Ken Fox wrote:>snip<
I was looking for an easy way of coming up with 14g doses and I guess that just does not exist. SO -- I took the bull by the horns and pulled out a 1/10g scale I'd bought on "suggestion" of Jim Schulman. Scales with this accuracy, although not really that expensive (this one was around $32), do not have a very high maximum weight; in the case of this scale, it is 300g. You risk damaging the scale by putting more than that weight on it, and in any event it won't work beyond that weight, so Tare-ing a PF (even a bottomless one) will not work if you want to weigh shots.

So, I decided to dose into a ramekin, which thankfully weighed less than 200g, and decided to use 14 g. I got both Maxs and the new Compak adjusted to that dose, and have now pulled Brazil Yellow Bourbon, Heather's WBC blend, and some Biloya out of these grinders at that dose out of my modified Cimbali Junior D, PID'd to deliver shots, currently set at a brew temp around 199F.

>snip<

I use a MyWeigh i1200 0.1g scale which handles 1200g - it cost me £85 + shipping in the UK - and my bottomless LS PF weighs 490g. It is faster than dosing into Tupperware which I used to do with just a 100g scale. Also, striking off the excess probably disturbs the nice distribution from a doser grinder less than transferring grounds from another container. If you are using a Titan grinder, it may well be worth the extra cost over a scale which cannot tare a PF.

Cheers,

Hugh

P.S. AnotherJim : Many thanks for investigating optimum doses from the Titan grinders, I'm trying 13.5g from the "relatively small Macap" MXK now and I think it's a definite improvement though I will experiment more later. Previously I had only tried doses from 16g to 19g before settling back at the 17g recommended in the HB review of the S1.

HughF wrote:I use a MyWeigh i1200 0.1g scale which handles 1200g - it cost me £85 + shipping in the UK - and my bottomless LS PF weighs 490g. It is faster than dosing into Tupperware which I used to do with just a 100g scale. Also, striking off the excess probably disturbs the nice distribution from a doser grinder less than transferring grounds from another container. If you are using a Titan grinder, it may well be worth the extra cost over a scale which cannot tare a PF.

Cheers,

Hugh

I have a somewhat similar scale with an accuracy up to 0.5g, which can handle 3kg if memory serves. It was made by Escali and cost around $55 from Amazon. It is considered a "lab scale." I think that 0.5g accuracy would probably be sufficient for these purposes.

What I'm finding, however, is that at least with conical grinders (and I'm considering the Max as a conical for these purposes, as well as my new Compak) that, at least with doses around 14g, I'm having no distribution or resultant channeling problems simply dosing into a ceramic ramekin and then transferring the coffee into the PF. I'm also finding that the process takes very little extra time as I need to do less cleanup afterwards from grinder mess as the ramekin sits nicely on the grinder PF forks, and catches all the coffee coming out of the dosers. It is also very easy to distribute 14g into a double PF basket with a teaspoon which takes about 2 seconds.

I don't know whether planar grinders would have problems with this approach, although I could pull out my old Junior or Cadet to test it. Jim has previously reported that the lower doses one encounters in Italy don't result in the channeling that we commonly see in places using larger doses of coffee.

ken

Wow. The Robur looks like you could not mess up a shot from the flat TDS. I think that alone would enough reason to get one in a commercial environment. The variance in shot from a very consistent and experienced barista vs. those from a novice who throws the dose all over the scale would be minor. I do see where this consistency may be a problem for a home user that likes to experiment with the profiles generated from dosing differences. I wonder how the Kony will fit in.

cannonfodder wrote:Wow. The Robur looks like you could not mess up a shot from the flat TDS. I think that alone would enough reason to get one in a commercial environment. The variance in shot from a very consistent and experienced barista vs. those from a novice who throws the dose all over the scale would be minor. I do see where this consistency may be a problem for a home user that likes to experiment with the profiles generated from dosing differences. I wonder how the Kony will fit in.

I'm beginning to think the Kony, with it's reported brightness, may be somewhat different from the MXK. If the Super Jolly is extra mellow, it may be that every grinder is going to taste bright in comparison. In any case, since I like the MXK from a usability standpoint, I've done quite a few private shots on it; and I'm not finding it bright in comparison to all the other grinders on the table.

I have no idea if the TDS data in any way predicts taste differences between grinders (it does predict taste differences between doses from the same grinder, but I'm not seeing that much carry over to grinder to grinder comparisons). If you want to use the TDS data to see which grinders are similar, here is the data laid out in terms of showing similarities and differences:



The horizontal axis shows the average TDS for each grinder, with the Mini the highest, and the Versalab the lowest. The vertical axis shows how much the TDS varies by dose, again with the Mini as highest, and the Robur as lowest. The closer the grinders are together, the more similar they are in respect to extraction.

However, it's interesting to find out that if you really sweat the details, and use a blend with high potential, getting it just right versus about right easily trumps which grinder you are using.

I did three rounds of blind tasting today with the Robur, the MXK, and the Compak, the three grinders which have people the most interested. Three shots is about the limit of what I can prepare and make in a 5 minute interval, so the earliest ground coffee doesn't suffer. In this case, I used a home roast with aged Sumatra and Yemen Ismaili. When the shot is just right, the taste unfolds into layers of complexity, the taste equivalent of finding yourself inside an Escher print. If the shot is off, all you get is wood and leather, yech.

As it happens, in each round, a different grinder was slightly off. It always lost that round. The difference between the off scores and on scores were greater than the difference between the winning and losing grinder. Here are the results, for the total, the flavor and body score are multiplied by two.

-------------------------------------------
Grinder & Pour    Crema Flavor Body   Total
-------------------------------------------
Compak 1, good     4.0   4.0    3.5    19
Compak 2, tight    4.0   3.0    3.0    16
Compak 3, good     4.0   4.5    3.5    20
   Total                               55
-------------------------------------------
MXK 1, good        4.0   3.5    3.0    17
MXK 2, good        4.0   3.5    3.5    18
MXK 3, loose       3.5   3.5    3.5    17.5
   Total                               52.5
-------------------------------------------
Robur 1, loose     3.5   3.0    3.0    15.5
Robur 2, good      4.0   3.5    3.0    17
Robur 3, good      3.5   4.0    3.5    18.5
   Total                               51
-------------------------------------------
Off Pours Total                        49
On Pours Total / 2                     55
-------------------------------------------

Having actually bought the thing, I'm relieved the Compak has finally won a blind test round. I was beginning to think I'd sprung for the lamest of the lot.

Jim:

Thanks again for the effort which you (and others!) have expended on behalf of this project. I remain intrigued by your findings as they unfold.

Although not part of this particular line-up, if you were to speculate, where would you guess a "Titan" flat burr grinder, such as the Major, might fall on the solubles yield graph?

As I queried earlier in this thread, I still have the "Does size matter?" (burr set, motor output, grind pathway) question in mind and wonder how said Major might stack up against the Robur (or Kony).

-Michael

k7qz wrote:
Although not part of this particular line-up, if you were to speculate, where would you guess a "Titan" flat burr grinder, such as the Major, might fall on the solubles yield graph?

As I queried earlier in this thread, I still have the "Does size matter?" (burr set, motor output, grind pathway) question in mind and wonder how said Major might stack up against the Robur (or Kony).

I don't know where big flat burr grinders would fall on solubles yield, or how sloped they would look; however, I'm guessing they'd tend to be more dose sensitive than the conicals.

But keep in mind that what we are looking at is not part of the grinders' design purpose. They are designed to produce a certain particle distribution at grinds suitable for 7 and 14 gram doses of Italian espresso blends. My guess is that all the grinders are pretty much the same when doing this. As far as the makers are concerned, you pay for speed and reliability, not better grind quality. These grinders are then exported to the US, where they are used at odd doses, intermittently, at grind settings not anticipated by the builders. I rather think all this testing is just to find out which of these grinders screws up the least when (mis)used like this.

Finally, it doesn't look like size matters for the gross extraction data - the Lux, a tiny grinder, and the Robur were virtually identical in my solubles measures. However, what is being extracted may be different even if two grinders have the same raw numbers. For instance, the MXK and M7KR, also virtual twins on solubles yields, taste rather different from one another. So while these figures are illuminating when comparing the same grinder at different doses, they might not be meaningful when comparing different grinders.

I'm sorry we can't give definitive answers yet. All our testing is raising new questions that need addressing before the simple one "which grinder tastes best" can be sensibly answered. However, it looks like the answer may be that they are all close enough in taste that ergonomic issues may be more important. This means questions like which grinder is easiest to use given the way you work, which one is most forgiving if you change blends often, which one is most perfectly tunable if you work with one blend at a time, etc. etc. are the ones that you should look at. My personal obsession with how the grinders taste is because I use a lot with light roasted, acidic coffees, and I'm jonesing for an espresso grinder that can bring up the same nuances I find when brewing these coffees. This is probably irrelevant to most potential buyers.

another_jim wrote:I'm sorry we can't give definitive answers yet. All our testing is raising new questions that need addressing before the simple one "which grinder tastes best" can be sensibly answered. However, it looks like the answer may be that they are all close enough in taste that ergonomic issues may be more important. This means questions like which grinder is easiest to use given the way you work, which one is most forgiving if you change blends often, which one is most perfectly tunable if you work with one blend at a time, etc. etc. are the ones that you should look at.

Thanks for taking a stab at my question Jim!

Reminds me of a sign that was posted in the UCLA Chemistry dept. some years ago (it could still be there I suppose!):

I fully realize that I have not succeeded in answering all of your questions... Indeed, I feel I have not answered any of them completely. The answers I have found only serve to raise a whole new set of questions, which only lead to more problems, some of which we weren't even aware were problems. To sum it all up... In some ways I feel we are confused as ever, but I believe we are confused on a higher level, and about more important things.

-Regards

And now for the entertainment portion of this review, see Jim's Titan Grinder Project: Can it Beat the Mazzer Robur?

That looks like a scale drawing of the Robur and Mini.

another_jim wrote:The conicals show either no slope (Robur and Lux) or very mild ones (the others). Given that the slope on the two flats has been replicated, this result seems like a significant difference between the two types of grinders.

I did a very small series of solubles yield tests today (graph is displayed below).

As you indicate, with the Robur grinder, changing the dose doesn't seem to make an appreciable difference in the solubles yield. Very strange.


Notes:
1. Green dots are 13.5 g doses, purple dots are 17.5 g doses. All were done in the same triple basket.
2. The horizontal scale is brewing ratio (grams dry coffee divided by grams wet beverage). Dots to the left represent "traditional doubles," as you move to the right the dots represent "ristrettos" and one "ultra-ristretto."
3. I should have pulled a LOT more shots, but the results are still pretty compelling: the 13.5 g dose shots fall right in line with the 17.5 g shots.