Home-Barista.com

RapidCoffee wrote:Perhaps you could try this? No refractometer required; just time your extractions.

I have. You'll note in my quoted text in your reply that "mine do, too", corroborating Luca's faster flow with WTD than without.

Also, I have used a similar nutation technique at times. I find that even gentle nutation can do a really nice job of compacting the puck. I'd be curious to see if my shots slow down with an added nutation vs without. Curious enough to try it and report back.

RapidCoffee wrote:You are welcome to posit other hypotheses.

I got nothing. Yours makes perfect sense from the flow rate standpoint, but I'm perplexed how the EY would go down if chanelling were reduced. It could be that the speculation around higher extraction being always, definitely, and without a doubt more even extraction (and also equally being better) could be just that - speculation...

pcrussell50 wrote:Anyway with my Monoliths (Flat and Max), and HG1, I found aggressive mixing to be mandatory, and there was still distribution to be done before tamp or tool.

My Super Jolly required WDT in the basket or a second trip through the doser to pour evenly. I never could understand why, but dumping the grounds back into the doser was as effective as WDT in the basket at curing my one-sided extractions. Having a couple well-aligned single dosing grinders on the bench right now, I don't find that any special puck prep is required. I do tap the portafilter to settle the grounds before tamping out of habit. I should try without that and see how things work out...

I don't have any data to go off of, just visually even pours and good coffee with minimal futzing. Can't really ask for much more, right?

Cheers!

- Jake

pcrussell50 wrote:On purpose built single dosing grinders like my Monoliths and HG1 there is little mixing done by the grinder because there is no mixing or homogenizing device in the flow path. I find that I get best pours from those grinders if I grind into a separate vessel and mix aggressively, before transferring to the portafilter. Think something like the LWW blind tumbler/shakers. Then once in the portafilter, there is still the need to distribute before tamping or using spinny grooming tool or both. I see mixing as a much more aggressive thing than merely distributing a pile in the portafilter before tamp or spin tool.

...

Anyway with my Monoliths (Flat and Max), and HG1, I found aggressive mixing to be mandatory, and there was still distribution to be done before tamp or tool.

-Peter

Peter, what thing is it that makes you say that aggressive mixing is mandatory? Is it based on shot-to shot variability? What the pour looks like? Or taste? And if it's taste, how do you find they taste different?

I'm piping up to ask this just because I have LU Max and I *think* I get better repeatability between shots by futzing around less. I haven't tried throughly mixing and I could, but I find that when you tip the cup into the PF it's usually slanted to one side, which, intuitively (and, again, data free zone from me here) I assume is bad.

RapidCoffee wrote: Observations:
* Baskets prepped with WDT take longer to extract to a given brew ratio. I've seen this consistently, ever since I developed this technique. This is probably due to reduced channeling. I do not understand why Luca got shorter pours with WDT preps.

Hi John, first up, thanks for doing this. Second up, thanks again for creating the python script for me a while ago to allow me to graph multiple shots against each other, which I haven't gotten around to using. I've got your latest scripts and have finally caved and bought office 365 for home, so hopefully I'll finally be able to get around to using it.

Anyway, I have at least one possible explanation:

I didn't really compare WDT vs no WDT. I compared a ridiculously OTT amount of stirring. It's like 30 stirs; probably 20-30 seconds to do it, and probably 28 stirs more than is necessary to reasonably WDT the basket with the tool that I used. What I wanted to do was to start off with ridiculously exaggerated differences, on the assumption that if I showed no difference between the ridiculously exaggerated technique and minimal prep, then at least no one would say I didn't stir enough.

If we want some random speculation, maybe WDT to the point where the clumps are broken and you have evenly distributed coffee, as opposed to a mound, has fairly far-reaching effects. However, maybe continuing to stir a lot beyond that point results in fines migration, via the mechanism Jonathan pointed out from the literature here:

https://www.baristahustle.com/white-pap ... migration/

One of the critiques of what I did on the Decent diaspora - and rightly so - is that it's not like I really had a particular stated hypothesis that I was looking to investigate. I was really wanting to work out if there was a difference at all that merited further investigation. But what I was wondering is if we would have all of the fines settle at the bottom and super jam up the extraction, but we in fact saw the opposite of this. Did I read the paper wrong and was it saying that the fines migrate to the top of a cylinder?

* Baskets prepped with WDT give lower refractometer readings (by ~1%EY). This corroborates Luca's results. My WAG: increased channeling on non-WDT baskets causes some overextraction, leading to higher Brix/TDS/EY

Sorry, I think we're in opposition on this, actually! My results are that the minmal WDT ie prongs about half way down, just push around the mound to make it flat in 1 stir had EY of like 22.5% or so vs the EY of 21.1% or so from the 30 stirs. So I'm showing maybe 1.5% greater EY from doing less. However, as I said, none of this is a comparison of WDT vs no WDT. It may be that 3-4 stirs or 2-3 stirs with prongs all the way to the bottom of the basket is fundamentally different from my 1 or 30 stir cases. This is more work that could be done.

luca wrote:Hi John, first up, thanks for doing this. Second up, thanks again for creating the python script for me a while ago to allow me to graph multiple shots against each other, which I haven't gotten around to using. I've got your latest scripts and have finally caved and bought office 365 for home, so hopefully I'll finally be able to get around to using it.

PM me for guidance, or if you'd like a version that also plots puck resistance:

no WDT on left, WDT on right

luca wrote:However, as I said, none of this is a comparison of WDT vs no WDT. It may be that 3-4 stirs or 2-3 stirs with prongs all the way to the bottom of the basket is fundamentally different from my 1 or 30 stir cases. This is more work that could be done.

Agreed, on all counts. Consider my experiment as further exploration into this area, which clearly needs more study (like pretty much everything in the coffee universe). Jake suggests that flat burr grinds respond differently to stirring than conical burr grinds. Decent Espresso is developing a "puck rake" for stirring the grinds, which looks like YAWT*, but is supposed to be used for stirring only the top 25% of the grinds. These claims should be tested. And please: publish methodology and data, not just anecdotal evidence!

* Yet Another WDT Tool (old UNIX hackers will get the joke )

luca wrote:Peter, what thing is it that makes you say that aggressive mixing is mandatory? Is it based on shot-to shot variability? What the pour looks like? Or taste? And if it's taste, how do you find they taste different?

What the pour looks like. Nasty ugly spritzy uneven pours. Bare spots all over the bottom of the basket. This went away when I started grinding into a separate vessel and mixing aggressively before transferring to the portafilter. Once in the portafilter, I perform WDT, more as an actual act of distributing than anything else: to prep the pile for tamping or spinny tool. (I use spinny tool).

I did not do taste comparisons between pretty pours and ugly ones. So I make no claims in that department. Nowadays I grind into separate vessel and mix aggressively, no matter what grinder I use.

luca wrote: I haven't tried throughly mixing and I could, but I find that when you tip the cup into the PF it's usually slanted to one side, which, intuitively (and, again, data free zone from me here) I assume is bad.

Agree here. Which is why I still WDT after transferring grinds from aggressive mixing vessel, into the portafilter.

-Peter

So, here is an n=1/2 test.

14g dark roasted aged Sumatra in a Rancilio "102" basket (stock 14g precision basket for Silvia from 6-8 years ago, I gather) ground at 2.0 on Ultra with burr lock at zero on the dial at 800 rpm.
Below are grounds as they fell from the slide into the basket.


After a gently side to side "shimmy" to even the bed:


Tamp:


Tamped puck:


And the messy, I mean messy pour:


I got 22g out in 28s.

This is faster than when I tap. I will pull the other half so that n=1 after I shovel my driveway.

The shots I pulled this morning with downward taps to collapse the bed were much slower. 36s for similar pours and were visually much better. Not perfect, but clean pours with no spritzing. This shot had 3 streams that never converged and made an awful mess of my cup. The flavors of this were bitter walnut vs leather and less bitter walnut on the other shots.

All this has me asking "what does any of this have to do with stirring grounds?"



Right. Baselines.

•So no taps - fast flow.
•Taps with no WDT - slow flow.
•WDT with taps - faster than slow flow.
•WDT with no taps - need to determine.

Ok. I'm still being productive. I think.

For what it's worth I've gone through quite a few different WDT techniques and tools in the last year. Starting with nothing and then a paperclip and then the keycap puller John recommended and now a Frankenstein tool with 6x 0.4mm needles rather than the 4x 0.75mm needles in the keycap puller. I've also just ordered the Levercraft version.
Here they are side by side:



What I was doing until recently was about 5-10 seconds of spirograph patterned stirring starting at the bottom and working my way up to the surface.
Now I've started using the finer tool I'm beginning with a bigger "whole basket mixing" for about 3-5 seconds and then the spirograph pattern stirring starting at the bottom of the basket, then progressing to half way up and then finally "raking" the surface which probably takes more like 20-25 seconds in total. I don't own a refractometer but my pours fill in much more evenly and finish much more consistently within 1-2 seconds of each other.

Here are a couple of examples of the bottomless extractions I'm seeing with the new tool. The first is a light roast Anerobic Geisha and the second a SO Colombian medium roast. The grinder is a Vario and the machine is an E61 with about 6-7 seconds of pressure ramp up limited to 8.5 bar. VST 18g and 20g baskets.
I don't think I'd want to go back to no WDT, at least with my setup. My Monolith Flat won't be here for 6-7 months but from what people have said the need for WDT doesn't go away.

Jake_G wrote:•So no taps - fast flow.
•Taps with no WDT - slow flow.
•WDT with taps - faster than slow flow.
•WDT with no taps - need to determine.

Ok. Same routine as before, but 20 stirs with the LeverCraft WDT tool, starting at the bottom, working my way to the top.

22g output in 29s, but a visually perfect pour. Taste was less bitter. Very smooth. Heavy walnuts and leather.

So

The matrix looks like this:
•No taps nor WDT - fastest flow, ugly pour, sharp/harsh flavors
•No taps with WDT - slightly slower but still fast flow, pretty pour, soft flavors
•Taps with WDT - Faster than slowest flow, pretty pour, soft flavors
•Taps without WDT - Slowest flow, decent pour, soft flavors

I will do some more bracketing with this coffee and see how much faster WDT pours are than no WDT when tapping is maintained, as tapping is part of my normal workflow.

I'll also verify my tasting notes.

Thanks Jake. Obviously there's many questions left to be answered. In particular, what is the impact on espresso extraction dynamics (and EY) of
* stirring grinds (to break clumps and randomize grinds)
* shaking basket (to level grinds)
* tapping basket (to settle grinds)
* nutation and "distribution" tools to level surface
* type of grinder (flat burr vs conical)
* beans and roast level

At least this is a start.