G23-026 Winnowing, or, Rube Goldberg does agriculture.

Posted on June 17, 2023

 

On the plus side, I bet you didn’t expect a blog post about winnowing.

On the down side, this entire blog post is about winnowing.

Image above: Winnowing Grain, Eastman Johnson, 1879, via https://www.wikiart.org/

Separating the wheat from the chaff, by any means possible.

When you see depictions of winnowing, it always looks so romantic.  In every culture from the ancient Egyptians to 19th century American farming, winnowing grain by hand is depicted as a light, pleasant task that ends the harvest season.

But here’s a clue.  Note that in the depictions above, people are doing a huge variety of things to winnow their grain.  Fanning, shaking, pouring, tossing, sifting, sieving.  Standing, kneeling, solo, in groups.  Basket-to-basket, hand-to-floor, hand-to-ground.  Indoors, outdoors.

Seems like people used every method imaginable to winnow by hand.  The only thing in common is that all these methods use the movement of air through grain, or grain through air, to separate the wheat from the chaff.

It’s only after you’ve tried doing it, for real, using those primitive methods, that you realize why.  None of that works worth a damn. At least, not for the modern individual who is used to processes that are, generally, somewhat efficient.

Here’s the problem.  The stuff you want to eat or keep — the grain or seeds — consists of tiny little objects.  Those are thoroughly mixed with bits of organic trash, of all sizes, that you picked up in the harvest.  You can’t eat it, in that state.  You can’t pick it clean by hand, in any reasonable amount of time.  You can’t float the trash off the seeds.

Instead, you maybe sieve it to pick out the biggest pieces of trash.  Then you blow some air through the remaining mess, or blow that mess through the air, and hope that more trash than edible seed floats away.

It’s not a one-and-done.  With each pass, in theory, more trash floats away.  And after N passes, you end up with something that you’re not too squeamish about grinding up and eating.  In the meantime, some portion of the seeds you were trying to save have been lost in all that tossing.

In any case, last year I failed Winnowing 101.  I tried a couple of semi-primitive methods to clean a batch of mustard seeds.  Basically, I spent some time pouring the seed-n-trash mix between bowls, in front of a fan.  I never did end up with something pure enough that I’d care to grind it and eat it.  And I lost the majority of the seed in the process, to boot.

This year, I’m growing maybe a hundred square feet of mustard plants.  The flowers are now fading and the seed pods are starting to mature.  So I’d like to harvest and eat the seeds.  Which requires taking another crack at winnowing them.

I learned a couple of lessons from last year’s failure.  Winnowing by hand is nowhere near as easy as some internet sources suggest.  At least, not with seeds as tiny as a grain of mustard.  But mainly, I need to find a better way to winnow.


A typology of winnowing machines.

As I started looking for a better approach, I soon realized that, even today, people use every conceivable method to winnow their grains and seeds.  Sieve it, screen it, shake it, blow it up, blow it around, use a fan, use a vacuum.  You name it, and some seeds somewhere are being separated using that approach.

I have come to realize that, in part, that’s because different seed/trash mixes work best with different methods.  Effective separation of (say) big, heavy dried beans from bits of pod, leaf, and stalk requires a completely different set of air velocities and pour rates, compared to separating (say) tiny little grains of mustard.

The immediate upshot of that is that:

  1. There is no one-size-fits all approach to winnowing.
  2. Somebody’s internet instructions for winnowing may or may not work for the seed or grain you are trying to separate.
  3. Good winnowing machines will be adjustable.

Let me now abandon all primitive methods.  So I’m not going to toss my seed/trash mixture in the breeze.  Nor am I going to toss it, in the open, in front of a fan.  Under the right conditions, those methods might work for some grains.  For mustard, I had little success.

What are the remaining options?  Restated, what do small-scale winnowing machines look like?


1:  Sieves and screens to pre-process and remove larger trash.

Winnowing — using air to separate seeds from organic trash — works well if and only if you’re trying to remove bits of trash that are small and light.  Big stems and stalks won’t be removed that way.  The obvious thing to do is to sieve or screen the seed/trash mix beforehand, to remove the big stuff.

A separate issue is that heavy bits of trash — stuff that’s as dense or denser than your seeds — will fall in with your seed.  Ever wonder why, when you buy a bag of beans, say, it says you need to watch out for stones?  (It’s a thing, google it.)  That’s because the winnowing process was unable to remove an object roughly the size of, and heavier than, the seeds being winnowed.  After doing my research, I now know there’s a separate “de-stoning” process, and that, as with winnowing, it’s not 100% effective.

For the home gardener, you can buy sieve sets or “riddles” with mesh of different sizes.  In my case, I think I’ll probably use whatever I have on hand that will work.  Anything that will allow the mustard seeds to pass through.

 


2:  Horizontal separation:  Single-pass past a blower

This is the same concept as dropping mixed seeds/trash past a fan.  But in this case, the situation is far more regulated than simply dropping seeds in the open air.

In these machines, a fan of some sort generates a horizontal stream of moving air of a given speed, and the seed/trash mixture is poured at a measured rate, through that.  Air speed and pour rate are adjusted to give as much separation as possible.  Ideally, trash is blown far away, seeds drop nearly vertically.

The first one shown is arguably the clearest illustration of this approach.

 

3:  Vertical separation:  Updraft tubes, with or without serpentine path.

With this approach, you pour the grain/trash mixture down some sort of tube, while air is blowing up the tube.   The air speed is set so that the seeds will fall, but the trash will rise.  Trash is then collected at the top of the air column, and seeds drop down to the bottom.  Some use a fan to blow the air through, some use (e.g.) a shop vac to suck the air through.  Some build in a zig-zag path to prevent clumps of material from dropping straight though and to maximize the extent to which the seed/trash material mixes with the moving column of air.

Those of you who are Mythbusters fans will recognize this as a variation on the setup used to determine the terminal velocity of a falling penny.  The air speed has to be regulated, in this case, to be below the terminal velocity of a seed, so that the seeds will fall.

4:  Other semi-mechanized methods.

4.1  Blow air across heavy seeds such as beans.  Pretty sure this would be risky for mustard.

5:  My choice:  Air popper without the heat.

Given the scale of what I’m doing, I kind of liked the machines made out of plastic bottles or pipe, and a blower of some sort.

When you get right down to it, updraft-style winnowers are really no different from an air popper, for popcorn.  The speed of the air in the popper is less than the terminal velocity of unpopped corn, but just slightly higher than the terminal velocity of popped corn.  As a result, the corn kernels stay in the bottom, but the fluffy popped corn rises.

So, my first attempt will be a variation on the second soda-bottle winnower above.  But instead of tilting the bottle, I’m going to put together a small “U” of PVC pipe.  The U will stand upright.  Soda bottle goes in one end of the U, heat gun (blower) goes in the other.  Start at low speed, drop the seed/trash mixture in, and re-process until all the trash is blown out.

Effectively, a hot-air popper, without the heat.

If not that, then maybe I’ll make the PVC pipe separator pictured above.  For one thing, I’m familiar with the material.  For another, the enclosed design and variable-speed fan seem like a good choice for these tiny grains of mustard.  For a third, the longer “barrel” of the device provides a greater length of air column over which to perform the winnowing, likely resulting in less seed loss.

In any case, a) for mustard, it has to work at least as well as pouring the seed/trash mix in front of a fan, and b) it’ll take me about 10 minutes to set it up.  The key is that the heat gun provides a powerful variable-speed fan.  The U tube keeps the seeds from falling into the blower.

Once the mustard is fully ripe, we’ll see how it goes.