Contents

I’m in the season of life where everyone is having kids, and I’ve been enjoying doing some light woodworking recently (I’ve discovered I need to regularly make or fix things with my hands in order to feel useful…).

So I decided to make some simple wooden toys.

I jumped straight to it, heading to the lumber yard, picking up some cool hardwoods, and cutting, routing, and sanding my first pieces. Then I got to finishing them, and the thought occurred to me: “infants will put these in their mouths. I wonder if I need to be especially picky about what I put on them.”

I started to research and it turns out the answer is a pretty emphatic “yes.” And not only that, but there are a whole bunch of other child-safety considerations the novice toymaker should be aware of. So I’m writing this post for those with the same idea and same lack of experience.

At the end I have a quick outline for one simple toy you can make (this etched cube) — more to come as I keep working on this project!

Safe wooden toy guidelines

Safety-wise, I’d break my findings down into four categories: woods, finishes, glues, and specs. Here’s the bottom-line summary of recommendations, but there’s much more detail below:

  • Woods: maple, beech, cherry, oak, ash, and birch
  • Finishes: 100% US-grown tung oil and beeswax, or mineral oil and beeswax
  • Glues: avoid glue in favor of single pieces or mechanical joinery; otherwise: waterproof PVA glue approved for food contact
  • Specs: no part of toy should be able to fully fit into a toilet paper tube; sand it all to at least 220 grit; round over the edges

Before diving in: is this all really necessary? After all, kids literally stick their mouths on the handles of grocery carts.

I think it is probably fine most of the time to make toys outside the parameters listed. Giving a kid an exotic hardwood cube, or one finished with boiled linseed oil, is not likely to instantly cause some devastating malady.

But I figure: if you’re making toys for babies, you might as well do it in the best possible way you can. Or at least, that’s what I want to do.

I’m not arguing that everything outside these bounds is dangerous; I’m arguing that if you’re making toys for your or your friends’ children, you might as well optimize for the lowest plausible exposure, the simplest chemistry, and — frankly — the best vibes.

If you have any corrections or suggestions, let me know!

Woods

Wood is wood, right? Not so.

Different woods can have a range of (typically) low-level toxic effects, including as “sensitizers” — meaning that repeated exposure can steadily induce an increasing allergic reaction.

In practice, with the exception of really toxic woods (e.g. yew or oleander) or a specific allergy or elevated sensitivity, you’re not likely to suffer adverse effects from having finished wood around your home. The risk for most woods is higher when working with the raw material — inhaling the dusts, touching saps, etc.

But for kids toys — which will inevitably be stuck in their mouths — it’s best to stay with the really safe ones.

Here’s a good compilation of reported effects from various woods. In general, for kids toys, a good rule of thumb is to stick with domestic hardwoods.

Aside from the toxicity, you also want to avoid woods that splinter easily, or are too soft, or whose splinters go septic easily (e.g. greenheart), or are chemically treated.

Some woods that seem good based on my research (do your own!) include: maple, beech, cherry, oak, ash, walnutSome people avoid walnut due to tree nut allergy concerns, but I’ve seen little to no evidence of issues from cured / solid wood. But if you really want to be careful, I’d skip it., and birch.

Sadly that means the beautiful wenge pucks I started — with plans to make different shapes and string them on a rope — will have to stay out of the playroom.

(And it probably wasn’t even great for me to work with the wenge in the first place!)

Finishes

Finishes are a whole separate issue. Most are not designed with infants chewing on them in mind!

You can, of course, leave the woods unfinished. If you sand them really well (see “Specs” section below), this can be a fine solution. But especially if you’re using a wood on the softer side and the toy is for an infant who will drool all over it, this could result in some degradation.

I’ll probably leave some of my pieces unfinished. But for others, I want a slightly more finished look, and the water-resistance that comes with it.

There is a lot of contradictory information out there on the internet about what is safest. Below are my personal conclusions.

Your first question might be: can’t I just use food-safe oils and finishes? My view: no.

Aside from endemic quality control issues in the industry, most “food-safe” rating is targeted towards indirect consumption — like food sitting on a cutting board that is finished with the oil, then you eating that food. This is obviously a bit different from a kid full-on putting the piece in their mouth and gnawing on it. So I don’t think “food-safe” is quite enough.

Then you might ask: aren’t there regulatory standards around this? Yes, there are — and that may be good enough — but I went a different direction.

The most relevant standard is Europe’s EN 71-3 (“Safety of Toys - Part 3: Migration of Certain Elements”). It’s a rigorous certification that looks at heavy metals and other compounds and sets limits on how much can be present in toy-related products.

I think you’re probably good to go if you pick an EN 71-3 certified finish. There are a few out there. But I’d rather go for something simple and natural rather than something engineered to meet a standard with certain properties. Part of the appeal of making wooden kids toys in the first place is that very simplicity and naturalness.

(Am I succumbing to the naturalistic fallacy here? Perhaps! But if I’m already putting this much effort into making toys, I’d rather bias toward simple, well-understood ingredients. And again: vibes.)

An issue with the EN 71-3 certification is that, given the cost to the producer, nobody seems to get it for simple / natural products, since it wouldn’t really show anything additional beyond the proof that the product is pure (which can be done more cheaply). I couldn’t find a pure oil that had the certification.

So, then you might say: I’ll just grab the organic coconut oil or the olive oil off my shelf!

Well, that’s not a great idea either. Most oils will go rancid over time, especially if exposed to heat and sunlight and the stimulation from kids playing with the toys. They’ll also just end up dissipating and you’ll need to re-apply to the toys.

What you’re looking for is “drying oils” — which is a bit of a misnomer. These types of oils soak into wood and then polymerize (their molecular structure actually shifts as it cross-links into a solid inside/on the wood). This process is more like “curing” than “drying.”

The issue is that most common finishing oils — including “food-safe” ones, per above — are probably not the best for kids. They often use intense industrial solvents and dryers to get their desired characteristics.

For example, boiled linseed oil is commonly recommended for cutting boards. It cures quickly and works well, and is derived from flax seed (also known as linseed). But it isn’t just “boiled” — they add toxic drier chemicals to it.

There are less-treated oils — like raw linseed or “polymerized linseed” (again: not boiled!) — that can work. But you need to read the label carefully for additives, and they cure extremely slowly (think: weeks!).

The best option I’ve found is 100% tung oil, which comes from the nut of the tung tree. It cures/polymerizes well, faster than linseed oil. And I found what I think is the only US-based grower and manufacturer of it! Tallahassee Tung Oil. They grow in Georgia, USA and provide a certificate of analysis and guarantee of purity for each batch.

You may also want to thin your tung oil (it can be pretty thick and sticky). If you do, Tallahasse Tung Oil also sells a 100% natural citrus solvent made from pure orange peel oil. I’m still deciding if this is worth it. I don’t see much risk by the time curing is done (the solvent will have evaporated), but part of me wants to stick with just the tung oil, even though it’s harder to work…

Aside from tung oil, you could also go with USP (pharmaceutical-grade) mineral oil, which of course, people literally drink as a laxative. I don’t love the vibes of this (it’s a petroleum byproduct), but more importantly, it doesn’t polymerize. It is chemically stable and inert, so it doesn’t go rancid, but it stays a liquid inside the wood and may dissipate over time. This isn’t so much a safety issue as it is a maintenance one — ideally you’d need to re-apply it over time.

So I think a polymerizing oil like tung is the best option, even if it is a bit more of a pain to apply and let cure. If for some reason you’re on a real time crunch for toy production (for the elves reading this in mid-December, I feel you), then mineral oil is a good move.

After the oil, I like conditioning the wood with some wax. The best option here is organic beeswax. You can either melt and apply it directly, or create a melted mixture of tung oil and beeswax (or mineral oil and beeswax) and do a layer of that.

Last: there are other whole categories of finishes, like surface-forming ones (shellac / resin finishes generally) and paint.

I personally don’t love the look of the surface-forming ones for kids toys, so I haven’t gone too deep into the research there. My high level understanding is that shellac is fine as long as the solvent used is clean — like pure/diluted ethanol. Many shellac solvents have other chemicals in them or are dissolved in denatured alcohol, so they should not be used.

(Did you know shellac is a resin that lac bugs secrete, dissolved in alcohol? I didn’t!)

And on paints — that’s a whole separate blog post. I’m mainly excited about natural wood looks right now, so haven’t spent time on the paints. But maybe will in the future.

A final recommendation — which really applies to all of this: for most any engineered product, you should be able to get a material safety data sheet (also known as an SDS or MSDS). These list everything relevant to health in the product. If you’re unsure, get the SDS! Reading these is so much better than skimming Google or ChatGPT results.

Glues

I’m mostly trying to avoid glues entirely and stick to single-piece toys. It makes it a fun challenge. There’s also a lot than can be done with simply wood-only joinery techniques, or hidden fasteners (be careful!).

But if you want to use glues, make sure that they are interior-only (i.e. not exposed to the kid’s mouth). I think the safest bet is waterproof PVA glues approved for food contact like Titebond III (which won’t soften when soaked in saliva) — just make sure they are completely cured before you hand the toy over.

There are also hide glues (made from animal collagen). These appeal to my naturalistic side, but are not as water-resistant and are weaker. So on balance, I’d rather use a small amount of interior-only waterproof PVA glue and ensure the toy stays together.

Specs

Aside from the materials, you need to make sure to make something that is actually safe in its finished form! That means big enough to not pose a choking risk, and no sharp edges.

In practice, the standard test for choking hazard is to take a cylinder that is 1.25” in diameter (the width of the mouth of the cylinder) and 2.25” long. This is about the size of an empty toilet paper roll. If any part of the toy can fit completely in the cylinder, it is considered a choking hazard. So just don’t make things too small!

Also, of course: no parts that could pinch fingers or hardware that might fall out and be a choking hazard by itself.

And then on the edges and faces: make sure to sand the toys very well before finishing. Depending on the toy, you’ll want to get to at least a 220 grit, if not 320+. As always, start at lower grit and work your way up, being thorough. They should feel very smooth to avoid splinters and bits getting in their mouth!

Beyond just sanding, it’s probably best to round over every edge and corner. I do this on a router/shaper table, and I like using at least a 1/4” roundover bit. 1/8” — or even just thorough sanding — is enough for safety purposes, but the 1/4” size gives it a nice rounder look which I think is nice aesthetically.

A first try: simple cubes

Aside from the wenge pucks above (RIP), the first thing I wanted to make were some simple stacking cubes. Here’s how the very first one came outNote that etching this detailed is likely to make little pieces that might come off when subjected to a child gnawing on it. For production runs I’ll likely do coarser and shallower prints with less risk there. :

Cubes are hard, it turns out! If you want them to stack really nicely, you actually need them to be… well… cubic, with all three dimensions near-perfectly equal.

Here’s how I made it and its first siblings:

  1. Get a few feet of nice 8/4 walnutFor those unfamiliar, 8/4 is pronounced “eight quarter.” Lumber is typically measured in quarter inches, so “8/4” is “eight of those quarter inch chunks,” or 2 inches. But the 2 inch measurement applies when the lumber is rough cut. Once you surface it on both sides to make it flat and even, it’ll typically take off about a tenth of an inch on each side, so it ends up being roughly 1-13/16” thick. Then you’ll work it and sand it at the end, so it’ll end up around 1-12/16” or 1.75” when all is said and done. from the lumber yard
  2. The lumber was “S4S” (all four sides were surfaced) but I ran it through the planer to make sure
  3. Set the table saw to 1.8” and ripped the lumber to that width — so now I had a couple long strips of wood that were roughly 1.8” square (one side from the original thickness, one side from the table saw)
  4. Took the strips to the router table with a 1/4” roundover bit and rounded all the edges. You could do this at the end with the finished pieces (and I still have to do some of that), but it’s easier to do the long sides than a bunch of tiny pieces
  5. Built a simple stop block jig for the miter saw that lets me repeatedly crosscut 1.8” pieces. People do all sorts of fancy stuff for this, but I’ve found something simple like in this video works great for my purposes. You may not even need to do the two layered pieces of wood since the pieces you’re cutting are so small — you can just clamp one piece to the right spot on the fence.
  6. After cutting all those pieces, you now have a pile of 1.8” cubes that are rounded over on some edges but not all
  7. I took those back to the router table and rounded off all the edges (make sure to get all 12 cube edges, not just the parallel faces!). Doing this with your hands is a sketchy, unless you want to accidentally route off a chunk of your finger. So I put the cubes in a wooden handscrew clamp (like this) with the to-be-routed edge protruding, and then I move the piece around with the clamp pressed against the router table.
  8. After this, it’s time to sand. Given the size of the pieces, I don’t like using an orbital sander. So I do a combination of 1) holding sandpaper in my hand and sanding it, 2) sticking sandpaper onto the table and rubbing the piece on the paper, and 3) using softer foam-backed sandpaper to round around the edges. I worked up to 220 — could probably go even further, but I’m playing around with a “rougher” / handmade look.
  9. Then I had an idea to use some AI-generated graphics on the sides of the cubes. So I went to Reve and played around for a bit, asking for letters with decoration in certain styles. I then removed the background of the resulting images (remove.bg is pretty good!). If I really want something great I’ll probably do more manual work at this stage, but this process was good enough for the proof-of-concept.
  10. I went over to the laser cutter and, after a lot of playing around with a test cube to place the design correctly, etched in the images. I was on an 60-watt Epilog Helix and 45 speed / 100 power settings worked well.
  11. I did some very light high-grit sanding afterwards to clean up the edges of the engraving.
  12. Then: time to finish. Since this was before I learned all my lessons above, I just brushed on some boiled linseed oil. The photo above is after one coat, but I’ll probably do another and then a waxing coat too.

Lots of room for improvement, but I’m pretty happy with this as the very first output! More toys (and mistakes, and lessons, presumably) to come. And I’ll try to remember to take more helpful photos as I go…

Looking for more to read?

Want to hear about new essays? Subscribe to my roughly-monthly newsletter recapping my recent writing and things I'm enjoying:

And I'd love to hear from you directly: andy@andybromberg.com