/TIA now at 2.5Gbps. I think the 10G
> > APD/TIA will follow the same trend.
> > 
> > As the circuit complexity you mentioned in your eralier email, the charge
> > pump for the APD/TIA is about 15% to 30% of the 2.5G PIN/TIA price depending
> > on the implementing approaches (maybe only about 5% of the 10G PIN/TIA).. The
> > price of the charge pump will not be changed with the data rate.
> > 
> > Best regards,
> > 
> > David
> > 
> > David Li, Ph.D.
> > Ligent Photonics, Inc.
> > 2701 Dukane Dr., Suite 102
> > St. Charles, IL 60174
> > Phone 630-513-7226 ext 15
> > ----- Original Message -----
> > From: "Hajduczenia, Marek" <marek.hajduczenia@SIEMENS.COM>
> > To: <STDS-802-3-10GEPON@LISTSERV.IEEE.ORG>
> > Sent: Tuesday, September 04, 2007 10:40 AM
> > Subject: Re: [8023-10GEPON] PIN-PD D/S power budget proposal
> > 
> > 
> > > Dear Justin,
> > > Thank You for the feedback.
> > > The question then remains: how trustworthy are the cost evolution curves,
> > which claim APD ~ PIN cost with sufficient production volumes? I must say
> > that we (I mean people who do not opt for any particular solution as for the
> > moment) are subject to conflicting confirmation - the APD proponents say the
> > cost will come down rapidly and each PIN levels, the opponents say it is not
> > possible. In the end of the day, I have to rely on someone's data since I do
> > not have first-hand information on this particular issue. In this situation
> > I do not know how to chose which camp to support and how to cast a vote if
> > there happens to be any motion on the floor.
> > > I hope we can clarify the issue during the September meeting - we will
> > have some time for discussion on power budget.
> > > Best wishes
> > >
> > > Marek Hajduczenia (141238)
> > > NOKIA SIEMENS Networks S.A., Portugal - R
> > > Rua Irm腴‰ Siemens, 1
> > > Ed. 1, Piso 1
> > > Alfragide
> > > 2720-093 Amadora
> > > Portugal
> > > * Marek.Hajduczenia@siemens.com
> > > http://www.marekhajduczenia.info/index.php
> > > (+351.21.416.7472  4+351.21.424.2082
> > > "C makes it easy to shoot yourself in the foot; C++ makes it harder, but
> > when you do, it blows away your whole leg." - Bjarne Stroustrup
> > >
> > > -----Original Message-----
> > > From: Justin Abbott [mailto:jabbott@gennum.com]
> > > Sent: ter艸┃feira, 4 de Setembro de 2007 16:30
> > > To: STDS-802-3-10GEPON@listserv.ieee.org; Hajduczenia, Marek
> > > Subject: RE: [8023-10GEPON] PIN-PD D/S power budget proposal
> > >
> > > Good Day Marek,
> > >
> > > It is quite correct that the 1G standard was based on APD's in the ONU,
> > > however, before deciding to follow suit I think it's very important that
> > we
> > > consider the cost issues that also plague the 1G module makers.
> > >
> > > I think everyone can agree that an APD OUN leads to a relatively expensive
> > > ONU, and in the 1G case, although APD cost have come down with volume, so
> > > have ONU prices.  Thus, even with APD cost reductions, module makers
> > > continue to struggle with the high cost of an APD ONU.
> > >
> > >
> > > _______________________________________
> > > Justin Abbott
> > > Product Manager
> > > Gennum Corporation
> > > Ottawa, Ontario, Canada
> > > Ph   (613) 270-0458 x2783
> > > Cell (613) 697-2066
> > > _______________________________________
> > >
> > >
> > >
> > > -----Original Message-----
> > > From: Hajduczenia, Marek [mailto:marek.hajduczenia@siemens.com]
> > > Sent: Tuesday, September 04, 2007 4:59 AM
> > > To: STDS-802-3-10GEPON@listserv.ieee.org
> > > Subject: Re: [8023-10GEPON] PIN-PD D/S power budget proposal
> > >
> > > Dear Hao,
> > > I would tend to agree with Your point of view. The only disadvantage of
> > APD
> > > based ONU that still holds at the moment is the related cost. However, if
> > I
> > > recall right, the previous 1G standard was also based on sensitivity
> > figures
> > > for APDs rather than PINs in ONUs (OLTs still use APDs?) and I was
> > wondering
> > > whether we cannot go the same way. The PIN sensitivity increase (as You
> > > described in the previous email - thank You for that, it was very
> > elaborate)
> > > and the technological developments You mentioned may eventually allow for
> > > replacement of the APDs with PINs. However, one thing is sure - if the
> > > prices are going to go down along with the ramp up in the production
> > volume,
> > > what are the other viable arguments of not having APD based ONUs? From
> > what
> > > I recall the launch powers are much lower, OLT transmitters probably do
> > not
> > > need cooling (except for PR30) making them more robust. The downside is
> > the
> > > more complicated electronics for ONU Rx but that was quoted as minor cost
> > > factor, wi!
> > >  th existing solutions and board designs which can be reused.
> > > Can anybody from the PIN-team comment on that ?
> > > Thank You
> > >
> > > Marek Hajduczenia (141238)
> > > NOKIA SIEMENS Networks S.A., Portugal - R
> > > Rua Irm腴‰ Siemens, 1
> > > Ed. 1, Piso 1
> > > Alfragide
> > > 2720-093 Amadora
> > > Portugal
> > > * Marek.Hajduczenia@siemens.com
> > > http://www.marekhajduczenia.info/index.php
> > > (+351.21.416.7472  4+351.21.424.2082
> > > "C makes it easy to shoot yourself in the foot; C++ makes it harder, but
> > > when you do, it blows away your whole leg." - Bjarne Stroustrup
> > >
> > > -----Original Message-----
> > > From: Hao Feng [mailto:h.feng@EUDYNA.COM]
> > > Sent: segunda-feira, 3 de Setembro de 2007 23:39
> > > To: STDS-802-3-10GEPON@LISTSERV.IEEE.ORG
> > > Subject: Re: [8023-10GEPON] PIN-PD D/S power budget proposal
> > >
> > > Hotta san
> > >   I mean mostly the high launch power related issues when PIN is used in
> > > ONU. At end of budget meeting in SF, we re-listed the technical issues for
> > > both EDFA/PIN and EML/APD solutions. The APD solution left only one issue
> > on
> > > cost for ONU. But PIN/EDFA solution has still several issues not to be
> > > clear. My point is we should move one solution forward, which solution has
> > > the less technical risk. The cost issue should be left for future. The
> > cost
> > > problem may be solved when the volume to go up as GPON case. Certainly, if
> > > some evolution on "low cost Rx" happens in future, it could replace the
> > > higher cost part without any problems.
> > >
> > >
> > > Best regards
> > > Hao
> > >
> > >
> > >
> > > -----Original Message-----
> > > From: Yoshifumi Hotta [mailto:Hotta.Yoshifumi@eb.MitsubishiElectric.co.jp]
> > > Sent: Monday, September 03, 2007 12:49 AM
> > > To: Hao Feng
> > > Cc: STDS-802-3-10GEPON@listserv.ieee.org
> > > Subject: Re: [8023-10GEPON] PIN-PD D/S power budget proposal
> > >
> > > Dear Mr. Feng
> > >
> > > I'm sorry but I cannot fully understand what is your questions are.
> > > Basically, I think Suzuki and his supporters had answered each problem
> > which
> > > is pointed out in the meeting. If you think there is " many remained
> > > thchnical concerns ", we are are very appreaciated to discuss in the next
> > > Seoul meeting.
> > >
> > >  From our point of view, all technical concerns for PIN based ONU are
> > solved
> > > by experimental data and technical study. Also, There is no doubt for cost
> > > advantage of PIN based ONU's.
> > >
> > > It seems for me some APD supporter are opposing PIN based ONU for reason
> > > that "PIN supporters proposal is doubtful". I think this kind of arguments
> > > takes us nowhere.
> > >
> > > Sencerely,
> > > --
> > > Yoshifumi Hotta
> > > Mitsubishi Electric R&D center
> > >
> > > > Hotta san
> > > >  There are a lot of information in the attached material to have been
> > > > presented on SF conference (3av_0707_suzuki_1.pdf).  The
> > > > 3av_0707_hamano_2.pdf has answered your concerns on SF conference,
> > > > such as high power EML issues. Could you focus on answering the many
> > > > remained technical concerns caused by high launch power? The
> > > > 3av_0707_hamano_2.pdf has answered all of questions about using APD
> > > solution.
> > > >
> > > > Best regards
> > > > Hao Feng
> > > > Eudyna Devices USA
> > > >
> > > > -----Original Message-----
> > > > From: Yoshifumi Hotta
> > > > [mailto:Hotta.Yoshifumi@EB.MITSUBISHIELECTRIC.CO.JP]
> > > > Sent: Thursday, August 30, 2007 6:32 AM
> > > > To: STDS-802-3-10GEPON@listserv.ieee.org
> > > > Subject: [8023-10GEPON] PIN-PD D/S power budget proposal
> > > >
> > > > All,
> > > >
> > > > In the last meeting, Suzuki had presentaion for PIN@ONU.
> > > > In discussion, because OLT max launch power is such high as +13dBm, we
> > > > had some questions about eye-safety from the floor.
> > > >
> > > > In the attached presentation, we clarify safety requirements which we
> > > > should support, and propose new power budget. We believe this one
> > > > could support both eye-safety and B++ with PIN-PD in the downstream
> > > direction.
> > > >
> > > > Having discussions on the reflector is appreciated, also if you would
> > > > like to support this presentation, please let me know.
> > > >
> > > > Sincerely,
> > > > --
> > > > Yoshifumi Hotta
> > > > Mitsubishi Electric R&D center
> > 
> > 

---
---------------------------------------------
HIROSHI HAMANO         Network Systems Labs.
FUJITSU Labs. Ltd., Kawasaki, 211-8588 JAPAN
TEL: +81-44-754-2641  FAX: +81-44-754-2640
E-mail: hhlsi@flab.fujitsu.co.jp
---------------------------------------------