Cisco Optics Podcast Ep 52. The smartest data center operators use multi-mode fiber - do you? (1/7)

[00:00:08] Hello everyone and welcome back to the Cisco Optics Podcast where we talk about pluggable

[00:00:13] optics for networks.

[00:00:15] Multimode Fiber has been the workhorse for fiber optic communication within a data center

[00:00:19] for years, as well as a choice for Ethernet backbones and office buildings.

[00:00:23] It has a rich history that goes all the way back to the first optical fiber standards

[00:00:27] when 100 megabits per second was a huge data raid.

[00:00:32] Anybody remember those days?

[00:00:34] This is episode 52 when we begin a new conversation with Hao Dong, Market Technology Development

[00:00:39] Manager for Corning's Optical Communications Business.

[00:00:42] We begin with his start in semiconductor lasers for telecom before switching to fiber lasers.

[00:00:48] Hao Dong is a Market Technology Development Manager for Corning's Optical Communications

[00:00:52] Business.

[00:00:53] Within this role he focuses on technology and market trends to identify, evaluate,

[00:00:58] and develop leading optical network solutions that form the backbone of today's connected

[00:01:02] world.

[00:01:03] Hao has over 18 years of industry experience and has held various positions in engineering,

[00:01:08] development, and marketing within the fields of optical components, fiber lasers, and

[00:01:12] optical communications.

[00:01:14] Hao holds a Bachelor of Science in Physics from Wuhan University and a PhD in Physics

[00:01:18] from the University of Connecticut.

[00:01:20] And now join me as I talk with Hao Dong.

[00:01:23] Hey Hao, how are you doing today?

[00:01:36] I'm good.

[00:01:37] How are you?

[00:01:38] Good thanks.

[00:01:41] So thanks for joining us.

[00:01:42] I know you're very busy.

[00:01:43] I'd like to start just hearing you tell us about your journey and optics, how you got

[00:01:49] into it, what you had been working on, what you're doing lately.

[00:01:53] Anywhere you'd like to start?

[00:01:55] Yeah, sure.

[00:01:56] Of course.

[00:01:57] Yeah, I actually start working in the industry I think about 15, 16 years ago.

[00:02:08] I got my first job as a product engineer in the semiconductor laser company in New Jersey

[00:02:15] actually.

[00:02:16] That was, yeah, that was back several years after the telecom bubble was gone.

[00:02:25] So the industry at that time was really slowing picking up.

[00:02:30] So you missed the telecom bubble bursting or did you experience it?

[00:02:34] Yeah, fortunately at that time I was still at school.

[00:02:37] So when I started everything, it's like I said the industry was slowly picking up.

[00:02:46] Many countries was getting ready to deploy 3G and it seemed like people were needing

[00:02:55] more and more fiber and lasers at that time.

[00:03:00] Okay, wow.

[00:03:01] 3G.

[00:03:02] Who can remember that that was once the hot new thing?

[00:03:06] Yeah, yeah, yeah.

[00:03:08] But I would say that the starting phase of the fiber to the home deployment, right?

[00:03:15] I still remember at that time I lived in New Jersey and the people we really switched

[00:03:22] to and totally you can see it makes a difference.

[00:03:27] So that's the first time we really, for real starting enjoying the convenience fiber bring

[00:03:32] to us.

[00:03:33] Yeah, so work wise like I said, you know, I was happy I, you know, I choose the telecom

[00:03:41] industry as my starting as my first job.

[00:03:45] I working for the semiconductor laser company at that time we were making a 10G, EML

[00:03:53] and TOSA.

[00:03:55] My job at that time was really, you know, working in the floor and very technical, working

[00:04:01] with the technicians and engineering, troubleshooting any possible optical issues where you came

[00:04:09] across almost on a daily basis.

[00:04:12] So you were doing the actual fabrication onsite of the semiconductor or did you have like

[00:04:19] that?

[00:04:20] Yes.

[00:04:21] Go ahead.

[00:04:22] At that time, you know, the company I think basically we did two parts of the work.

[00:04:28] One is to, you know, MOCVD to grow the chips and then packaging the chip together we call

[00:04:37] it the front end and back end, right?

[00:04:39] You know the difference.

[00:04:40] Actually no, explain that to me please.

[00:04:43] Yeah.

[00:04:44] So you make, you know, for the chip making by basically you grow the materials starting

[00:04:51] from the chemicals, right?

[00:04:53] That's what we call it the front end and when you have the chip then you need to go through

[00:04:58] a lot of testing and environment process.

[00:05:02] And eventually you make the selection and the package everything together into the

[00:05:07] dials and the packaging stage that's I think we call it the back end.

[00:05:13] Okay.

[00:05:14] So the front end is what happens first, the actual fabrication of the semiconductor and

[00:05:19] then you have to put it into what some kind of substrate wire bonds I'm guessing.

[00:05:26] Exactly.

[00:05:27] Yeah.

[00:05:28] I think it's called the chip on the chip on the chip on the base or chip on the chip

[00:05:37] on the sub mount.

[00:05:38] Yeah.

[00:05:39] Okay.

[00:05:40] Yeah.

[00:05:41] Okay.

[00:05:42] And then back end.

[00:05:44] Yeah.

[00:05:45] And then as you can imagine when you got the chip, when you packaged the optics together

[00:05:51] into the dials, you still have to go through a lot of the testing and the testing is,

[00:05:57] you know, it could be very lengthy and a complicated process, right?

[00:06:06] Sorry does the packaging include fiber attach, fiber coupling?

[00:06:11] Yes.

[00:06:12] Yeah.

[00:06:13] It involves a lot of fiber alignment because if you imagine when you have the chip,

[00:06:21] if you have the chip and then eventually your job is to launch the light out of the,

[00:06:28] from the semiconductor laser and launch the fiber into the fiber, right?

[00:06:32] There will be an isolator in there to make sure there is no back reflection.

[00:06:37] And also you need to make sure you have the lens, you know, you have the lens system to

[00:06:42] calm into the light, right?

[00:06:45] And the final stage is to work on, to basically to work on the alignment to make sure you

[00:06:52] have a higher alignment efficiency, which is also very important.

[00:06:58] And that's all actively aligned I'm guessing where you have to measure the power and

[00:07:03] tweak as you're optimizing the power?

[00:07:06] Yeah.

[00:07:07] I, you know, there were a lot of details, different common ways of doing it.

[00:07:12] You could do the positive, you know, the active alignment or passive alignment, but most likely

[00:07:19] you will need a lot of cameras in there, watch the camera and it's a, at least at that

[00:07:26] last day, at that time, the work is still very, you know, manual and, of course,

[00:07:35] I'm sure, you know, in these days, probably a lot of work will be done by automation.

[00:07:42] Yeah.

[00:07:43] Yeah.

[00:07:43] But at that time it was like, yeah, I also remember at that time it was very manual,

[00:07:48] very complicated because it was a young industry.

[00:07:52] Automation hadn't, hadn't really been developed yet for that particular process, right?

[00:07:59] Yeah, exactly.

[00:08:00] So, yeah, that was my first job, you know, it's...

[00:08:04] Oh, sorry.

[00:08:04] I interrupted you before you got to the back end.

[00:08:06] So...

[00:08:09] Right, right.

[00:08:11] Yeah.

[00:08:11] So as you can imagine, there is a lot of testing work to do and as a product engineer,

[00:08:19] I also did a lot of testing.

[00:08:22] So like I said, that was my first job.

[00:08:26] I really enjoyed the learning from the work.

[00:08:30] I learned the LabVIEW program myself.

[00:08:32] So, you know, what was your...

[00:08:34] LabVIEW.

[00:08:34] Yeah, LabVIEW program back then.

[00:08:36] Oh, I love LabVIEW.

[00:08:37] Yeah.

[00:08:37] Yeah, when you have some idea, think about that, you know, to put everything together,

[00:08:44] hook up those equipment, write the LabVIEW program, you know, make it happen, right?

[00:08:50] So that was really satisfying for me.

[00:08:55] Yeah, yeah, I can totally relate.

[00:08:57] Yeah, so that's how I end up in the telecom industry.

[00:09:04] But after several years, I switched to another company.

[00:09:09] Basically, it was kind of like a 180 turn.

[00:09:15] I started working on the fiber laser.

[00:09:16] So it's totally non-telecom even though it's still fiber based.

[00:09:24] Yeah, so...

[00:09:25] And this was a high power fiber laser?

[00:09:30] I saw in your LinkedIn, it was kilowatt single mode.

[00:09:36] Yeah, exactly.

[00:09:36] It was a kilowatt single mode CW laser as you can imagine.

[00:09:41] It's not post...

[00:09:41] If it's post laser, I think, you know, kilowatt...

[00:09:48] I mean, the laser...

[00:09:50] Avagabrage power, right?

[00:09:51] Average power, yeah, exactly.

[00:09:53] The fiber laser was for cutting and the welding applications.

[00:09:56] So kilowatt is typically a CW based laser.

[00:10:03] So...

[00:10:05] But in a day, it worked.

[00:10:07] Back in my grad school days, we used to do mode locking of fiber lasers.

[00:10:13] So we would think more about the peak power.

[00:10:15] Average powers were still, you know, less than a watt.

[00:10:20] Yeah, yeah, exactly.

[00:10:21] Peak powers were, yeah, certainly the kilowatts.

[00:10:24] Right, but I think, you know, probably we were talking about the same,

[00:10:29] you know, similar stuff doing the similar work, right?

[00:10:31] Splicing, fiber-splicing, fiber processing, handling, that kind of thing.

[00:10:38] So it might daily work the most...

[00:10:41] I would say the most important thing is fiber-splicing.

[00:10:44] You have to develop the recipes, right?

[00:10:48] Slicing recipes.

[00:10:50] And be careful of almost all the single details.

[00:10:54] People always say the devil is in the details.

[00:10:58] I think it's, you know, for me, until I started working on the fiber lasers,

[00:11:05] I really understand what does it mean by that.

[00:11:09] Because when it comes to the high power,

[00:11:12] you've got to be careful with all the single details.

[00:11:16] You can't overlook some even, like minor things, small details.

[00:11:21] Otherwise, you know, catastrophic failure does happen, right?

[00:11:25] Fiber-succes is just a bit scary.

[00:11:29] At those power levels, things can melt, I guess, right?

[00:11:35] Oh, it's a fiber-fused.

[00:11:36] Or at least get damaged.

[00:11:38] Fused?

[00:11:39] Yes, yes.

[00:11:40] Yeah, when the fiber-fused happens, it's total damage.

[00:11:47] So there is no...

[00:11:48] What do you mean by fiber-fused exactly?

[00:11:51] It's burnt.

[00:11:51] The fiber...

[00:11:52] Basically, the root cause of the fiber-fused is the high temperature, right?

[00:11:57] When the temperature goes much higher than the polymer coating can handle,

[00:12:03] it's not an issue for glass, right?

[00:12:05] For the glass, you know, it's a silicon material.

[00:12:07] It's typically, it can withstand the much higher temperature.

[00:12:12] But it's not the...

[00:12:14] It's not the same case for the coating.

[00:12:16] For the polymer coating, when the temperature goes up to, say, 200 degrees,

[00:12:21] you're at 300 degrees or something like that, it gets burnt.

[00:12:25] So that's why you have to handle the, you know, the temperature very,

[00:12:30] very carefully.

[00:12:31] The mechanical design, the thermal processing is very, very critical.

[00:12:37] So how did you keep the polymer coating cool?

[00:12:40] Did you...

[00:12:41] Is it air-cooled or any liquid cooling?

[00:12:44] There are many different ways.

[00:12:45] It depends on what you need.

[00:12:46] It depends on the power level.

[00:12:48] Sometimes I think the common practice is what are cool.

[00:12:51] You need the cold plate below the fiber and below the fiber plate.

[00:12:55] And also you need a package of fiber in a well-managed way, right?

[00:13:02] For example...

[00:13:02] The thermal management?

[00:13:04] Yeah, the fiber management, thermal management,

[00:13:08] how do you lay the fiber on the plate?

[00:13:12] You know, how to package in your fiber to make sure the contact,

[00:13:16] that kind of thing.

[00:13:17] So yeah, it's quite tricky and there are a lot of details

[00:13:23] when it comes to the thermal management.

[00:13:26] You know, it's funny we're talking about this now

[00:13:28] because this is a big topic in data science.

[00:13:41] That was the first part of my new conversation with Hao Dong.

[00:13:44] Next time we'll get into both single mode and multi-mode fiber.

[00:13:48] We have a new website.

[00:13:50] It's optics.podcastpage.io.

[00:13:53] You can either listen there or use the same podcast platform

[00:13:56] you've been using all along.

[00:13:57] Please subscribe.

[00:13:59] Better yet, leave a review,

[00:14:01] especially if you've been using Apple Podcast.

[00:14:04] Remember, we're part of the Cisco Podcast Network

[00:14:06] where you can find other great Cisco podcasts too.

[00:14:09] We also have educational videos on YouTube.

[00:14:12] Just go to youtube.com and search on Cisco Optics.

[00:14:16] Thank you for listening.

[00:14:17] This is Pat Schild, Product Manager at Cisco Optics.

[00:14:19] The next episode is part two of my conversation with Hao Dong.

[00:14:23] Until next time.