EEVblog #1032 Part 3 - John Kenny Keysight Interview

back to the design of these sorts of things how BIG's your design team on various leaders I saw eater money on ten or twelve people got you know in for analog engineer is three one one or two Mechanical Engineers and for firmware and maybe a software person Oh firmware and software different absolutely software like as in PC software you're talking that software and then is this kind of bridge gap where you have a web browser technology which is sort of a hybrid okay right interested and software is one of the areas that we're investing in more heavily than any other area we just started the new Center in Atlanta in Atlanta Georgia and we've started a relationship with Georgia Tech and we're growing a huge Software Center there to expand our standard software across all of our products God is software yeah not everybody wants a front panel and when you do more sophisticated solution oriented measurements that have multiple pieces of equipment you know the front panels not enough so take a pixie system there's no front panels and you're you're solving a much larger problem you need software to make that really fly is every product going towards web browser like a web page built in what's the correct term Alex re-emerged to have some sort of presence on the web right it does also because now you can go in and you can access the web page for the instrument right and that's how you can control it and basically one of the things that I'm working on actively is if you go through all the different products that look the same and you look at their web presence what they come up with it varies on product to product because this design team thinks this is what the web should look like and this design team thinks this is what the web should look like and this design team doesn't have enough money so they did the minimum necessary and and it's it's not good for our customers yeah okay we've had cases where we went overboard and spent over a million two million dollars on our web interface um Wow and it's amazing but we could have done it with a software package for a third of the cost got it the only reason they like it in the web is that the the field engineers don't have to install the software before they demo it for the customers correct let's go on they phone and go straight to the web address oh yeah that's not a reason to spend two million dollars yeah okay and the problem is everybody thought the web was gonna not have to be updated the way was always gonna stay the same well you can't get Java can you you can't get Flash Kenya so it hasn't that worked out to be as good and it's evolving just like every other part of the PC is evolving so it hasn't been a zero it's thing we thought it was gonna be so we're pulling the web back to more of a common level that's gonna be more standardized across our products and more more rapid for us to deploy more easily still give you the local control over the web which people want but not as full-blown is some products have done it's not a stripped-down which is not good enough for from us some other products I mean one product we make the web control consists of a winner you can type Skippy commands yep that's well it's yeah it means the other guys cheat but not from that but we still that's an easy thing to integrate but that's all the integrator you know he meets the LXi standard but that's about it well we were most of our products were trying to standardize on it's where you have a view at the front panel and you can press buttons and the display looks just like the front panels and that turns out to be for most people they just want to remotely control the product and and verify that it's doing what they expected so that's what I expect I expect to see something well know there's a group of people who want it because the web has giant you know CRT you can do much more than you do on a limited size for a panel yeah and you can even add additional level high level constructs that you wouldn't put in the product cuz it was hard to use on a small display that's the area that I think there's more debate can we do more of that and my feeling is from an efficiency standpoint let's get everything common and then let's move the whole thing as a phalanx forward and make it better more efficiently instead we have some groups going all the way here some groups going to here it just doesn't lend itself toward efficiency and what really frustrates me is we have a concept the name that Ron came up with in their management they have called one key site we want our customers to have a consistent one key site experienced across all of our products that's as important as week we kind of we're a bit cynical us engineers in their market in that we think that's a bit you know that's a bit corporate II silly but it's if you if it wasn't there you'd probably miss it you're probably it's probably under appreciated by us I think most people is if you talk to them if they own an Apple product right now that doesn't mean apples perfect no there's lots of areas that there's issues with apples in went but for the most part it makes people brand loyal and it they do have a consistency inequality for the experience that they're very big on you know I'm not saying that we're trying to clone Apple but we are trying to take some of that same shared experience and make it working system so we'll find this user interface on most of your benchtop products so like these we have a box coming your way with the new in 36 3 11 12 or 13 I know which one they're going to send you and there are screens on that that are identical to the screens are awesome for things like setting up the land for setting up the file manager well I have different experience on each one it's a file manager you know and that's that's something is the example that we literally have four or five different implementations on this exact screen because it was developed here versus here versus here uh-huh that's not acceptable yeah we're paying for something three times a night of course yeah that's pretty much the customers hate that yeah they learn how well no I don't yeah I want to be able to use the same I expect the same interface and you're not getting three yes but that's being worked on absolutely excellent how quickly can you spin fixed firmware issues and things someone reports a firmware issue on a product how what's the mechanism to fixing that if a firmware issue comes up the the hard part is getting it to the people who can fix it yeah because we're a big company it's typically doesn't or a support group the first thing they have to go is to add valance truly a mistake and sometimes that communication takes a while cause this is not the customer wants to just sit on the phone on hole waiting for the support person to describe and document it that can take the longest amount of time when he gets did a design team depending on how hard it is to fix the bug that can take a couple days it can take a couple weeks if it's a case of a memory overflow that only happens on a rare occasion and that's the hardest thing that might takes two three four weeks to get it you know catch it with its pants down so to speak once you find it it takes very little time to fix it and it takes typically a week or two to guarantee that we didn't break something else we do exhaust the testing before we ship a firmware update out because the worst thing you can do is fix it and try something else so that's that delays us getting out there we have very very thorough regression testing and even that doesn't have perfect coverage so we like people we'll pass it around to people to make sure that you know other people that make sure we didn't break something else in the process is that an automated testing do you have automated tools in place like to exercise all these functions remotely and just go through every possible well company for the bus make sure the buses all the different buses you know GPIB LAN USB but front panel is very difficult to automate many people know you can use vision yeah bet you can push the buttons from that you can push the button but you actually have to love the LXi you have to look at it and with vision systems it's very complex and time-consuming to program vision systems to simulate all that stuff and catch it and you know ok so it's different then someone's loss to look at well how is it different and to figure out what that means there's no artificial intelligence I'm gonna make all that happen so front panels are probably the biggest headache when it comes to automated testing and catching things on the bus stuff is much more amenable to automation and yeah also though a voltmeter you're measuring voltage we have to make special fixtures that feed different references those zip codes of different types so and actually the hardest one believe or not to write regression is power products because you have both sourcing and loading and measurement it's nasty and different types of sequencing like so for example most of our power products have the ability to put out sequences and with different time delays and all that so it's they can get very complicated to regression testing the 3458 a mm-hmm there's rumors that maybe a new ones in the works and are you still going for those sort of you know the top of the line and cowl lab type instruments is that still on the 3458 ok so most reviewers products yes today have not been eclipsed by anybody for absolute performance we know that's really important and we will not let our customers down ok that's not all I can say ok being worked on this I did it's been considered we will not let our custom nothing to do not let your customers do well done all right I'm with old products like that I mean how long's that one being around as long as the 34401a released I think you're right in my case with her yes I miss release yeah 2030 it's better one Wow um do you have you lost talent I mean clearly you're going to lose talent if people know like you know like at a height at such high in there's so much subtle magic in there that so I actually managed that how does working on ya know the first time we tried to do a replacement on it uh-huh and it was fascinating for me to learn kind of subtleties they have to go through on that and I learned a great deal about you know high involvement managing those two guys both those two guys are still with the company they still with coming out the four guys the four horsemen that developed the original Fianna none of those four guys are still with the company one of those four guys is with one of our biggest competitors that make a mess right okay and he's an amazing engineer that we lost and he's still developing with them he's still doing quite a good things for them the 58 has some of the most challenging aspects to it that the subtleties of that one of our R&D managers who was running the group at the time came over and one of the poor guys was still with the company had handkerchief laying over the circuit board when he was testing and he says why are you doing it I says because the subtle differences in the air temperature going over to some of the parts we're moving tenth of a degree and it showed up as noise you could you could see you could say it you could see the fluttering of the air on top of the parts was changing DC yes so he said hard to say you've still got some talent oh yes well they are in the manager that managed those four guys is still with us all right and he's busy it has been busy training younger guy who's he's been there 15 18 years um he developed the 34 almost kind of site yes almost uh apparently it went into this yes and both of those guys are still with us and we're looking at how we can take that to the next level do we have a seven and a half digit model now and the seven and a half is very close to 58 performance Wow I mean in some areas we we know some of the stuff we'd have to do that next level so we came very close to having a complete new 58 design but part of the challenge with our old management team was they said you're just replacing the revenue we had before there's no growth in that there's no new profit in that and we were part of the life sciences cash cow Lee's bidding one but I don't wear a test and measurement company exclusively there's a lot more willingness to to be having a premier product in our catalogue that we understand got it and that's one of the great things about being a pure test and measurement boy is we're much more focused on what matters the reference inside this and others are they get in difficult to get no no that's an interesting story though yep the reference that's inside the 3458 is was developed by us 30 years ago and we came to the conclusion that we're not the right people to make it so we gave the designs linear technology this is the LT Zed 1000 correct yep okay you guys Tamela gulped it and went out to them and we have a big burning system in our factory that we mount them in little boards and we burn them in and we sell a graded version created dry for the other of results now the seven and a half digit one this one this is the seventh deal yep uses the LTZ 1000 it's what it does it allows us to get to seven-and-a-half digits because the drift this doesn't have all the hooks that a 58 has it doesn't have the level of a cow it doesn't have the the AC measurement technology that is unique on the 58 some of that stuff is is more straightforward we had already started to work on redesigning that and stay tuned okay we're in the process of looking how to take things to the next level there's still a lot of new measurement technology that can be done a voltmeter is a fundamental building block for a lot of what we call synthetic measurement technology and we're looking for how we can grow into other measurements and that's one of our competitors down in Texas does a great job all they do with synthetic measurements right because their their granularity is a little card about this big they pop into a frame um we have focused on the standalone they focused on that there's a significant opportunity in that area for us and we think you're gonna see more of that over time interesting what what sort of volumes would those high end instruments sellout I wouldn't imagine there's like how much would they have to smell for to get to make it worthwhile all the years and years of developing such a high-end product with such a limited market its which is basically cow laughs and I was limited to I think not know in one of the bigger opportunities for the 58 it's kind of surprising when you look at high-volume consumer products throughput is a huge factor for these people where you know the the volt meter even if it's a ten thousand dollar box like the 58 it's the cheapest piece of equipment in the rack you start putting a spec in or a high-end space than that there's thirty forty thousand and you've got material handling equipment trying to move this stuff in and in a few seconds yeah you put a in a half did you vote me during that system you don't have to change ranges ever and you still get an accurate reading going on six stomachs much faster taste through port much for the better you don't have to change ranges you can laugh about a hundred volt range no matter what single comes in an accurate reading yeah on a six-and-a-half you might have to change ranges and that takes time and it takes time and the server is down your product that's right so we have some people using the 58 Ford's throughput benefits we have other people who build large multi based systems they put a 58 at the bottom of the rack and they recertifying cal everything else in there at everyday high because they want the system to be robust and we sell into military aerospace it's a very popular configuration so it's not all just Cal labs in fact frankly Cal ABS we've we've lost a lot of that business to our big competitor in you know yeah they've really taken over a lot of the Cal they really focused on Cal as a as a big deal we get less of that then we used to get for sure and these other things you know this kind of transfer standard where they mounted in Iraq they just calibrate the 58 and then they put it back in the rack and it calibrates everything else got it because it's so accurate that's that works you know you might have one of these they might have some some pixie cards some other things you know an arm a function generator this the 58 is incredible RMS a you see measurements it can verify all those things in the rack they just route everything through it it does a confidence check and a calibration and it's less expensive and smaller then can we expect to see an eight-and-a-half digit bench one like this rather than Iraq meal I guess if there's benefits to through port I mean I mean seven half digits is almost there but if you can get the extra I don't know exactly what my know that there that's an interesting question I don't know if there's as much new benefit to having it on a bench we're gonna you're dealing with data right yeah it's more automated system rack hmm stuff yeah again the AC measurement technology in the 58 is something that has a lot of benefit in this day alone like this so we are looking at how can we bring that that kind of technology forward but the the pure 8 and 1/2 digitus there was a lot of debate to do a 7 and a half in this format okay we flipped that battle over and over and yet finally people said you know how much will we get out of it and they said wheat the technology is so common we think we can do it fairly quickly it's one of the tricks if you can make the effort low people stop argument - yeah every place could be really hard to you you know down to the bone yeah because it's gonna cost you 12 million for a project to do it for this thing this was a follow-on to the 60 and 61 and he took a lot less time because they planned for it oh so when you're first designing this you thought maybe we might do a 7 - later they knew the 7 and a half was on the roadmap okay we knew I was coming they didn't lay out all at one time but still because we had the front panel we had all the other pieces they could do it in a lot less resources so instead of having a team of 12 I think it was a team of four come here me did it 10 percent of his time because the package was already done exactly it's just a matter of change and engineering the 7 1/2 digit measurement solution everything else is the same yeah so it's a lot quicker a lot smaller project team and as a follow-on and all even on the marketing literature and documentation was reusable all the screens are the same yep exactly so the firmware was minimal change got it do you see do you laugh at the competition that try and do like six and a half digits and their specs aren't even close or I come or some of the Chinese can never laugh at anybody no no no but as some of the Chinese competitors getting there I think the biggest mistake we make is to ever underestimate our eyes yeah I'm talking about this this morning you know in the Toyota and Honda came into the US market with terrible terrible inexpensive cars mm-hmm so did Hyundai yep and we almost now they produced the most reliable and they're the largest selling car in the world Hyundai is the fastest growing car company my world all right ignoring your competitors bigger smallest I'm always a mistake you know being arrogant is it was our big mistake we made for many years God is something that will not be repeated if a few of us have anything to say about it excellent you don't get to be the best and you get arrogant you won't stay the best for very long so we can talk about is through one of the things that I really was happy to be able to do this with you is the kind of shift in how people learn about products yeah and you know one of the things I tell people so for example developed the 36 3 11 12 and 13 I demanded that all the projects on the project team go to your blog I said all the criticism you know but all the things they get right - I was thought yeah I don't want to know what they got wrong I don't know what they got right I'm gonna learn what they figured out that I can use yeah Steve Jobs that are to steal right absolutely I look at something what you provide to the industry is much like a movie critic you you saved me from watching their movies her okay and at the same time you teach me what to look for and what the craft of a great movie is so we really appreciate what you bring to the market and the fact is you're much more accessible than those millions of customers that you represent so it's something that as we've tapped in more to social media we think that you are really a valuable resource for all of us in the test and measurement history you make our industry a better industry for it you you are an objective purveyor of how good products you I think you're gonna make all of our products better you're gonna make their presence big art that's the intention right and that's something that I think it's really exciting a little scary too because we don't have any control of it oh oh it's not exactly it's it's that standpoint it can be a little frightening I did not come to this today now you've got to take the hits with the you know they've got you know if your product fails like a you know the demeanor for example the you remember the the the 1272 yes meter ended though soldiering issues and they've gone through but I'd like to think that people value the response that's right that you give if you if you just ignore it then they're not gonna give I don't say hey yet we admit it's a problem we're working on it everyone's gonna think better of the company you know there's a there's a doll cliche that it isn't what you do when you fall down it's what you do get out yeah and to us you know what you bring to the market is really an invaluable resource that makes us better thanks and it's something you can harness that and get a perspective cuz when you say what you say if your readers don't agree with you we read that too so yeah and it really helps us create a perspective that it would be very difficult for us to get any other way so you're actively looking to be on the forums and things and in an official company we have way all of our project team members them for the low cost products I demand they read all your forum stuff on all the products that compared to our product fantastic we also buy all of our competitive products and tear them down yeah first we evaluate them in a black box way and we've instituted a practice where they each of the engineers on the project has one of our competitors products and they actually have to use it ah during the project on the bench interest every few months we sit down and say now what do you feel now and then happen some things that they got really right you try to incorporate them in our product and other things that they didn't get right and we try to say well how can we do it better what don't you like about it and did we end up tripping into that doing it the same way no that's you know let's move the bar up and so I think you've helped make that possible it's a fantastic so that's how we're gonna stay ahead well we have some interesting stuff in all the different products one of the ones I find is the most amazing is what we call true form yep okay true form is really and I told you before it's a trade secret but there's aspects of that I can talk around with I'm leaving the trade secret if you look at a traditional dds style function generator DDS is great because it's cheap mm-hmm but it's got all sorts of problems it's got jitter it's got it doesn't work well for other types of square waves for example square bees are horrendous with jitter DDS technology true farm has allowed us to dramatically simplify the structure of a function area I don't know if you're aware but in the older function jitter is they would not generates queries with the DAC right David generally I right away yeah me we'd run that into a comparative it's just you have a limited like two or three days and slew rates and you would have like a ten percent minimum duty cycle now with trueform you can get less than one percent duty cycle and you can get in infinitely variable slew rates and that makes the products so much more usable on your bench yep and we're actually looking how can we take that technology and integrate it in more and more and more of our products because it is so so much better for testing and use it's one of those things that we first did it we learned about it from the RF guys because they've done they're trying to create spectral purity with their weight on generators and that's all they focused on but they this is one of these that I didn't get involved in but it kind of showed what could happen one of our guys from the team that develops functionaries actually went to California to our RF site and learned about it you know at a lunchtime Congress said he was off steel on that yeah yeah he was smart enough to realize what a ground for function dairies brought it back it takes a ton more FPGA yeah and in fact in our hundred and 20 megahertz part we actually had to do the waveform in pipeline in four paths get the gigahertz down because F PGA's don't want to run past about tonight and yes some yeah we couldn't afford an ASIC cuz the volume on the high performance one is lower so he actually figured out a way to pipeline the whole thing as before pass and then mix them all together with a delay line at the end nice and it gives us an incredible benefit and some of our competitors have picked up a little gloss kind of Saia sieglin of doing a they do a one that they claim is innovative is that just a copy of no it's not a copy because we don't there's nothing to copy we don't write what we did it nobody's the even but the intention isn't they copied in what they've done is a much simpler version that's not as sophisticated it's a good solution this is allows you to generate small duty cycles in a large right in a large memory do you know what siglent own I believe is just basic interpolation all right okay we've done is much more sophisticated than interpolation it's a much lower distortion much lower jitter much more fine-grained and we're not telling him what we did because what once we tell him they could do it too yeah although it was interesting like I said in the high frequency function area once they knew how to do it because they did it on the lower frequency when it was it's just one path the hard part was putting it in a giant FPGA and figuring out how to do four paths in parallel and and line them up at the end that was really pretty tricky stuff but the that's the kind of stuff I when you say digital there's I don't think we've even really come close to fully fleshing out all the things we can do digitally to make up custom measurement equipment even better I mean I think the Scopes that they're doing spec hands now that's an example where they say hey what's the difference and we're doing more and more and more of those kind of things what you're gonna see next is you're gonna see where if you have two volt meter's in a function you have a low frequency network analyzer yes that's right but you can't do that today very well no it's a big clunk in speed you can't synchronize the three yep you can't synchronize things but there's no reason you can't in pixie you can do it but then you don't have the positive user interface you don't have it all linked together so there's a lot more to come how would you link them in a technical aspect how would you Lincoln what would you link them through the network interface orders that have to have a dedicated physical well we developed you largely well I Triple E 1588 many years ago we actually developed it in our labs and then gave it out to the world for free like we did with GPIB 1588 is time sync over land right and it's used for video it's used in everybody's house they don't realize it it's not how they do streaming video and you're out the TV and stuff and how they distribute streaming video in the enemy that was an IDI quesada agilent there was developed as I think he'll the Packard oh how that was high a what's been around a long time okay it was one of those deals we did GPIB you have to pay a dollar to get a license I know you're out there but that's a four huh I was a little bit miffed when I found out how much we gave way on it but it's now built in all the microprocessors and laying out stuff we've developed a method to take it a step further we combine it with an FPGA and we can get pica second-level synchronization with 1588 all over the network over the network Wow okay how do you get Pecos and send another trade secret and you're not gonna tell us a secret pretty amazing stuff the labs guys that come up with the problem with 1588 is we just put it in our products and told you you figure it out yeah it would be so complex they couldn't write it's really difficult so you made the application layer these application layer to do it and frankly one of our Rises as a companies our software is not that up to seonyu right so now we're renewing our focus on software and really putting a lot of money and time on the software so we can start getting out some of these higher-level solution focus the the big term in the company today is solutions not hardware software not just hard ways you know it's hardware software and people make solutions that's really that's doing the other diet I was characterizing the power supply on the bench and I wrote them all down by hand you know I didn't because I couldn't be bothered or true figuring out how to automate them you know they're all got LXi connectivity and everything yeah I'm not gonna said bit if there was a solution that would I could just download from the website oh just tie these together and it just worked yeah I would have used it it's gonna take us a while to fully flesh out that outcome but that's in a path world yeah if there's no question about it that's really where we're going because that's the kind of problems that people face today they don't have time to figure this stuff out and yeah you know you look at the emergence of IOT and 5g and some of these new emerging technologies there's gonna be more electronics in everything you own I mean everywhere you turn when you open the room the lights turn on because there's a sensor that senses you coming in and all about that's just the beginning there's so much more that's coming with all these new emerging technologies and they just don't have the time to do it the old-fashioned way and they're often on engineers that's you're doing what they're doing design that's right it's just like they're like they're gluing together our do nose or us Braveheart yeah yeah yeah and they're doing amazing stuff because they're using all this off the shop that's right hardware and software but how do you characterize that stuff in more complex ways and that's really the challenge for us is to create higher-level integrated solutions with multiple boxes and it's a huge challenge because it forces us to stop being little sites that developed independent products and starting everything like one keysight and that's why our management team has really shifted investment focused dramatically toward the software toward the solution side even to the point of how we market and sell our products we now have what we call solution teams and centers of excellence and I'm responsible for a third of the company as the technical side of centers of excellence and my boss is the manager for all this group and we have four solution teams tied to key industries right so we have one automotive an Energy Group we have one general-purpose an Education Group we have a wafer test semiconductor test and and board tests and those groups just focus on solutions in those industries and what a motive is I think the one the most fascinating ones could let's face it 10 years now you're gonna be reading a book where the car takes you where you want to go right you know it's a better title that self-driving vehicle and the technical challenges behind autonomous vehicles are just it's off the charts crazy yeah what they're gonna do you know Google and other people are doing these self-driving vehicles did they map everything out you know what they're finding out they're finding out that it don't work so good yeah when it's raining when there's construction and solving those problems is is incredibly difficult they're relying on networks to go back to central computers I was just I was driving across Sydney they have that Sydney is impossible to drive in at the best of times it was terrible for us driving home at night it was torrential pouring rain and it's construction everywhere they're chopping and changing lanes on a daily basis and I went not like this is just a nightmare from a self-driving car point of view it's just terrible and they know that and they know that but they still this is an enabler for them to sell you and I'm even more expensive far or even you're gonna rent time on a cars cars is so expensive we're not gonna own them and they're gonna have how do you feel about a car that the other guy didn't put air in his tires yeah when you're driving the to 80 miles an hour and oton of his driving lane yeah so the lot of things are gonna change in ways that I don't think people are ready yet for you but the technology that's going to emerge because that's what people want it's gonna be it's gonna be crazy do you think it's kind of tight longer than most people think cuz everyone's told you know yeah it's gonna be here like that I think in limited varies it's gonna be sooner but I think it's gonna take a longer to get all the bugs with you one of the things they're talking about now is they probably have to put transducers in the road oh yeah I want you get to that level it's the HOV lanes are gonna be coming all right oh no driving lanes that's not the first playthru saying you're gonna sing it they're gonna repurpose the in chobe lanes and you're gonna pull into like a gating area with an entrance ramp cars gonna stop and then it's gonna pop you in it right yeah you're gonna be tailgating the cars I've got it and then you write so you break the the wind barrier so you get efficiency at higher speeds and when it'll be like airplane crash is when an accident happens a lot of people gonna die but it'll happen very very importantly exactly that's gonna be one of the I think the one of the most amazing transformations that's gonna change everything we know but even the way that electronics is in our lives one of the things that we've been looking at is counterfeiting right and your Ray Ban sunglasses yeah yeah prints a 60% of all Arabians are counterfeit Wow okay yeah so they're talking about putting electronic chips in every pair oh so they they can tell if it's counterfeit yep and then they can just copy that as well yeah it's hard because they're trying to be one step ahead that's right yeah is there a point where that one step ahead thing just you might as well you're better off given up and just figuring out a better way to but how do you feel about doing medicine when it's fake yeah no no true well playing pacemakers they're wrong artificial pacemaker better today okay you know so there's some security issues that you may not want to ever see a fake but there's goblins in that area too with drugs and and food and you know they're talking about putting you're gonna get meat and it's gonna have a tracer in the meat and you swallow and you're gonna take it out the other way and that's that's what's gonna happen and say you

Channel: EEVblog

Views: 18,744

Rating: 4.9026546 out of 5

Keywords: eevblog, video, ieee 1588, lan sync, lan timing, trueform dds, competitors, ltz1000, voltage reference, testing, 8.5 digit multimeter, keysight, web ui interface, web interface, 3458a, automated testing, john kenny

Id: fpstmpm_rFM

Length: 32min 33sec (1953 seconds)

Published: Wed Oct 18 2017