Pre-infusion using flow control vs. pre-wetting - Page 2

Need help with equipment usage or want to share your latest discovery?
vit

#11: Post by vit »

Jake_G wrote:Indeed there are.
Actually, the flow rate into the puck decreases as the pressure increases. In fact, the reason the pressure increases is because the flow decreases. Remember that pressure drop across a valve is proportional to the square of the flow. If your water debit is 4ml/s at a pressure drop of 9 bar (9 bar at the pump, zero at the shower screen/puck), then when the flow is 2ml/s, the pressure drop will only be 2.25 bar (meaning 6.75 bar present at the puck). The beauty of the flow control device is that you can change this relationship by adjusting the valve, but if you leave the valve fixed, the flow will always decrease as the pressure at the puck increases.
Yes, you are absolutely right, I corrected it above. I was thinking about that a while ago and it looks something was on the wrong side of equation so to say ... Anyway, it is not constant ...
Jake_G wrote:Indeed there are.
Also, the air above the puck does not reduce the flow rate of the water into the puck. Not sure where you're getting that info. Water inrush occurs unimpeded at whatever the water debit is set to until the puck begins resisting the flow. It is at this point, when the water forms a liquid seal at the top of the puck, that whatever remaining air in the brew chamber begins to compress, and this can occur as soon as the flow rate exceeds the wicking rate of the puck. There is some evidence that suggests 4ml/s is about the peak that a dry puck can absorb without generating pressure. I suspect this is highly dependent on the anatomy of the puck in question, but something in that range certainly passes the sniff test. Flow higher than this wicking rate can build pressure prior to the puck being fully saturated, and this is why preinfusion is a thing.
I suppose this depends on chosen PI flow rate, thickness of the puck etc ... However, I don't agree about 4 ml/s not generating any pressure (for "normal flow basket" at least). In my case - using Flair at my usual PI flow around 1.5 ml/s (which would translate to about 3.2 ml/s for commercial basket) pressure rise first few second is indeed very low, but after about 10s pressure is progressively rising to around 2.5 bar - if I leave no gap between piston and water. If I leave about 5 ml air gap, pressure rise is considerably slower and after 10s is around 1.4 bar, because air also has to be compressed so its volume is reduced ... so on ordinary espresso machine with some air above the puck, it does have influence I would say, as pretty much the same happen (maybe some air escapes to the puck first second or two, the rest stays above water and gets compressed, so that volume of water from the pump is used to raise the water level above the puck instead of going into the puck). It's however harder to say how the flow curve looks like without some calculation/simulation for particular case. So I would say that puck saturation is not something that happens when the water gets to the bottom, but is a kind of gradual process starting slowly at the first second of PI

User avatar
Jake_G
Team HB

#12: Post by Jake_G »

vit wrote:However, I don't agree about 4 ml/s not generating any pressure (for "normal flow basket" at least). In my case - using Flair at my usual PI flow around 1.5 ml/s (which would translate to about 3.2 ml/s for commercial basket) pressure rise first few second is indeed very low, but after about 10s pressure is progressively rising to around 2.5 bar - if I leave no gap between piston and water.
Sure. But remember that a machine with no boiler will behave quite differently than a pump-driven machine moving hot water into a dry basket. The 4ml/s claim is based on how much water flow it takes to create a water seal over the top of a dry puck. In the case of the Flair or Robot, this is achieved the moment you pour water into the cylinder, so you always start with a puck that has a slug of water sealing the top of the puck. With pump machines, you start with drops of water raining onto a dry puck, so the initial wetting of the puck occurs under dramatically different circumstances.
vit wrote:so on ordinary espresso machine with some air above the puck, it does have influence I would say, as pretty much the same happen (maybe some air escapes to the puck first second or two, the rest stays above water and gets compressed, so that volume of water from the pump is used to raise the water level above the puck instead of going into the puck).
Remember though, when 1ml of water enters the headspace between the shower screen and a dry puck, the volume of the headspace is decreased by 1ml. So, depending on the initial flow rate of water over a dry puck, more or less air will remain trapped in the headspace at the moment a water seal forms on the surface of the puck. Prior to that moment, the indeed takes the only exit available, which is down through the puck. It seems counterintuitive, but it is no different that pushing air out of a Turkey baster or a syringe.

Again, the dynamics with a Flair or Robot are very different because the water seal is already there, so no air can be evacuated through the puck until all the water is pressed through. Pump machines, and even traditional lever machines really don't have this phenomenon. PI on your Flair is all about how quickly you move that column of water through the puck, but on traditional machines, it is all about how quickly you establish the equivalent column of water.

To that point, Chris' suggested technique (which I agree is a very good technique, I simply disagree on the semantics that it is the "best" under all circumstances) is really all about establishing that column as quickly as possible and then moving to a phase where the shot dynamics mimic what you have the moment you place the piston into the brew chamber on the Flair.
vit wrote:So I would say that puck saturation is not something that happens when the water gets to the bottom, but is a kind of gradual process starting slowly at the first second of PI
How is puck saturation anything other than the moment the puck is "saturated" with water? Again, the Flair and Robot are wonder illustrative tools of what PI means since the fill stage occurs the moment you pour water into the brew chamber from a kettle. This clearly separates the idea of preinfusion from filling the headspace, whereas with a pump machine, preinfusion is misunderstood as being everything that happens before the puck is delivering coffee into the cup. I think of preinfusion (incorrectly, but it's how my brain works) as the process of moving water through the puck prior to generating peak brew pressure on it. So the Flair example works perfectly. Push slowly (how are you measuring flow rate on the Flair, anyway? Do you have graduated marks on the lever or something?), and move water through the puck at low pressure until you have drips. Once you have drips, the puck is saturated and you can push harder and generate peak pressure. Alternatively, you can push hard right off the get go and generate peak pressure prior to saturating the puck. This will result in a MUCH slower flow through the puck over the remainder of the shot and my hypothesis is that this is because the column of water over the puck was used as a "super tamper" on the dry puck, compressing its structure and blocking the flow of water. If you wait until water has been introduced into the puck (I would argue this is the process of preinfusion), the water serves as a stressed member in the structure of the puck and prevents its collapse under pressure, hence the faster flow rate and corresponding finer grind size that is used with this approach.

Anyway, this is why I take the approach that preinfusion is all about pressure and less about flow. My goal is to avoid collapse/conpression of the puck until it is sufficiently "supported" by water becoming part of the structure. The flow rate at which this occurs is not particularly knowable with most pump machines, but holding the pressure at some arbitrarily low point until the basket is dripping is easy to do.

Cheers!

- Jake
LMWDP #704

User avatar
RapidCoffee
Team HB

#13: Post by RapidCoffee »

Jake_G wrote:Anyway, this is why I take the approach that preinfusion is all about pressure and less about flow.
Perhaps this is largely semantics, but I take the opposite approach. Puck pressure is essentially zero during "preinfusion", so you can only talk about line (or pump) pressure. When the puck saturates, pressure rises and becomes a meaningful parameter. Before that... not really.

The DE1 allows you to adjust "preinfusion" by changing the flow rate into/through the puck. Puck saturation time changes more or less linearly with flow rate (i.e., the puck saturates twice as quickly at 8ml/s as 4ml/s). This provides a more useful and reproducible method for modifying extraction than pressure adjustment.
John

User avatar
Ursego (original poster)

#14: Post by Ursego (original poster) »

Jake_G wrote:Anyway, this is why I take the approach that preinfusion is all about pressure and less about flow. My goal is to avoid collapse/compression of the puck until it is sufficiently "supported" by water becoming part of the structure.
Let's say that the moment has already come when "the puck is sufficiently "supported" by water becoming part of the structure", and the puck is ready for a strong stream of water to brew. What does it matter how it was achieved - under pressure of 2-3 bars (pre-infusion) or with no pressure at all (pre-wetting)? I think, that approach is "preinfusion is all about fully wet puck and not about what did you do for that" (that is my approach too!).

User avatar
Jeff
Team HB

#15: Post by Jeff »

Yes, it does matter.

As you can tell, it is likely that the answer as to what is "best" will depend on your machine, grinder burrs, coffee, water, extraction profile and preferences in flavor.

A more actionable question might be

Should someone starting to explore espresso be concerned about the choice of how basket fill is done?

I think there, as long as you're wetting the face of the puck reasonably evenly with good repeatability†, they will all provide a good starting point for later explorations.


† I'm not a fan of "raise the E61 lever half way" on a reservoir-fed machine.

User avatar
RapidCoffee
Team HB

#16: Post by RapidCoffee »

Ursego wrote:What does it matter how it was achieved - under pressure of 2-3 bars (pre-infusion) or with no pressure at all (pre-wetting)?
The rate of "preinfusion" does matter. Longer, slower PI seems to soften the puck, and results in higher flow rates during pressurized (~9 bar) extraction. Shorter, faster PI results in slower flow rates. There is a noticeable difference in taste and body/mouthfeel. I prefer slower PI for lighter roasts (usually harder to extract, and benefit from a finer grind), and faster PI for more classic espresso blends.
John

vit

#17: Post by vit »

Jake_G wrote: Remember though, when 1ml of water enters the headspace between the shower screen and a dry puck, the volume of the headspace is decreased by 1ml. So, depending on the initial flow rate of water over a dry puck, more or less air will remain trapped in the headspace at the moment a water seal forms on the surface of the puck. Prior to that moment, the indeed takes the only exit available, which is down through the puck. It seems counterintuitive, but it is no different that pushing air out of a Turkey baster or a syringe.

Again, the dynamics with a Flair or Robot are very different because the water seal is already there, so no air can be evacuated through the puck until all the water is pressed through. Pump machines, and even traditional lever machines really don't have this phenomenon. PI on your Flair is all about
how quickly you move that column of water through the puck, but on traditional machines, it is all about how quickly you establish the equivalent column of water.
Yes, I'm aware that dynamics in Flair/robot is different. First, there is some "0 bar preinfusion" after you start pouring the water, mount the piston to cylinder (or basket to the stand in case of Robot) and before you actually start pushing the piston (and water) - in my case this phase is 10-12s long (could be a bit shorter, but can't be avoided) plus some other not that important details. This doesn't happen in "regular" machines. And obviously air between water and piston (if there is a gap - usually there is some) stays there and gets compressed

In "regular machines" puck is obviously dry at the beginning and there is some gap between the puck and shower screen, then when you start the preinfusion water starts spraying onto the puck (with variations regarding number and distribution of holes among machines), some of it being absorbed into the puck and some starting filling the gap between the puck and shower screen. Obviously lower the flow, bigger part of the water is actually being absorbed into the puck at close to zero pressure. So there is obviously some time before a film of water is formed above all puck surface. But ... what would actually force the air above the puck to escape downwards ? Water droplets exit the screen and hit the puck; if let's say flow is that low that all water is absorbed by the puck, air above will stay where it is. When there are enough droplets in that gap to start pushing the air downwards (by their displaced volume) between droplets, they will connect with each other and seal the surface. Ok, if this whole process is very uneven for some reason, that might take some time, maybe even a few seconds, but do we actually want this to happen ? It would create a lot of micro channeling so to say ... Or if the gap is only 1-2mm, I believe big part of air may actually escape downwards (in that case amount of air is small anyway), but I thought it was bigger in most machines or am I wrong ?
Jake_G wrote:how are you measuring flow rate on the Flair, anyway? Do you have graduated marks on the lever or something?
I put a millimeter scale on the piston. 1mm per second is around 1.5 ml / s. Just aligning millimeters passed with ticking of the clock (actually made a custom video clip with numbers and sound for that purpose, which is used for controlling duration of cylinder preheating as well) :D For lighter roasts using 2 mm per 3 s or 1 mm per 2 s ...

PIXIllate
Supporter ♡

#18: Post by PIXIllate »

RapidCoffee wrote:Perhaps this is largely semantics, but I take the opposite approach. Puck pressure is essentially zero during "preinfusion", so you can only talk about line (or pump) pressure. When the puck saturates, pressure rises and becomes a meaningful parameter. Before that... not really.
These are my thoughts exactly.

User avatar
Jake_G
Team HB

#19: Post by Jake_G »

I think it really comes down to a change in terminology.

With levers, preinfusion is that thing that happens before you let the piston do work. Likewise with line pressure feeding an HX or a double boiler, preinfusion was always that thing that happened before you turned the pump on.

The DE1 has changed the vernacular. I think that the "fill rate" is a thing that matters, and it is indeed totally related to flow and not at all pressure, since the pressure is zero during fill. If you fill really slowly, you can preinfuse at nearly zero pressure using the same, or nearly the same fill rate. However, I maintain that there is differentiation between filling and preinfusing and that the difference is pressure-based and not flow based. For instance, if one were to fill at 4ml/s and then ramp to full pressure when the headspace is filled, there is basically zero preinfusion. A pump machine with a 4ml/s water debit will do this with no input from the barista whatsoever. It's just "a shot".

However, if this same fill rate is used to bring the pressure up to say 3 bar and then the puck is held there until it starts dripping, I call this phase preinfusion. This is what a line-pressure-fed machine would do that has a water debit of 4ml/s with the brew valve open to the water main pressure, which is the classic definition of preinfusion. When the basket is dripping, you kick the pump on and you exit preinfusion.

In both of these cases, the fill rate is the same 4ml/s, but in one case, brew pressure is reached right away, and in the other, preinfusion pressure is maintained at 3 bar until the puck is fully saturated. Having pulled lots of shots both ways, I can tell you that the resulting pours are very different.

The fill rate is very important, but it is not definitive of what preinfusion is. But thanks to the DE1, the vernacular has taken on. So fill rate becomes preinfusion flow rate, which is fine, I guess. Especially since now every machine has preinfusion, since they all have a flow rate. :|

So what of preinfusion pressure? Well now that's called "bloom" because that makes as much sense as sweating like a pig does. Slayer coined the term "bloom" in their early literature to tie back literally to a pourover bloom; add the water slowly and allow the puck to bloom, meaning to off-gas and expand. It makes sense that this could happen whilst trickling water into the puck. But now, to Bloom simply means to hold at a zero flow, regardless of the pressure. Early on, at least the flow needed to be zero. But now, the idea has spread to including holding the puck at some arbitrary pressure. How "blooming" at 3 bar makes any sense at all is beyond me. That's just the classical definition of preinfusion.

So yeah. Different naming conventions.

The Slayer approach is novel because you can accomplish complete puck saturation before the pressure ramps up to full pump pressure without needing a discrete change in the flow delivery method. Many recipes have partial saturation with a 1.5ml/s flow rate maintained for only 5-8s before ramping up to pressure, but you can easily stay in pre-brew long enough to build full pressure if you grind fine enough. Whether or not you should is another conversation entirely. Slayer made it a point to call this "Pre-Brew" and it really should be something different than preinfusion, because a Slayer has no method of doing traditional preinfusion.

Plumbed machines like a Synesso Cyncra are setup so you can hold line pressure preinfusion as long as you like. 5s, 10s or whatever suits you. The point being that the difference between preinfusion and brewing is the pressure at which they occur. The flow rate really isn't part of the equation with classic espresso machines; it just is what it is, and it is very dynamic.

In both cases, by controlling the time spent in either a pre-brew flow rate or a preinfusion pressure, you can control how the resulting shot flows. Longer time spent in either "pre" condition leads to a finer grind to maintain the same shot time. But the approach taken on the Slayer isn't preinfusion, it is just a normal pump machine with a shockingly low water debit...

Perhaps I need to get with the times and update my vocabulary, but it is really unfortunate that works that have had historically valid meaning suddenly mean something different and now there is a new word or several words that must be used in specific context to convey the idea of the historical word that made perfect sense - even if it didn't.

Cheers-

- Jake
LMWDP #704

PIXIllate
Supporter ♡

#20: Post by PIXIllate »

Jake_G wrote:I think it really comes down to a change in terminology.

Perhaps I need to get with the times and update my vocabulary, but it is really unfortunate that works that have had historically valid meaning suddenly mean something different and now there is a new word or several words that must be used in specific context to convey the idea of the historical word that made perfect sense - even if it didn't.
Thanks Jake. There's a lot of good information there and definitely some things I'll think about. I actually don't think this is all semantics and it probably should be reconsidered in the broader historical context of the technologies/machines where the terms where first used to discuss and describe.

It should also be noted that after a lot of technology advances (DE1...) there does seem to be a strong trend amongst advanced users towards a very classic lever style shot. Everything old is new again, but now with an app and a graph.

I agree, I don't like the word bloom when it comes to espresso but it does seem to be in common use. I'll often say bloom/zero flow. Zero flow/pressure hold is probably better.

Or maybe I'll buy a lever and learn to stop worrying and love the bomb.