The Hollander Beater: Researching the Hollander Beater with Claude
So significant advancements have happened since I have started trying to build the Hollander Beater.
- My time has become very limited and life has been giving me lemons.
- Claude and the usecases of AI have become significantly easier to use.
- I have gained significant experience in handling power tools.
So maybe now is a better time to build a more professional, yet affordable Hollander Beater.
One of the core aspects of building the beater was conceptualising a method to design and build the roll which would be strong enough to withstand the strain.
As well as durable enough to be used long enough without breaking. I have decided against many ideas, however the most effective way in both cost and time would be to design it in FreeCAD and then have it CNC cut by professionals. Unfortunately this is the most expensive part of the whole project. But if there is a part of the project that I wouldn't mind spending money on it would be this, since it is a critical component and may be worthwhile investing in.
Saying all of this I have no experience with CAD, and even though I am a techie. It would require a considerable amount of effort and learning before something could become viable. But I am hoping that with the advancements of Claude I would be able to find a way to design this in CAD which would be able to be cut and manufacturer in a reasonable amount of time.
So let me give this a go:
Some other information worth noting.
- This is still the best source of information I have about Hollander Beaters.
https://www.youtube.com/playlist?list=PL6fkyjWkcvKGKSvlipar9ZvxegsLm01fa
- VG Paper has a blog which does go into quite a lot of depth of problems they ran into when building theirs initially which I think is also highly valuable.
https://vgpaper.com/blog2.html
Instructions for Claude:
Prompt
I want to design a hollander beater for a DIY project at home. I am thinking of using wood and sheet metal with pop rivets. The fibre glass the side walls. Can you give some advice.
Claude's Response
Great topic — a Hollander beater is a genuinely achievable DIY project, and your material choices (wood, sheet metal, fibreglass) are well-suited to it. Let me give you practical advice alongside a diagram of the key components.
Here's a breakdown of the key design and construction considerations for each part:
Tub (the main body)
A D-shaped or oval tub with a central dividing wall (the midfeather) is the classic form. For your wood + fibreglass approach:
- Build the tub frame from 18–22mm marine ply or solid hardwood. The oval shape is best achieved by laminating thin strips (3–5mm ply) around a former, or by cutting curved staves and joining them.
- The midfeather can be a solid wood panel, notched into the tub walls and braced underneath.
- Once the wood structure is sound and dry, lay 2–3 layers of 200–300g/m² woven fibreglass cloth with epoxy resin (not polyester — epoxy is more water-resistant and bonds better to wood). Wet sand between coats and finish with a gel coat or epoxy topcoat. This creates a watertight, smooth interior that won't absorb water and is easy to clean.
- The tub needs a drain bung at the lowest point, and a water inlet.
The roll (beater roll)
This is the most mechanically demanding part. A typical small roll is 150–250mm diameter and 200–300mm wide. Your sheet metal + pop rivet approach is workable for a light-duty beater:
- Fabricate the roll cylinder from 1.0–1.5mm mild steel sheet rolled into a tube, with end caps cut from thicker plate (3mm). Pop rivets can join the longitudinal seam and end caps, but weld if you have access — it's far stronger and watertight.
- The bars (the bars on the roll surface that do the actual beating) are typically mild steel flat bar (e.g. 25×6mm) pop-riveted or bolted around the roll surface at slight angle. Space them evenly — 8 to 12 bars is typical for a small roll.
- The shaft needs to run through the centre on proper pillow block bearings mounted to the tub sides or a wooden bearing bridge above the tub. Alignment here is critical — any wobble will destroy the bearings quickly.
- Drive with a motor via a V-belt and pulley. A 0.5–1kW single-phase motor is usually sufficient for small batches. Aim for roll surface speed of roughly 3–5 m/s.
The bedplate
This is a flat bar of steel (say 50×10mm mild steel) fixed to the tub floor directly beneath the roll, with a small adjustable gap between it and the roll surface. This gap controls how aggressively the fibre is beaten — too close and you cut the fibres; too open and you do nothing. A simple adjustment mechanism using bolts through slotted holes works well.
Fibreglass tips
- Surface prep is everything — sand the wood to 120 grit, wipe with acetone, apply an epoxy seal coat before the cloth.
- Work in sections if the tub is large, overlapping cloth by 50mm.
- Inside corners are a weak point — lay a 50mm tape of glass first before the main laminate, to avoid voids.
- Once cured, use a gelcoat or two-part epoxy topcoat for a smooth, cleanable surface.
A few practical cautions
- The roll creates significant vibration and torque — mount the whole unit on a heavy base (a thick timber sled or steel frame bolted to the floor) to prevent walking.
- Pop rivets are fine for the roll bars but make sure you use stainless or monel rivets — standard aluminium rivets will corrode rapidly in a wet paper pulp environment.
- Keep the roll-to-bedplate gap adjustment accessible from outside the tub.
Would you like more detail on any particular part — the roll fabrication, the drive system, or the fibreglass layup process?
Prompt
How can I design a hollander beater in CAD or are there already open source CAD drawings for hollander beaters.
Claude's Response
Good news — I found the VG Paper math and tub pages, which are clearly from the same series you're following. Here's a full picture of what exists and what your best path forward looks like:
Are there ready-made open source CAD files for a Hollander beater?
The short answer is: not really. The VG Paper author explicitly notes there are no sites with exact plans for building a Hollander beater — it was up to them to work out all the mechanics and ratios from scratch. This is echoed across the papermaking community. Only rare individuals who know their way around a machine shop study the workings of a beater and attempt to build their own. No dedicated CAD files exist on GrabCAD, Thingiverse, or similar repositories for this machine.
The closest thing to open documentation is the VG Paper blog series you're already using, plus the Peace Paper Project — they set out to design and build a DIY Hollander beater for under $1,000 using locally sourced materials, with the finished result intended to be published as an open source file on their website. It's worth checking drewmatott.com and the Peace Paper Project site to see if those files were ever released.
Your best approach: draw it yourself in CAD
Since no ready-made files exist, you'll be drawing from the VG Paper dimensions. Here's the recommended workflow:
CAD software to use (all free)
For a project like this — mixed wood, sheet metal, and fibreglass — you want parametric 3D CAD so you can adjust dimensions as you iterate. The best free options:
- FreeCAD (freecad.org) — fully open source, parametric, runs on Windows/Mac/Linux. Best for this kind of fabrication project. Has a Sheet Metal workbench for your rolled components, and a Part Design workbench for the tub.
- Fusion 360 — free for hobbyists, more polished UI than FreeCAD, excellent sheet metal tools. Cloud-based with some file format restrictions.
- OnShape — browser-based, free for public projects, very capable. Good if you want to share your design with the papermaking community afterwards.
For 2D drawings only (if you just want cutting templates and flat patterns):
- LibreCAD — free, simple, good for dimensioned 2D plans you can print and use as templates directly.
How to structure the CAD model
Based on the VG Paper approach, model it as separate components in an assembly:
The tub body is the most complex part — VG Paper looked at stack-laminating Baltic Birch plywood to get the curved shape, and also explored coopering the tub together from angled boards like a barrel, using cheaper 3/4" fir, planning to fibreglass it for waterproofing. The stave angles went through several iterations — 22.75°, 15°, and finally around 10°.
Model this in FreeCAD using the Part Design workbench: start with the oval/D-shaped profile as a 2D sketch, then use the Loft or Sweep tool to generate the tub walls.
For the drive and shaft, VG Paper worked out the pulley ratio using the formula d1×n1 = d2×n2, starting from a 1725 RPM motor with a 3.5" pulley (standard on an evaporative cooler motor), aiming for the drum to spin between a target RPM range. Model the pulleys parametrically so you can tweak the ratio without redrawing everything.
For the tub volume, VG Paper calculated that a tub 42.25" long, 13.5" wide and 6" tall gives about 14.8 gallons, targeting a 2 lb dry material capacity with 10–12 gallons of water. Put these as named parameters at the top of your model so you can scale the whole design.
Would you like help working through the specific dimensions for any component — tub, roll, bedplate, or drive system — so you can start the CAD geometry?
Next Steps
In the next blog post I outline the steps I took when designing the roll in FreeCAD with Claude.