Drip coffee test with 4 coffee grinders, from cheap (GB£39) to mid-range (GB£559) prices.

To be clear at the outset, I think if you are spending money on anything, then you have to be happy with your purchase in every respect. You (and your partner) want to be pleased to have it on display and you want to feel that you have purchased something of quality, that will last.

Anything that is flimsy, excessively noisy, or messy may not inspire confidence in its longevity and even if it objectively ‘does the job’, it can still be difficult to muster, or maintain any enthusiasm for it.

In terms of a perception that pricier grinders might potentially shorten the learning curve, or provide short cuts, or require less intervention to achieve a good result, it’s important that we remind ourselves that basically, grinders all work the same way – we turn a dial to open or close the burr gap (though for stepped grinders the resolution of the steps might be a factor in usefulness). Having a popular grinder can be helpful in polling other users on a reasonable range of settings for a brew method.

All that said, I am not a masochist, I prefer using the most expensive grinder I currently own, for its convenience and many of the points I note above. But, my second cheapest (GB£65) electric grinder produces grinds that make slightly different, but overall, just as enjoyable brews. I’m still not likely to recommend the cheaper one to someone that can afford a more substantial build, unless that someone just needs a small, cheap solution for travel/caravanning/boating etc.

I find myself adjusting the pricier, stepless, electric grinder most often. Maybe because there is a range of ’close enough for rock’n’roll’ settings? Whereas the stepped, cheaper, electric grinder locks you in to fewer, more defined, workable options.

So, speaking now specifically about the sensory aspect, regarding grinders for manual brewed methods and domestic auto drip (whilst temporarily setting aside aesthetics, feel and workflow), at what point do people really think that price begins to change the game?

I ask because I currently own several grinders. I have owned and/or used many more. From cheap to fairly expensive hand grinders (GB£200+), to mid priced, all purpose/brewed electric grinders and, quite honestly, I can’t see that price bears any relationship to how much I enjoy cups. (I’m referring here to widely used, ‘known good’ examples of type, rather than some previously unknown, suspiciously cheap items – I did once take a punt on a bizarrely cheap, steel burr, hand grinder that was made to unbelievably bad tolerances…but I never did really put it through its paces).

To test this concept, I made 40 brews in total:

10 brews each, with…

4 different grinders, using…

5 different coffees and…

2 different brew cones, fed by a Melitta Aromaboy.

All brews were dosed at 12.5g coffee to 185g brew water, brew cones were preheated prior to brew start.

For each grinder 5 brews each were carried out using a Hario V60 01 cone and Moccamaster #1 filter papers. Another 5 brews per grinder, using a Fellow Stagg X cone and Kalita 155 papers. I have previously established, ‘no significant difference’ between the two brew cones, with respect to either extraction yield, or subjective preference, when used with the Melitta Aromaboy.

The grinders used were:

Hario Mini Slim+ (MSS-1DTB) 40mm ceramic burr, hand grinder (used with a Hario, rechargeable grind stick), set to 9 clicks from fully tight.

De Longhi KG79 electric 40mm flat, ceramic, burr grinder, set to 3 (2 clicks clockwise from the stop).

Fellow Ode V1.1 electric, 64mm flat, steel, burr grinder, set to 1.67-2.00.

Niche Zero electric 63mm conical, steel, burr grinder set from 65.5 to 67.0.

There was no significant difference in the average extraction yield across the four grinders. The overall average for 40 brews being 20.26% with a stdev of 0.789%. There was, however, a difference between the Niche Zero and the Fellow Ode v1.1 in terms of my subjective liking, for this condition. (In previous exercises, at coarser settings, this has not been the case for the Ode, so I am reluctant to make any solid determination based on this particular example).

The benefits I see are essentially in terms of the workflow, practicality and ease/enjoyment of use. For example, I wouldn’t try to grind 80g of coffee with a Hario Slim (well, I have once, never again),. Similarly, I wouldn’t want to listen to the tortured scream from a GB£65 electric grinder every time I wanted a cup, nor for any length of time if grinding a large dose.

I sifted samples from all four grinders at 400 & 1400 Kruve, whilst there were no surprises here (Ode v1.1 had the tightest distribution, Hario Mini Slim+ the widest) there was no obvious relationship between distribution extremes and preference scores.

I did a further sift at 900 & 1100Kruve for the Ode v1.1, Niche Zero and Hario Mini Slim+ and I was struck with how similar the median grind size was, with a span of 3% at 900Kruve (50-53%). Unfortunately, in the meantime I had accidentally removed the top burr from the De Longhi KG79 and, given the slop in tolerance for this unit, I didn’t have confidence in achieving a like-for-like sift result, with this grinder for 900 & 1100Kruve.

Measuring the objective parameters of brews, then parsing out trends from one brewer to the next, the whole question seems extremely difficult to answer, or to make a specific recommendation without ‘hands-on’ experience of the combination in question?

Just to be specific, when scoring, I’m talking about overall enjoyment of cups, rather than saying that I can’t tell one from another. Also, I am not saying that sensory benefits aren’t possible…indeed, if people are certain that they are, I’d be very keen to hear why/how and to see how these perceptions were determined?

Hario V60 01, Hario Pegasus 01 & Fellow Stagg X brewer comparison, with Melitta Aromaboy.

I regularly hear that, ‘such and such’ design of pour over/drip brewer extracts more than another style of brewer. It is often presented as if this is accepted fact. I have logged about 2200 comparable, dialed in, drip/pour over brews and have found it really hard to find any evidence of this.

As I frequently use my Melitta Aromaboy to brew single cups, I can use a range of brew cones for my twice daily cup (my cup height is 62mm, allowing for the use of various cones, that would otherwise be too tall for the Aromaboy server). So, I logged some brews to see if this assertion panned out, when used at the same grind setting for each of 10 coffees (but with adjustments as deemed necessary from bag to bag) and at the same brew size (avg. 12.47g coffee to 186g brew water). Using the Melitta Aromaboy negates any bias in pouring style and ensures a consistent water delivery rate.

I started out with 4 brew cones:

Hario V60 01 ceramic with Rombouts #2 papers. Pour time (from the time of switching the brewer on) 131s, stdev 3s (x20 brews).

Hario Pegasus 01 truncated, Melitta style cone, with Kalita 101 papers. Pour time 130s, stdev 3s (x20 brews).

Fellow Stagg X (flat bottomed), with 155 Kalita papers. Pour time 131s, stdev 3s (x20 brews).

Kalita Wave 185 (flat bottomed) in steel, with Kalita 155 papers. Pour time 130s, stdev 2s (x10 brews).

All brew cones were preheated in a 1L Pyrex jug with boiling water, prior to brewing.

Long story short, there was no statistical difference in extraction between the brewers.

There were subjective differences in the perception of brew quality – The Kalita Wave 185 was shelved halfway through because the taste proved to be lacking and I just couldn’t face any more disappointing brews.

By the end I had 60 brews, 30 ground on a Niche Zero, 30 on a Feldgrind. 10 brews per coffee for each brewer & grinder combination.

Of 98 Melitta Aromaboy brews in total, the average EY was 19.96%, average brew time was 157 seconds, subjective liking score 7.11/9.00

Included in that 98 were the 60 test brews. For these, the average EY was 19.93%, average brew time was 154 seconds, subjective liking score 6.98/ 9.00

The biggest difference in EY, when using the same grinder, was with a Niche Zero – Fellow Stagg X (19.97%, stdev 0.63%) vs Hario Pegasus 01 (19.60%, stdev 0.66%). Not statistically different.

Comparing the Niche Pegasus 01 brews to the Feldgrind V60 01 brews (the sample with the highest EY) saw an average 20.14%EY for the V60 01, with a stdev of 0.72% – no statistical difference to the Niche & Pegasus 01 (p=0.147).

No statistical difference in subjective liking score either, but if you’re curious, the V60 01 with Niche (7.2/9.0), V60 01 with Feld (7.1/9.0) & Stagg with Niche (7.1/9.0) all scored about the same. Pegasus 01 with Feld & Niche both scored 6.8/9.0. My hunch here is that the surface area of the grinds bed is greater with the Pegasus 01 and flow through the bed is a little less even. Likewise, for the abortive Kalita 185 test sample…a greater surface area and a shallower bed here (I do have a glass Kalita 155, but I have not had any on-par brews with this, whether using the Aromaboy, or otherwise).

It is certainly possible that intervention with a shower screen could enhance the steel Kalita 185 (which has given par brews with a goose-neck kettle, historically) & Pegasus 01 brews, but I wanted to specifically test ‘apples to apples’.

My last 100 bags of filter (drip/pour over) coffee.

A while back I decided to make notes on 100 bags of coffee. I tend to make a lot of notes on grind settings, method, time, extraction & overall taste score, but for this bunch I focused on recording the following parameters:

Roast level (my colour perception by eye).

Whether the bean was representative of the stated notes. I used a 5 point scale: 1- Not at all; 2- Not really; 3- Maybe/maybe not; 4- A little (at least one note present); 5- Definitely (two, or more, notes present).

Cost per 100g (as per my purchase size).

Extraction yield.

Subjective overall taste score, using the Quartermaster’s, 9 point, hedonic scale. The coffees are sorted in order of preference, so on the chart you can see the scores declining from left to right, my preferences in origins are also visible.

I used only ‘known good’, dialled-in brew methods, switching between V60 01, V60 02/Switch, Bartleet 3 hole Melitta style cone, Kalita Wave 185, as well as a Melitta Aromaboy brewer used as stock and also with V60 01, Hario Pegasus 01, Fellow Stagg X & Kalita 185 cones, for smaller capacity brews.

As an aside, it is my strong opinion that, if you need multiple attempts to brew a representative cup, you are not dialled-in. (Of course, fine tuning brews to accentuate specific attributes is desirable, but if you are not 90% there with a ‘known good’ method, then it is not a ‘dialled-in/known good’ method.)

Grinders used include Niche Zero, MbK Feldgrind, Timemore Nano, Fellow Ode V1.1, Wilfa flat burr grinder, Comandante X Trail, OE Lido E/ET/2 (with either Kinu brew burrs, or Etzinger AP burrs).

Brews for a given coffee were not repeated with the same brewer and grinder combination. Each brew was essentially one ‘live take’.

The aspect in which the coffees were the most consistent was extraction yield (gold line on the chart), the standard deviation was only 3% of the average value (19.8%). Coffee brewing is very consistent, if you are, with respect to brew size, grind setting & pour speed/rate.

The next least variation was in terms of subjective overall taste score (grey line on the chart, declining from left to right), the standard deviation being 9% of the average score (7.22/9.00).

Whether the coffee was deemed, by me, to be representative of the tasting notes (green line) had a standard deviation of 23% of the average (3.88/5.00). So overall, coffees generally displayed at least one of the stated notes, as far as I could tell.

The roast level had more variation (brown line), odd as I only ever buy light, filter roasts (1/5). It would be reasonable to expect some medium light roasts (2/5), but there were a fair few mediums (3/5) and even a medium dark (4/5) amongst them. Standard deviation was 32% of the average, of 2.19/5.00.

Now the biggest variation, price (blue line) with a standard deviation of 42% of the average (being GB£7.84 per 100g). The average was quarter way between the cheapest (GB£2.72/100g & #43/100 regarding preference) and the most expensive coffee (GB£23.33/100g & #86/100 regarding preference). Of course, the most expensive coffee was a Gesha.

Further observations on, “Pour over brewing with one recipe & 5 different grinders.”

If you take the 5 grinds above (for Hario Switch 02, Drip Assist & medium crepe Japanese white Hario filters, 40g poured every 40s to 200g total, first two pours in the centre of the Drip Assist), the variations (stdevs/averages) at Kruve sieve intervals are as follows:

400 Kruve – 19% variation.

700 Kruve – 15% variation.

900 Kruve – 14% variation.

1100 Kruve – 25% variation.

1200 Kruve – 33% variation.

1400 Kruve – 55% variation.

…So the smaller end is generally more consistent between grinders than the larger end of the distribution, but the average grind size seems even tighter. Doing a rough, ‘back of a fag packet’ analysis the d16 (16% smallest ground mass) variation is around 9%, d50 around 6% d84 around 11%.

The variation in grind distribution Q factor (square root of the span of +/- 1x stdev) is only 3%. Average Q factor for all five grinds is roughly 1.56.

So basically, the differences in average grind sizes seem fairly small, as do the differences in grind size distribution quality (Q).

The two grinders that scored lowest, to my preference, are those that ended up being set to the coarsest grind (Comandante X25 & Wilfa Uniform), rather than having similar Q values to each other, nor those significantly different to the preferred three grinders..

Hario V60, Alternative white filter papers – Rombouts.

In January 2023 Hario introduced what is now the only option in white V60 filter papers that they offer. These latest papers are recognised by the, now familiar, tab but also, by a smooth inner surface in conjunction with a medium crepe outer surface.

The current packaging sports a green stripe design, but I have also received these papers in the more familiar pink & black livery. I have been informed that they now ship with new Switch brewers, which may have plain packaging for the filters.

To quickly recap, there are now (or, have been) four different filter papers that have been marketed as ‘Hario white V60, VCF-##-W:

1. The heavily creped, tabless filters that seem identical to the Cafec Abaca filters, though Hario Global deny any knowledge of a relationship with Cafec and say they do not use Abaca in their filters. Most commonly seen in the pink & black boxes containing 40 filters.

2. The Dutch made filters from Filtropa, with tabs, embossed with ‘V60’. When these were released people noted longer draw down times. These had an ‘H’ suffix and were most commonly seen in shrink wrapped plastic.

3. There was also another Japanese made filter paper, again tabbed, medium crepe both inside & out. These were most commonly seen in the pink & black, loose bag packaging, but I have seen them supplied with the Hario Switch and in the pink & black boxes of 40 filters. These were the papers that I used mostly and gave the best results.

4. The current, post January 2023 papers (smooth inner) tend to give muddy tasting brews with low perceived acidity and muted notes. Therefore, I have been keen to find an affordable, well stocked, replacement.

I dabbled with the Cafec filters but results weren’t that encouraging. Plus the differing versions, each of which have two designations, seem a pain, as does the high domestic shipping cost in the UK

I also tried the LilyDrip Lotos, a device that converts V60/Switch to Kalita flat bottomed filter use. I think this is more of a ‘get you home’ in an emergency fix, rather that an everyday solution. Compared to Kalita Wave brews, Hario V60/Switch with the Lotos always seemed siltier & drier.

I decided to compare the Rombouts No.4 (folded to fit V60/Switch) against the Chemex full circle white filters (trimmed to fit) and the ‘medium crepe inside & out’, Japanese, tabbed, white Hario filters. I did 10 brews with each of the 3 filters, the same 10 coffees used in each case.

I had a slight preference for the Rombouts & medium crepe papers compared to the Chemex, but then I’ve never been blown away by the Chemex filters (they’re better than the green stripe Hario). If you’re good with the Chemex filters, then you may not feel the need to read further.

My previous regime with the medium crepe Hario filters was to grind 13.8g of coffee at 3.33 (9% 400Kruve) on my Ode V1.1 grinder (equivalent to 10% 400Kruve at 74.5 on the Niche Zero). The brew regime was to pour 40g of brew water every 40s to 200g total, each pour taking roughly 15s. The first 2 pours in the centre of the Hario Drip Assist 02, the remaining 3 pours in the ring.

For the Chemex & Rombouts filters, I ground coarser (6.67 on Ode V1.1, 6% 400Kruve) & stretched out the pour rate to 30g every 30s (210g total). Rombouts received 2 pours in the centre of the Drip Assist, then the remaining 5 in the ring. With the Chemex filters, I started with 3 pours in the centre and then the remaining 4 pours in the ring.

Edit: 04/02/25 I have also settled on setting 94.5 on my Niche Zero for this method, this grind setting also yields 6% at 400Kruve.

I had a slight preference for the Rombouts & medium crepe Hario papers compared to the Chemex, but then I’ve never been blown away by the Chemex filters (they’re better than the green stripe Hario). If you’re good with the Chemex filters then you may not feel the need to read further.

Averages over 10 brews, each with a different coffee, using the Rombouts filters, gave the following results:

Average extraction yield: 19.5%, +/- 0.69%

Average brew time to dry bed: 3min 42sec, +/- 20sec.

Average subjective liking score: 7.4/9.0 (80% if more intuitive).

Compared to, 10 brews with the same coffees, using my preferred medium crepe, Hario papers, pouring 40g every 40s, with the finer grind:

Average extraction yield: 20.2%, +/- 0.90%

Average brew time to dry bed: 3min 20sec, +/- 15sec.

Average subjective liking score: 7.3/9.0 (79% if more intuitive).

No significant difference in extraction yield (nor subjective liking score).

Ceado Hoop – Consistency & brewing concerns.

I noticed that the brew water stopped coming out of the ‘spray’ holes pretty shortly after filling the hoop. So I added a filter, rinsed the brewer & paper and filled the brewer with 253.5g of hot water in about 10s (no coffee). At 25s the water had run through the ‘spray’ holes into a cup and flow significantly slowed. By 45s all the water that could pass had dispensed via the spray holes…maybe (read on).

Only 210.5g in total made it out of the brewer and the trapped ~40g is still in there now, many minutes later. This is also pertinent to the Whole Latte Love video where they see the Hoop retain 50.5g of brew water for a 17g dose, this is likely more than the coffee bed itself would retain.

Myself, I have seen between 37g and 46g retained with a 13.8g dose (a liquid retained ratio varying between 2.68 and 3.33). I’d normally expect a difference of a couple of grams at most between brews with a Kalita Wave. A difference of +/-9g out will prevent consistent brewing.

This also then led me to wonder if any remaining water leaks out the screw thread, bypassing the coffee, when the water drains entirely from the Hoop (as it can do, presumably, if not set totally flat). It strikes me that the brewer is a bit finicky when it comes to properly sealing the tower & filter against the hoop. Water droplets visible on the tower threads, after brewing, seem to confirm this is the case. It could also be a factor the long brew times.

Pour over brewing with one recipe & 5 different grinders.

I have been very keen to understand what makes one grinder ‘better’ than another in a specific role. So for this excercise I subjectively compared my preference scores for 5 grinders that I use for pour over brewing.

This is not ‘science’, nor is it intended to be, neither am I trying to convince you of anything. This is just a record of some brews, dialled in to a single recipe, using 5 different grinders and my observations of the results in the context of brewing day-to-day, drinking the entirety of every cup brewed.

Prior to starting the the sequence, 4 of the grinders were dialled in for the best results I could get, except for the Feldgrind V1. If a brew suggested a noticeable & potentially fixable issue, I made small adjustments as I saw as necessary. The Feldgrind has been set between 33 (2 turns+5) and 35 (2 turns+7) for more years than I can remember and is a grinder that I look at as a ‘normal reference’ at a typical, pour over grind setting.

Here are the representative grind curves achieved by sifting samples with a Kruve sifter – each sample was sifted with 3 sets of sieves. Each sifted sample was a fresh ground sample for each sifting (no re-sifting of the same grinds) and the same coffee was used for all the sifting. Each grinder’s approximate grind distribution standard deviation factor (the square root of 84% grind size divided by 16% grind size) is shown in the brackets.

The recipe – The pour over recipe for all brews was as follows:

Hario Switch 02 with pink & black packaged, loose bagged, Japanese Hario 01 white filters. These are the, now discontinued, tabbed filters with a medium crepe on both inside & outside surfaces.

Brewer preheated with around 150ml of boiling water with the valve shut. Drained, then 13.8g of coffee added, Drip Assist 02 placed on the brewer, starting with 2 pours of 40g every 40s (each taking about 15s) in the centre, then the final 3 pours of 40g/40s (to 200g total) in the ring of the Drip Assist. The Switch valve was open throughout brewing.

Both the Ode & Comandante have the note. “no knock”, as both retained a degree of static. after grinding there was no knocking of the grinders/knockers prior to transferring to the brewer and the overdosed whole bean (half a gram minimum for the Ode, min 0.2g for the Comandante) was trimmed & placed into the brewer.

The Feldgrind had very little static, dosed to 14.0g minimum, but still the free falling grinds were transferred to the brewer. Grinder then tapped & shaken to clear before next grind.

The Niche was just dosed (14.0g min) into the grind cup. After dosing the brewer, the grinder was run & lid was clacked a few times to force out any small amount of chaff/fines prior to the next grind.

The Wilfa has a habit of trapping a bean or so, so was given a quick wobble/shake when starting the motor, the free falling grinds added to the brewer, then the grinder run & wobbled again to clear any chaff/fines before the next brew.

The finest grind is the Ode V1.1, the coarsest is the Wilfa, the Comandante is the 2nd coarsest. Niche Zero & Feldgrind are close to the same average size (between the Ode & Comandante average grind sizes), the Feld having a little more below 400Kruve and a little less over 1200 Kruve – but very close to the Niche Zero overall.

50 brews were completed, 10 per grinder, each of the 10 with a different coffee.

For 1,831 pour over brews with various brewers & grinders, my average score is 7.41/9.00, average EY 19.9% (stdev 1.12%).

Overall, for these 50 brews: Average score 7.44/9.00. Average extraction yield 20.6% (stdev 0.82%). Average brew time 3:14 (stdev 12s).

Fellow Ode V1.1, setting 3.33 no knocking, any fines & chaff clinging to the grinder and spout excluded from the brewed dose: Average score 7.70/9.00. Average extraction yield 20.9% (stdev 0.74%). Average brew time 3:16 (stdev 14s). Comments: 1 brew ‘juicy’, 1 brew ‘delicate & a little filmy’, 1 brew ‘slightly tart’. In a previous test of 24 brews with this grinder, the 12 where I brewed including knocked out fines scored 6.75/9.00 vs 7.42/9.00 for fines & chaff discarded from dose.

Niche Zero, setting average 74: Average score 7.50/9.00. Average extraction yield 20.1% (stdev 1.18%). Average brew time 3:23 (stdev 16s). Comments: 1 brew, ‘clean, delicate, a shade filmy’. 4 brews noted sweetness. 1 brew, ‘drier, high mouthfeel’. 1 brew bang on supplied notes.

Made by Knock Feldgrind V1 with 38mm burrs, set to 34 average: Average score 7.50/9.00. Average extraction yield 20.6% (stdev 0.75%). Average brew time 3:14 (stdev 11s). Comments: 2 brews noted tannic notes, 3 mentions of siltyness, 1 ‘fairly sweet’, 1 tangible supplied notes. This grinder had the most particles under 400Kruve.

Wilfa flat burr, avg setting 24.3: Average score 7.30/9.00. Average extraction yield 20.4% (stdev 0.53%). Average brew time 3:10 (stdev 8s). Comments: 3 mentions of silty sensation but 2 of which still with brightness, 1 mention dry tannins, 1 tangible notes (as supplied by roaster), 1 mention of ‘fairly sweet’.

The stepped adjustment resulted in quite significant changes in the brew – setting 23 made brews overly silty for my taste, 25 tended to make for under-extracted tasting brews. This grinder had the overall coarsest grind & joint least particles under 400Kruve, so I don’t see that the texture/siltyness/powdery mouthfeel is due to an excess of fines/small particles. This grinder had the most particles over 1100, 1200 & 1400Kruve.

Comandante X25, average setting 20.9: Average score 7.20/9.00. Average extraction yield 21.1% (stdev 0.47%). Average brew time 3:07 (stdev 5s). Comments: 2 brews mention muted acidity, or notes. 1 brew delicate, 1 brew thin, 1 brew under, 1 brew sour, 2 brews filmy/silty.

As with the Wilfa, steps are quite large with a profound impact on brews – setting 20 made for chewy brews, 22 under & watery. This is the second coarsest grind in this run.

In the case of the Wilfa & Commandante, brews may have been improved by grinding coarser and breaking the pour down into more pulses, but I wanted to keep the method common for all grinders. I feel the Ode benefitted from being able to exclude chaff & fines, which was much harder to do on the other grinders.

Hario V60 02 with Drip Assist, 13.8g coffee to 210g brew water.

The brewer and filter paper were pre wetted and pre heated prior to brew start.

Grind for this brew allows: 9% to pass through Kruve 400 sieve. 12% to pass through Kruve 500 sieve. 22% to pass through ASTM/ISO 600um sieve.

Start with 30g poured into the centre (each pour taking 10-15s duration).

At 30 seconds pour up to 60g in the centre.

At 1 minute pour up to 90g in the centre.

At 1 minute 30 pour up to 120g in the ring.

At 2 minutes pour up to 150g in the ring.

At 2 minutes 30 pour up to 180g in the ring.

At 3 minutes pour up to 210g in the ring.

For 17 brews, each with a different coffee, average dry bed at 3:28 +/- 7 seconds. Average extraction yield 19.3% +/- 0.4%. Average subjective liking score 7.6/9.

Variation #1 – V60 02 & Drip Assist, finer grind, allowing: 12% to pass through Kruve 400 sieve. 17% to pass through Kruve 500 sieve. 29% to pass through ASTM/ISO 600um sieve.

Start with 40g poured into the centre (each pour taking 10-15s duration).

At 40 seconds pour up to 80g in the centre.

At 1 minute 20 pour up to 120g in the ring.

At 2 minutes pour up to 160g in the ring.

At 2 minutes 40 pour up to 200g in the ring.

For 10 brews, each with a different coffee, average dry bed at 3:19 +/- 14 seconds. Average extraction yield 19.8% +/- 1.0%. Average subjective liking score 7.6/9.

Reasonable expectation in single-dose grinder, dose consistency.

Now, I reaĺly have no desire to rain on other people’s parades, but the frequency with which people claim a dose consistency (AKA retention, even if it is the incorrect term technically) of less than 0.2 grams is a little bizarre.

In several tests of 10 grinds per grinder (you need 0.01g resolution to do this) the stdevs in whole-bean dose fall around 0.04 to 0.06g. Extrapolate this to 95% confidence levels and a stdev of 0.1g in whole-bean dose seems reasonable (I personally expect to get thousands of doses from a grinder, not just cherry pick 6/7 out of 10) .

For your grinder to accommodate 0.1g consistency and reflect it in the output, it would need to know when to add & subtract grinds to achieve 0.1g dose consistency. (Which doesn’t much matter anyway because chucking another bean, or two into the grinder to achieve a consistent, trimmed dose is no hardship at all).

0.2g dose consistency in terms of output is actually very impressive and should be appreciated as such. If you are going to make an outlandish claim at least make it more exciting than an arcane parameter that no one really cares about. Maybe claim an invisibility cloak for the grinder, so that your significant other can’t see what you’re wasting money on? 😁

A frustrating foray into pour over clarity.

After witnessing much hearsay on the benefits (or otherwise) of certain grinders, regarding the brew clarity they contribute, I though I would log a few brews with each of three of my grinders:

Wilfa Unicorn…oops, sorry I mean Uniform.

Niche Zero with NFC disc.

Orphan Espresso Lido 2, fitted with a Kinu brew burr.

I started out with a best guesstimate on grind size with each grinder, then tweaked settings based on flavour, extraction and perceived siltiness in the cup.

The brew regime was to grind 13.8g of coffee, pour 210g of brew water, brewed with a Hario V60 02 augmented with a Hario Drip Assist. I poured 30g of brew water every 30 seconds (each pour taking ~15 seconds each), the first three pours were straight down the centre, the last four pours were around the ring/doughnut…whatever you want to call it.

I elected to rate ‘clarity’ on a five point scale as follows:

Representative of tasting notes: 1 – Not at all; 2 – Not really; 3 – Maybe/maybe not; 4 – A little; 5 – Definitely

If a brew exhibited one of the bag notes, I’d score it a “4”. If more notes than that were detected then I’d score a “5”. None of the coffees displayed all the notes listed in the same brew. I would also note that one of the tastiest coffees scored low on this parameter as it typically didn’t match any notes from green suppliers.

The quality of coffee I have generally received lately seemed to make this endeavour rather moot (will I never learn that multi-roaster subscriptions typically have no QC), nevertheless what did I find?

• If the roast is off, it doesn’t matter what you grind with.
• If the notes are off, you may still enjoy tasty cups even if they don’t taste quite like the supplier says they will.
• Clarity, if I have experienced it here, is a double edged sword, because you might like one note particularly, another not so much…but if clearer, then they stand out more, preferred or otherwise.

To summarise, the Wilfa Uniform displayed the least clarity (2.6/5 and a preference score of 6.6/9), cups felt the siltiest, but setting the grinder coarser than 34 increased sourness and relative under-extraction. Preference for brews with this grinder was a little lower than the others, but generally in the past I have been happy with the brews from the Wilfa.

Brews with the Niche Zero and Lido 2 (a Mk.I Lido 2, originally supplied with Italmill burrs and retrofitted with a Kinu brew burr) had similar preference scores (7.1 and 6.9/9 respectively). The Lido 2 displayed the better separation of flavour notes (3.1/5), which was fine except one of the coffees had a slightly bitter edge and this was most apparent with the Lido and completely absent with the Niche. On more than one occasion the Niche produced the highest perception of sweetness, despite a relative lack of flavour separation, though the clarity score averaged out the same as the Lido 2 over the ten coffees, at 3.1/5.

So, what do you want? Whilst I find flavour separation interesting, day-to-day a high preference score and highest perception of sweetness is most desirable. I don’t need to identify a specific flavour in a coffee (which, let’s face it, is somewhat accidental) to know whether, or how much I enjoy it.

If typical/majority grinders do not produce ‘brew clarity’ and only a few specific grinders do, is it a feature, or a bug?

Here are my perceptions, broken down by grinder:

I then thought I’d carry out some sifting with a Kruve sifter to see if there was any insight there…

Here is a cumulative plot, although I inverted the percentage at 1600 Kruve to make the plot more intuitive to read than an ‘S’ curve…

I have to say that even I was a little surprised to find that all three grinders ended the exercise with essentially the same percentage at 400 Kruve. (The same coffee was used for all sifting.)