Simulation of wind gusts for crosswind stability

While reading on human-powered vehicles (HPVs), one of recurring themes seems that all HPVs tend to be unstable in crosswinds. To me, this seems to be the main limitation of HPVs and lightweight EVs: they all sound nice in paradise climates like the Bay Area, but in more miserable climates only the most dedicated ones ride HPVs all year round. If I ever attempt to build an HPV, it could be an interesting quest to build something more weather-worthy than the state of the art. I would be willing to sacrifice 25% of the top speed for a factor of 2 improvement in handling the side wind (not sure how to quantify the handling in side wind though).

Just found this 2015 Ph D thesis X Zhang; the intro reads very well. The intro opens with a very promising statement: "Current developments in vehicle engineering have been showing a trend of faster, more energy efficient and more comfortable vehicles with a high capacity of passenger transportation. These trends require that the vehicle has a much better handling stability. However, the demands for light-weight constructions and higher driving velocities are in conflict with crosswind stability."
The Ph D adviser was this Prof:

While reading the Intro, I came across a sentence that indicated that nonstationary wind conditions are frequently encountered upon exiting tunnles. I think, this fact can be used to simulate wind gusts with simple vehicle models, perhaps, even motor-free. Here in the UK one could go to Wales in shitty weather (which is 1/2 of the year, and 3/4 of the time in the winter), find a downill-pointing tunnel on a non-biusy road, and test the handling of the model at the tunnel exit. Repeat as many times as needed by towing the model back up. Not sure how one quantifies 'handling' thoughs. The model would need to have good breaks and steetring.

DIY carbon disk wheels

If one wanted to make composite wheels as a DIY project, I imagine, it makes sense to get deepest Alu rims available, to increase the carbon-Alu gluing surface. It probably also makese sense to test the self-built wheels using static load first, then on an existing bike. Sounds like a good project to get one's feet wet in CF building.

Here is a good discussion on DIY carbon disks:

The thread has a link to an excellent slide show of a full carbon recumbent build by Sanderson. Link live as of July 2017:

Source of affordable Titanium parts

If one had to build an HPV on a budget, yes make it lightweight and strong, using off-the-shelf aircraft grade materials seems reasonable. I've previously used parts from Aircraft Spruce in my DUY projects.

If one had to make a double A-frame type suspension, here is a manufacturer of Ti bearings and rod ends. The idea would be to combine them with Carbon fiber tubes, or with Ti tubes procured from suprplus stores like Titaniium Joe.

Purifying proteins in a walk-in -20 freezer room

The lab has been trying to express a particular type of fusion protein in bacteria, and this series somehow tended not to make it to the end of the first Ni-NTA elution in one piece. They seem to like to fall apart into two-three pieces. We did everything right otherwise (protease inhibitors, 4C, fast protocols, etc). I am not a protein prep guru, but this time the degradation seemed excessive.

So I figured... what the hell, let's try to do the entire initial prep, including NiNTA elution, in a -20 room. We have one at the Institute, and it's never used for anything exciting anyway.

Not sure if it worked yet, but the experience was interesting. Not sure if what I did was right, but here are my observations so far.

I started out by preparing 1 gallon of 50% glycerol in a styrofoam container. I placed a metal eppendorf holder in it and pre-chilled it in the -20 room for a day. This was supposed to be a -20 bath, in case I decided to bail from the -20 room into a 4C room nearby. In hindsight, this turned out to be a very good idea, except for the 50% glycerol bath. Not sure why I decided to keep it liquid, a block of ice precooled to -20 would have done the job just fine. I also could have combined this with dry ice on the bottom, but I did not want my solutions to freeze.

Placed a spare beaker of 50% glycerol, about 2L, next to the main bath. That turned out to be useful for cooling the sonicator tip -- again, in hindisght.

Placed a centrifuge, a shaker, and a sonicator into the -20 room. Checked them next morning, everything seems to switch on. Verified the temperature of the -20 bath, it was indeed -20. Kept the sonicator controller outside and ran the cable inside (did not want condensation to ruin the main board of the controller).

Prepared a lysis buffer based on 50% glycerol, pre-chiled it to -20, it remained liquid. Prechilled about 20 eppendorves in the -20 bath. There was no lysozyme, as this is -20, so I figured it was useless anyway.

Washed NiNTA agarose beads with the buffer in batch mode. Spinning worked fine, agarose beads settled fine at 1000 g. Found that, when at -20, the NiNTA beads are not keen to be resuspended by flicking, only pipetting up and down with the 1 ml tip worked.

 I got myself a down parka with a hood and a hat and a pair of ski gloves. Good luck opening eppendorved in thick gloves. This is the first time I wished I had one of those eppendorf-opening tools.

I had about 500 ul of 4 cell pellets in 4x50 ml falcons. I thawed the pellets to 4C (as they were still in PBS, that would never resuspend at -20), then mixed with 1 ml of the -20 lysis buffer, placed it all in the -20 bath. In this case, the fact that the glycerol remained liquid, helped. If it had been a block of ice I would not be able to dip the falcons with cells in it.

The next step was done in the -20 room, where I brough a 1ml pipet and a box of 1 ml tips as well. It started looking like a small polar research lab.

Atttempted to resuspend the cells in 1 ml of lysis buffer by pipetting up and down. That did not go well (expected) -- everything was viscous and snotty as hell. Blasting it with 6W of ultrasound helped: after this  I managed to transfer most of the stuff into fresh eppendorves, it was till snotty from the genomic DNA. In hindsight, 1 ml of 50% glycerol-based lysis buffer was too little, too viscous, and unmanageable. 2 ml would have made it better, and I could probably still fit everything into 2 ml large eppendorves.

I blasted each eppendorf with the lysates 4 times more with30" and 6W of ultrasound, to the point they got non-snotty. by that time, my fingers were numb, and I was not enjoying myself. Holding eppendorves with gloves was mostly OK, but opening them not so much, so I kept one hand bare, and that almost resulted in frostbite.

Interestingly, the sonicator did not mind being continously on at 6W at -20, I guess the temperature sensor does not go off. So I just kept it on (the controller was oustide) and held eppendoves one by one, with the tip inside. Gradually, the solutions cleared, so I could see the tip inside. Cooled the tip by dipping in 50% glycerol bucket between eppendroves -- that came quite handy.

After 15 min of this I was getting concerned about my fingers freezing, so I placed the 4 tubes in the centrifuge. Adjusting the time to 15 min was fun: my fingers kept freezing to the buttons. I managed to start it, and then got the hell out of there to warm my hands. But the most annoying step was done. Not sure if it stayed at -20 inside from the heat of spinning, should have tested with a tube with water (if ice remains ice, it was -20).

In hindight, 6 blasts would have been better, or using 10W. I am saying it, b.c. after spinning I still got ~50% of unlyzed cells.

It took two spins to clear the solution. I transferred the -20 bath into the 4C room. It went up to -15 in about 20 min, not bad. Mixed about 500 ul of lysates with 100 ul of washed Ni beads and left on the shaker (rocker, to be exact) overnight. Transferred the -20 bath back into the -20 room.

Will see what I've got next day. It was not too bad, after all. Having one solid -20 ice bath, one liquid 50% glycerol -20 bath, and one beaker of clean 50% glycerol for sonicator wash would have been ideal. Also, next time, if this is ever done again, I should use 2 ml of the lysis buffer for this: the lysates get extremely gooey and non-manageable. Not sure if this can be done in larger volumes, unless one splits 20 ml of lysate into 10 eppendorves (larger centrifuges do not get to -20).

Here is a picture of the -20 bath in the 4C room (the latter felt nice and warm compared to the -20). The lowest reading on the thermometer is -10,  and it's another 8C below that. The block holes for tubes are shallow, but they are eppendorf-shaped (I have a custom drill bit for that). Since my NiNTA beads were only 100 ul in volume, they stayed nice and chilly.

The level of glycerol went down with cooling... not sure if it's the shrinkage, or if there is a leak in the goddamn styrofoam plastic bucket. Actually, therre are two containers: the plastic styrofoam is inside of a bigger styrofoam one in which I filled the gaps with packaging peanuts and sealed with Alu tape on top.

Update 2 days later: washing, elution etc worked great. Washed 100 ul of beads with 1 ml of 50% glycerol-based buffer, 4 times, eluted with 150 ul of 500 mM Imidazole, got about 1 mg/ml of all proteins, in 150 ul, from about 50 ul of culture. Clean, no signs of degradation at all.

Will try to repeat with more challenging proteins. 

Building large molds for velomobiles on a budget

While researching methods on velomobile building, it appeared to me that the most epic stage is building the shell. If done properly with composites, one needs to sculpture a mold. For a one-off project, it's a huge undertaking that takes a lot space and produces a mess. So one needs a garage and a angel spouse, too. On top of that, once you are done (if ever) you throw away the mold (unless you want to open a velomobile business, and why would you want that?) The advantage of building a 'proper' mold, of course, is that you can make any shape you want.

So I was thinking... if one limits oneself to the simplest shapes, like a blimp, one could get an inflatable baloon from a party store. Inflate it, hang it, coat it whith wipe-on resin, let it set for a days (enthiasts are not in a hurry, since htey are not paid for the hobby, rememner). Then use the casted resin as (now stable) mould. Or, even use it as the first layer of the would-be shell. E.g. epoxy with CF shavings may do as the first layer.

Here is a video where this has been done... not for a velomobile shell, but for a comparable shape/size. Not sure how much this "StyroSpray 1000" costs, he seems to be using gallons of it. I am sure there are relatively cheap resins that can do the same job.

Costing on StyroSpray from about 10 years ago, found in this forum:


Here is the info they sent me in March of last year, so I don't know what changes to pricing may have occurred since then.

Quote Originally Posted by StyroSpray
Thank you for inquiring about our products,

I will mail you a free information package which contains
a coated sample of EPS foam for your evaluation.

Pricing for the StyroSpray 1000 is $55.05 per gallon.
It is sold in 2 gallon kits, $110.10 for the 2 gallons.
We also offer a smaller 2 quart trial kit for $29.00.
The price for the optional StyroPrime foam sealer
is $26.97 per gallon it is also available by the quart
for $8.25. Coverage for the sealer is 150 square feet per gallon.

StyroSpray 1000 does not contain water or solvent, so your coverage will
be much greater than other coatings you are using.

A gallon of water based paint is usually about 40% water.
A gallon of house paint which costs $35.00 actually
costs $49.00 when you factor in the lost water which evaporates away.

With StyroSpray nothing evaporates away. We do recommend a heavy
coverage of 50 to 60 square feet per gallon for outdoor applications.
StyroSpray is nonhazardous and nonflammable for shipping purposes, it contains no VOCs.
To place an order please call us at 800-766-3832 / 713-943-8451, we accept all major credit
cards. We are the direct manufacturer of this product and we ship direct to your door via UPS.

click here->
to download the StyroSpray instruction guide.
I will also mail you a printed copy of the instruction guide with the sample.

Can you make a velomobile for under 2K?

Contacted Begorett about the projected costs of their vehicle.

They intend to offer it at an initial price higher than the Quest by Velomobiel, hoping it will gradually go down. I have reservations.

The Quest is aruably the most commercially viable vlomobile, made by a team of 3 (according to their web site), and yet it's still a tiny niche market, at a price of 8000 eur for the 4-wheel version. Thus, we can expect the Begorett cost at around 10000 eur, and it's a non-starter. While I applaud the enthusiasts who push the field forward, I think a fundamentally new strategy is needed to bring velomobiles to the wider market.

No matter how you put it, when the economy is in poor shape, spending 10000 eur on a bicycle with a shell around it cannot be justified. My Ford Fiesta costed me that much, come on now.Yes, it's green, cool, etc. but  10K is a year worth of the typical europeans' expendable income. I realize that volomobiles are built by manual labour which cannot be scaled, thus, one pays for the manufacturer's time. I wonder if they instead could sell their plans for 100 eur to enthusiasts, and let the enthusiasts invest their time building the vehicles at their own pace. With world-wide under-employment, most people have the time, not the cash, on their hands. I would pay 100 eur for a good instruction manual, as it would help me avoid obvious mistakes and time wasted on trial-and-error.

If the velomobile is ever to become the vehicle of the people, it has to be affordable for the people. The way I see it, currentlt, it's a vehicle for the eccentric rich. I put the price arbitralily for under 2K. But all of us who ride racing bikes know that 2K is dirt cheap. Yet, 90% of people think that a 0.5K insanely expensive. But with 2K you can at least claim that it can pay back in fuel costs in a year of commuting.

Some examples how veolomobiles can be made happen for under 2K. :

1. Sell plans and manuals. I would pay 100 Eur for a good step by step manual.

2. Rent out the moulds and vacuum pumps, let the consumer pay for one-off consumables. I would pay 100 Eur for the use of a mould for 3 days, as it would suck to spend months making one, and then it's a one-off use, since I only need one vehcile.

3. Use aerodynamic, but simple, shapes which can be produced with mass technology. Sure space-age curves look cool. but making the respective moulds is a feat in scuplure. What's wrong with a rotationally symmetric ellipsoid. Probably not much worse than the perfect airfoil, by 10%? I could probably pay a wood turner to make an egg-shaped mould out of a solid log?

4. Use carbon fiber. This sounds counter-intuitive, isn't carbon fiber (CF) the most expensive material ever? But carbon fiber is also very low tech, one does not need welding etc equipment, it's all epoxy, vacuum bagging, patience, and time for trial and error. For trial and error -- see 1 (if there is a good manual, most of it can be avoided). CF parts is costly b.c. it's all manual labour. In my experience, the fabric and the epoxy are 10-20% of the costs of the final parts. Enthuiasts have their hands and the time available. CF tubes can be made from CF sleeves, and sleeves are 5x more affordable than ready made tubes (e.g. Soller Composites). Finally, CF is easy to repair. I fix my CF nordic ski poles with CF tape and epoxy all the time (whereas most people throw CF poles away when they break, and they break all the time in races -- at 300 eur a pair).

5. Organize building classes. If I bought my own materials, I would pay an additional 500 Eur for a week-long class, if I knew I would end up with a ready vehicle by the end of it (or at least the shell, I can manage adding on my bike components on my own, ebay is full of used good quality parts). You can combine 10 people in one class -> 5 grand for a week of the instructor's time. I say, not bad.

All of the above could be done in collaboration with a composites company. E.g. Easy Composites in the UK is a great business.

Increasing the resistance of human-powered vehicles to side wind and bad weather

This was compiled based on reading discussion forums on HPVs, as well as information from other fields, e.g. wheelchairs, car tuning, airplane design etc. Some thoughts may be entirely wrong/outlandish.

Goal: make human-powered vehicles roadworthy in bad weather conditions. The arguments also apply to any light-weight vehicles, e.g. if human power is replaced with E-drives.

it's understood that most of the below may compromise the straight-line speed performance, but let's not worry about this for now.

A human lying on the ground flat does not get blown sideways, at least not < with 50 mph winds. Yet a typical HPV gets blown around with such winds, and is unsafe to ride.

Rain resistance

  • Silicone-based repellent for the windshield. Can be applied each time, or as needed. There are transparent teflon-based plastics, but they are too expensive, probably easier to vapor-deposit fluorinated or chlorinated silicone.

  • Windshield wipers. This never seems to be done in commercial HPVs.

Side-wind resistance:

  • 4 wheels.

  • Aggressive negative camber. This can be prototyped by 3D printing small scaled models and building a model wind tunnel. The negative camber increases the footprint and streamlines sideways.

  • If 4 wheels are used, may need a direct shaft and a differential rather than a chain drive

  • Reducing the overall side-profile of the shell, w/o bringing it too close to the ground (as to not compromise visibility). Most HPVs seem fish-shaped, but, perhaps, streamlining in the second dimension is also important. Or, is it not realistic to avoid bringing it close to the ground.


  • Fatter wheels (e.g. from a small Vespa) with round profile (for camber)

Shell construction

  • Velomobiles are made as monocoques ->  lighter weight, no frame inside, more space inside.

Controlling the temperature inside

  • There is a notion that controlling the temp of just the face and the head is key, whereas the body is more tolerant to discomfort (heat/humidity).

  • If cooling is needed... can I take some dry ice?

Windshield condensation

  •   heat the windshield, blow air. Air has to be dry though.

Velomobile concept from designer Joseph Campbell