I guess many remember that in my recent calculations of the 
specs under which lifter can be self-powered (lifting its own power-
supply) there was a parameter max_dens for a maximal density of 
collector/corona wire array per meter. Who don't know what I am 
speaking about, take a look at "Complete engineering model of autonomously powered 
corona-discharge propelled aircraft" section at http://sudy_zhenja.tripod.com/lifter_theory.

Anyway, the calculations for Savioiur-type power supply gave value
of max_dens~ 0.3 gm/m and I thought this value quite difficult to 
achieve. However, today I got my hands on a piece of an Al vapor-
deposited mylar film used in winded capacitors. This is incredably 
light and strong, and its conductivity is quite high (1 cm distance 
contact gives ~10 Ohm). Now, its area specific density is 
0.6 mg / cm^2. This means that 1 m of 1 cm wide stripe of such a film 
will weigh 0.063 gm ! This is twice better then the max_dens 
caclulated. 
Remembering that 50-gauge corona wire weights almost nothing, we can 
see that 50 gauge wire/ mylar Al-plated collector combination can 
satisfy self-powered lifter density requirement quite easily. This 
film is also widely available - any dry winded capacitor (low voltage 
one) would have 100 m of it, and they are very cheap.

I looked more closely on possible design based on Al-coated mylar 
collector. The design could be like this (Exact dimension 
calculations depending on voltage and particular electrode density 
can be made with MathCad worksheet you can download from my site,
http://sudy_zhenja.tripod.com/lifter_theory):

- Electrode array lenght L= 800 m
- Saviour power supply with Li-ion batteries (~600 gm, 50 kV, 150 W)
- collector is made from twisted to tube-form 1cm wide Mylar Al-
coated film from condensors
- Corona/collector distance 70 cm
- Corona/collector couple arrays have distance between each other 10 
cm (total area would be 8m x 8m if flat)
- Corona/collector arrays are stacked, 8 arrays one above another. 
This reduces the base cross-section to 3m x 3m, and hight of the 
device is 6 m.

I belive to privide the lightest possible frame and more easy handwork, it is useful to employ stressed barrel configuration.
To visualise this idea more clearly, I made a draft drawing (click on picure to increase)

Anyway, in the beginning some smaller prototype could be build which 
would test the main design parameters:
- the twisted as spiral to tube-form 1cm wide Al-coated mylar film
as collector
- 70 cm distance between electrodes
- 50 gauge wire

To clarify the idea additionaly, I add here my answers to the questions about this design, asked by Rolland:

>1. As I understand the design you are proposing, the "basic" lifter 
>element is a AL coated mylar tube with a circumference of 1 cm and a 
>length of 3 meters. Thus the tube diameter would be 1/Pi cm or 
>approximately .32 cm. Correct?.

If we use circular (barrel) design rather then square design, one stack (or layer)
will include several wire/collector circles with different diameters. The lengh of each circle will be
L[i]=2*pi*(R-dr*i) where R is the biggest radius, dr - distance between wires 10 cm, and i=0..ceil(R/dr)

The largest radius can be approximately found from summar lengh  L of all circles in a stack as

The L should be found as Ltot/Nstack. I found that for Ltot 900m the Nstack = 7
(seven stacks) gives radius ~2 m with hight 5.7 m.

>2. A series of these tubes (around 31 of them) would be "laid" 
>parallel to each other and approximately 10 cm apart (measured from 
>the centers of adjacent tubes). This then forms the "bottom" square 
>of our 3m by 3m by 6 m lifter. Is this correct?
Yes, in square case. For barrel case see above.

>3. At 70 cm above these parallel tubes would run 31 parallel 50 gage 
>steel wires, one wire above each tube. Is this correct?
>
>4. At some distance above the steel wires, (presumably greater than 
>70 cm ?) , we would start over again with 31 parallel tubes oriented 
>in the same direction as the bottom layer and above them at 70 cm the 
>31 wires. Is this correct?

I believe that the best configuration would be with exchanging wire/collectro polarity, like this
(from the top):
- wire +
- 70 cm distance
-collector (-)
- 10 cm distnce
- wire (-)
- 70 cm distance
- collector (+)
- 10 cm distnce
- wire (+)
... so on.

Saviour already tested similar stacking, In his case the corona-wire for next level was replaced by sharp edge of 
foil- collector from "upper" level. I put a actual wire of same polarity 10 cm below the collector because 
our collector is made as at tube to prevent counter-current, and so has no sharp edges.

>I am sure I have missed something, because to get 8 such layers, I 
>would come out with a total height greater than 6 meters. Maybe you 
>can clarify.

h= Nstack*(10cm+70cm)
For barrel configuration h = 5.6 m

>How do you propose to maintain the shape of the mylar tubes? Are 
>they wrapped around a tubular core (Styrofoam? , Aerogel?) or around 
>a hollow straw-like plastic tube? I think they have to be very rigid 
>over a 3 meter length. Is this correct? 
For barrel configuration the mylar tybe will be in contact with "guiding tehters" at quite small intervals.
I think they can be glued to the tethers at contact points with acryle instant glue. As for supporting the tubular form of twisted 
Al-coated mylar film - I believe it can be avoided. First, the mylar film is quire rigid, and holds tube form
if 1 cm wide stripe is twisted. At the other hand, when we switch ON high-voltage, electrostatic repulsion will
foce the stripe surfaces to be as far as possible from each other, so forcing it to take tube form.

>If so, have you worked out 
>the weight, so we have a reasonable idea what one tube with its 
>internal "supporting element" and one "corona" wire would actually 
>weigh?

We have some free space for additional "pay-load" such as frame, because the Al-mylar is
way under-critical (0.06 gm/m vs max_dens=0.3 gm/m). Calculation shows that maximal payload (additional to 0.66 kg 
Saviour power-supply) which lifter with 0.06 gm/m electrode density can carry at V=100kV is 0.367 kg - so that is the maximal weight of the
frame we can afford. In this (worst) case of heaviest possible frame we will have electrode lengh 8.7 km. So I guess we need to do
much better then 367 gm frame. For further calcs I will assume 100 gm.

But how light we can actually  make the fram depends on material and design. It needs some experiment to find out 
fram dencity per volume before I can put this numbers into calculation.

I believe we can use very thin nylon silk-cloth tubes (~0.5 cm diamter) impregnated by epoxy as light composite material for frame, and 
thin nylon silk thread as guiding teathers. 

>Have you tried to work out an estimated total weight of the lifter 
>with power supply, even with "best case" assumptions regarding tube 
>weights and the weight of a minimal supporting structure? 

If I assume that whole frame and threads weigh together 100 gm (payload), the resulting design parameters 
are like this (calculated for 100kV):

Note that adding additionoal payload gives a need to increase efficiency, as power supply is not changed.
Therefor d1 increased from 70 cm without payload to 90 cm with payload.

Now, same caculations for 50kv case:

 

As you can see, the electrode lengh increases at lower voltage, but total hight can be significantly 
reduced - which means that at lower voltage we have higher volumetric force 
density, which might in turn allow to use lighter frame (frame density is volume related).

>Are you of the opinion that a series of lifter elements would generate 
>about the same amount of lift if they were either run side by side or 
>if they were "stacked" one on top of another (at some appropriate 
>spacing). That is, does lifting power just ADD as the number of 
>elements increases no matter how they are oriented? 

Mostly I think it just adds. Saviour recently mentioned good idea that with pressure
k decreases, so efficiency ef=d/(k*V) could increase if higher stacks increase pressure
for lower stacks. However, my equations assume that air is not moving. For high wind rates, comparable with
ion velocity (~300m/sec) all equaitons are no longer valid (ion flight time between electrodes will become shorter) 
so such simple logic might not work and more exact calcs considering air velocity are needed.

>I agree that using your ideas, we could proceed in a step-by-step 
>fashion, measuring and testing at each phase. First build a single 
>tube or partial tube element and measure results, then a series, 
>(maybe 5 or so) then a small stacked array, etc. 

Yes, and also testing of + -- ++ -- stacking would be interesting.

Regards,
Evgenij