Hi,

Ken Fox's suggestion about dosing seems to be working for my Cimbali M29 with the stock double baskets. Some coffees, such as Colombian Popayan, I cannot get to extract tolerably at all unless I distribute *under* the basket rim before tamping. Distributed even with the rim or above, they block, they channel, they taste awful, whatever the grind, whatever the tamp.

Other coffees, such as Indian Mysore, are far more tolerant, but even they work better, on this machine, if dosed no higher than the basket rim.

So my question is ...

What is it about "overdosing" that makes channeling more likely? Is it that any unevenness in the puck runs up against the screen, which then cracks the puck during lock-in? Otherwise, I'd think that an "overdose" is nothing more than an extra layer on top of an "optimally" thick puck, and therefore *less* likely to channel.

Thanks,

Eric

The taller the puck in relation to its breadth, the greater the pressure required to send water through the coffee grounds, all else staying the same. And the greater the pressure, the more likely the water is to "discover" the weak spots, as it were, either in the cake or at the side-wall.

Regards
Timo

Hmmm... that confuses me a bit. The pump keeps the pressure constant, doesn't it? If we forced a fixed volume per unit time through, then I could see that the pressure would be higher for a thicker puck.

Or do you mean that any existing channel will seem that much easier to get through in comparison to the thicker, non-channeled surroundings?

In any case, does it make sense that a thicker puck would *have* more channels, or that any channels it would have would be more exploited, or both?

Eric

Illy writes (in The Chemistry of Quality, § 8.5.4) about basket dimensions and "cake shape", specifically the ratio of height to diameter, remarking that when the height of the basket exceeds one-fifth the width of the basket, "excessively high pressure" is required for optimal extraction. I have noticed, when using manual levers, that more pressure is required to push water through pucks that are taller -- even when the dose remains the same.

The pressure is a result of a number of variables, and the flow through the coffee medium and out into the cup alleviates the pressure. Any factor that slows the outflow will cause pressure to build. It might take more than 9 bars for the water to percolate through the coffee. Depending upon how the pump is configured -- can it ramp up to, say, 11 bar? is there an over-pressure valve and if so at what point is it set to crack? -- you will either choke the machine, or the water will be pushed through the puck, or out the OPV. The pumped water seeks the path of lesser resistance. When a weak path opens up, the greater the pressure at the time, the worse the gusher.

Overdosing and heavy tamping are not protection against channeling. Overdosing and/or heavy tamping invite channeling.

Regards
Timo

If you overdose too much, the puck will scrape on the shower screen creating cracks in its surface. You could also create density differences in the puck by compressing part of the puck with the shower screen while another portion remains uncompressed.

There is also a headspace issue. More headspace allows a more gentle infusion IMHO. That gentle infusion gives you more forgiveness. Remember water will seek the path of least resistance and exploit any defect in the puck. Then there is the migration of the fines in the puck. The added cushion from the headspace may allow those fines to migrate to the point of least resistance which would basically self heal the defect in the puck. That may also explain how a shot that starts with a little pinhole channel may suddenly plug up and you end up with a good shot despite the less then spectacular start. Just a theory from a mad scientist espresso guy.

ericpmoss wrote:What is it about "overdosing" that makes channeling more likely? Is it that any unevenness in the puck runs up against the screen, which then cracks the puck during lock-in?

Eric, it might not be the overdosing itself that makes channeling more likely, but the reduced headspace that results from cramming more coffee into a given size basket.

This could have several effects:

(1) If the overdosing is severe enough, the top of the coffee cake will be damaged simply inserting it into the group. This will create fractures which evolve into channels.

(2) When there's lots of headspace, the coffee at the top has room to move around and "self level." Weak spots have a chance to fill in. When there's little headspace, the coffee is probably more restricted in movement because of contact with the shower screen and self-leveling doesn't happen.

(3) Also when there's lots of headspace, it takes an extra second or so to fill. This more gentle buildup allows the coffee grains to expand and interlock before they're subjected to the harsh brutality of 9 bars hydraulic pressure.

The stuff i've mentioned above can pretty easily be tested by comparing the results of using a fixed dose in shallow and deep baskets. This should distinguish between the effect of a thick or thin dose and the effect of a big or small headspace.

There are two issues/problems here. Your post addresses mostly #1 (below) but I personally think that #2 (further below) is even more important.

(1) For reasons already explained above, and perhaps others not yet expressed, going beyond a certain fill level in a PF basket and/or reducing the "headspace" below a certain minimum, the above being machine-dependent, will result in fracturing of the puck or separating of the puck off the sidewalls of the PF basket and allow fast passage of water that is bypassing the puck itself. We call this channeling in these parts and it is fairly obvious even to a newbie with a bottomless PF. As you improve your barista skills you can get around this problem, but getting around this problem does not prevent issue/problem #2, below.

(2) Espresso is an emulsion produced by forcing hot water at high pressure through a puck of ground coffee. There are different things one can extract from the coffee and in fact what is extracted differs during the period of extraction. What is extracted also differs by how much water is forced through how much coffee over what period of time. When you put more coffee in the way of the same amount of water, you will as a result extract differently than you would extract from a smaller amount of coffee. Those things that are more soluble will be preferentially extracted, all other things being equal, when one uses more vs. less coffee. As a result, the taste will be impacted by seemingly innocuous changes such as increasing from 14g (typical Italian dosing) to 18 or 19g (typical N. American dosing).

It is not surprising that different sorts of coffee will have different impacts on their flavors if one "updoses" vs. using a standard dose, and you might ultimately arrive at the idea that you should roast differently in order to get a certain type of result from a larger than from a smaller dose.

My observation is that if you use lightly roasted coffee (where the roast is terminated before the onset of 2nd crack), and if the coffee is good and suitable for use in espresso, a lower dose will emphasize the varietal flavors that you are paying for in having bought a good coffee. More darkly roasted coffee might taste "better" if you like that sort of thing, in a larger dose.

ken

Based on the following observations, I have some doubts about the "fractured puck" theory: I sometimes pull a shot without any tamp whatsoever -- the surface is already "fractured", and there's plenty of head-space so the screen is not compacting the coffee grounds -- and no channeling occurs if the coffee has been given time to swell adequately under the gentler pressure of preinfusion, before pressure ramps up.

Regards
Timo

Ken Fox wrote:.... As a result, the taste will be impacted by seemingly innocuous changes such as increasing from 14g (typical Italian dosing) to 18 or 19g (typical N. American dosing).

.........

My observation is that if you use lightly roasted coffee (where the roast is terminated before the onset of 2nd crack), and if the coffee is good and suitable for use in espresso, a lower dose will emphasize the varietal flavors that you are paying for in having bought a good coffee. More darkly roasted coffee might taste "better" if you like that sort of thing, in a larger dose.

This looks contradictory to me, because a typical italian roast is darker than the roast i've seen in N. America.

Illy writes also in The Chemistry of Quality, § 7.5.1, that "Minimum Values (5 gram of coffee per cup)..... ..... are only barely tolerable when dark roasted coffee is used, for it has a higher content of soluble substances."

Greets from Jasper
(barista for Kimbo, dark coffee from Naples )

It really depends Jasper; the traditional Northern American roasting style is indeed lighter than the Italian style of roasting. However, the "3rd wave" roasters tend to roast darker, as does *$.

I don't however think that a darker roast benefits from a higher dose per sé, I think it depends on the dose vs the supposed size of the basket. For example, I get better extractions when I dose 18gr into a 21gr basket than I would with a 21gr dose. I think this has to do with the extra headspace which my Silvia is always in dire need of