Large solar array support beams

[s/v Jedi]

Too late to edit:

The 1708 is equal to 650gsm of carbon fiber for strength, but only equal to 300gsm of CF for flex. But that should still be enough I think…

[Lee Jerry]

Originally Posted by s/v Jedi

Thank you! I think many others will follow with the wish for large solar arrays and this thread will really help for diy

You need to remember that these (probably) aren’t absolute numbers due to all of the assumptions made to cover all of the variables. The purpose is just to give you an idea for sizing. It's better to use them for “comparative engineering,” which means that you look at the effect of change in an input (say, beam size) and see the resulting change it gives in deflection (or strength).

I have looked at materials that I can easily source from a local shop and Jamestown Distributors and have this proposal:

Beam with integrated stud for the first 1000mm, then unsupported for the next 1100mm.

For the core I plan to use three layers of 3/4” thick Coosa board, so 57mm wide. Keeping the 2:1 ratio you like (I normally use 1.618, anyone guess why?

I'm sure you have some PHIlosophical reason.

...but that’s not enough height for this application) then the core height is 114mm.

You could increase the core width to 3" (76mm) and use your gilded fraction to get a height of 4.9" (123mm).

But I want to taper the beam. Starting at 114mm I would like to taper that down to 57mm at the end (so the end is square).

With the fiberglass skins added, this approaches 60x120. The Coosa tensile strength is insignificant; it can barely carry itself.

The glass/carbon fiber layup schedule, I’m trying to compensate a bit for the tapering as well as core dimensions:

1. 1708 (equals 650gsm carbon fiber biax)
2. CF twill weave 203gsm
3. Extra layer like 4
4. Extra layer like 4
5. E-glass cloth 6oz equals 145gsm carbon std weave

Total layup equals 1400gsm of carbon fiber. I selected a twill weave to get it to form nicer over roundover edges and I see it has the same strength properties as regular weave.

How did you get the “equals carbon" values?

I think you are way, way over estimating the effectiveness of the fiberglass in this application (i.e. mixed with carbon). The problem is that even though they have similar strength (~400-450 MPa for fiber, ~500-600 MPa for carbon) and density (~1.9 g/cc for fiber, ~1.6 g/cc for carbon), the modulus is very different (~25 GPa for fiber fabric, ~70 GPa for carbon fabric). Thus, the carbon is ~3x stiffer than the glass. What this means is that since the two will be bonded together, when loaded they will see the same strain (i.e. elongation), but very different stress (tensile load). I have two analogies to help explain.

Consider two springs, Spring A is soft (low spring constant of “k”) and Spring B is stiff (high constant of “3k”). When the springs are deflected the same distance, Spring B exerts 3x the force of Spring A.

Or more of a sailing reference, compare two lines, the first is high tech like Vectran, the second basic 3-strand twisted nylon. Secure to the same anchor points, say cleat on dock and cleat on boat, with the same initial tension and then push/move the boat – which line is taking the load?

While I would (and have) advise against it, if you were to include the glass in the layup, the equivalent to carbon would be more like 172 gsm and 60 gsm for the 1708 and E-glass, respectively. Therefore, the total equivalent as defined is only 840gsm and you will have to add some more layers of carbon.

For unidirectional CF on the edges, I can easily source 318gsm from Jamestown Distributors/ TotalBoat and I wouldn’t mind using two layers instead of one. Where in the laminate schedule would it be best to add these two layers? Separately or together and between layers 3 and 4?

I would either put them as the furthest outer layers (of carbon) or alternate cloth - uni - cloth -uni - cloth, but I'm not sure which, if either, is better.

Also, I plan to use a 1/2” round-over router bit for the edges and think I can put the unidirectional very neatly over the edge. I wouldn’t mind to use one layer 100mm wide and lay it over the top so coming down 20mm each side and the second layer 120mm wide to overlap the first layer by 10mm on each side. This would almost double the unidirectional CF at the top.

For the bottom edge I plan to only cover the roundover itself, making it flush with the sides. With CF a bit weaker in compression, should I add another layer to compensate? The top would get a total of 220mm width unidirectional and three layers on the bottom would total 180mm width.

I strongly recommend that you make the beam symmetric top to bottom, i.e. the same layup, so that the neutral axis is in the middle (or as close as possible). I think that is a bigger factor than the relative tensile - compressive strength ratio in terms of where failure will occur. I can elaborate if desired.

The factor of safety >10 is mind blowing. Does this really mean it would take 500kg to break it? This means testing is relatively safe, as I would only be testing deflection without risking breaking it.

I am really curious where these proposed beam dimensions and layup schedule would get me and if you can do the math for the tapering :-) I think the tapering doesn’t matter for strength because it follows the load curve but it will have a big impact on deflection. I hope to have countered that with more unidirectional primarily and a little more for the rest to balance it out, although I have a feeling it may need another layer of 200gsm CF…

Yes, that's basically what the FoS means (if the assumptions and calcs are close enough). W.r.t. the testing, remember that the deflection is with regard to the deformed shape of the beam only and does not account for any slop in the system (such as tension building in lashings).

I'll have to get back to you on the effect of the taper.

[s/v Jedi]

Originally Posted by Lee Jerry

You need to remember that these (probably) aren’t absolute numbers due to all of the assumptions made to cover all of the variables. The purpose is just to give you an idea for sizing. It's better to use them for “comparative engineering,” which means that you look at the effect of change in an input (say, beam size) and see the resulting change it gives in deflection (or strength).

Exactly. I will also document my build and tests so that others get a good starting point of data. Most will be using the same size panels and similar arch constructions.

How did you get the “equals carbon" values?

I corrected using a factor for the difference in strength and flex.

I understand your example with the two ropes but thought it could be compensated by using three of the same nylon ropes instead of just one (850gsm instead of 300gsm), but I guess this isn’t enough to make up the difference.

I will have to find a new supplier then because neither my local shop nor Jamestown have biax CF…

I would either put them as the furthest outer layers (of carbon) or alternate cloth - uni - cloth -uni - cloth, but I'm not sure which, if either, is better.

I strongly recommend that you make the beam symmetric top to bottom, i.e. the same layup, so that the neutral axis is in the middle (or as close as possible). I think that is a bigger factor than the relative tensile - compressive strength ratio in terms of where failure will occur. I can elaborate if desired.

Yes, bring the unidirectional as far apart as possible. Putting it between woven layers may give it more sheer strength. Probably the new supplier I’m gonna find also has the higher weight woven CF so I’ll redo this when I find specific materials.

I see the need to keep it symmetrical. On the bottom it will be a problem at the roundover because the distance is different and fibers will buckle, so I need to stay in between the roundover edges, which reduces the width of the unidirectional reinforcement.

Apart from repairing a spinnaker pole, I have never worked with carbon fiber, but from watching others I think it may be possible to lay a twill weave over a 1/4” radius corner, which optimizes the width for the unidirectional fibers.

Yes, that's basically what the FoS means (if the assumptions and calcs are close enough). W.r.t. the testing, remember that the deflection is with regard to the deformed shape of the beam only and does not account for any slop in the system (such as tension building in lashings).

I'll have to get back to you on the effect of the taper.

Dyneema stretches in two ways: constructional stretch which is the braid being pulled flat, which with a lashing is taken out during tightening the lashing and creep which should add length over time. I have experienced creep in my lifelines but not before with lashings. This is something to keep an eye at for sure.

For the taper: I am curious if it can be compensated for with 300gsm extra unidirectional. I think the side walls don’t need extra fiber for the taper? For designing sails they have software showing in colors where stresses are high and low and I guess similar software exists for designing beams, trusses etc. It’s fascinating

[s/v Jedi]

I have a hard time finding a supplier for carbon fiber biaxial cloth. Most don’t have it at all, some list it as sold out etc.

I am now considering using regular cloth turned 45 degrees? But that would create a seam…

I see some on eBay but fear they will ship folded etc.

[s/v Jedi]

I just realized that ai could use a regular weave 3” wide tape and wrap it, which gives the +/-45 degree fiber orientation…

[Lee Jerry]

Originally Posted by s/v Jedi

I corrected using a factor for the difference in strength and flex.

I understand your example with the two ropes but thought it could be compensated by using three of the same nylon ropes instead of just one (850gsm instead of 300gsm), but I guess this isn’t enough to make up the difference.

Yes, you could compensate that way and make up the difference (roughly). But why would you? Why have three lines each working at ~30% of strength to match with the other stiffer one? It would be better to either:
- use two Vectan lines if stiffness is important,
- or pick between two Vectran or two nylon lines if stiffness is not important (since they have approximately the same strength).
It seems that mixing them gives you the worst of both without any of the benefits, IMHO.

I will have to find a new supplier then because neither my local shop nor Jamestown have biax CF…

I assume you're in the USA. But the Easy Components company in the UK referenced above shows it available; couldn't hurt to inquire about cost and time to ship.

For the taper: I am curious if it can be compensated for with 300gsm extra unidirectional. I think the side walls don’t need extra fiber for the taper? For designing sails they have software showing in colors where stresses are high and low and I guess similar software exists for designing beams, trusses etc. It’s fascinating

It looks like a linear taper to 1/2 height (with constant width and layup) will increase deflection by 50% compared to a straight beam.

You could compensate (at least somewhat) by increasing the amount of carbon. However, I think it would be easier / cheaper / better to just increase the initial dimensions of the beam by 15% - 20% and then get roughly the same deflection. That is a 115cm beam tapering to 57.5cm has roughly the same deflection as a 100cm straight beam.

N.B. For both comments above, this applies at least around the beam size, loading and strength (construction layup) of beams we're looking at in this application. I'm not sure how they would apply generally to significantly different cases.

[fourlyons]

Originally Posted by s/v Jedi

I have a hard time finding a supplier for carbon fiber biaxial cloth. Most don’t have it at all, some list it as sold out etc.



I am now considering using regular cloth turned 45 degrees? But that would create a seam…



I see some on eBay but fear they will ship folded etc.

Jedi you might contact Composite Envisions in Wisconsin. I just bought a bunch of Hexel IM2 Uni, intermediate modulus carbon at a very good price. They told me that for some reason there's been a big increase in the price of Toray recently, which made the higher performance Hexel actually cheaper. When I looked, they had many different carbon (and other) fabrics. I just checked, they do have biax in stock.

[s/v Jedi]

Originally Posted by Lee Jerry

Yes, you could compensate that way and make up the difference (roughly). But why would you? Why have three lines each working at ~30% of strength to match with the other stiffer one? It would be better to either:
- use two Vectan lines if stiffness is important,
- or pick between two Vectran or two nylon lines if stiffness is not important (since they have approximately the same strength).
It seems that mixing them gives you the worst of both without any of the benefits, IMHO.

I assume you're in the USA. But the Easy Components company in the UK referenced above shows it available; couldn't hurt to inquire about cost and time to ship.

It looks like a linear taper to 1/2 height (with constant width and layup) will increase deflection by 50% compared to a straight beam.

You could compensate (at least somewhat) by increasing the amount of carbon. However, I think it would be easier / cheaper / better to just increase the initial dimensions of the beam by 15% - 20% and then get roughly the same deflection. That is a 115cm beam tapering to 57.5cm has roughly the same deflection as a 100cm straight beam.

N.B. For both comments above, this applies at least around the beam size, loading and strength (construction layup) of beams we're looking at in this application. I'm not sure how they would apply generally to significantly different cases.

Thank you

I have been studying a lot and working out construction details. I think I just got it worked out and of course completely different once again. I think I’m ending up where you basically told me to start

I read about the importance of biaxial fibers for the web but as well for the core. Also, the biaxial should wrap around the corners to transfer forces better. This isn’t easy to make and I think I have decided to:

1. make a continuous beam 57x114mm core, non tapered and with separate strut. This allows me to make the beam easier as I can just wrap the cloth around it.

2. laminate the core from three layers of 3/4” Coosa board with biax in between the laminations. Coosa board lacks biax fiber.

3. I can insulate from metals using G10 fiberglass pads that are also shaped for the stainless steel tubes and these can be glued to the beam so I can skip the wrap in fiberglass cloth. I can use shims with these pads for aligning everything.

4. I finally found a supplier with 12 ounce biaxial: https://fiberglasssupply.com/12oz-do...-fiber-fabric/

I am in Europe now but will be building these beams in Florida.

5. I can buy carbon fiber pipe 1” i.d. for the struts, with end fittings to help make the connections and transfer the load of the pipe to a center attachment.

6. I will use carbon fiber pipe 0.75” i.d. for lining the holes in the center of the beam that hold the lashings. Also, fiberglass washers glued over the pipe openings for reinforcement and shaped to prevent chafing of the lashings.

I’m gonna think it all over again, make new sketches etc. so that’ll take a couple days.

[s/v Jedi]

Originally Posted by fourlyons

Jedi you might contact Composite Envisions in Wisconsin. I just bought a bunch of Hexel IM2 Uni, intermediate modulus carbon at a very good price. They told me that for some reason there's been a big increase in the price of Toray recently, which made the higher performance Hexel actually cheaper. When I looked, they had many different carbon (and other) fabrics. I just checked, they do have biax in stock.

Oh that’s a good source, thank you! It seems ai need to search aerospace suppliers rather than marine

They have everything.

[s/v Jedi]

Originally Posted by Lee Jerry

I assume you're in the USA. But the Easy Components company in the UK referenced above shows it available; couldn't hurt to inquire about cost and time to ship.

Okay, I worked it out, found all the materials and did the puzzle.

For the core, I will build it a little longer, then after lamination, cut each end to get into good laminate and the 210cm length, then finish the caps with a small piece of carbon fiber veneer, finished with a flush trim bit in the router.

The core itself is a lamination of three plies of 20mm Coosa board. This has a shear strength of 400-500psi. If this isn’t enough, I could add fiberglass plies in between the Coosa boards.

After laminating I will clean it up and use a 1/2” radius round-over bit for the long edges to help wrapping the cloth around the corners.

This creates a 120x60mm core. I will build a jig that holds the core horizontally between a couple big nails so that I can turn it while wrapping the fiber.

For the carbon fiber I went for the lighter weight cloth and do an extra wrap. This allows me to balance out the seams on each side panel. Here is the schedule:

1. 193gsm plain weave
2. 193gsm plain weave
3. 200gsm biax
4. 200gsm biax
5. 404gsm uni top and bottom only
6. 193gsm plain weave
7. 404gsm uni top and bottom only
8. 193gsm plain weave

Total 1,172gsm sides and 1,980gsm top and bottom. The cost of the carbon fiber will be $800 from https://compositeenvisions.com/produ.../carbon-fiber/

Can you check deflection and safety factor for the 110cm cantilever beam? I’m hoping for 5mm deflection and a safety factor well over 10

[nfbr]

I recommend glass mounts for where panels attach to carbon or you’ll have rapid electrolosis.
My 3.6kw carbon arch weighs 55kg with 7ft legs on a catamaran.

[Lee Jerry]

Originally Posted by s/v Jedi

I am in Europe now but will be building these beams in Florida.

Where and when in FL? Or more importantly, indoors or out? FL summers, which has started, are hot, which can make it difficult with epoxy. If working outside, use the slowest hardener your chosen system offers (West System has an "Extra slow hardener"). And probably plan on working mornings or evenings, while mixing small batches.

Originally Posted by s/v Jedi

Okay, I worked it out, found all the materials and did the puzzle.

For the core, I will build it a little longer, then after lamination, cut each end to get into good laminate and the 210cm length, then finish the caps with a small piece of carbon fiber veneer, finished with a flush trim bit in the router.

The core itself is a lamination of three plies of 20mm Coosa board. This has a shear strength of 400-500psi. If this isn’t enough, I could add fiberglass plies in between the Coosa boards.

After laminating I will clean it up and use a 1/2” radius round-over bit for the long edges to help wrapping the cloth around the corners.

This creates a 120x60mm core. I will build a jig that holds the core horizontally between a couple big nails so that I can turn it while wrapping the fiber.

For the carbon fiber I went for the lighter weight cloth and do an extra wrap. This allows me to balance out the seams on each side panel. Here is the schedule:

1. 193gsm plain weave
2. 193gsm plain weave
3. 200gsm biax
4. 200gsm biax
5. 404gsm uni top and bottom only
6. 193gsm plain weave
7. 404gsm uni top and bottom only
8. 193gsm plain weave

Total 1,172gsm sides and 1,980gsm top and bottom. The cost of the carbon fiber will be $800 from https://compositeenvisions.com/produ.../carbon-fiber/

Can you check deflection and safety factor for the 110cm cantilever beam? I’m hoping for 5mm deflection and a safety factor well over 10

That sounds like you've come up with a good plan.

I think you'll be right in the performance you're looking for. I'm getting ~4.7mm on deflection and >15 bending stress FoS. And I expect that (or I've tried to be) conservative with the calcs - deflection should actually be lower.


What are you planning for the strut attachment? More specifically, any additional material locally at the 1m mark (on the applicable sides) to distribute that load?

[s/v Jedi]

Originally Posted by nfbr

I recommend glass mounts for where panels attach to carbon or you’ll have rapid electrolosis.
My 3.6kw carbon arch weighs 55kg with 7ft legs on a catamaran.

You can read my mind

The panels will be mounted in fiberglass angle which is then lashed onto the carbon fiber beams.

Also, the beams mount on stainless steel tubes of the arch that is built on Jedi’s stern. There will be G10 fiberglass pads in between that adapt the round tube shape to the flat underside of the beam to spread the load as well as galvanically isolate the two.

Tell me more about your carbon arch! My array mounted on top will be 1.76kW so you double that on a cat I assume you have eight panels? How did you mount all that, do you have pictures?

[s/v Jedi]

Originally Posted by Lee Jerry

Where and when in FL? Or more importantly, indoors or out? FL summers, which has started, are hot, which can make it difficult with epoxy. If working outside, use the slowest hardener your chosen system offers (West System has an "Extra slow hardener"). And probably plan on working mornings or evenings, while mixing small batches.

Jedi is on the hard in Port Canaveral. Luckily we have a little hideout place in Titusville and one of the first things I did there was convert two garages to air conditioned workshops

One is normally a car garage and here I have a large workbench, table saw, bandsaw, drill press etc. so it’s where I make a lot of dust and the other is normally a storage/ utility room with a golf cart garage where I made a 4 x 8 assembly/ paint station with dry racks etc. so this is the clean room. Both had A/C installed right after we bought it

I always used West System as this was easy to get in the Caribbean but now in Florida I have been buying more TotalBoat epoxy because it comes in plastic containers that stay nice. I recently even tried TotalBoat 3:1 epoxy resin with West System fast hardener because I was in a hurry and took the risk and like people told me it worked great.
For most work now I use TotalBoat 3:1 epoxy with their slow hardener and have a couple gallons at hand most of the time (buy when they have a special ). This gives me a full cure overnight in the A/C so that’s good for this project as well.

That sounds like you've come up with a good plan.

I think you'll be right in the performance you're looking for. I'm getting ~4.7mm on deflection and >15 bending stress FoS. And I expect that (or I've tried to be) conservative with the calcs - deflection should actually be lower.

That is great news! I had good hopes that the extra layer of unidirectional would achieve this

What are you planning for the strut attachment? More specifically, any additional material locally at the 1m mark (on the applicable sides) to distribute that load?

I plan to buy G10 3” wide by 1.5” thick bar stock and cut/mill supporting pieces like mounting pads and the strut attachment from it. For the strut attachment I also hope to extend the support from the strut further up the beam.

I do not know the angle that the strut comes up with yet but imagine it to be similar to our hydraulic boom vangs. From the G10 bar stock I want to cut triangles that have the longest side up against the beam, the shortest side facing the strut and here a 90 degree corner towards the third side that spreads the load further along the beam. The shape of a boom vang attachment.

At the back end, at the sharp corner, i will cut a piece off to get a flat surface perpendicular to the beam, so that I can glue a stop block to the arch to keep the triangle in place. Between the beam and the G10 block I plan to put “carpet tape” as a cushion to deal with any irregularities between the two surfaces and keep it a little in place, with one or two Dyneema lashings around the beam as the main fastener. Maybe I will glue a couple small side rails to the underside of the beam for the block to sit in between.

Between the triangle and the stop block I can put shims for tuning. Also, where the strut mounts to the arch tube, I plan to have a stainless steel M10 threaded rod for length adjustment.

As I am sketching this, I found that none of the lashings interfere with anything so I can skip the holes through the center of the beams and simply lash all around them. This will double the length of the lashings but I don’t expect any trouble there and can simply re-tension when needed.
I will need to drill at least one hole through the triangle block for the lashing.

I am amazed about the strength of the tubing from DragonPlate. Here is the end connector that I want to thread for M10: https://dragonplate.com/1-end-connector-blank
The other end, at the G10 block: https://dragonplate.com/1-male-clevis-connector
will engage with clevis jaws that I will machine in the G10 so I can put a 1/4” clevis pin through.
Here is the tubing: https://dragonplate.com/axially-opti...tube-1-id-x-96

[Grith]

Lots of fancy stuff being done here. Congratulations.
Personally I just added 540w of thin frame hard panels over my existing Bimini by adding an additional rear stainless tube and using thumbscrew clamp style mounts to securely attach the three panels rear edge and the same clamps over the canvas covered front Bimini tube to attached the front.
I then replaced the front tensioning straps for the Bimini with matching stainless tubes providing a more solid front support.
Combined with my pre-existing 180w identical tilting panel attached to my rear targa bar I now have 720w of unshaded and easily removable thin frame hard panels on my 28 foot trailerable yacht.
This is in addition to my tiltable cabin roof top panel attached with the same clamps allowing port starboard tilt adjustment at anchor and my dodger top flexible panel which moves up to on top of my bagged main.
Total is just under 1kw on a 28 foot trailerable yacht set up for extended coastal and inland waters expeditions.
To date the whole setup has survived a several hour crossing with winds over 35knots with an exceptionally short steep swell on Lake Alexandrina here in South Australia.
The Bimini mounted solar array power two lithium EcoFlow Delta portable power banks allowing me the equivalent of being connected to shore power whilst out remote cruising. These are in addition the yachts usual AGM house starter bank running all the regular systems charged by the other two panels.
We cook via induction, our hot water is an 800w 240v element electric storage system (outboard so no heat exchanger) , fridge and separate freezer are electric compressor type (again no inboard for eutectic) and we also charge our torqeedo light winds auxiliary batteries which also doubles as our dingy outboard and a variety of other occasional 240v appliances.
The EcoFlow units have been connected to our shore power system in our fuse box with an electrician fitted change over switch.
The main other unusual expense was the Ferris plastic thumbscrew clamp style panel mounts. They do however mean that all the panels can be easily removed and stowed below if really bad weather threatens.
The old iPhone shown below (degraded battery so permanently on charger cord) is providing a dedicated remote charging/discharging and state of charge monitoring and output remote switching for the EcoFlow units which are located in plastic boxes attached to access hatches under the dinette seats on both sides of the yacht. The green lit switch is the hot water service on and heating. Input and output and state of charge can be read on the photo.