Upgrades

Now that we have figured out what happened on our final Ivanpah run, it's time to get on with the list of upgrades we need to install before the next round.

1: Upgrade prop shaft ratchet and reinstall.

While we knew we could make do on the Ivanpah trip without this device, there are several good reasons to have it installed:

• 
  It allows the propeller to be easily spun separate from the transmission drive. This is handy when we're working on the prop hub or want to align the prop with the tower to tie it up when not in use.
• 
  It allows the pilot to grab a big handful of brake without overloading the drive (one of the reasons it was there from the start)
• 
  It will be needed to allow us to keep the inertia of the propeller from feeding slack line to the spool drive system which we still may use for the official NALSA record runs.

2: Upgrade the drive axle shaft ratchets.

It appears so far that the drive axle ratchets are stout enough for the job at hand (as long as I remember to engage both of them), however we have been surprised at the speeds we have reached and simply aren't sure how much faster we will go. With increased speed comes increased HP demands -- we're likely going to have steel hubs made and use bigger ratchets just to be sure.

3: Install the rest of the pilot adjustable variable pitch system.

As you saw from the video, the drive up to wind speed is excruciatingly slow. This is because currently our pitch must be set for the entire run before the start and so obviously we use a setting optimized for the higher end. It's like starting your car in a very high gear -- not the best way to get off the line. Again from the videos one can tell that once the vehicle gets past wind speed, the prop hooks up and the acceleration becomes comparatively brisk.

NALSA rules will not allow us to have a variable ratio transmission -- with such it is possible to turn the propellers rotational momentum into short burst of speed, thus it is not allowed. However, when the prop shaft and drive axle shaft are connected through fixed gearing, there is no way to use a variable pitch prop in a way that allows acceleration through energy recovery so variable pitch will be allowed.

As noted previously in the blog, we made the prop hub variable pitch capable -- there were two reasons we didn't finish the process and finalize the linkage:

• 
  I was intent on demonstrating that not only could a vehicle be built that could go DDWFTTW, but I was also intent on demonstrating the simplest version of one that would self-start and go from a standstill to well above wind speed. In Ivanpah we were able to do this easily. Looking forward, if we want to do it again we can simply unhook the VP system.
• 
  Accepting the invitation from NALSA to attend the Ivanpah event put a crunch on our construction schedule. By the time we knew about that event through the invitation, what was going to be a three month build had to be finished in just two months. We had to make some time tradeoffs and one of the items that we could afford to temporarily axe was the VP system.

4: Aerodynamic improvements.

The higher speeds we have achieved have quadrupled the aerodynamic drag as compared to the project goals. To maximize our record speed, we need to make fairings for various components. These should include:

• Deck over and under between the main frame rails. This will make the front fuselage essentially a hollow box. Currently air can spill in and around this area increasing our effective frontal area. We only want to have to split the air once.
• Install some aero improvement around the pilot. This might be a recumbent bike fairing or some sort of stretchy 'body wrap' that goes over the pilot and velcros down to the sides.
• Change the pilot seating configuration slightly to get the pilot less in the airflow.
• Fair the pylon uprights -- a rather straightforward task with several ways to go. Flexible mylar sheet over ribs or just one layer of glass over wire-cut foam.
• Treat the rear axle assembly frame and rear of vehicle -- again, rather straightforward tapering down to nothing.
• Wheel spoke covers. We'll have to custom make these to match our wheels -- perhaps mylar again cut to roll into a cone or again maybe fiberglass over a foam cone and solvent away the mold.
• Prop spinner and perhaps even an aero treatment behind the prop thrust box.

5: Paint and logos

• We'll be hitting the whole thing with a coat of paint to take the edge off the ugly. Perhaps the BUFC (big ugly cart) will simply become the BFC.
• We'll be having several Joby Energy and Google logos printed to be displayed proudly and conspicuously on the cart.

We'll get to it over the next few week and keep everyone up to date on the progress

A tiny bit of carnage and a week long mystery

We've spent the last week thinking about, collecting evidence and discussing the end of our last run in Ivanpah - where our highest speed run ended with mechanical failure while the vehicle was still accelerating.

First I'll describe the behaviors and damage, post a few pics and then get to our final (we hope) hypothesis.

In the below blog entry titled "Flat out -- the end" at the 2:30 mark of the accompanying video, the vehicle has begun to exhibit some strange behaviors (unseen in the video). Rick has just reported passing through 43mph and though not reported over the radio, is feeling some rather large, low frequency power surges -- strong enough to rock the entire chassis.

About the same time, out the driver's window of the chase truck, I've noticed that the chain tensioner has begun to swing drastically forward and aft -- not in random bounces, but in large, regular arcs about its lower pivot. This motion has reached a point so dramatic that I am reaching for the radio to tell Rick to abort so we can have a look at the drive train.

Bang!! At 2:39 on the video, the propeller loses it's connection with the drive axle and we can see that the top sprocket has been bent, spitting the chain off it's teeth. The run is over and we pack for home.

As we pack the vehicle away, we notice that the right ratchet hub has some damage -- not enough to keep it from functioning, but damage none-the-less. I carefully check the left hub for damage and find none (later, you'll find out why I wish I hadn't checked the left one).



On the way home, Rick was sort of giving me more detailed narative about what he was feeling and thinking during the run. Playing it through in his mind he couldn't separate the moment when he applied the brakes from the moment the chain came off. I was surprised because I had assumed he had applied the brakes *because* the chain came off. Turns out he started braking because he felt the surging had reached a critical point and wanted to abort the run. As soon as he braked, he felt the sprocket let go.



Let's go back to a few days before when we broke the ratchet hub at the New Jerusalem runway push tests -- the one where we stripped the ratchet while pushing it at a speed we really never expected to reach. As you remember, we decided to simply lock off (bypass) the ratchet on the prop shaft for this trip. It wasn't required for the chain drive (it's designed into the system assuming a spool drive, something we're still aiming for) but locking it out does have one possible downside -- braking can now shock load the drive system to far higher loads than the wind can. Previously, one could grab a handful of brakes and the prop would just freewheel to a stop -- now the chain has to also take the load of the slowing propeller.

After downloading and reviewing the GPS data that had been logged, we began to consider the possibility that the sprocket didn't let go under power, but under braking. Rick and been quite gentle on the brake application, but the GPS data clearly showed a far more rapid drop to zero than the climb to speed.

When we got back to Mountain View, as we unloaded the truck I noticed something very strange -- the right side tire had extreme wear as compared to the left side tire. Now, as previously mentioned, let's remember that we likely ran the vehicle in anger no more than five miles and above windspeed perhaps two of that -- hardly a high mileage vehicle.


Another intersting aspect of the tire wear was that it wasn't consistent all the way around the circumference -- the knobs were in far better shape at say 3 o'clock than at 7 o'clock. It was as if the tire had been out of round and yet it spun true as could be.

To further the mystery, Ken noticed on the video that even at lower speeds, before we passed and went in front of the vehicle, there was significantly more dust coming out from behind the right tire than the left tire.

We considered a number of hypothesis but none fit all the data. We did know that due to the torque reaction of the propeller, the right drive tire would have been loaded vertically less and less as the speeds and prop rpms rise. This would have led to slip on the right side -- except that our ratchet hubs don't allow for any power transfer from side to side like a differential would -- under most circumstances it behaves as a straight through axle. Also, why did the side that was more lightly loaded (and thus unable to tranfer as much power) end up with the damaged hub?

Well, today Rick produced the hypothesis that fits all the data perfectly.

(first, a bit of background on how the ratchet hubs on the drive axle are used: Our vehicle has a very wide stance. The hubs allow for one side of the drive to freewheel and catch up if the rig is turning. We don't want the axle to 'bind up' if we are steering to stay on a DDW course and lose the hard gained momentum. *Also* (and this is key here) we reverse the ratchet hubs after the run to allow us to move the vehicle around and even tow it back upwind, without the prop spinning.)

In short, it appears that before the run, the Car Chief on duty (or not on duty as the case may have been) failed to either reverse the left side ratchet, or at least set it to the wrong direction. This means that we made our final run with a one wheel, right-side-only drive vehicle. This explains the asymmetric dust trail. It also explains the right side ratchet hub damage and pristine left side - the right was double loaded while the left not at all. The surging was the result of the right side getting lighter and lighter (remember the torque reaction of the prop) until the tire would lose traction ... briefly suspend prop thrust ... regain traction ... repeat. These rather violent surges are what tore the knobbies off the right side tire -- but not quite all the way around ... just where the tire first hit before regaining traction.

Now, we'll never know for certain if this hypothesis is correct. Because I played with the left side ratchet before we left Ivanpah (remember I said I regret this now?) we can't check to see what position it was in during the run. But if we go out to our next tests and blow through 45mph without any such surges, we can be pretty sure Rick nailed it.

There is actually quite a bit of relief all around -- it's frustrating to not have all the pieces fit and now we can move forward fixing the problems.

Just so you know, we're not going to reveal who the Car Chief was at the time of this incident -- it we did he might just stop updating the blog. :-)

The money shot

Check out the streamer above the nose of the vehicle and also just to the left of the prop. What's wrong (or right) with this picture?

All this work to get a streamer and a flag to oppose each other with only wind power. :-)

Who are you again?

We'll take a bit of a break over the next couple weeks and get reacquainted with our familes. :-) After that we'll get back busy as heck polishing up the craft for the next round.

Here's the rough plan:

We were happily surprised with the performance of the vehicle. It truly exceeded the speed we had set as our project goal in it's very first run even though it ran out of room on the playa. The second and third runs were progressively faster to the point the vehicle is going faster than it was designed for. We'll have to rethink several parts of the vehicle in response to this very good news.

To increase the speed even more and set the very highest record the vehicle is capable of we need to now work on the aerodynamics of the chassis. We now expect to be bucking apparent headwinds of 40mph or more (up from the 20mph we designed for). Fairings for the chassis, pilot and wheels will make a significant difference at those speeds. A prop spinner and nacelle on the tower will also help.

We have to rethink drive wheels and tires -- we went to Ivanpah with a brand new set of knobby mountain bike tires and in the course of no more than 5 total miles of 'under power' distand and less than 2 miles of greater than windspeed running (where the horsepower requirements get stiff) we cleaned the knobs clear off of the right side of the cart. The right side is the side the torque reaction of the prop is trying to lift.

Still a lot of work to do to get a ratified record in the books.

Sweet success.

On Monday, March 22, 2010 our project with San Jose State University overwhelmingly succeeded in demonstrating 'direct downwind faster than the wind' on the Ivanpah dry lakebed south of Las Vegas, Nevada.

There were numerous NALSA (North American Land Sailing Association) officials in attendance, and one NALSA BOD member (Bob Dill) was there for each of these runs, and collected some of his own rough wind and ground speed data. All collected video and data corroborate the personal witnesses.

To be clear, this was NOT a NALSA sanctioned event but was presented as a demonstration to the NALSA BOD that the vehicle was capable of ddwfttw. Post demonstration, the team is working out the details with NALSA for rules and instrumentation related to any upcoming official NALSA ratified tests and records.

Pending NALSA BOD approval, we expect to be able to certify a record according to these upcoming rules in the range of 3 times the speed of the wind.

More to come

We've got more video from different runs that I'll get to later. We drove all night to get back and I haven't been to bed yet. It's time.

JB

Flat out -- the end

Even when you're happy, what's wrong with a little more happiness?

We kept increasing our prop pitch (we started below design intentionally) and our speeds until we blew out the prop shaft sprocket.

We'll fix it. :-)

Wait for it.

We were hoping we could get some good video of the vehicle going from a standstill to well above windspeed with a fixed pitch propeller. We have now done so. It takes time for the prop to hook up, but when it does (just above windspeed), it takes off nicely.

(that's Bob Dill of NALSA and Iron Duck fame checking the wind behind the vehicle before it starts)

There's a first time for everything

A rare sighting of a ddwfttw vehicle in its natural habitat

Off to Playa

New shoes

Off with the pavement tires and on with the desert knobbies.

Push it real good

Here is a shot of our top speed run with the prop (40mph).

Notice that 'popping' sound as the vehicle goes by the camera. That is our second prop shaft ratchet letting go. Last time (on the dyno) the failure mode was the lightened hub itself. This time, in a strengthened hub, it was the ratchet itself.

We've taken the prop shaft ratchet out since we don't need it for now. With the hub still OK, we could have popped in another ratchet (I still have the one from before) and just kept it to 35 and below, but we've opted to just bolt it up tight for this trip. The ratchet was there for when/if we build the spool drive and isn't needed with the chain so we'll lock up straight

Add that spinny thing

After we were happy with the chassis dynamics, we threw on the propeller and slowly increased the speeds.

Here we get up to 30mph

Bob and Weave

Before we allow the vehicle in the grasp of the wind on an open surface we wanted to do a few vehicle dynamics tests. With that in mind, we spent much of Saturday at the New Jerusalem taxiways, first slowly increasing the speeds of the vehicle without the prop installed run by run.

These runs in the video are our later ones and have Rick at about 50mph and trying to upset the cart through steering inputs.

Brakes, steering and general stability proved excellent. There is a bit more brake disc squeal than I was expecting ... I'll do some research on the bike forums and see if there is an answer.

She ain't heavy

When it comes time to do the record runs post shakedown, we may want to go with a very light pilot. Rick's wife Monica has volunteered and as is the procedure for all hired gun drivers, she came in for her seat fitting this evening. :-)

She can't be more than 95lbs dripping wet and she sure would be aerodynamic.

Braking wind

On the eve of our vehicle dynamic testing at New Jerusalem taxiways, I fabricated, installed and cabled up the primary brake lever. The adjustable stops at the base of the lever keep it within the limits of the brake assemblies at the rear.

These brakes will be the ones normally used to slow the vehicle -- the brake set on the front wheel is just a totally independent backup used only in case of emergency.

Two is better than one

The calipers on the drive axle rotors are now mounted. Tomorrow I'll install the last caliper on the prop shaft, fabricate the brake lever and cable them all up.

Wide load

We added the two extra bearings, replaced the aluminum axles with extended steel ones (added 15lbs over the aluminum ones unfortunately) and then load tested the rear assembly.

We're happy with both its strength and also the bit of 'suspension' that the slightly flexible chrome moly cantilevered axles give us.

Up to this point we've kept the vehicle in its narrow configuration so it would fit on the dyno and take up less space in the shop. We widened it out for its upcoming first open surface tests.

Nosing around

In prep for our trip to New Jerusalem on Saturday, Rick put together a pair of mating surfaces to push the vehicle around on the taxiways.

Done in front

Fender mounted, footrests installed, steering stops machined and welded in place, brake cable and guide ran and caliper adjusted.

We'll be taking the vehicle out to the New Jerusalem taxiways early Saturday morning where we can push it around in no wind to get a feel for its dynamic behavior. We'll be in the wider rear axle config for the first time at that point. We'll test brakes, etc with planned runs up to 40-50mph. We can then return and spend Saturday evening / Sunday morning making any adjustments we feel we need.

Steering complete

We finished up the steering linkage tonight. All that is left up front is to run the brake cable, loctite everything and install a pair of footrests.

Front caliper mounting

I'll work on the other three mounts tomorrow evening, but it was nice to get the first one out of the way and learn about the adjustment system of this particular model.

Travel ready

This evening Rick worked on the saddles and travel mounts for the prop and the tower. The propeller will be removed from its shaft and layed under the tower as shown. We'll pad everything up with pipe insulation once its all fabricated.

DDWFTTW for hours on end

We got off a full weekend of dyno testing.

** Full range of speeds (up to 140% of design)

** Full range of AOA (from zero tip to 20+ tip)

** Full range of HP input (up to 2x design).

No runs, drips or errors.

Are we 100% happy? -- that never happens - especially with a limited budget, but we're feeling very good about being ready for the trip to Ivanpah.

From what we learned on the dyno we'll finalize our chain drive system. Additionally we're going from the weight saving aluminum axles to steel. We came to suspect that the aluminum versions weren't going to hold up as well as we were told they would. Don't want to risk a break of an axle in the desert. Steering linkages and brakes will be my focus the next few evenings.

The very good news from this dyno round was that again, at every AOA, and every RPM (within the limits of our motor of course) we were pulling forward with enthusiasm.

Sit down!!

Rick's new masterpiece in use by my lovely partner Kathy.

Really, he only does macrame' to relax

Who knew?

Just happy to be a part of all the madness

Ken Milnes (long time VP Engineering for Sportvision) has kindly stepped up to take one thing off our plate -- an RTK survey of the Ivanpah playa. Ken is the inventor and primary developer of much of our proprietary RTK capabilities here at SV and as such is perfectly suited to setting up the survey system.

The purpose of the survey is of course to know the vertical profile of the lake bed -- this will help us make sure we're not polluting the results by running downhill. We of course expect the bed floor to be pretty damn close to perfect, but it's just good to know.

A few days ago he pulled one of our in-car systems out of the Indy mobile production unit before it headed out for the season. The hardware we'll be using on the playa will actually later be installed and running in one of the racecars during the upcoming Indy 500. We'll try to keep it below 250mph. ;-)

Here Ken is holding a telemetry antenna that will be used to communicate with the base station beacon during the survey.

Sometimes slower is better

With the straight line, directly downwind nature of the craft, we have no need for fast response course correction. In fact we want extreme straight line stability in our steering.

With that in mind we installed an industrial damper aft of the wheel.

A little more meat on the bone

After getting another kart hub in, I fabricated a ratchet hub to replace the one we broke on the prop shaft. Same as the last one except I didn't mill off the flange. It was this thinned down material that sheared off last time.

Down the stretch

We've had an unspoken (blog wise) goal for the last two months -- the vehicle running on the Ivanpah lake bed during the last full week of March. It's not a random date we picked ... the NALSA (http://www.nalsa.org/) is holding their premier yearly 'Americas Cup' landsailing event on those dates and in that venue.

Ivanpah is the playa where almost exactly a year ago the sensational Greenbird (http://www.greenbird.co.uk/land-record) set a new world wind powered land speed record of over 126mph. NALSA is the ratifying organization for such records here in North America.

Starting last year, we have been in discussions with the NALSA BOD regarding the possible addition of a DDWFTTW landsailing class and record ratification procedures. In January they suggested that we bring our vehicle to the event in Ivanpah where a good number of the BOD members will be. This would give them an excellent opportunity to see the vehicle first hand for themselves and provide a more hands on knowledge base for further discussion.

We accepted the challenge even though it meant moving up our construction schedule by almost a full month. We've been charging hard to make this date and it's all looking good at the moment. One week from now we should be well into the first open surface shakedown runs of the vehicle.

In a step beyond the dyno testing, we hope to accumulate another level of understanding regarding the vehicles dynamics which will help us prepare for our record runs later in the spring. The weather has been a bit wet on the playa so we won't know how much actual running we will get until we arrive. Fingers crossed for good weather and a solid surface.

Snap, Crackle, Pop

Soon after fixing the shear pin, we blew up the most highly loaded of the ratchet hubs. The ratchet head itself was fine, it tore out of the housing that I machined.

Several good things here:

A: No damage other than to the hub itself
B: It's easy for me to machine another more substantive hub
C: The propeller was already absorbing almost 2x the design HP before this happened.

All in all an easy fix and under field conditions we don't expect to ever load the system this high again.

Shear pleasure

Our first minor casualty was a shear pin in the prop shaft. The components in this part of the chain are under the most torque load of all. We quickly replaced the soft pin with a harder grade 8 variety and I commented that now we could well break something more expensive. Turns out I was right.

Let's show some restraint here people

Enough excess thrust was generated *after* achieving windspeed, that the dyno was dragged down the parking lot several feet. To keep it in one place, we finally tied it off to the pick truck. The dyno is not on wheels btw, but skids.

This result is completely contrary to the oft quoted theory that no thrust can be generated at windspeed

Uphill, both ways

During some of the testing we loaded up the front of the vehicle and tilted the ramp to force the vehicle to climb as it advanced to the stops. By positioning the stops to start the vehicle axle aft of the dyno axle, the rear wheels were also forced to climb.

The vehicle advances under all conditions.

Before the storm

A few pics from the setup and prep before firing up the wind.

Picture updates to come

In addition to the videos I just posted, I've got pictures as well. I'll get to that after dinner.

JB

Prop shaft break

We moved the pitch to 20 degrees and ran the ratio up to 20/65. First we sheared the grade 2 'butter bolt' that engaged the hex drive to the ratchet hub. We pretty much had the motor maxed out power wise (6.5hp) when that happened so I was hopeful when we replaced that with a grade 8 bolt that we wouldn't break anything else. However before we fired it up this time I said to my buddy Jerry "well, with that soft shear pin gone, now we can break expensive stuff".

It almost came true.You'll see in the video that I fucked up and engaged the idler on the motor drive waaay too fast and I'm sure that's where I blew the side out of the ratchet hub (carnage pics on the blog later). It only lasted a short time and I never got close to max throttle before it blew up.

High speed dyno test

At about 115 rpms and still with the AOA at 15 degrees. Design speed is in the 120 range

First dyno test

This is the first spin up of the cart on the new dyno setup. We were at 15 degrees AOA and were running our slowest gear (15/65). Design point is 20 degrees(tip at static) and 23/65.

Even with the Honda engine nearly idling the vehicle immediately pulled to the front and held against the stops. (for those who aren't up on what that means, it means that the vehicle is going ddwfttw at a very low windspeed -- a good sign of our overall system efficiency).

Pickup work

There were several bits that had been being put aside that we worked on this evening. We made the final thrust box stiffening brackets and installed them. Rick installed the RF board and wired the load cells in the hub and on the chassis. I drilled the prop shaft for a safety pin. We applied Loctite to about a thousand bolts and cap screws. On and on.

Bright and early in the morning we'll be spinning the rig up on the dyno and we'll find out how good or bad we've done both with our drive system and propeller.