Speculations on espresso body and mouthfeel

Want to talk espresso but not sure which forum? If so, this is the right one.
User avatar
RapidCoffee
Team HB

#1: Post by RapidCoffee » Sep 07, 2019, 6:20 pm

I've had the privilege to review two espresso machines for Home-Barista.com: the Breville BES920XL and the Decent Espresso DE1+. In both instances I was not able to match the body and mouthfeel* of my Spaziale Vivaldi S1 (although clarity and flavor separation was greater on the BDB and DE1).

This is a bit of a head-scratcher. After all, if the puck "sees" water at the same pressure and temperature, the resulting extraction should be identical on different machines... but this is not the case. What other factors might influence the extraction? There are many possibilities: different groupheads, baskets, pumps. Unlike the BDB and DE1, the Spaz is a fully plumbed single group machine with rotary pump, 53mm grouphead, and nonprecision baskets. Water debit is high (~10ml/s), headspace is low, and pressure ramps up very quickly. Could this have something to do with body/mouthfeel?

I recently ran a small experiment on my Spaz, to test this body/mouthfeel hypothesis.

First, I pulled three shots using a 15g dose, with the grind dialed in to a 1:2 brew ratio in ~35s:
1) standard double basket
2) double basket with paper filter (to remove fines, presumably ruling out nonprecision baskets as a factor)
3) triple basket (to increase headspace)
Extractions were timed for 30g liquid, and tasted for body/mouthfeel.

Image
double baskets on outside, triple IMS basket (same hole pattern) on inside

Image
thin paper filter used on one double basket

Ground 45g+ Klatch Belle Espresso on my Robur, mixed it well, dosed each basket with 15g:
Image
note greater headspace on triple basket

Extractions all looked good:
Image
even paper-filtered espresso pours nicely

Image
results in cup (L->R: triple, double, paper filter)

Time to extract 30g: triple 34s, double 36s, double with paper filter 35s

Taste:
Extractions were sampled after all pours were complete, ~1 minute apart, without blinding.
Belle Espresso is a classic espresso blend, roasted medium-dark, flavor profile dominated by liqueur aromatics.
Allowing for temperature variation, the double and triple pours were indistinguishable.
The paper-filtered double had a subtle flavor shift, which I can only describe as slightly muted and smoother. (Sorry, I'm not great at describing taste.)
Body and mouthfeel was excellent, and essentially identical, on all three pours.

Second, I pulled two shots to test the impact of preinfusion (PI). Same scenario as last time: ground 30g+ Belle Espresso and dosed two double baskets. Then I timed two 15g->30g extractions on the Spaz, one with no PI and the other with PI set to 8s. On the S1V2, PI delays start of the pump, and allows line pressure PI to saturate the puck. My line pressure is set to 3 bar and pump pressure to 9 bar.

Time to extract 30g: PI 0 took 35s, PI 8 took 31s

Taste:
PI 8 had similar body but lower mouthfeel, compared to PI 0. This difference was noticeable straight and with sugar, but imperceptible with milk.

I then repeated these tests using a medium-light roast (Firebird Espresso from Charlotte Coffee Company), and got the same results. The paper-filtered espresso tasted subtly different (muted flavors), and the biggest reduction in mouthfeel came from preinfusion.

Preliminary conclusions:
1) Headspace made little difference.
2) Paper filter muted flavors, but did not have a dramatic impact upon body and mouthfeel. (Perhaps the puck itself acts as a fines filter?)
3) Preinfusion leads to faster flow rates after the puck saturates. I don't understand the mechanism (people claim that preinfusion "softens" the puck, whatever that means), but IMHO preinfusion is a key ingredient in making lighter roast coffee palatable as espresso.
4) Preinfusion seems to reduce mouthfeel, but not body. This makes sense if mouthfeel is associated with emulsification of oils (which requires high pressure), and body with total (in)solubles. With its high water debit rotary pump (and low headspace), the Spaz ramps up pressure more quickly than is possible on the BDB and DE1.

I consider these results very preliminary, and hope others will chime in with experiments of their own.

* I consider body to be associated with both soluble and insoluble material in the extraction. For example, in the world of brewed coffee, French press is higher in body than filter brew. Mouthfeel is subtly different, associated with coating the mouth, presumably due to lipids that emulsify under high pressure. Mouthfeel is much greater with espresso than non-espresso brewing methods.
John

User avatar
peacecup

#2: Post by peacecup » Sep 08, 2019, 7:32 am

I use the smallest-diameter baskets around, the 45 mm Caravel/Ponte Vecchio. Putting 16g in one of these, compared to a 58 basket, makes a much deeper and narrower puck. I've always loved the mouthfeel from these machines. Perhaps the basket size difference is playing an important roll in your findings.
LMWDP #049
Hand-ground, hand-pulled: "hands down.."

OldNuc

#3: Post by OldNuc » Sep 08, 2019, 8:36 am

Length of extraction column is a definite critical parameter in any extraction.

User avatar
AssafL

#4: Post by AssafL » Sep 08, 2019, 8:47 am

RapidCoffee wrote: 3) Preinfusion leads to faster flow rates after the puck saturates. I don't understand the mechanism (people claim that preinfusion "softens" the puck, whatever that means), but IMHO preinfusion is a key ingredient in making lighter roast coffee palatable as espresso.
This is pretty simple. The faster flow is due to less density in the puck.

Short PI leaves air in the puck. When pressure is applied the air compressed and puck density increases, slowing flow.
Scraping away (slowly) at the tyranny of biases and dogma.

User avatar
RapidCoffee
Team HB

#5: Post by RapidCoffee » Sep 08, 2019, 12:02 pm

peacecup wrote:I use the smallest-diameter baskets around, the 45 mm Caravel/Ponte Vecchio. Putting 16g in one of these, compared to a 58 basket, makes a much deeper and narrower puck. I've always loved the mouthfeel from these machines. Perhaps the basket size difference is playing an important roll in your findings.
OldNuc wrote:Length of extraction column is a definite critical parameter in any extraction.
I'm aware of the role that height plays in column chromatography. That's why I compared a 15g dose in the Spaz 53mm basket to an 18g dose in standard 58mm baskets (yielding pucks of equivalent height).

Updosing may increase mouthfeel, perhaps by reducing headspace and allowing for a faster pressure ramp up. I was a surprised that the triple basket (same dose, same puck height, but greater headspace) still had good mouthfeel. Spaz water debit up may be large enough that it makes little difference.
John

User avatar
JohnB.
Supporter ♡

#6: Post by JohnB. » Sep 08, 2019, 12:19 pm

If you are losing mouthfeel with PI maybe you need to use a finer grind. Every shot with my Speedster starts with a 12 second pi ramp from 0-8.75 bar. Typical shots run 45- 60 seconds & produce thick, creamy shots with excellent mouthfeel.

From the Speedster Infusion section of Kees website:
The importance of pre-infusion generally still seems to remain under-estimated. Machines with a slight variation in brewing temperature tend to offer a larger spectrum of different flavours from the coffee grounds. On machines offering a very strict temperature a well-executed pre-infusion becomes imperative, otherwise they might not be able to extract the full range of complexities hidden in the coffee.

Properly infusing the grounds makes for an even wetting and swelling of the grounds. This careful wetting and slow swelling ensure the chance on channeling becomes much smaller. This process also locks in the fines, preventing them from travelling downwards and clogging the filter. Even more importantly however, pre-infusion allows a warming and softening of the fats and oils in the coffee grounds, ahead of the actual extraction with high pressure. This clearly results in a higher yield of solids in the cup. These fats and oils are the carriers of the delicate aromas and flavours.
LMWDP 267

OldNuc

#7: Post by OldNuc » Sep 08, 2019, 12:23 pm

Hard to tell but total column volume may also be a factor. Vibrating and gear pumps have pulsing pressure output while a centrifugal pump has very little to none and this may be a cause for different extraction results.

User avatar
RapidCoffee
Team HB

#8: Post by RapidCoffee » Sep 08, 2019, 12:27 pm

AssafL wrote:This is pretty simple. The faster flow is due to less density in the puck.

Short PI leaves air in the puck. When pressure is applied the air compressed and puck density increases, slowing flow.
Perhaps this is correct, but I'm not aware of any supporting evidence for puck density increase due to compressed air.

There's little doubt that the Spaz rotary pump ramps up pressure much faster than the vibe pump machines. With maximum flow (8ml/s) and preinfusion turned off, the DE1 still takes 7-8 seconds to saturate the puck and produce maximum pressure:
Image

Dan posted an interesting graph many years ago, showing pressure profiles for two different E61 variants and an Elektra A3:
Image
As Dan notes, "the A3's profile in the middle stands out as starkly different from the other two - the pressure rockets up in barely one second, which is consistent with the rapid onset of its extraction in actual use." The Spaz is similar to the A3 in this regard. And if this rapid increase in pressure is significant, then the Elektra (like the Spaz) should also have heavy body/mouthfeel. Any Elektra A3/T1 owners care to chime in?
John

User avatar
RapidCoffee
Team HB

#9: Post by RapidCoffee » Sep 08, 2019, 12:32 pm

JohnB. wrote:If you are losing mouthfeel with PI maybe you need to use a finer grind. Every shot with my Speedster starts with a 12 second pi ramp from 0-8.75 bar. Typical shots run 45- 60 seconds & produce thick, creamy shots with excellent mouthfeel.
I was unsuccessful in matching Spaz body/mouthfeel on the BDB and DE1, despite numerous attempts involving different baskets, grinds, doses, and profiles (including PI settings). Can you compare PI vs. no PI on your machine? If mouthfeel does not change, then there must be other factor(s) involved.
John

User avatar
NelisB

#10: Post by NelisB » Sep 08, 2019, 12:58 pm

JohnB. wrote:If you are losing mouthfeel with PI maybe you need to use a finer grind. Every shot with my Speedster starts with a 12 second pi ramp from 0-8.75 bar. Typical shots run 45- 60 seconds & produce thick, creamy shots with excellent mouthfeel.
The Speedster is king of body and mouthfeel. No, Emperor!