Published on February 19th, 2014 | by Mike Barnard24
Google’s Makani: From Regulatory To Technical To Wildlife Challenges
Almost a year ago, Google took a 100% ownership position in Makani, an airborne wind energy company. It had previously held a $15 million position in the company, but after the tragic death of one of its founders, Makani’s future was less certain. Now it has deep pockets in Google X, the division devoted to breakthrough technologies.
So, how likely is Makani to disrupt wind generation?
Close assessment of Makani’s public documentation shows that their approach to airborne wind energy has several serious challenges, some in common with other airborne wind energy approaches, some unique to Makani:
- Unlike current wind turbines, their devices cannot be placed on existing, working farmland or areas with any secondary uses
- They must be placed a minimum of 1.5 km from public roads and power lines
- They may not get permission to fly with unmarked, unlit tethers, and their approach won’t work with increased tether drag
- They must be spaced further apart than equivalent wind generators, amplifying secondary use issues
- Maintenance will likely require shutting down portions of the wind farm due to safety concerns
- Significantly increased maintenance cycles will likely reduce capacity factors to HAWT levels or below
- They will be unable to operate in a greater range of adverse conditions and will be more susceptible to damage when grounded
- Avian mortality will likely be higher than for current wind generation technology
It is unlikely that they will be able to put up more than a handful of onshore wind farms given the restrictions, but offshore in tropical and subtropical climates is somewhat more promising. Some of these restrictions might be eased after years of zero failure production.
Despite these significant limitations, they make the interesting claim that:
AWTs have increased siting flexibility (as opposed to conventional wind turbines)
That said, regardless of their serious shortcomings compared to conventional wind turbines, the Makani model may prove economically viable in some circumstances. It does use much less material, require less onsite construction and harness better winds for higher capacity factors. Transportation of materials is easier, and in-factory construction of complete components ready for shipping could reduce costs if the system is ever built in scale. If it’s viable at all, it might have a favourable lifecycle cost of electricity in some places.
What is the Makani model?
The allure of higher altitude wind is straightforward. Wind speed increases with altitude for the most part. The energy in wind that can be harnessed by a bladed wind turbine increases by the cube of velocity, which means that as the wind gets faster, the energy available in it increases extremely quickly. If you can get a wind generator into that higher, stronger, steadier wind, you’ll generate more electricity. This is why the biggest innovation causing wind turbines to get more powerful has been putting them on taller masts. Airborne wind energy intends to go even higher.
The second innovation is where airborne wind systems and wind turbines part ways. Wind turbines blades got longer so that they could catch the energy in a greater swept area of wind. Since surface area increases with the square of blade length, putting longer blades into stronger air multiplied a cubed factor by a squared factor with historically excellent results. And partly because the wind was steadier higher off the ground, capacity factors increased as well, so they generated a larger percentage of their possible output annually. Horizontal axis wind turbines (HAWTs) have moved steadily up the scale and now there are 5 MW onshore and 7 MW offshore wind turbines with capacity factors of 40% and more, and even bigger units coming offshore. There are onshore sites in the USA and Brazil that see 50% capacity factors.
Makani, among others, has gone a different route.
They built a hard-wing, tethered plane that flies in circles like a kite in the wind. The wing will have eight much smaller wind turbines on its leading edge. They put aside the squared advantage of the larger surface area for the cubed advantage of higher velocity wind that they could get by using the wind’s force to fly in circles very quickly. While they haven’t released specs on velocity publicly, my calculations suggest that they will need to fly their projected 5 MW airborne wing at about 130-140 kilometres per hour in their targeted 265 m radius circle. This strikes me as a bit aggressive given radius, tether drag and eight turbines slowing the wing as they generate electricity, but I’ll assume they’ve done the modelling and believe with iterations that they can get there. The velocity is based on a couple of assumptions in the calculations I performed, especially blade efficiency, and they could probably reduce the speed a bit if my assumptions were off. The velocity is validated somewhat by the minimum flight speed of their current Wing 7 of 30.5 m/s or about 110 kph per Chapter 28 of the Springer AWE book by a member of the Makani team.
Take off and landing will be handled like a helicopter, vertically under the eight propellers. The tether will be winched in and out as needed.
So Makani passes a few sniff tests. They have a working prototype. The physics of what they are trying to do is well understood and was well before they decided to exploit it. In fact, the seminal paper defining this approach was written in 1980 by Miles Lloyd of the Lawrence Livermore National Laboratory.
So what stopped people from doing it before? Mostly will, materials science, and computers. The will part is simple; it’s really tough to build autonomous flying vehicles that will work in a lot of different types of weather and not kill people. Fly-by-wire passenger jets and drones have decreased the apparent difficulties substantially. Materials science is simple too; the tether has to be extremely strong and extremely light, as does the wing. Pultrusion of carbon fibre to the rescue, along with more commonly understood carbon fibre approaches for the wing. The increasing power of small devices and tiny sensors has reduced barriers for automation as well.
So you’re thinking, what’s the problem? The physics make sense, the components are there, Google owns the company, what could go wrong?
Well, a few things in general, but a few specific things for Makani’s model as documented in their Tethered Aircraft Concept of Operations (TACO) submission to the US Federal Aviation Authority which was necessary in order for them to test fly their prototype in daylight hours along with notices to airmen (NOTAMs) every time they tested it. The TACO document will need to be updated and made more robust as they proceed, and many parts represent Makani’s wishes, not what aviation authorities will agree to.
It’s fascinating reading, and the first thing that made me sit up was their suggested setback for their proposed 600 KW onshore device of 1.5 kilometres or just under a mile. That’s right, a 1.5 kilometre setback for a 600 KW device. That’s a lot given that a 600 KW HAWT is considered tiny these days, and in Ontario under Regulation 359/09 would qualify for 550 meters. In South Australia under their noise setback regulations, a 600 KW HAWT would possibly be even closer.
But it gets worse for Makani. The maximum 550 meter setbacks for HAWTs of the same are from inhabited dwellings. What are the requested setbacks for the Makani devices?
A setback of 5,000 ft. (1.5 km) from any trunk power lines and public roads should be used
That’s right, Makani’s request, likely aggressive, is that the smallest commercial device they want to make should be almost a mile from roads and power lines, never mind people’s homes. And this is from every edge of any wind farm built with the Makani devices. That eliminates the use of these devices in farmland completely, and allows installation of these devices only on non-arable, remote land.
For context, the state of Victoria in Australia put a politically motivated two kilometre setback from inhabited dwellings in place and almost no wind farms have actually been approved in the state since as a result. The effect for Makani will be similar.
Why, given this kiss of death for onshore siting, would I suggest that Makani was actually being aggressive?
Let’s examine the device a bit more carefully, and compare it to a 600 KW HAWT.
The onshore targeted wing is projected to be a 28 meter (92 feet) wide, 1050 kilogram (2314 lb) device. It will be traveling at perhaps 130-140 kph (80-85 mph) when operating as designed at distances of 440 meters (1443 feet) from its mooring point. It will have eight rapidly spinning blades approximately one meter long on its leading edge. There will be a virtually invisible (more on this later), electrified, carbon-fibre tether of extraordinary strength, traveling at 130-140 kph near the wing and approaching zero speed but with enormous force near the mooring point.
If all goes well, then that device will stay attached to the mooring point by the tether, the tether will always be pointed up into the sky and the device will never get anywhere near the ground except when it is acting like a helicopter and moving very slowly. An enormous amount of testing will be done to ensure that this is almost always the case.
But accidents happen. Software, even safety-certified software, is imperfect. The atmosphere is going to be throwing every weird condition at this flying device and its tether. Some are guaranteed to operate incorrectly, more in the early years until all conditions are understood and accounted for, but also more late in life as components fail and if maintenance is slipshod.
And at that point it’s quite possible that this large, heavy, fast moving device with its invisible and very strong cable is going to be sweeping along the ground or plunging straight toward it or have broken free of its tether and be sailing downwind toward whatever is there. Or potentially worse, be flying downwind dragging its very strong possibly still electrified cable behind it.
Makani, high-altitude wind energy researchers, airborne wind entrepreneurs, and the Federal Aviation Authority are well aware of these potential occurrences. They are all working on multiple redundant safety systems to mitigate these risks.
But it’s unlikely that the FAA will allow a setback of only 1.5 kilometres until the software and system has proved itself in production for a few years. And it’s unlikely that insurers would insure these devices for liability at 1.5 kilometres until they’d been in operation for several years. There’s only so much that engineering and testing can prove.
So the setbacks of 1500 meters, or almost a mile, are likely an aggressive request by Makani. They’ll be longer. Maybe twice as long. Maybe three times as long initially. And they will likely stay that way for a decade or more.
That’s probably a good thing actually, because the sound these things make will likely be more annoying than HAWTs. Watch this YouTube video of testing of the 20 KW prototype starting at about one minute.
This is the current Wing 7, 20 KW prototype from what looks like a couple of hundred meters crosswind. Note the serious increase in noise as the device flies closer to the camera. There are two primary sources of noise here. The first is rotor buzz. The faster blades spin, the more noise they give off and these are spinning very rapidly. The second is tether noise as the unaerodynamic round cable is whipped sideways through the air at high speed, which is very familiar to anyone who has heard taut ropes or guy wires singing in the wind. (Attempts decades ago to build aerodynamically sheathed tethers failed and no progress has been made on this front). Now scale that noise up by a significant factor as the 600 KW device will be much bigger, faster moving and the cable will be thicker. Every time the wind shifts toward a home downwind of a Makani farm, the noise at that home will increase much more than for a similar shift at a HAWT farm for the simple reason that the Makani devices will shift with the wind and be much closer to the home.
The minor annoyance for some people of the whoosh of HAWT blades will likely be significantly outweighed by this noise, so longer setbacks will probably be a blessing in disguise. Makani calls for a 1.5 kilometre noise setback in their FAA submission as well, and this is equally untested or validated. It might have to increase as well. In their FAQ, they have this to say about noise:
The noise level of the Airborne Wind Turbine (AWT) should be comparable to that of traditional wind turbines. The sound emitted from the AWT is higher frequency than that of a conventional turbine and high frequency sound diminishes quickly with distance. At the operational altitude of the AWT, the noise level is expected to be consistent with conventional wind turbines;
The noise characteristics in terms of frequency are unproven as yet. It’s certainly true for the small prototype, but for the much larger devices and their tethers, this is likely not true.
The second thing that leapt out a me was this request of the FAA by Makani. I’ve included the entire section of the document as it is very instructive.
The first thing to note is that if they have to put ribbons, reflective balls or lights on the tether, their wind generation approach won’t work. Full stop.
The second is their disingenuous claim that the tether is really just the same as the guy lines on a stationary radio tower. This is disingenuous for two reasons. The first is that the lights on a radio tower stay in one place, and are perceived by pilots as stationary objects. The lights on Makani device will be moving, and more likely considered at first glance to be those of an aircraft. The second is that the guy wires on a radio mast don’t move at 130-140 kilometres per hour.
The expectation is, once again, that Makani will be testing their device and ensure safety and reliability, and that permanent NOTAMs will be in place for pilots. However, it’s hard to see why the FAA should consider that a 440 meter invisible, ultra-strong, electrified cable whipping through the sky at 130-140 kilometres per hour potentially anywhere in a half-globe with an area of 178 million cubic meters would be a reasonable comparison to a static guy wire, at least not until several years of safe operation have occurred, and possibly not even then.
There are likely other ways around this, and Makani will likely be forced to undertake them, but off the top of my head the most likely is massive floodlights pointing up into the sky at cables whose poltruded carbon fibre has been altered to include highly reflective elements. Instead of invisible cables whipping through the sky at night, brightly lit 440 m cables will be whipping through the sky at night. Once again, those setbacks are looking to increase rather than to diminish.
The third thing that caught my eye was this statement:
one tether length from any other turbine in the downwind half of the hemisphere
The anchor points of the Makani devices will have to be a little over a tether length apart, which makes complete sense. That means that downwind and side-to-side, for the onshore solution, these devices will have to be 440 meters from each other (actually a bit more, but let’s stick with tether length). They can’t be closer because of the risk of cables wrapping around one another in mid air or slicing one of the wings in half or a wing slamming into a downwind anchor mast at 130-140 kph. They must be 440 meters apart.
All good and well, but how far apart would a wind farm of 600 KW HAWTs have to be? Well, the rule of thumb is 6–10 blade diameters apart in the direction of the prevailing wind to avoid interference, and perhaps half that at 90 degrees to the prevailing wind because of suboptimal generation when the wind quarters will be made up by more wind turbines.
600 KW HAWTS with typically 25% capacity factors have a rotor diameter of around 40 meters on average. That means downwind spacing of 240-400 meters, and crosswind spacing would be 120-200 meters. The tether length of the Makani 600 KW device is 440 meters. Assuming a 10 km by 10 km wind farm, a little math tells us that there would be about 2.4 times as many 600 KW HAWTs at maximum spacing in a wind farm as Makani devices at minimum spacing.
Now, the Makani devices will be flying up where the air is stronger and more consistent, so they will be getting better capacity factors. But 2.4 times better, or a 60% capacity factor? Unlikely, especially as it isn’t just the wind velocity that’s a factor (more below).
And this is using HAWTs no one would bother to use because they are too small. Let’s take a hypothetical case that’s more realistic, of 2 MW HAWTs and consider again a 10 kilometre by 10 kilometre wind farm. The 2 MW HAWTs have 82 meter rotor diameters as one example and modern ones typically expect 35%+ capacity factors.
Doing the math results in about 1.9 times generating capacity at HAWT maximum spacing and Makani minimum spacing. Once again, the Makani devices will have better capacity factors in theory, but a 1.9 times better capacity factor? They would have to have a 67% capacity factor to produce the same energy as a farm of 2 MW HAWTs.
And, of course, this is ignoring the ability to fill up probably another kilometre in every direction with more wind turbines. That would bring the potential HAWT generation to about 2.7 times the potential Makani generation, requiring Makani to have a capacity factor of 95% to be equivalent. The tech is interesting, but not better than any form of generation of any kind in the world.
So Makani has some challenges competing in a straight utilization of non-arable land for maximization of generation. However, in theory they are going to be a lot cheaper and easier to install than wind turbines, so if the costs ever come out, it’s quite possible that they will be viable economically in areas where there is absolutely no other use for the land.
Tulip farm in the Netherlands. Image courtesy of http://lemonopinion.wordpress.com/2012/11/12/potd-like-a-coloring-book/
However, HAWTs work brilliantly as additional revenue sources on farmland, taking up less than 1% of the land for bases, tracks and related infrastructure and returning very significant per hectare cost gains to farmers. Makani might be able to compete on a pure economic basis in completely unused land in the middle of nowhere, but can’t operate intermingled with farm operations by the company’s own FAA submission.
But we’re not done yet.
Capacity factor comes from a number of things. Minimum winds for operation. Maximum winds for operation. Suboptimal winds where generation is below maximum. Cold and hot weather. Lightning, thunderstorms, and other extreme weather. Maintenance downtime requirements.
Regarding minimum and maximum winds Makani has an advantage due to the advantages of tether tension vs mast and blade shear forces. The technology at one level really is superior, and Makani has launched, landed, and flown a prototype to prove parts of it. Suboptimal winds are questionable. Makani’s device needs a certain amount of wind just to stay aloft, while a wind turbine’s blades can just turn lazily in lower winds. Similarly, at winds approaching maximum, wind turbines continue to produce, but it’s unclear what the peak winds that are viable are.
But let’s first consider temperature ranges.
In below zero temperatures, airplane wings require de-icing in order to take off. Makani is completely silent on this point on their website, in their FAA submission, in their YouTube videos, and in their interviews. There isn’t any reason to believe that the Makani wing can take off and fly safely with ice on it when other flying devices can’t. And it’s not just the wing. The cable is quite likely to ice as well, creating additional drag and weight that will prevent the wing from flying, and could impede winch performance as well. And then there is snow and ice build up on the winch and launch cradle. It’s very possible that Makani’s device simply won’t be operable in subzero temperatures as a result. Snow, freezing rain, sleet, and hail are also likely to impair performance substantially.
What about HAWTs in similar situations? Well, for the most part they just sit there and generate electricity. There are 1600 or so of them in Ontario, Canada alone, which is known for the beauty of its freezing cold winters; and more in Alberta, which is even colder; and more in Sweden and Denmark, also known for being cold and snowy; and even more offshore in the North Sea, which isn’t famous for being the French Riviera. Normal HAWTs work well down to -20 degrees Celsius (-4 degrees Fahrenheit), and cold-weather HAWTs operate down to -30 Celsius (-22 Fahrenheit). They work in snow, sleet, freezing rain, etc. In some circumstances, ice buildup might degrade performance a bit before it falls off, but it’s extremely rare that a wind turbine is stopped by ice. (And ice throw is a non-issue at setbacks required for wind farm noise.)
What about heat? In the Middle East, airports don’t operate in the daytime during the summer months because the temperatures make air density so low that landing passenger jets — which are vastly over-engineered for safety and flying conditions — cannot land safely. Once again, Makani is silent on the upper temperature range of their flying generator. This is likely to be a smaller factor than cold, but still present.
As for HAWTs, they work just fine at the equator in the desert. In fact, Saudi Arabia is putting in a wind farm and solar panels to help desalinate sea water.
What about lightning, thunderstorms, tornadoes, hurricanes, and other extreme weather? Utility scale HAWTs mostly just sit there when the winds get too high. The blades are feathered and braked and that’s about it. An infinitesimal percentage of the time, something fails and one HAWT in a farm goes boom, usually an older, poorly maintained one. Makani’s model will require the device to fly safely down in increasing and turbulent winds, dock safely with the cradle, then ride out very high winds without damage as the various component get shaken against one another. Small, light planes are typically tied down during wind storms, and they are not in a metal cradle on a metal post exposing the full surface of their wings to the gusts broadside on.
At least one analysis by an airborne wind proponent says that airborne wind energy will have to be out of the air during lightning storms due to the characteristics of the devices and lightning. Makani says:
Makani will follow standard practices within the aviation and utility industries for surge protection, shielding and grounding. The wing will meet the lightning strike standard defined in military specification MIL-STD 464-A.
This doesn’t state that they will meet these standards while the wings are flying, but it’s possible that’s what they mean. It’s possible that when they get further into this, their device won’t be able to fly when lightning is expected, but will be safe when docked. Time will tell for transient lightning, but it’s unlikely that the Makani device will be flying safely during severe thunderstorms, or during less severe ones either.
What about maintenance downtime? In general, the more tiny moving bits a system has, the more maintenance it requires. The harsher the environment, the greater the maintenance required. The more it is susceptible to grit, branches, etc, the more maintenance is required.
Makani’s devices have lots of small moving bits in the winch, the mast, the launch cradle, and the wing itself, which are mostly exposed to the elements and grit. The combination is highly likely to increase maintenance downtime.
By comparison, HAWTs have a big mast, three big blades and a bunch of enclosed mechanicals. They require maintenance checks every few months, have guaranteed uptimes from manufacturers of 95%, and usually see 98% availability (this isn’t capacity factor or amount of time they are generating, but the time that they are not out of service).
It is highly likely that Makani’s devices will require more maintenance than HAWTs. On the other hand, due to their smaller size and relatively lower cost — this is an autonomous carbon-fibre 28 meter wingspan plane after all, so it’s not going to be cheap — it’s possible that having a couple of spares around so that one can be swapped out while maintenance occurs will be fiscally viable and time efficient for the onshore devices. Given that the wing and cable are 1300 kg (2900 pounds) between them; the winch, mast, and cradle will likely be in the same range; and the wingspan is 28 meters (92 feet), this won’t be a trivial exercise either, but it is at least possible to consider it compared to the equivalent operation on a HAWT.
But on the note of maintenance, how will workers’ safety boards and insurers consider maintenance workers driving vehicles into a Makani device farm where invisible, carbon-fibre, electrified cables might be sweeping around at 130-140 kph close to the ground? Or a malfunctioning carbon fibre 1050 kg wing with eight props might drop out of the sky? At least for several years, it is likely yet again that all maintenance would need to be performed while the farm is at least partially shut down. Imagine that, to do a maintenance check on a HAWT required a third of the wind farm to be shut down for the duration.
Finally, a little about wildlife. While the Makani device will be better than fossil fuel generation for wildlife, the reality is that they require a thin, invisible cable to be flying from ground level to 350 meters (1000 feet) for the onshore model or 650 meters (2000 feet) for the offshore product. That cable will be flying through a cone at a speed of around 130-140 kph near the wing, and slowing as it nears the ground. This is likely to be much more difficult to perceive and avoid than any wind turbine blade for birds. The best evidence is that many species of birds including seabirds and many raptors simply adapt to wind farms and avoid them, as there are highly visible masts and blades that they perceive as they do jutting islands and trees. There is no evidence to suggest that this is true for Makani’s relatively small masts, long tethers, and flying wings. The tethers will cover a much, much greater volume of airspace than wind turbine blades. It is difficult to not see greater avian mortality as a likely outcome, with resultant challenges for safe siting.
So, enough about the onshore device. It has much more severe restrictions in terms of nearness to human infrastructure and homes than HAWTs. It can’t be interspersed with farmland as HAWTs can. It likely can’t be used during anything resembling winters. It might not be possible to use it during summertimes in certain countries. Maintenance will be required more often and will probably cause much more extensive disruptions to generation for the foreseeable future. It won’t generate more power than an equivalent HAWT farm, never mind a serious, modern HAWT farm. They’ll probably kill more birds per MWh. And if they can’t convince the FAA that an invisible, carbon-fibre, electrified cable cutting through the air is the same as a stationary radio antenna guy wire, it won’t work at all. A Makani farm might still generate electricity more cheaply on full lifecycle costing, but it has a lot of disadvantages that make this less likely.
Under these restrictions, it’s hard to see more than a handful of Makani onshore farms being built worldwide. And, to be fair, Makani along with most airborne wind generation organizations agrees that offshore is where the devices are really targeted. It’s not like they aren’t aware of many of the challenges outlined above.
What about the offshore Makani device?
First off, let’s consider an apples-to-apples comparison of a 30 km by 30 km wind farm with 5 MW devices. Due to spacing requirements again, HAWTs would have about 42% more capacity in the same area, so that’s the capacity factor increase Makani’s devices would have to achieve to be equivalent on that coarse and not particularly useful scale. As modern offshore HAWTs typically run in the 40%–50% capacity factor range, Makani’s devices would have to run in the 60–70% range to be equivalent. This is a more viable gap than for the smaller onshore devices, but still unlikely to be achieved.
One obvious point out of the analysis is that as they scale, the comparisons become more favourable for Makani (barring other challenges). 10 MW Makani offshore devices might be the sweet spot, if they are technically feasible at all.
But feasibility of even the 5 MW device is questionable.
Launching and landing a 9,900 kg, 65 meter wingspan device in offshore winds will be much more difficult than landing their current 60 kg, eight meter wingspan device in calm onshore winds. Ask the Osprey operators how easy and safe it is to transition a very large aircraft from forward flight to hovering. And then getting the 65 meter wingspan device into a cradle perhaps 30 meters above the sea without crashing it into the cradle and breaking things when the cradle is likely to be moving as well is non-trivial.
There is also the delicate matter of ditching. When a Makani device hits the water, as they will occasionally, they are going to sink rapidly. Not because the wing is particularly heavy or dense, but because the 3,660 kg carbon fibre and aluminum cable is going to plummet into the depths and drag the wing down with it. Adding flotation to the cable isn’t an option. Adding flotation to the wing might be, but it will be yet another layer of complexity, cost, and maintenance challenges and it will likely degrade performance. Adding a breakaway option on the tether attachment to the wing is viable, but is another point of failure, especially under the loads projected.
The maintenance problem in offshore conditions with 10 meter plus waves, salt spray corrosion and fouling, and difficult access will be multiplied. Offshore icing is a large problem; the North Sea wind turbines work just fine, but they are massive, simple machines by comparison. The tether length is 1060 meters (3500 ft), and at the end of each tether is a 65 meter wingspan, 9,900 kg wing with 8 props flying at 130-140 kpm. There are no conditions under which it would be feasible to fly anyone in to maintain these devices as is done with offshore HAWTs except by shutting down significant portions of the wind farm. And under what conditions would any ship’s pilot or ship insurer consider it reasonable to pilot a boat under a Makani farm where one or more of the devices are malfunctioning? Just as with the onshore farm, it’s entirely possible that maintenance inspections and work will require large portions of the farm to be shut down entirely for the duration. Finally, swapping a 65 meter (213 foot) wing out for repairs at sea is a non-trivial exercise requiring a large and specialized craft by itself. This is technically the equivalent of replacing a blade on a wind turbine.
Makani has a long way to go before operation of an offshore wind farm of these devices is shown to be cheaper per unit of electricity than an equivalent HAWT farm.
And they still have to get aviation authorities to agree that long, invisible, carbon-fibre, electrified tethers buzzing along at 130-140 kph are acceptable, or the entire solution won’t work at all.