My long and rambling path to preinfusion/pressure profiling - Page 24

Need help with equipment usage or want to share your latest discovery?
User avatar
AssafL

#231: Post by AssafL »

tonythewonderful wrote:Stage two: rotary pump, metering valve, water reservoir
Yes I did it! After outboardinf the vibe pump, and thinking I was going to have piece and quite, I was dissatisfied with the amount of vibrations and buzzing of all the cups during each shot.
Water reservoir -> Rotary pump -> metering valve -> check valve -> boiler 
    ^--OPV #1 (9 bar) <--|        ball valve                           |-> OPV #2 (11 bar)-> drip
Essentially the same idea, this time OPV#1 is absolutely necessary, as the rotary pump would probably just burst all the nylon piping without it.
Most rotary vane pumps sold for espresso have a recirculation valve as part of the pump body itself which would negate the need for OPV1. Not that an OPV won't work (especially if fed back to the reservoir) but:
1. Cheaper - They are very cheap.
2. Volume - they are matched to the pumps water debit - so that water can spin around the pump happily.
3. Reliability - the original pumps were designed for vending machines and beverage kiosks where reliability is commercial/ industrial grade. So they work for a very long time even within espresso machines elevated temperatures.
4. Conveniently adjustable & Less messy. One small pump attached to a larger motor.
tonythewonderful wrote: By the way, my understanding of OPVs and bypasses is that they are all essentially the same - a check valve with adjustable cracking pressure. Am I right? Because I am using the same type of valve from swagelok to achieve all that functions. For simplicity I will call everything OPV, as it is too long to write "a check valve with adjustable cracking pressure"...
I would assume that in theory - yes. But from an engineering perspective one should sweat the details. I am not implying that a single Swagelok adjustable cracking pressure check valve can't do it all - just that it should be evaluated for each of the use cases independently. So a few notes:

1. Usage "duty":
a. What on H-B is referred to as OPV (protecting the steam boiler) would blast once or twice and should probably be replaced. They are not meant for steady operation. They are a safety component. The reason being that the pneumatic components - if they burst - are dangerous (scalding, large volume increase due to flash boiling, etc.). They are calibrated at factory (and sealed for anti-tampering) since finding a 3 bar air source isn't handy for most. Mis-calibration can be dangerous.
b. What we refer to as "Expansion Valve" (as well as the vacuum breaker - but in reverse) - are meant for constant operation. They always drip (and hiss) and get accumulated debris and gunk and deposits. They are usually built to be replaced and/or serviced multiple times. The expansion valve is usually adjustable by the user based on the manometer.
c. For a circulation valve one would want a large orifice that can handle the pump's water debit. Furthermore, it is going to spend most of its life letting through lots of water, so lifetime fully open is important. I don't know if there are special specifications for this, but my assumption is that they would tend to clog less (as the water would wash away the gunk that would harm an expansion valve....

2. Apart from usage "duty", there will be other aspects of the mechanical design of each which may be different. As an example, the safety "OPV" of LM is encased in a tin container. Sort of like a tuna can. So if it blows the steam is captured and redirected through a hose to the drain box. Still fills up the apartment with steam (happened twice to me) - but it has no negative effects on the inside of the machine. Check valves will be "inline" (with BSPP/NPT fittings on both sides); whereas vacuum breakers and expansion valves will at most have a slip on hose to the drain. Also, the adjustment range is different.

3. Adjustment ranges: The Safety OPV should reliably blast at 2-3 or 4 bar. An expansion valve may not even open at those pressures but be designed to have its resolution enhanced in the 8-12 bar range.
Scraping away (slowly) at the tyranny of biases and dogma.

User avatar
Jake_G (original poster)
Team HB

#232: Post by Jake_G (original poster) »

AssafL wrote: 1. Usage "duty":
a. What on H-B is referred to as OPV (protecting the steam boiler) would blast once or twice and should probably be replaced. They are not meant for steady operation. They are a safety component. The reason being that the pneumatic components - if they burst - are dangerous (scalding, large volume increase due to flash boiling, etc.). They are calibrated at factory (and sealed for anti-tampering) since finding a 3 bar air source isn't handy for most. Mis-calibration can be dangerous.
Since we're "sweating the details" :wink: , it's worth noting that the valve you describe is technically a pressure release valve (PRV) but the HB community generally calls them a safety valve. OPV is a term used almost exclusively for the adjustable bypass valve used primarily on vibe pump machines among the HB community. It is the same, but different to your 1-c recirculation valve only in the fact that recirculation valves on rotary pumps are generally built in, as you pointed out. The purpose is identical.
tonythewonderful wrote: How come? I always thought the brew temperature should be between 195F - 205F.
Also, does this mean that with a saturated group, the boiler can be run at brew temperature (or very close to it)?
I'll let Assaf answer that in detail, but the group neck is the highest point in the boiler and is completely flooded with brew boiler water. Solenoid-actuated saturated groups are surrounded by brew boiler water, with only a short loop of tubing running through the external solenoid and back into the group neck to control brewing. I believe the offset between boiler temp and brew temp is only a few degrees most.

Cheers!

- Jake
LMWDP #704

User avatar
AssafL

#233: Post by AssafL »

tonythewonderful wrote: How come? I always thought the brew temperature should be between 195F - 205F.
Also, does this mean that with a saturated group, the boiler can be run at brew temperature (or very close to it)?
Ugh. Fahrenheit...

I just typed 190F into google and it turns out to be a warmish 85C. I've never been outside 89-95C so indeed 195-205F or so.

Sorry....

In a saturated group the boiler actually runs a bit above pull temperature. The Gicar controller has an offset parameter set by the manufacturer that corrects for that difference (in my machine it is 1.6C offset). Obviously it is set for some room temp and will be in error at freezing or summer temps.
Scraping away (slowly) at the tyranny of biases and dogma.

tonythewonderful

#234: Post by tonythewonderful »

Thank you, Assaf, no problem.
I've learnt the hard way that it is always a good idea to check that I've got my basics right. That would be quite embarrassing for me (albeit not at all surprising) to find out that I've been brewing my coffee at higher than recommended temperatures. Phew!

About the OPVs. Swagelok calls them Safety relief valves, with a very strong emphasis on "safety", since they should guarantee to open fast and wide enough to relieve the pressure and save everybody around. Swagelok also warns everywhere not to use check valves as safety relief valves, because check valves cannot guarantee any of these.

With the bypass valve. It was offered as an option with my rotary pump, which I turned down, as I intended to use the adjustable check valve which I already had.

This is amazing that saturated group offset is only 1.6C compared to 14C (25F) for e61.
this means that with the saturated group you do not have any problem with the brew boiler flash boiling when the brew valve is opened ... hmm... In theory this means that you can [edit: fill the head space and] saturate the puck with as low flow as the metering valve allows.

User avatar
AssafL

#235: Post by AssafL »

tonythewonderful wrote: This is amazing that saturated group offset is only 1.6C compared to 14C (25F) for e61.
this means that with the saturated group you do not have any problem with the brew boiler flash boiling when the brew valve is opened ... hmm... In theory this means that you can [edit: fill the head space and] saturate the puck with as low flow as the metering valve allows.
Yes. Us saturated group owners are very fortunate in that respect. But don't let perfection get in the way of commerce: we now have the option of ogling the Decent that can profile temperatures. For those cases where having 3 parameters to play with just feels too much of a constrained. :)

Has anyone considered making an insulation sleeve for the E61 group?

Perhaps it won't get you under 212F (100C) - but it may lower the temp so that there is less flash boiling and thus less vapor pressure. Perhaps, as an experiment, try to wrap in hi-temp wool (fireplace etc.) and see if it can help reduce the offset enough to make it a worthwhile effort...
Scraping away (slowly) at the tyranny of biases and dogma.

Tonefish

#236: Post by Tonefish »

AssafL wrote: Has anyone considered making an insulation sleeve for the E61 group?
Now there's an idea. It's just about the right shape and proportions for the head of a garden gnome ... hmm?
LMWDP #581 .......... May your roasts, grinds, and pulls be the best!

User avatar
Jake_G (original poster)
Team HB

#237: Post by Jake_G (original poster) »

Chapter 18: What's going on in there, anyway? Part 4

It's June, 2018 and I've been spending the past few months really enjoying my machine since adding line pressure preinfusion. I've been thoroughly enjoying the shots of medium-to-light roasted single origins from South America; mostly Borboun (yellow and pink) and Catuai (yellow and red) from Brazil, Columbia and Guatemala, mixed with the occasional Pache and Java, and some modestly darker roasted espresso blends. I'm continuously learning more about how long to rest different varietals post roast before freezing, and learning how to optimize the shot parameters for different roasts. It's fun. It's delicious, and it's addictive.

For the last many months I've been outlining my plan to add a needle valve to give Slayer-like preinfusion along with pressure profiling capabilities akin to La Marzocco MP machines. The term I've adopted for this control strategy to accomplish these goals is "Variable Water Debit". Marketed machines using a variation of this approach include the Dalla Corte Mina and the Lelit Bianca, among a few others. I've made the argument that fancy electronic controls and PID loops are not necessary for accurate control of extraction pressure if you understand the fundamentals of the fluid dynamics between the pump and the puck and select a control valve that provides the "appropriate" range of flow within a user-controlled range.

Water Debit:
Stupid term. Valuable measurement. Understanding the water debit of your machine is an extremely valuable exercise as long as you know what pressure your pump is producing when you measure it. As noted in Jim's E61 Group Pressure Correlation to Brew Pressure thread, the water debit flow, coupled with the pump pressure when flushing yield the Coefficient of Flow (Cv) of the piping between the pressure gauge and the shower screen of your machine. The Cv is a very valuable property of your machine and is closely tied to the size of the gicleur installed in your group.

For a fixed gicleur, we get a flow curve that follows a profile like the chart below:

This is based on my measured water debit data and gives me some interesting information to put to use. First off, there is the right side end point of the curves. For my machine, this is simply the water debit measurement. The pump is cruising at just over 9 bar with the bypass on my rotary pump in full effect against the back pressure generated by my relatively small 0.5mm gicluer. The puck pressure is really the pressure at the shower screen, which is zero (atmospheric) when there is no portafilter in the group. Sliding to the left on the graph, you get into espresso shot territory and see some data that is usually hidden because most machines don't have a pressure gauge after the gicleur. Towards the tail end of my shots, the flow rate climbs up into the 2 to 2.5 g/s second range, as measured by my scale under the cup. At 2.5 mL/s on the graph, you can see that there is a pressure drop across the gicleur of just under 1 bar. Since the puck pressure is equal to the pump pressure minus the pressure drop, it is around 8.5 bar. Not a big deal, but based on the physics of water flow through an orifice, the pressure drop with my 0.5mm gicleur should be measurable. So, what if I put in a larger gicleur? How about a smaller one? Well, as long as the pump compliance allows it to keep up, a larger gicleur will have a larger water debit, and less pressure drop. Conversely, a smaller gicleur will have a smaller water debit and a larger pressure drop at lower flows, as shown below:

The image above is what a variable water debit machine can accomplish. Bianca is designed such that when the paddle is in the fully left hand position, there is no flow at all; imagine if the blue vertical line were slid all the way to the left. No flow, no pressure. Then, imagine the paddle is opened slightly and the blue line sits at 1.5 mL/s. This is a representation of a machine with a water debit of 1.5mL/s or 45g in 30 seconds, as we commonly measure water debit. You've just mimicked Slayer's pre-brew. This profile can be seen below:

The pressure gauge will read 0 until the head space is filled and the puck starts limiting the flow. As you can see, the puck sees half of the pump pressure when the pressure drop and puck pressure curves intersect. This happens around 1.1 mL/s. When this happens, the puck starts getting compressed and the flow is further restricted until the flow is restricted far enough that the puck pressure hits around 9 bar. At this point, the puck flow is at its lowest of around 0.2 mL/s. This is the natural point to end pre-brew. In a Slayer, you would move the paddle and the 3 way valve would take the needle valve out of the picture and the 0.7mm gicleur would take over. Not much happens to the puck pressure when this transition occurs because the flow is so low. With Bianca, you would swing the paddle to the right and achieve the exact same results as Slayer does with their solenoid. The puck won't even register a difference between the two approaches.

OK. I think this helps draw a picture of how variable water debit can impact the rate of preinfusion, while allowing the full-flow extraction of a traditional machine once we're satisfied with our preinfusion, whether that be by reaching a certain pressure, seeing first drops, or getting so many grams of coffee in the cup. But how does this relate to pressure profiling? For that, we need a few more charts, and it gets a little less concise. It gets less concise because the puck dictates the flow through the espresso machine, and the needle valve will have a pressure drop across it that is a function of the Cv of the valve and the flow rate through the valve. Lowering the Cv of the needle valve will increase the pressure drop across the valve, and lower the pressure at the puck.

But if the flow through the puck drops as a consequence, the pressure drop through the needle valve will decrease and the pressure at the puck will increase...

No, this won't result in wild oscillations of puck pressure and flow, but it does mean that it will be hard to estimate the resulting pressure of a certain needle valve position because it depends heavily on the flow through the puck at that particular moment in time. As such, it will take active manipulation of the paddle to maintain target brew pressure throughout a pull. Alternatively, this means that there is a natural declining pressure profile that can be achieved by metering the flow right on the edge of what the puck would flow at full pressure...

Chew on this:

As I said, less concise...

Two key features of interest on this 3d map are the left edge and the back wall. The left edge represents the water debit as the needle valve is brought from closed, to wide open. Puck pressure is zero when measuring water debit, and as the needle valve Cv on the right hand scale increases, so does the water debit. The back wall represents the pressure vs flow curve with the needle valve wide open. Each colored strip represents a range of flow rates. Looking again at the back wall, you can see that as the puck chokes the flow rate down to 1 mL/s or lower, the puck pressure rapidly approaches pump pressure (shown here at 9 bar). If you follow the orange strip around, you can get a feel for where you would need to adjust the needle valve to achieve any pressure you want when the flow is between 0.5 and 1 mL/s, as it might be during the first 15 seconds after the cone develops. Below you can see how this spatial representation of pressure, flow and Cv might be used to guide us to a target paddle position.

Let's say we want 6 bar of pressure at 1 mL/s:

Basically the process here is to draw a straight line up from 6 bars on the pressure axis tracing the "Flow vs Cv" curve. Then follow the orange curve that aligns with 1 mL/s on the back wall. These two curves intersect on the "Pressure vs Flow" curve for the Cv that we want to target. When you trace the curve back to the Cv axis on the right, you see the relative paddle position that will get you the pressure you want at the flow you want targeted flow. You also see how the pressure will rise or fall as the flow rate changes if you were to leave the valve at this position for an entire shot. Following the curve all the way to the left edge, you can see that the water debit should be around 1.75 mL/s for a needle valve position that gives you 6 bar when the flow is 1 mL/s. The exact Cv isn't important, but what is important is that the paddle will be somewhere between fully closed and halfway open, and a quick water debit check should give a solid guideline of a repeatable position to give the results we're after, or close to them :). You can also see that when the flow chokes due to puck compression at full saturation with the valve in this position, the pressure will peak at close to 9 bar.

If you wanted to keep the peak pressure from ever approaching 9 bar, you would need to choke the flow even more than Slayer does, which could mean a crazy long preinfusion time. But there's nothing stopping you from achieving a quick preinfusion and then slamming the valve down to nearly closed to prevent the pressure from spiking all the way up to pump pressure. That is, you can achieve a super slow 25-30 second ramp up to full pressure, or a super quick ramp up to a lower pressure, or anywhere in between by recognizing the relationship between valve position, flow rate and pressure drop and manipulating the valve during preinfusion to achieve your target pressure.

Cheers!

-Jake
LMWDP #704

User avatar
Jake_G (original poster)
Team HB

#238: Post by Jake_G (original poster) »

Hey all!

As I recently posted in my build thread, I ordered all my parts to complete the installation of my Sprofiling Valve.

Today I did my first test fitment of the valve and I think I'll be ready to pull the group and begin the install as soon as my 4mm Teflon tubing shows up.


After looking closer at the depth of the gicleur chamber behind the cap, I decided to face a SS 8mm bolt and mount the 5mm needle valve on that. It provides a very good fit, but my gicleur cap is hollow, with insufficient meat to provide a sealing bore, so I will have to do a crush fit on the o ring sealing the needle valve to the cap, which is not ideal. It's not a problem, but it means a bit of trial and error to ensure the o ring seals when the cap is tightened against its crush washer. I'll live :P

Soon, I'll be able to put all this conjecture and bantering and techno-babble into something more recognizable to human beings. Also, I have a surprise arriving for my Super Jolly tomorrow (Thanks Dennis!). I busted out the refractometer (optical, Brix. I know, I know...) for the first time this morning in preparation... I'll post more on that in the grinder forum when the time is right. :wink:

Cheers!

- Jake

**Edit**
Mazzer Super Jolly SSP Burr Upgrade: Thoughts and Results
LMWDP #704

Tonefish

#239: Post by Tonefish »

Exciting times ahead for you Jake! :mrgreen: Watching with anticipation and knowing I'll learn something!! Thanks for going Bold, Sharing it, and Cheers!! :D
LMWDP #581 .......... May your roasts, grinds, and pulls be the best!

roastini

#240: Post by roastini »

Hi, all. Here is some data the folks reading this thread might find interesting.

I've been using my Decent DE1+ and its advanced profile feature to preinfuse at moderate volume (3 mL/s) into a 16 g. puck until I detect just a bit of back pressure (1.3 bar), and then letting the puck bloom for 30 seconds (I have the machine pump as necessary to maintain 1.0 bar of pressure - but in my most recent shot this literally meant no pumping at all during the bloom). After the bloom, I pump as necessary to achieve an estimated flow into the cup of 1.5 mL/s, and stop my shot at about 40 g. in the cup (2.5 : 1 ratio).

I don't have a refractometer, but using a somewhat similar approach and a similar grinder, Scott Rao has reported a 24% extraction. That suggests that post-shot my portafilter might have been down to about 12 g. of coffee grounds (with the rest in the cup).

I weighed my PF before and after my shot, and the post-shot PF weighed 30 grams more, but the puck was sloppy wet. I gently poured off the excess water, which reduced the weight difference to 21 g. Assuming that ~4 g. of coffee was extracted into the cup, that implies ~25 g. of water in the puck, which is consistent with the presumption that a puck can absorb 2x its weight. (If you assume an extraction of about 23%, you end up pretty much exactly at 2x for the water:coffee ratio in the final puck.)

Nothing new from that data, really, but I find it interesting to see what information can be pulled out of the DE1+.