"I've made a version of your (Aether/Sylph) winnower, but it's not quite working.  I'm still getting husk in my nibs and nibs in my husk.  Help.  And I really don't understand what the point of that other valve is that isn't changing the vacuum.  Can you explain that more?" Great question.  I keep forgetting I don't have the manuals on line.  The short answer (long to follow) is you have to tune the winnower.  Just like you have to tune an instrument, each winnower has be to be tuned AND re-tuned as needed.  So, the following is right out of the manuals. My winnower is designed to winnow roasted and cracked cocoa beans. It has two adjustment valves to allow you to ‘tune’ it to your particular cocoa bean and vacuum. All cocoa beans are not alike. You know this as they all have different flavors. Likewise, they require slightly different settings on the Sylph and these settings MAY CHANGE over time.* Also, the Sylph, regardless of its magical name, is not magic. The quality of the nibs you get out is only going to be as good as the quality of beans you put in. This means if you have flat, under-fermented and/or mucilage covered beans then you will most likely get flats and other pieces of ‘non-nib’ in with your nibs. This is a consequence of what you put in, not a matter of the Sylph not functioning properly.**

First off, no winnower is perfect. A small amount of nib in your husk (as powder) and/or a small amount of husk in your nibs can be expected. Industry standard is less than 1.5% husk in nib (that would be nearly 2.5 oz of husk in 10 lbs) If you were to weigh it out and look at it, it looks like a LOT of husk, so when you see a little, don’t panic. On average, with well prepared, properly roasted beans, you can expect less than 0.5% husk in your nibs and virtually no nib in your husk bucket.

Tuning the Sylph occurs in two steps: setting the Vacuum Adjustment Valve and then setting the Discriminator Valve. Think of them as coarse and fine adjustments respectively. Please do not change both valves at one time as you can then not tell which valve affected your results.

As a starting point, open the Discriminator Valve all the way and open the Vacuum Adjustment Valve ½ way.

Test with fully cooled roasted beans.

Process 10 oz of roasted beans (see operations manual for step by step instructions) and calculate your recovery: Weight of nibs / Weight of beans (10 oz).

  • 0.80 or 80% is great. If the nibs look good and there is virtually no nib in the husk, you are done. Go forth and winnow.
  • 82% is either perfect or you have husk in the nibs. If it is the later, increase the vacuum a little (close the valve) and test again with another 10 oz.
  • Less than 80% - generally there is nib in the husk waste. Decrease the vacuum a little.

Continue adjusting the Vacuum Adjustment Valve and testing until you hit a maximum recovery of nibs with minimal husk content. At this point there will still be some nibs in your husk waste or husk in your nibs.

Note: ‘Flats’ do not count as husk if you see them in your nibs. You can’t ‘adjust’ these out – they have to be screened out.

Setting the Discriminator Valve

At this point, you have adjusted your Vacuum Adjustment Valve to the point where there is either husk in your nibs or nibs in your husk. Adjust the Discriminator Valve depending which is the case.

Husk in your nibs – Close the Discriminator Valve 1/4-1/2

Nibs in your husk – Open the Discriminator Valve 1/4-1/2

Process another 10 oz of roasted beans and calculate your recovery: Weight of nibs / Weight of beans (10 oz). Compare this recovery to your last recovery.

If there was husk previously in your nibs, you want your recovery to go down as that shows more husk is being removed. If there is still too much husk in your nibs, close the Discriminator Valve another 1/4 - 1/2 and test again with another 10 oz of roasted beans.

If there was nib previously in your husk, you want your recovery to go up, giving you more nibs. If there is still too much nib in your husk, open Discriminator Valve another 1/4-1/2 and test again with another 10 oz of roasted beans.

Through all of this, keep in mind to LOOK at the nibs and husk and if they look good, you are done.

Troubleshooting tuning:

If you are at 80% recovery or above, stop adjusting and proceed to winnowing.

If you are way above 80 % recovery and a lot of husk is in the nibs, adjust your Vacuum Adjustment Valve closed (although this should not be the case as you should have closed it enough to have no husk in your nibs).

If you are way below 80% recovery, examine your husk waste. If there is nib present, there are three possible reasons:

1) You cracked while the beans were warm and they powdered too much – re-test and calibrate with cold roasted beans.

2) You need to open the Discriminator Valve more.

3) If the Discriminator Valve is already fully open, you need to close it to 1/3 open, open your Vacuum Adjustment Valve a little, and test again.

And of course, at any point you are happy with the distribution, stop. I like to process 10 oz each time because the math is easy, the amount is representative, and you can see very quickly whether the tuning change you made had a positive or negative effect.

After a while, you should be able to ‘eyeball’ your tuning without all of the math, but it’s rather helpful at the beginning while you are learning.

* Over time, the dust filter on your vacuum will accumulate dust and the vacuum pressure will drop. As this happens you will start to get husk in your nibs. The solution is to simply close the Vacuum Adjustment Valve a little to increase the vacuum pressure until you clean the dust filter out. Don’t touch the Discriminator Valve!

** If you find you are using a bean with a lot of flats, (and there are some great tasting beans out there with flats so I do understand) you can screen your final nibs to remove these as they are generally much larger than the nibs. A ¼” hardware screen tacked onto a frame works very well for this.