Recumbent & velomobile FAQ

Door mux op donderdag 18 september 2014 09:51 - Reacties (12)
Categorie: Fietstechniek, Views: 1.263

I have recently gotten quite some questions - both from people on the streets and on the internet - about my bicycles. I've gotten a surprising amount of questions from foreign readers - that's why this blog is in English. I mostly use recumbent bicycles and my daily driver is an Alleweder velomobile. Here's all you need to know about (my) recumbent bikes

http://tweakers.net/ext/f/7JieJS9GmMaBGhtJIA6f6x2g/full.jpg

Q: Why recumbent?

A: A better question would be: why do people ride upright bicycles? What is the advantage of having more wind resistance, poorer vision and lower speed for the same effort? It just seems daft.

Of course, there are important reasons why only few people ride recumbents:
  • The first, and most widely publicized one, is because the UCI (international cycling union) at the start of the 20th century banned recumbent bicycles from all major sporting events. This meant recumbents got a bad rap (they 'cheated' by being better and/or they were 'dangerous' because they went so fast) and anyone aspiring to ever take part in a competition wouldn't be able to enter on a recumbent.
  • The second big reason is because most recumbents after the '70s haven't been built for practical use, but more for either long-distance traveling or sport. Like many upright sports bicycles, it's rare to find popular recumbents with practical, upright-bike-like cargo capabilities or even something as simple as a bike stand.
  • The last reason, which is simply a consequence of the niche-ness of recumbents, is that it's a foreign, weird looking device. People often feel embarrassed in their first hours or days on a recumbent. It's different, you have to get used to that and many people (think they) can't.
As for positive reasons why recumbents are a good idea:
  • They are much more comfortable to the butt, back and neck. Like, 100x better. I am not exaggerating. It's a positive delight to ride recumbents for hundreds of kilometers on end. The only thing holding you back is your legs and possibly your bladder.
  • Vision is much, much better. (see question about vision)
  • You go faster and/or require less effort to move and/or can lug along more stuff
Q: Why do you ride a velomobile?

A: For me, a combination of cost and environment. Most people of my age and beyond tend to have at least one car in their household. This is both expensive and, more importantly, I am pretty critical about the environmental and societal impact of cars. This is not the place to rant, but:
  • Cars sit around doing nothing >95% of the time. Not just that, they require prime real estate to do that. It's one of the most wasteful purchases in terms of utilization factor.
  • Cars have the lowest transportation mass efficiency of any vehicle; a car routinely weighs an order of magnitude more than its payload. Compare this to aircraft (about 0.7-2 times as much payload as vehicle weight) or bikes (5-10 times as much payload as vehicle weight) and it's clear how wasteful this kind of design is.
  • Cars have relatively poor energy efficiency. Internal combustion engine powered cars use about 300-700Wh/km to move around at an average speed of about 60km/h. Even the best electric cars around use much more than 100Wh/km. Bikes, even when powered with the grossly inefficient motor that is human muscles, weigh in at - at most - a couple Wh/km. Orders of magnitude better.
  • Cars require incredible amounts of infrastructure to work: hundreds of square kilometers of roads, buffer zones, sound walls, associated infrastructure. Bikes require much less infrastructure per traveller.
TL;DR: I really don't like how inefficient cars are, especially when compared to bikes. There are some nuances to this issue that I might go into in a later blog though (people make cars out to be evil environmental cataclysm-engines that ruin everything, but it's not that bad). Anyway, I looked around to see how I can delay having to buy a car as much as possible. I have my own business in-house which means I don't have to commute, and when I do need to move significant distances I figured a velomobile would be the obvious choice that is not a car.

Other reasons to own a velomobile are:
  • Because they are even more streamlined than recumbent bikes, wind doesn't influence you at all anymore. Your speed (and thus commute time) is extremely consistent throughout the year and wind conditions.
  • It's warm and dry inside. Can be a disadvantage in hot climates, but in the Netherlands it's usually appreciated
  • More place to store stuff.
  • You are better protected, which is a definite advantage considering the statistical dangers of riding a bike a lot at high speeds

Very useful as a trolley as well!

Q: Did you build it yourself?

A: The Alleweder was originally a kit velomobile, and as such it does look like something that is very amateur-built. It's got a lot of little dents, imperfections, crooked rivet placement etc. etc. It's not nearly as nice-looking as modern fiber reinforced plastic monocoque designs.

That being said: no, I didn't build it myself. I bought it from - as far as I can tell - the third owner, making me the fourth owner. The bike is probably either from 1996 or 1999, so at least a decade and a half old. For such an old bike with such a heavy usage pattern, I'd say it has held up very well.

Q: Is it hard to ride a recumbent bike?

A: Of course, riding a velomobile isn't hard at all. Getting in and out is a bit harder than other bikes (you do have to be able to lift yourself out with your arms, so this is not suitable for the elderly), but other than that there is no balancing. It's a three-wheeler. Easy as can be.

Two-wheeled recumbent bikes are a different story. Short wheel-base bikes like my M5 Blue Glide are the easiest to ride: they respond very much like any other bike. Most people can instantly ride on this bike without falling. Because of the smaller front wheel and heavier front steering assembly it does feel different ('twitchier'), but it doesn't take long to get used to that.

http://tweakers.net/ext/f/cuAqqO0TDfVUCPXkdkZV3xNI/full.jpg
My M5 Blue Glide, a short wheelbase recumbent bicycle

Low-riders and long wheel base bikes are a bit different. These respond really differently and it usually takes about a day of exercise before people feel confident enough to go on the road with these bikes.

Then there is my favourite recumbent; the Flevobike. This is a totally different beast. You don't steer with a traditional steering wheel, but with your legs. The frame bends in the middle. This leaves your hands free to... well, relax. Maybe brake once in a while (the steering wheel on a Flevobike is just there to hold the brakes and shifters). This bike takes ages to learn to ride on, and longer to master. It took me about 2 weeks to get even remotely comfortable riding it on public roads. The frame can be detached in the middle, making it a foldable bike that you can take with you in most public transport. It's maximum awesomeness, but so hard to ride.

http://tweakers.net/ext/f/fc12p1X5mI4ICs0LBx4Mv1nO/full.jpg
Me on my Flevobike when I just got it

Recumbents exist in so many shapes and sizes that it's hard to be exhaustive here, but these are the major types.

Q: Is the neck position comfortable?

A: Most people see people on recumbents and think the neck position is uncomfortable. Quite on the contrary; two things are at play here: 1) your neck muscles are made to keep your head upright no matter what the orientation of your torso is and 2) it's actually more relaxed to bend your head forwards a little (i.e., looking at the ground when you're on an upright bike) than to keep it in a straight line with your torso. So yes, it is comfortable.

This doesn't mean everybody is magically comfortable when they first try out recumbents. Some recumbents have very extreme positions that are just inherently uncomfortable of course, but this is usually only the case for track racing recumbents. More commonly, people that first try out recumbents try 'too hard' to keep their head in a certain position and their neck muscles getting tired as a result. If you just relax, it should be fine.

Q: Does it go fast?

A: Depends. Like upright bikes, there are fast and slow recumbents. For instance, the small-wheeled Flevobikes aren't really made for speed. The double suspension sucks a bit of power and I've purposefully outfitted my Flevobike with slow-but-indestructible tires and an internally geared hub to make it a very dependable backup bike.

All recumbents have a speed advantage over upright bikes, especially with headwind. This is just physics at work: less frontal surface area means less air drag. But none of my bikes easily break 35km/h, including my velomobile. I like the comfort more than the speed.

Some people do have the really fast bikes. A friend of mine cruises at about 45km/h in his Quest velomobile. Recently, a Milan velomobile was spotted in the Dutch province of Zeeland breaking 80km/h on a regional road - and maintaining that speed for quite a while.

Q: Why don't you have a little flag thingy on the back?

A: Because it's useless. Like helmets ;)

No, but really: the hypothetical idea behind a flag on a bike is to improve visibility of the bike. This fails because of two very important reasons:
  1. The flag has a vanishingly small visible surface area as compared to the rest of the bike, improving the visibility by almost nothing
  2. The flag has absolutely no surface area in the most important direction: as viewed from the front of the bike.
http://1.bp.blogspot.com/-_BgQFViMhmc/UfQkeJmgyvI/AAAAAAAAAJs/Z3tBHfuhooY/s1600/P1010037.JPG

Human vision is sensitive to a combination of contrast, surface area and movement. In order to make yourself seen, you must try your hardest to have contrasting colors to your surroundings, have the maximum amount of surface area and move with respect to the background. A flag doesn't help. And let's be honest, being a recumbent bike doesn't particularly help with frontal surface area. That's the whole idea behind the bike.

Add to that the fact that these flags are all but invisible in poor vision conditions and that I had a fucking accident just last friday in perfect vision conditions because some poor old lady just didn't look at all... And there are much more important things to do when it comes to traffic safety:
  • Get properly bright battery-powered (or capacitor-supported) bike lights. Always have them on in everything but the best of viewing conditions. Lights work MUCH better to stimulate our contrast-sensitive brain parts than reflectors or clothing.
  • Ride defensively. Give way to other traffic, don't run red lights.
  • Know your surroundings. Get some mirrors, they add SO much road awareness. Look ahead as far as possible and plan your actions in advance.
Of course, all the carefulness in the world won't save frequent cyclists from getting in accidents and I've had a few (most of them one-sided, e.g. slipping off my pedals and flying off my Flevobike head-first). Be aware that cycling is one of the most dangerous methods of transportation and that you have a responsibility to yourself and other traffic to be at the top of your game when you step into or onto a bike.

http://tweakers.net/ext/f/AD2zx0XMRVkP35NgyzIgg4oE/full.png
SWOV, 2011. Risk of death (left) or major injury (right) per passenger kilometer. From left to right: pedestrian, bike, moped, motorbike and car. For people that ride their bike a lot, bicycles are fairly unsafe as compared to cars

Q: How well can you see from so low down?

A: Better than on my upright bikes. As I mentioned before, your head has a tendency to tip down on an upright bike, making you very aware of the ground beneath you, but not necessarily of the road ahead. This is not a problem on a recumbent; you're just always looking ahead. The comfortable position also makes for better sideways vision, I find. But most importantly: because it's harder to turn your torso around to look behind you, almost all recumbents are fitted with bike mirrors.

Get bike mirrors now. They are awesome.

Once you get accustomed to using bicycle mirrors you can really appreciate what makes them great, and how handicapped your fellow upright cyclists are without them. You have so much more vision. You can see people about to overtake you from hundreds of meters away, avoiding dangerous situations. You will never be startled by anything again (well, except for cars hitting you in the side).

Besides the mirror and head position thing, vision is usually good. On the higher recumbents (like my M5) you tend to be exactly on the same level as car drivers, looking them in the eye. I find this to be a better position than on an upright bike, because human vision is constantly looking for things that look like faces, and my face is easy to find when it's right in the center of a car's vision.

Lower recumbents do suffer from reduced vision, especially in places where there are obstacles between you and other traffic that are just a bit too high to look over. In these cases you really do need to ride defensively and slowly.

In general though, infrastructure in the Netherlands is pretty good and this is a rare occurrence. Municipalities are required by law to cut bushes and other road separations to less than 45cm exactly to avoid this problem, and most roads are in compliance.

http://tweakers.net/ext/f/XR26hBiaBL69mud6zxylvBK7/full.jpg

Q: Is it expensive?

A: Like any other bike, you can make it as expensive as you'd like. Or as cheap as you like. I mostly have really cheap bikes. My M5 was about 250 euros when I bought it second-hand, my Flevobike was ¤45 but needed about ¤60 in repairs and my Alleweder velomobile was roughly 1000 euros including repairs - again, all second hand.

But there are plenty of recumbents that cost more than a car. One of the most technologically impressive two-wheelers is the Flevobike Greenmachine, which costs about ¤5000 new. Some velomobiles can set you back close to ¤10000 if you buy all the options.

That being said, especially the more expensive bikes tend to depreciate very little, if at all. For a while, second hand Quest velomobiles sold for 20-50% MORE than their list price, due to high demand and low supply. Even now that this has been fixed, the difference between new and second hand is less than taxes. Only the really old ones can be considered 'cheap'. A recumbent bike is a good investment and much cheaper than a car, even when the initial outlay is higher.

As for maintenance and parts: this depends. Original parts are, because of the small market, expensive. I just paid 40 euros for a single idling chainwheel. It's a really good one, I'm happy with it, but if this were the normal bike market this would have cost me less than ¤10. And be prepared to do your own maintenance, because most bike shops won't touch recumbents.

ASN Live: the missing bits

Door mux op dinsdag 16 september 2014 16:38 - Reacties (15)
Categorie: -, Views: 1.720

Vorige week was ik te gast bij ASN Live, een talkshow-achtig evenement georganiseerd door de ASN Bank. Ik had hier eerder al over geblogd en getwitterd, en ik heb van veel van mijn lezers begrepen dat ze ook hebben gekeken naar de livestream van dit evenement. Maarja - met effectief maar een paar minuten on-camera spreektijd heb ik niet echt alles kunnen vertellen wat ik wilde vertellen. En een hoop van jullie hebben het ook gemist. Hierbij dus een recap en aanvulling op een enerverende avond.



De ASN Bank is uiteraard een bank waar ik groot fan van ben; ik heb er immers de Wereldprijs gewonnen. Maar het winnen van een geldprijs is niet echt een goede reden om een bank aardig te vinden. ASN doet een hoop meer - excuse the pun - voor de wereld van morgen. ASN Live is één van die dingen. Als piepkleine bank met een vrij klein, weinig imposant kantoortje in Den Haag is het lastig om als klant of investeerder feeling te krijgen met de bank. Er zijn uiteraard bijeenkomsten waar je als belegger of investeerder kunt zijn en ieder jaar wordt een groot evenement georganiseerd - Morgen Vandaag - maar dat is het wel zo'n beetje. Met ASN Live gaat de bank verder het land in om wat PR te doen en een idee te geven van waar ze zich mee bezig houden.

ASN Live in Zwolle was hun tweede evenement van dit soort en was opgehakt in drie gesprekken: ééntje over een aantal duurzame bedrijven (o.a. Nederlandse Spoorwegen, Peeze Koffie), ééntje met buitenlandcorrespondent Olaf Koens over mensenrechten in Rusland e.o., en als laatste een gesprek met een ASN Battle- en Wereldprijswinnaar. Ik was uitgenodigd om te kletsen over de wereldprijs en mezelf.

Dus, wat had ik allemaal te zeggen?


Allereerst; de rest van de gesprekken waren best interessant. Om het maar heel cru te zeggen: voor een hippie-bijeenkomst doet ASN altijd goed hun best om redelijk concreet te blijven en interessante dingen te zeggen. Maar ik neem aan dat jullie dit blog lezen om te weten wat ik te zeggen had dus ik ga nu maar even over mezelf praten :P

Het gesprek begon met de vraag wat ik nou eigenlijk doe. Tja... da's niet echt een vraag die geschikt is voor een sub-5 minuten format :P Ik begon carrièretechnisch als techblogger, kreeg wat bekendheid, won wat prijzen en heb nu een bedrijf waarmee ik MADPSU aan de man probeer te krijgen in veel verschillende vormen.

Tussendoor wordt ook e.e.a. gevraagd aan de dame naast mij, iemand van een organisatie genaamd 'Iemand Thuis', die mensen werft om bij ouderen thuis te komen voor gezelligheid en kleine klusjes. Nogal een contrast met mijn onderwerp, en dit was op zich best een leuke opzet.

Hier staan mijn projecten opgeslagen ;)

Er werd gevraagd wat ik concreet nu doe. Ik trek één van mijn chineesbakjes uit de kast en laat mijn USB-sensor zien: een printplaatje dat heel nauwkeurig energie van USB-apparaten en/of mobiele telefoons meet. Dit om het verbruik van USB-apparaten en mobiele apps te kunnen optimaliseren. Ik probeer het verder uit te leggen maarja, dat past niet echt in de tijd ;)

Als laatste wordt er nog gevraagd over wat ik aan de Wereldprijs heb gehad, en zoals jullie waarschijnlijk wel weten is dat een hoop. Ik kom om in het werk, vooral na de publiciteit die de Wereldprijs heeft gegeven. 10/10 would win again.

Tot slot wordt er ook nog gevraagd over mijn ligfiets - een velomobiel waarin ik mij pleeg te verplaatsen. Ik vertel dat dit te maken heeft met mijn doel om zo lang mogelijk autoloos te blijven, en dat dit goed werkt. De velomobiel is een geschikt alternatief voor de auto.

Mijn alleweder onopvallend naast andere fietsen in een stalling

Wat had ik allemaal te vertellen, maar kon ik niet vertellen?


Deze blogpost kost nu al 30 minuten om te schrijven, en ik ben nog niet eens in de buurt van de helft van wat ik wilde vertellen. Eigenlijk is televisie/video een heel slecht medium voor het overbrengen van écht gedetailleerde informatie zoals ik eigenlijk wil. Maargoed, laten we eens beginnen bij de meest gestelde vraag ooit:

Hoe zit het nou met MADPSU? Wanneer kan ik er één kopen?

Goed, stapje terug. Wat is MADPSU? Wat wil ik hiermee bereiken? MADPSU is een apparaat dat je in je computer stopt en dat:
- De voeding vervangt
- Je verbruik (heel gedetailleerd) meet
- Aanpassingen kan maken aan de stroomvoorziening om verbruik te optimaliseren
- Data aggregeert en mij of een sysadmin in staat stelt om vergelijkingen te maken tussen heeeeel veel verschillende computer-onderdelen, en op basis daarvan advies kan doen welke onderdelen je moet kopen

Dit is een heel erg complex probleem en er zitten vooral aan de laatste twee punten aardig wat haken en ogen voordat dit werkt en ook echt nuttig is. Niet in het minst privacy-problemen en potentiële risico's bij het bouwen van een non-compliant computervoeding. Kort antwoord op de vraag waar MADPSU is, is dus: voorlopig nog nergens.

In plaats daarvan heb ik mij de afgelopen maanden tot jaar gericht op het oplossen van het probleem: hoe meet je nauwkeurig, snel en gedetailleerd data zodat je er iets nuttigs uit kunt halen? Resultaat daarvan zijn twee producten: de USB-sensor en PC-sensor. We zijn nog op zoek naar pakkendere namen ;)

http://tweakers.net/ext/f/NVB2kcd9z41I6kPIku7uxKOH/full.jpg

Even tussendoor: meetgereedschap is best zeldzaam

Voordat ik het daadwerkelijk over m'n nieuwe speeltjes ga hebben wil ik kort wat vertellen over hoe weinig meetgereedschap er eigenlijk is. MADPSU is geboren uit de observatie dat meetgegevens onmisbaar zijn voor energie-optimalisatie, en dat eigenlijk geen enkele computer zijn eigen verbruik fatsoenlijk kan meten. Een aantal laptops kunnen de coulomb counter in hun batterij meten, en het verbruik van CPUs kan doormiddel van RAPL of de Intel Energy Checker SDK worden uitgelezen, maar dit gaat op zijn best een paar keer per seconde en de fout in deze metingen is significant - iets waar ik het later dit jaar zeker nog eens over ga hebben (één van de studenten die ik zijdelings begeleid op de HvA is bezig met een rapport hierover).

Mijn nieuwe bedrijf, GreenInvents B.V., is dus ook opgezet rondom dit 'gat in de markt'. Wij maken apparatuur die betrouwbare, nauwkeurige en breed inzetbare meetapparatuur (en de software om dit effectief uit te lezen) maakt. En we willen deze apparatuur zodanig breed neerzetten dat op een gegeven moment reviewsites en bloggers onze USB-sensor en dergelijke apparaten allemaal hebben liggen en gebruiken om het verbruik van... van alles te kunnen meten! Nieuwe tablet op de markt? Goh, eens kijken wat het rustverbruik van dat ding is. Custom ROM erop: hey, het stroomverbruik is lager, hoe komt dat? Nieuwe draadloze muis van Logitech ter review aangeboden, eens kijken wat de ontvanger verbruikt om te schatten wat de impact hiervan is op de batterijtijd van je laptop. En zo verder.

Dit is informatie die momenteel gewoon weinig beschikbaar is. Als je enigszins serieus het verbruik van elektronica wil meten moet je al gauw aan de slag met multimeters en andere zaken die zelfs high-profile websites zoals Anandtech pas sinds heel erg kort beschikbaar hebben - en op zijn best gebruiken voor hele grove 'gemiddelde' metingen. Anyway:

USB-sensor

USB-sensor rev.3

Laten we het eens hebben over de USB-sensor - gezien deze op korte termijn al een pilot gaat draaien bij de HvA: Dit is een apparaat dat het verbruik van USB-apparaten kan meten. Wat is daar zo speciaal aan?

Nou, het is nog best moeilijk om écht netjes te meten, zeker in het geval van USB-apparatuur. Sommige USB-apparaten - bijvoorbeelde de Logitech Unifying-ontvanger - trekken maar enkele tientallen µA, terwijl andere apparaten op een charging-poort bijna 2A kunnen trekken. Om een goede resolutie te behouden over een breed meetbereik zitten er twee daadwerkelijke 'sensoren' (shuntweerstand+meetversterker) op de printplaat: één voor grote stromen en één voor lage stroomsterktes.

Dat niet alleen; ik heb ook mijn best gedaan om het ontwerp en de firmware zo in elkaar te zetten dat onnauwkeurigheden in de meting zo goed mogelijk weggekalibreerd kunnen worden, zodat de sensor sub-0.1% absolute nauwkeurigheid heeft na kalibratie.

De belangrijkste feature, en hetgene dat veruit het moeilijkste te implementeren was, is de hoge meetfrequentie. De sensor meet met 125kS/s, met minder dan 1/10e van een sample jitter. Dit samen met de frontend zorgt voor een meetbandbreedte van ca. 20kHz analoog. En dit is op zijn beurt weer snel genoeg om de karakteristieke stroomverbruik-pulsen bij USB-interrupts (ca. 1kHz) te zien, alsmede grappige dingen zoals de backlight PWM-frequentie van mobiele telefoons (doorgaans ca. 2kHz). Om dit soort dingen te bekijken zit in de bijgeleverde software ook bijvoorbeeld een FFT-functie.

FFT van de USB-sensor zonder aangesloten zooi - zie ook de sampling artefacts dichtbij Fs

En wat doe je hier dan mee? Er zijn twee belangrijke redenen waarom ik zoveel moeite heb gestoken in supernauwkeurig, stabiel en snel meten:
  1. Dit is een moeilijk probleem dat ik zo snel mogelijk voor elkaar moet krijgen, want deze firmware en hardware gaat nog in vele toekomstige producten gebruikt worden. Een slechte start betekent dat ik in de toekomst een hoop dubbel werk moet doen.
  2. De eerste paar batches zijn bedoeld voor de educatieve markt, waarbij de beoogde markt alles van middelbare scholen tot universiteiten is. Met name HBO's en universiteiten willen gegarandeerde performance.
Het belangrijkste doel van de USB-sensor is het meten van het verbruik van apps en mobiele apparaten in het algemeen. Gezien het laadcircuit soms enigszins in de weg zit om echt nauwkeurig te meten kan hiervoor ook een batterij in de witte connector worden gestoken en het verbruik door de batterijpolen worden gemeten. Anders meet je vooral hoe hard de telefoon de batterij probeert op te laden.

Daarnaast kan het door ontwikkelaars van usb-apparatuur worden gebruikt voor de analyse van verbruik van hun product.

De PC-sensor

MADPSU en andere producten

Dit is in feite MADPSU zonder dc/dc converters - je hebt hier dus nog een gewone voeding bij nodig. Ik zie de PC-sensor als opstapje naar een echte MADPSU, mede omdat MADPSU zelf nog op het moment té hoge test- en certificatiekosten heeft. Voorlopig dus alleen een sensor - ergens in 2015 volgt de volledige voeding.

De PC-sensor meet evenals de USB-sensor op hoge snelheid (iets lager dit keer - 20kS/s) het verbruik van alle stroomlijnen op het moederbord, de SATA-poort en de CPU. Hier kun je vervolgens mooie grafiekjes, FFTs, heat maps en andere karakteristieken uit halen om het verbruik van je computer zowel op korte als lange termijn te analyseren.

Het belangrijkste wat me is opgevallen tijdens de ontwikkeling van de PC-sensor is hoe ontzettend groot de impact van software op het verbruik is. De kleinste veranderingen in settings of simpelweg het gebruiken van een computer (zonder nieuwe dingen te installeren) kan tientallen procenten verbruik schelen. Om ontelbaar verschillende redenen. Dat is iets wat simpelweg onzichtbaar is voor de meeste gebruikers en ontwikkelaars, maar met behulp van een sensor die continu data verzamelt is het opeens zichtbaar en kan er wat aan gedaan worden.

Naast computers kan de PC-sensor - die bestaat uit een interfacebord en het daadwerkelijke sensorbord - ook worden uitgebreid met dingen als een PCIe-sensor (waarmee ik in de afgelopen maanden data heb verzameld over mijn nVidia Geforce GTX 750Ti) en andere toepassingen. Dit is allemaal op het moment nog in prototype-fase, dus ik zal hier niet teveel over vertellen voor het geval ik me bedenk in de komende maanden :P

Andere onderwerpen: waterstofauto's

Iets anders wat tijdens de vragenronde ná het evenement ter sprake kwam was waterstofauto's. Ik heb dit al eens laten vallen, maar ik heb tussen 2007 en 2009 elektronica ontwikkeld voor (en algemeen meegeholpen met) Formula Zero Team Delft, een studententeam dat een waterstof-aangedreven elektrische racekart maakte voor de fictieve 'Formula Zero'-competitie. Deze raceklasse is er nooit echt helemaal gekomen, maar dit project gaat wel nog steeds door onder Formula Student met ook dit jaar weer een afgevaardigd team van de TU Delft*.

De Forze I

Tijdens ASN Live vroeg iemand waarom ASN Bank niet meer investeert in de waterstofeconomie. Gezien ik vanuit FZ wel het e.e.a. weet van dit onderwerp en elektrische auto's een grote interesse van mij zijn had ik daar wel wat over te zeggen. Kort antwoord: waterstofauto's gaan nooit werken, the math just doesn't add up. Lang antwoord: ik heb een gigantisch blog liggen waar ik in detail uitleg waaróm waterstof een slecht idee is voor mobiliteit. Ik zal m'n best doen om die binnenkort eens op te poetsen en een eerste deel van te posten.

*zodra iedereen, waarmee ik vooral mezelf bedoel, tijd heeft ga ik samen met Infant en DaWan, twee andere ex-Forze Delft-leden, nog eens een blog bouwen over alle coole dingen die we in die tijd hebben gedaan

Zo, da's wel weer genoeg voor deze post


ASN Live was erg leuk en ik kan niet anders dan zeggen dat ik het erg naar mijn zin had. Leuke avond, leuke onderwerpen en een goede reactie van publiek IRL en op de stream. Bedankt aan iedereen die heeft gekeken!

De komende maanden worden ook erg spannend voor mij; de USB- en PC-sensor gaan 'in het echt' gebruikt worden in in ieder geval twee pilotprojecten aan de HvA, en hopelijk kan ik ook nog meer mensen, bedrijven of instellingen vinden die willen meehelpen met het helemaal productieklaar krijgen van zowel hard- als software. Als de hard/software voor de USB-sensor in de loop van oktober het goed lijkt te doen ben ik ook van plan om een dozijntje rond te sturen naar... nouja, iedereen die denkt er iets aan te hebben en bereid is om feedback te geven. Met de PC-sensor en de volledige MADPSU moeten we nog wachten tot 2015 voordat hij de wijde wereld in mag.

mux is morgen te gast bij ASN Live!

Door mux op maandag 08 september 2014 20:25 - Reacties (17)
Categorie: -, Views: 3.006

Ooit in een duister verleden heb ik eens de ASN Wereldprijs gewonnen met MADPSU. Dit was, zoals jullie wellicht hebben gemerkt aan het opvallende gebrek aan blogs, een moment waarop ik veel bekendheid en vooral heel veel werk kreeg. Benieuwd naar hoe het nu met me gaat?

Morgenavond vanaf 19:00 organiseert de ASN Bank een talkshow over duurzame ontwikkeling genaamd 'ASN Live'. Hiervoor zijn verschillende mensen uitgenodigd die zich in de afgelopen tijd bezig hebben gehouden met duurzame ontwikkeling in de breedste zin - van sociaal-ethische kwesties tot duurzame energie. Ik ben hier ook uitgenodigd om te spreken over mijn ervaringen met de Wereldprijs - een prijs die in het leven is geroepen om kleinschalige duurzame ontwikkeling te bevorderen - en over alles wat hieruit is voortgekomen.

Ik zal hier ook e.e.a. vertellen over mijn nieuwe bedrijf GreenInvents B.V. en de eerste producten die wij gaan uitbrengen op het gebied van energiemeting en -besparing in computers. Als alle techniek dit toelaat wil ik hier ook een kleine demonstratie van geven. Het gesprek waar ik in deelneem is van 21:00-21:15.

Het evenement wordt gelivestreamd op morgen-vandaag.nl/livestream. Ik zou zeggen: komt allen!

Car culture in the U.S. and Europe

Door mux op maandag 28 juli 2014 12:00 - Reacties (24)
Categorie: -, Views: 4.277

A video published yesterday of a car plowing through the San Diego Zombie Walk at ComicCon caused immense controversy on reddit and other social media sites. I'd like to take this opportunity to point out that: this accident is a great example of American car culture. And of European car culture. Here we go.



Here's what happened. During the past weekend, San Diego was flooded with all the weird and wonderful people that like comics, animated series, cosplaying and such fun things. Central San Diego was pretty much on lockdown; roads near the event grounds were closed to accomodate the huge daily in- and efflux of convention attendees and most other roads were also pretty much clogged up with pedestrians. To add to the traffic, as per tradition every year a Zombie Walk was organized. People in zombie outfits walked (or Rascalled) a couple blocks through the city center to show their undeadness to the world.

Then came a deaf driver who probably hadn't seen this happen at all in his life. He needed to be somewhere, but in his way was this Zombie Walk. So first he stopped about a foot in front of the crosswalk, honked his horn to signify that he wanted to go through - but with this kind of mass of people that kind of fell on deaf ears. Punny.

So he inched his way through the mass of zombies, and people started to be a bit aggravated by his shenanigans. They started touching his car and eventually sitting on the bonnet. That's when the driver saw flashes of the coming zombie jesus apocalypse, hit the gas and ran over a couple of people to get to where he had to be. Understandable, right?

The controversy

To me, as a Dutch citizen and part of our culture, it's absolutely unimaginable to even start thinking that any aspect of what this car driver did is acceptable. Around here, it's engrained in law and culture that you never endanger weaker traffic participants. A great deal of our infrastructure is based on this implicit assumption that nobody in their right mind would even start running over a pedestrian or cyclist, even if that person really hates zombies and the pedestrian is dressed up like a zombie. Here's an example of such a construction for cyclists:

http://users.skynet.be/Stam/lefevere/Blog/FietsOpstelvak.jpg
(funnily enough; this is actually a picture from Belgium) Cyclists at an intersection get positioned in front of all motorized traffic so that they get priority on the intersection and motorized traffic cannot intercept or surprise them in any way

If you ask anyone here, that driver at the Zombie Walk is 100% guilty of attempted murder with a deadly weapon or at least the highest degree of hit and run. He clearly had lots of room to back up and wait for the crowd to disperse and he clearly didn't have any reasonable opportunity to drive forward. The angry reaction of the pedestrians and their touching of the car was possibly threatening and not very nice, but expectable for somebody honking their horn and forcing their way through a crowd of squishy bags of meat with a 2-ton motorized lump of metal.

But then I look on reddit and see... Lots of people defending the driver! Comments range from (quotes from Reddit):
Maybe im fucked up, but this was justice for me. Just fuckin let the guy pass, it won't take up more than 10 seconds of your time. Then that fuck sitting on the front of the car ughhh. Fuck these people.
to
You don't drive your car into people. How is there any debate here?
Controversy all around, with almost a down-the-middle split of people who either (partially) condone or dismiss the driver's actions.

Why? It's a culture thing

This is a great example in my opinion of the difference between American and European car culture. First of all, let me be clear that this isn't some black and white thing. This isn't a wrong and right thing. It's just a big difference in how we view traffic.

I live in the Netherlands, we have an exceptionally pervasive pedestrian and cyclist culture and everybody who drives a car is probably also a cyclist and/or pedestrian at least a couple times a week. Everybody understands each other's movement patterns and grasps how we need to treat each other on the road. Not all of Europe has such a large cyclist and pedestrian movement necessarily. And we move a lot for being such a small country; we have a great big car culture as well. The Dutch travel on average about 15 000 km per year in a car - depending on which US state you live in that is actually more than American car use. And we own more cars per household than the U.S. And we have waaaay more car infrastructure per square kilometer. We also drive faster on average. The stereotype of Americans doing everything in their car and being absolutely car-mad isn't actually true; we trump you in most respects.

But what makes American car culture so unique, is that it is absolutely car-centric. A pedestrian or cyclist is viewed as fundamentally less of a traffic participant. They should always make way for cars. Infrastructure is made with cars and only cars in mind; with pedestrian crossings used sparingly and with cyclists being forced into the fringes almost everywhere. And this is really what the comments on Reddit and other social media sites exemplify; a completely different way of thinking about traffic.

And this really stems from the fact that Americans are often almost exclusively car drivers and nothing else. Bikes are used by enviro-nutters (and cycling is very dangerous), walking is for going next door and nothing else. There's much less of an opportunity for empathy if you don't regularly experience the mechanics of other forms of transportation.

Discussion

I'd like to conclude with the remark that this isn't all I necessarily have to say on the matter. This is just one aspect of this incident that I thought was interesting to highlight. The rest of the incident is just another news story; mildly interesting now, but will be readily forgotten once the next thing comes along.

There's still one more issue, and that is driver ed(ucation). It's well known that the U.S. standards for driver's exams are pretty low, especially compared to the Netherlands or even more extremely Denmark or Finland. Some comments highlight this, saying that the driver maybe isn't aware that pedestrians always have the right of way even if they are jaywalking (which is also true around here - I always like to say that all of the traffic infrastructure is a legal place for pedestrians). Also, the driver was honking his horn, which should only ever be done to signify immediate and significant danger. Clearly, the driver was not in danger. It's also very strange that he would think (presumably) that it would be faster to drive through the crowd than to let it pass and THEN drive on, because after the incident he had to talk to police and wait around for many hours before he got to go. The Zombie Walk only took about 15 minutes to pass.

But again, this isn't particularly interesting to me. I just wanted to talk about culture and how it fundamentally shapes our thoughts. It makes the difference between having the opinion that something should be charged as manslaughter with a big jail sentence or a civil offense with a small fine.

ASRock Q87TM-ITX and B85TM-ITX 'review'

Door mux op vrijdag 11 juli 2014 19:46 - Reacties (9)
Categorie: -, Views: 9.355

Now this is thoroughly interesting. I have a motherboard that has been announced but isn't in stores yet, with the UEFI of a motherboard that hasn't even been announced yet. I have no idea why the dear Hardware Gods have given me this crystal ball view into the future, but I am here to share all the details with you.

So this is something totally different for once. I usually do electronic design engineering; I design things like VRMs and consumer electronics. I have even dabbled in computer hardware design since my previous blogs were so successful. But I never got early hardware - from anyone, really. Even hardware OEMs that ask me to do engineering for them never give me any preproduction hardware; they are extremely protective of everything they do. I realize that I am probably going to be phoned or e-mailed by ASRock today, because I am going to spoil their marketing somewhat. I have something they haven't announced yet. But who in their right mind wouldn't jump on this opportunity?

Yes, this is my desk.

How this happened



Earlier this week one of my earlier builds - Fikki3 - started acting up. Well, it has been acting up for a while; it's now a pretty old Core i3 530 and it is struggling with modern video codecs (Youtube/Twitch playback) and has essentially zero graphical power. It's aging and needs to be replaced. Having much less time nowadays to do mods, I decided to make a stock system based on the ASRock B85M-ITX which I have tested about four months ago to be one of the most power efficient motherboards out there.

I was to combine this motherboard with a Haswell Pentium and Maxwell (Geforce 750/750Ti) graphics card to create a sub-15W (idle) gaming machine with muscle to spare.

So I looked around in the second hand market to see if I could score one of these boards for cheap; I succeeded. I found an advertisement with a B85M-ITX for ¤30+postage. That's a steal for such a nice board.

When I got the board in today I noticed the package was remarkably thin. Part of this was because it was quite badly packaged, but it was mostly down to the board not actually being a B85M-ITX. It was a B85TM-ITX. Note the extra 'T' in the name.

The TM-ITX is for Thin Mini-ITX, a motherboard specification for small, extra thin motherboards that fit inside one of those all-in-one chassis. I have a long-standing love for these boards. I was very excited.

But wait, I discussed about this board around about the writing of my Best Buy Guide (BBG) 4 months ago, and back then it was only just announced at CeBIT. Usually these boards don't come out until at least half a year later; it took the B85 non-thin ITX board more than half a year to arrive from the moment it was announced. So I did a quick Google search and yes, it isn't actually available yet. This motherboard does not exist. So what is it like?

Around the B85TM-ITX



B85TM-ITX Top photo

B85TM-ITX Bottom photo

I'm sorry for the slightly crooked photos. This is the board. It's very interesting. Not because of the particular components on it, or anything else immediately obvious. No, there's something more subtle going on. Look at the text on the board that says 'B85TM-ITX', just under the SODIMM slots. Doesn't that look.. off to you? The color is different. Is that a sticker?

It's a tarp! I mean sticker11!

Why yes it is. This is actually an H81TM-ITX PCB. And now that I look at them side by side - that doesn't seem far off. The H81TM-ITX is basically identical; the only differences are that:
  • The DC/DC converter under the mPCIe slot is populated
  • The USB 3.0 internal header is populated
  • Slightly different MOSFET component choice (although this can just as well be a board revision change
I've had a play around with the H81TM-ITX before, and it was a pretty nice board, although not at all power efficient. I'm amazed they use one of their least power efficient designs for the thin ITX market. In my tests the H81TM-ITX in idle used a bit over 15W DC (or about 19W AC with a decent power adapter) and the B85TM-ITX is absolutely no different. It's a power hog as far as low-power boards go. To give you some insight into the actual state of art: a decent mini-ITX Haswell board shouldn't go over 6 or 7W DC nowadays - or about 10-12W AC with a PicoPSU and decent power adapter. Needless to say; this is just accounting for the bare minimum setup: the board, a processor+HSF, some memory and an SSD.

Anyway, looking around the board I can't really spot any differences with the basic H81TM-ITX. It still sports the same ALC892 audio codec for sound, with it seems slightly lower quality coupling caps, although again this can be a revision change:

Realtek ALC892 audio codec

The network connection has an Intel i218-LM Gigabit Ethernet controller/PHY instead of the i217V on the H81TM-ITX. Wait, why? Why an LM chipset? Those are literally only used if you want vPro, otherwise the -V chips are way cheaper... Hm, this is a weird board:

http://tweakers.net/ext/f/KwDJg39MpBrKTT9GQir2Vhy9/full.jpg

HDMI/DVI is driven by the ASMedia ASM1442, which is the first chip I saw with a datecode: 1332, or the 32nd week of 2013. This indeed confirms that this is not a new design and they did just put a sticker over the H81TM-ITX and called it a day:

http://tweakers.net/ext/f/trIIOHlZ1CiwPmCxp9xuL8bK/full.jpg

And the 'Super' I/O - jeez, these things are almost completely redundant now - is filled in by our good friend Nuvoton with their NTC5573D:

http://tweakers.net/ext/f/zXgae5og0QvrlakMpl6rjb63/full.jpg

The CPU VRM is a Richtek RT8889A, which is a budget VR12.2 3-phase controller - kind of surprising for a low power board, but this is probably also why it's not actually power efficient at all:

http://tweakers.net/ext/f/de3LQmpztcIkrFonwgypf0E9/full.jpg

On the bottom side, there's really not that much to see. Just a Chronotel LVDS driver (and an ST RS232 driver):

http://tweakers.net/ext/f/nqzYZFyGm4ZzwBHCClS2gnHu/full.jpg

And a Realtek ALC109 headphone amp:

http://tweakers.net/ext/f/CH4Oxo3o6dbDSlwjsJp5Y95u/full.jpg

So... this is kind of a letdown. It's just a H81TM-ITX with extremely minor changes. Well, I was a bit disappointed until I took off the heatsink from the chipset and...

Around the B85TM-ITXQ87TM-ITX



That's a Q87 if you ask me

So... ? Well, if you look up the S-spec on this chip - SR173 - you will find that this is a Q87 chipset. On a board with a B85 sticker over an H81 silkscreen. This just keeps getting juicier. Because whereas the B85TM-ITX was already announced, the Q87TM-ITX hasn't yet.

And now things start to fall in place. That i218-LM vPro ethernet controller doesn't belong on the B85 chipset. B85 doesn't support vPro, but Q87 does. This is actually a Q87TM-ITX board. Or maybe some weird Frankenstein board where they took a production H81TM-ITX and slapped on a Q87 chipset and i218LM - I don't know.

But what really stands out to me is the fact that the date codes on this board are all roughly the summer of 2013. Almost a year old. Why would they wait a whole year, wasting most of Haswell's release cycle, before releasing a board that apparently is so incredibly easy to make? So many questions!

Anyway, conclusions are for later. Now let's fire the board up and look at the BIOS. Yep, there it is. This is a Q87TM-ITX:

That's very beta. Version 0.0.1?

With UEFI version 0.0.1L, that's very beta. Interestingly though, it works pretty well. Like the H81TM-ITX, the board has fucked up ACPI tables that cause the CPU never to enter package C3/C6/C7, but this is a fairly widespread phenomenon with many motherboard vendors. I've had the same issues with a Biostar Hi-Fi B85N 3D; a friend of mine even tried to report the issue with them but this never really led to a BIOS update. Anyway, I'm rambling.

The only other slightly interesting thing I came across was a mention of wireless functionality; I assume this has to do with the miniPCIe slot and some UEFI-level support for wireless network functionality. No other remarkable things to be honest.

Conclusion



I must say, I'm confused and intrigued. This is most of the reason why I am writing this blog post. I actually e-mailed the second hand seller of this motherboard to ask where it is from. He/she replied and said to have gotten it at an inventory auction following ASRock Netherlands B.V.'s pending move from Wijchen to... somewhere else. By the way, ASRock's Dutch office is moving (not sure if this is news, but I can't find it anywhere else).

One of the possibilities is that this board has been sent out a bit early to system integrators, much like the DH61AG was available more than two months in advance for system builders. That way, feedback on the particulars of these boards (especially driver support and electrical interfaces to the LVDS connector, as well as other power-related things) can be figured out before launch.

The board design itself shouldn't really be a surprise. All motherboard manufacturers basically make one PCB and then selectively populate it to create different skews. I am a bit surprised, though, that apparently they don't utilize any Q87 feature on this board other than vPro. No extra SATA ports, no TPM, no second gigabit. Kind of a missed opportunity.

This does, however, enable us to do some educated guesses on pricing. H81TM-ITX differs by the chipset (H81: $26, Q87: $47), ethernet (i217-v: $1.72, i218-LM: $1.92) , a USB 3.0 header (<$0.10) and a licensing cost for vPro. With the street price of H81TM-ITX at ¤65 currently or ¤54 without tax, and with about 40% gross margin on these kinds of products, I would expect the Q87TM-ITX to run about (54 + ((47-26)+(1.92-1.72)+0.1)x1.4->EUR) + tax = ¤92. Possibly a bit more because of the relatively low volume (and subsequent higher margins) on this skew and additional licensing costs. Awesome. I'd love a true DQ77KB-successor!