My long and rambling path to preinfusion/pressure profiling - Page 42

Need help with equipment usage or want to share your latest discovery?
Mesmer
Posts: 71
Joined: 7 years ago

#411: Post by Mesmer »

What's the buzzing sound until engaging the pump, also during the blooming?

User avatar
Jake_G (original poster)
Team HB
Posts: 4293
Joined: 6 years ago

#412: Post by Jake_G (original poster) »

Bunkmil wrote:Thanks for the info!

Then I guess the profile would look like this on my DE1 :

1.5ml/s until 3bar
30s bloom
8bar for 15s
1ml/s until the end of the shot
That sounds about right. Although I'm not sure I ever actually hit 3 bar. At 1.5ml/s I've just hit saturation at 19 seconds assuming my machine is similar to Assaf's GS/3 relative to volume needed to get drips. Also bear in mind that my 1.5ml/s is the starting point. My pressure will rise slower than a DE1, because the restriction in flow from the puck as it fills up -well- restricts the flow :P

As such. The dynamics of pre-brew are different on a needle valve machine than they are with any other method of flow profiling. My peak pre-brew pressure is line pressure at zero flow, and It's zero bar at the measured water debit. So, where the DE1 pumps at the preinfusion rate until a trigger is met, my machine starts at the preinfusion rate and the flow decays as the puck saturates and pressure approaches line pressure. My goal is to let all this magic happen and then choke the flow before beading occurs on the basket. I waited too long in my video to cut the flow, but the resulting shot was still very good, both in terms of flow stability (puck not eroding and gushing) and far more important, flavor.

Not sure if the advanced shot editor let's you have a sloping flow profile, but if it does, I would start at 1.5ml/s for like 10 or 15 seconds to fill the headspace and then have the flow target drop to zero over the next 10 seconds and then hold zero and bloom for 30. This should get you a more gradual bloom and it will act more like my machine does naturally. I'm not saying this will make better coffee. But it's what my machine does if I pre-brew with a fixed water debit prior to blooming.

Give it a try and let me know what you think!
Mesmer wrote:What's the buzzing sound until engaging the pump, also during the blooming?
Ha! That would be my 25 year old Sirai brew solenoid. What you're hearing is 120Hz hum due to a worn out shading ring in the pole piece of the solenoid valve armature. When the coil on a solenoid valve energizes, the holding force on the electromagnet diminishes every time the AC waveform passes through zero (twice every cycle, so it's not 60Hz hum). In a new or even healthy "advanced age" valve, the shading ring makes contact with the plunger in the solenoid when its energized and allows the magnetic field that is induced in the plunger to create an out of phase current through the plunger and into the ring. This current creates it's own magnetic field that keeps the plunger seated and eliminates the buzz when it's working.

Shading-Ring
Understanding Electrical Solenoids

Mine is working well enough that the valve doesn't chatter, but it's working poorly enough that you can hear the buzz. One of these days I'll fix it :wink:

Cheers!

- Jake
LMWDP #704

Advertisement
Bunkmil
Posts: 358
Joined: 10 years ago

#413: Post by Bunkmil »

Jake_G wrote: Not sure if the advanced shot editor let's you have a sloping flow profile, but if it does, I would start at 1.5ml/s for like 10 or 15 seconds to fill the headspace and then have the flow target drop to zero over the next 10 seconds and then hold zero and bloom for 30. This should get you a more gradual bloom and it will act more like my machine does naturally. I'm not saying this will make better coffee. But it's what my machine does if I pre-brew with a fixed water debit prior to blooming.

Give it a try and let me know what you think!

Cheers!

- Jake
There isn't that much that the DE1 can't do :D! So yes I can set a declining flow step. I actually just tried it and it worked great! The headspace that needs to be filled is a little higher on the DE1 than on most machines so I ended with a 23s @ 1.5ml/s followed by a 7s decline step to 0.0ml/s. The pressure didn't get above 1bar with this method and the bottom of the basket was wet by the end of the blooming stage.

Really liked my shot!

Mesmer
Posts: 71
Joined: 7 years ago

#414: Post by Mesmer »

Also tried shots like this and indeed they are quite interesting. Much like a concentrated espresso. Even visually you can see more body and more oils in the coffee. Taste wise all the flavors are merged together, like say a conical grinder.

One think i can't understand is why these shots are not overly bitter, as you would expect since there's a long and thorough extraction.

User avatar
Jake_G (original poster)
Team HB
Posts: 4293
Joined: 6 years ago

#415: Post by Jake_G (original poster) replying to Mesmer »

Dan,

You've forced me to draw a picture! (Well, edit a picture, shamelessly stolen from here.)
Below is beginning stage of a "typical" extraction where full brew pressure is reached prior to the basket saturating with brew water.

The small section above the grey line is the "Wet Puck". There is 9 bar on top and a dry puck underneath it. This thin and wet puck sits atop the dry puck below and presses on it with the full brunt of the 9 bar water pressing on it. There is nothing else supporting it other than the dry puck below, so the wet puck atop effectively tamps the dry puck below with a force equal to brew pressure times puck area. This Super-Tamp results in a very difficult life for the water as it tries to pass through the dry puck. I've created 3 fingering flows below the grey line, which can be thought of as channeling. However, any time you see a spot or two of flow on the bottom of a basket, thisis what you're really seeing; it's the tips of the fingering flows before they spread throughout the rest of the puck, which they generally do. Water desparately seeks a path through the compacted puck and once it establishes flow, the flow through those areas first traveled increases. Generally, there are more than three of these fingers, and they coalesce relatively early in the shot. The thicker the wet puck, the better the extraction of the wet part, because unlike in the dry puck, the flow has no preferred path through the wet puck.

Even if the shot pours well, these early flows still extract first and they are the parts of the puck that blonde first and over-extract when you pull long shots. So, why does blooming "fix" this? Well, here's why. Suppose you limit the available pressure as shown below:

Then lets just wait and let the puck wick up the water above it and bloom with no flow at all. So, the puck soaks up some water and releases some CO2. The CO2 rises up into the headspace and gently forces more water (from the headspace) through the puck. Additionally, diffusion allows the more dry areas of the puck to equalize with adjacent more wet areas and you end up with a relatively homogenous saturated puck. As long as your initial charge of water was sufficient to saturate the puck, I believe the puck will self-charge during bloom and ensure that at the end of the bloom phase, you will have a fully saturated puck with a relatively dry headspace of air and CO2 above it:

Now, you are free to extract your shot, but there are a couple interesting things to observe at this point. First, most of your shot is in the puck at the end of bloom. Your job now is to extract your shot from the puck and then dilute it to your desired brew ratio. Second, with a completely saturated puck, you can extract at just about any brew pressure you want to without fear of choking the machine. The water supports the physical structure of the puck and keeps it from collapsing on itself, which causes channels. While there's no such thing as perfection with respect to flow through a puck, starting with a completely saturated puck is a MAJOR benefit when it comes to channeling and obtaining a more even extraction. At this point, you start the extraction in earnest and introduce your solvent (the water) to the top of the puck and push out the filtrate (your espresso) from the filter (the puck, now evenly stripped of it's soluble compounds thanks to the long and arduous procedure of infusing the cake with an initial charge of water and allowing it to bloom and saturate and equalize until we're ready to extract.

This whole process is exactly why I prefer to time my blooms such that the basket wets by the end. If it wets mid bloom, I probably charged with too much water during the infusion process, or my grind was too coarse. If it doesn't wet during my planned bloom time, then the opposite of one of those conditions must be true. If I bloom longer and the basket wets, it means I ground a bit fine, but I had plenty of water in the charge. As long as I wait until the basket bleeds, the flow should not be choked, as the longer you wait, the "softer" the puck gets. While there are natural limitations in play here, I find that I can generally work within whatever the bloom tells me as long as it's 20 seconds or longer. Any shorter and I"m doomed with a gusher, which I can pull slower with reduced pressure and make it ok, but it won't be great... If I wait and wait and the basket never wets, that just means that I didn't get enough water in the initial infusion. No big deal, but not Ideal. Generally speaking this doesn't happen as I have plenty of tactile indicators to guide me in getting enough water in before I bloom. My machine is funny in that I can actually hear the heaspace filling up and there is a corresponding shift in the tone of the water through the needle valve that tells me when it's time to kill the flow.

The next natural progression here is extraction, which I hardly need to paint a picture for, but what the heck ;)

As long as you don't botch the puck prep, you can pull this kind of a shot off with a rediculously fine grind and it will pull evenly at whatever extraction yield you want without getting those nasty over-extracted flavors coming through. Since the whole puck is extracted more evenly, there are no areas in the puck that get taken to that extreme level of -over-extraction that you get when areas of a dry puck fracture and channel. You can basically stop the shot after all the filtrate is stripped from the coffee puck and clear water is coming out and it will still be a good shot. Whether or not the coffee you are pulling tastes best when pulled like this is another question altogether. In general, anything taken near second crack is probably not going to be great when extracted this far, but that's not a hard and fast rule. What I can say is that I haven't found this approach to be preferable when pulling coffees that have any ashy or roasty flavors to them. Those coffees seem to be best pulled fast and cool.

Cheers!

-Jake
LMWDP #704
★★ Quite Helpful

pcrussell50
Posts: 4010
Joined: 15 years ago

#416: Post by pcrussell50 »

And the slow but growing body of knowledge for us early profilers takes another step forward here at HB.

Bookmarked ^^^ for continuous future reference.

Well done, Jake.

-Peter
LMWDP #553

Mesmer
Posts: 71
Joined: 7 years ago

#417: Post by Mesmer »

Indeed, thanks for the nice explanation.
I tried shots like these after first installing the pump dimmer on my machine. They were still very concentrated but with small amounts of harshness as well. I was running the pump at a low flow until beading and then 9 bar. Starting grind size just about where a normal shot would not flow at all.

This morning i tried to take in account the amount of water during pre-infusion so that coffee would bead on the bottom of the basket at the end of blooming. I'd say the shot is much sweeter, with a lot more body. Funny how you can actually see the oils on the bottomless portafilter with this method. Also when cleaning up if feels a lot more oily, even though i'm using light roasted beans.

For people going this route, besides the modding or lots of $$$ required for a profiling machine, the grinder is also going to be a problem. By design most traditional espresso grinders would clog because of the fine grind needed.

Advertisement
User avatar
Jake_G (original poster)
Team HB
Posts: 4293
Joined: 6 years ago

#418: Post by Jake_G (original poster) »

Long and Nerdy Diatribe regarding Needle Valve Profiling:

Ok, I've said most of this before in bits and pieces, but there is enough momentum around folks modding their machines with needle valves in the brew path that I think it's fair to rehash a few sticking points relative to the strategy as a whole. Let's talk a bit about what we are doing, why what we are doing is doing what it's doing, and how we measure what we are doing, when we are doing it.

What's a Needle Valve, Anyway?
In short, a needle valve is an adjustable gicleur. We talk a bit about jets, restrictions and gicleurs from time to time, arguing which is best. The stock 0.8mm restrictor or the 0.6mm on an LMLM; the 0.7mm on most E61s and for the primary brew path in a Slayer; or none at all on old Lineas and countless other commercial machines! Enter the needle valve. Yep, Slayer did it first-ish. Let's have TWO jets! It's pretty cool. But why not just have the best of ALL worlds and be able to change the jet size on the fly?

Not Pressure Profiling, Not Flow Profiling
Some folks refer to Bianca and nodded BDB machines as flow profilers, and I'm guilty of using that label from time to time, but it's not any more accurate than pressure profiler. Needle valves do something different than changing the speed of a gear pump (volumetric) or changing the cracking pressure of an OPV (true control of max available pressure). What needle valves do is unique because they control and alter the relationship between available brew pressure and flow through the puck. In so doing, they also limit the maximum available flow rate, based on the pump pressure.

Variable Water Debit
Water debit is a silly concept, but it's a powerful measurement. We think of it as "max flow", but it's really so much more than that. We experience the water debit of our kitchen sink faucet every time we turn it on. Depending on what we're doing, we may turn it on full blast and fill up a jug as fast as possible, or we may turn it down to a trickle if we're using the water stream to clean something delicate. This is essentially all a sink faucet ever does. We open the valve and the high-pressure water behind the valve passes through it and leaves the spigot at atmospheric pressure. So why do we care so much about turning our espresso machine into a kitchen faucet?

Iterative and Interconnected Properties of Pressure and Flow
Water debit describes what happens when we allow all of the available pump pressure to drop across the gicleur or needle valve, with no puck in the way to generate any back pressure on the downstream side of our restriction. So what happens as the puck generates some back pressure? Well, increased pressure on the downstream side of our valve means there is less pressure drop across the valve. Less pressure drop means less flow. How much less is a function of the size of our restrictor and can be predicted based the measured water debit! In the case of fixed restrictors, like a 0.6mm gicleur, there is a nice curve that shows you how much flow to expect at what pressure drop across it. So, if we choose a valve position that yields a water debit equal to a 0.6mm gicleur, we will have a pressure-flow relationship identical to a 0.6mm gicleur. Pretty cool, but how do I use this to profile a shot?

Well, ever wonder why no one uses a 0.1mm gicleur? On top of it likely clogging constantly, it would take forever to fill the headspace, you'd have to grind crazy fine to build any pressure at the puck, and as the puck erodes, the pressure would drop like a rock. But let's say that you wanted to grind super fine and pull a Slayer style shot, but didn't want your pressure to fall off. Slayer does this by switching out of pre-brew, but the needle valve gets an even better result. Here's how:

Let's say you finish your super slow pre-brew and your puck is saturated and you're reading a peak pressure of 9 bar on your puck gauge. Your backflush pressure is 9.5 bar, so you know that at the lowest flow in the shot, just as the basket starts to drip, you have a 0.5 bar pressure drop across your needle valve and a flow of about 0.5g/s into your cup. So you start to open your valve. The pressure can't go up much, because even if there were zero flow, it would peak at 9.5 anyway. As you open your valve, the relationship between pressure and flow changes, and as the flow starts developing, your brew pressure hangs in there, staying pretty close to 9 bar as flow picks up to 1-1.5g/s into the cup. You keep this going until the puck begins to allow more water to flow through it as soluble compounds are stripped from it. Since you presumably don't want to let the flow get carried away, you begin to close off the valve in response.

Here's where it gets iterative. Closing the valve again changes the relationship between flow rate and brew pressure, but the puck also significantly impacts this relationship, so who wins? Here's the dance that just happens, without any input from us other than progressively closing the valve while the shot carries on:
Let's say 2g/s was making it into the cup at 9 bar. That's again 0.5 bar pressure drop across the valve, but this time at 2g/s instead of at 0.5g/s when we were in pre-brew. We close the valve and immediately two things happen. When we close the valve, that 2g/s now requires a higher pressure drop, so the brew pressure drops from 9 bar to 8.5 bar for a 1 bar pressure drop, up from 0.5 bar. But now we've decreased the brew pressure on the puck, and as a consequence, we less water flows through the puck, let's say it drops to 1.5g/s. Iteration 1 comes back to the beginning.

Now the same valve position that gave us 2g/s with a 1 bar pressure drop now is only flowing 1.5g/s, thanks to the puck. Thanks puck! Well, 1.5g/s is less flow than 2g/s, so the pressure drop will be lower as well, which means instead of the brew pressure being 8.5 bar, it's actually closer to 8.75 bar. But now, you see, the puck has a little bit more pressure on it, so instead of flowing 1.5g/s, it's going to flow 1.75. Iterstion 2 is complete, and the whole cycle begins again until eventually we end up at a new brew pressure and flow rate that the needle valve and puck are both happy with. You could say the needle valve limits the flow and the puck sets the pressure, or you could argue that the needle valve limits the pressure and the puck sets the flow. Both are true. Neither are true. The needle valve controls the relationship of the brew pressure and the flow rate.

The iterations I'm describing are nothing more than apparitions from trying to calculate what the brew pressure will be in these circumstances. In reality, you move the valve and the whole system responds immediately and moves to the newly commanded equilibrium position, as set by the valve position and the state of the puck. There is no oscillation of brew pressure, no up and down of flow rate. You close the valve, and the pressure and flow decay together. Open the valve and the pressure increases. Whether or not the flow follows suit is a function of the puck, but generally if the puck is flowing in the espresso range, there will be a directionally logical response to flow whether the valve opening is increased or decreased.

That's all for now.

Cheers!

- Jake
LMWDP #704
★★★ Very Helpful

User avatar
another_jim
Team HB
Posts: 13871
Joined: 19 years ago

#419: Post by another_jim »

Sweet! Thanks for the writeup
Jim Schulman

User avatar
AssafL
Posts: 2588
Joined: 14 years ago

#420: Post by AssafL »

Yes indeed. Jake's last two posts are Required reading. Nothing less...
Scraping away (slowly) at the tyranny of biases and dogma.