Positive pressure, PID Pavoni: world domination begins - Page 2

[matadero210]

Your TC reading is almost 3°F hotter than actual. So the water is cooler than the readout. There is no unambiguous evidence that you are burning the coffee. The harshness you're tasting could be the result of insufficient rest. 2 or 3 days post-roast is a relatively short resting period.

Right you are.

When you let the machine come to temperature with the inlet valve open (wait for 20 minutes, say), what does the temp read? If, after 20 minutes, you lift the lever once and only once (or pull 50ml with an empty basket), what does the temp drop to? Does the heating element come on? If so, does the TC overshoot? By how much?

Here's a single dry pump at 60s after 20minutes of idling, with 92.5 C as the set point. Also, while I was at it, some steam recovery curves. Note that the boiler is rapidly cooled because the steam pressure is much less than 20psi, so there is a big inrush of cold water. When the new regulator arrives, I think this will improve since I'll set the feed to 10psi.



enjoy,
raj

[timo888]

Is a dry-pump a flush without the basket in place? How much water is flushed?

(EDIT: Or, now that I recall your earlier posting, is that the pump where you are simply cycling water from the group?)

Have I correctly identified with the highlight the section of the graph that deals with the flush or pump?

It looks as though at around the 5-second mark the flush/pump begins; the temperature drops about 1°C in 10 seconds, reaching its nadir at the ~15 second mark, whereupon it recovers in 5 seconds, overshooting the setpoint by about 0.5°C. Then immediately after the recovery and overshoot, there is another drop of 1°C at about the 20-second mark, and a recovery to the setpoint in the remaining 5 seconds of the 30-second interval.

Also, could you clarify how the transition to steam works? Do you turn the inlet valve off and draw some water to create room for steam, then switch the temp to high?




Regards
Timo

[matadero210]

Hi Timo,

Sorry for being unclear: the dry pump means raising the lever until some (tiny) infusion begins and then lowering the lever. This preheats the piston and cylinder. It is as your EDIT suggests. Almost no water is released.

I think your highlighted area is off: I showed 1 minute of idling (which does include some un-pretty cycling), then I pulled the dry shot at 60 seconds. That produces a very large cooling transient that would probably go 2-3C if the heater didn't come on. It does (around 70s) and produce a bit of overshoot (at 80s). By 120s the overshoot has moderated and we are back to brew temp (92.5, I think). The fluctuations at 0-20s are the PID acting poorly.

As per my granita post: I recently pulled 1L of shots using some crap beans from Trader Joes. After 3-4 shots, there was no further need to heat the grouphead. Here's how I conclude that:

1. This series of shots uses a 10sec pre-infusion followed by a 20sec pull. During the pre-infusion, cooled water from behind the piston returns to the chamber and activates the heater. I observed the heater coming on during pre-infusion 2-3 times.

2. Then, the heater started coming on only after the shot was half pulled. This system inlets cold water into the boiler only during shot pulling. Thus, if the head is hot, the heater should only come on during the shot. This is what I observed after the first few shots.

I have on my to-do list to repeat the above experiment with the heater de-activated during the pump. This will give a more accurate # than the 2-3C I cite above. This drop * 800g should tell us how many calories are lost to the head. Is this what you are getting at?

Steam sequence: exactly as you pose. 1. turn off the inlet, 2. open the steam wand with cup or use the group to drain water, 3. turn on steam control (just turns on heater, bypassing PID). In 2-3 minutes (a bit faster if the steam knob is kept closed) there is plenty of steam. I can get to nice, dry steam in 90s if I drain the water from the cold water part of the circuit, so I'll be installing a dump valve there, I guess.

Once done steaming: 1. close steam valve, 2. disable steam override, 3. open inlet. Because the boiler is at <20psi during steaming, inlet water enters and cools the boiler. This is the transient shown three times (once with the regulator sticking a bit). With a little practice, I can get from steam to brew temp in 30s, but sometimes it takes a full minute.

raj

[timo888]

I was taking a break from doing my taxes and my head must have been still reeling. I thought the sixty-second mark was where you were doing the steam tests. Thanks for the clarification. It is surprising that ~45ml of water from the piston cylinder should cause the boiler temperature to drop so precipitously after 20 minutes. Does the temperature drop occur on the upstroke of the lever during the dry pull, or on the downstroke? If on the downstroke, the cooling is analogous to the destabilizing effects of auto-fill, the result, at least in part, of an influx of water from the mains. If on the upstroke, then we have a complicated scenario. Since my head is already mush from reading IRS regulations and instructions, I hope water from the mains is the culprit.

Regards
Timo

[matadero210]

For the dry-pulls, the cooling occurs (I think) on the upstroke. Why: on the upstroke, water that has cooled in the head is pushed back into the boiler, and the same quantity of hot water is injected into the inlet supply tubing toward the expansion tank. After 20 min, the head is, say 60C, so a pressure drop of (95-60)C * 45ml/0.8L=1.8C is to be expected.

The same thing starts to happen during a pull: on the upstroke, piston water goes to boiler; and boiler water goes toward expansion tank. But when the infusion happens at the top of the stroke, the boiler water that was displaced toward the expansion tank returns to the boiler as 45ml enters the piston. No net water from the mains has entered yet. Now, during the pull, the back of the piston is sucking water in, and 45ml of cold water is drawn from mains. Assuming the tubing is either warm, insulated, and non-mixing, things are fine. To the extent that there is cooling of the boiler water in the expansion tubing, some instability will happen.

The expansion tubing is fairly narrow, so while mixing is no likely, cooling definitely is. In practice, assuming I've got the group hot, the cooling of the boiler occurs on the down stroke, after the piston is already filled. Thus, it seems that the draw-water instability is not an issue.

r

ps: while I have your attention, Timo, I wanted to ask about pressure control/gauge mods. Alchemist's mod seems great, but I'm not too handy on the lathe. I have something much simpler in mind--a torque wrench. Given dimensions, we can convert bar in the piston to torque (I get 1 bar = 34.5 inch-# = 40cm-kg for my pre-M 1.74" dia piston with 1.0" lever arm between the pivots). So, with a torque wrench applied to the lever arm--treating the lever arm as a torque extender--will allow me to control/measure the piston pressure. Drawbacks compared to Alchemist's mod are errors due to friction, general coolness (always a factor), and the fact that I'll learn pressure with different angle/grip/hand position since I'll be using the torque wrench, not the normal lever. One advantage is that pressure meters are notoriously inaccurate, while a torque wrench can easily be 4% (mine) or 1% (for $$$). The simplest way to do this is to tap a 6M screw thread on the lever fork. Then I can put a screw in with a nice hex head to which I can attach my dial torque wrench. Any comments/suggestions?

[timo888]

I wanted to ask about pressure control/gauge mods. Alchemist's mod seems great, but I'm not too handy on the lathe. I have something much simpler in mind--a torque wrench. Given dimensions, we can convert bar in the piston to torque (I get 1 bar = 34.5 inch-# = 40cm-kg for my pre-M 1.74" dia piston with 1.0" lever arm between the pivots). So, with a torque wrench applied to the lever arm--treating the lever arm as a torque extender--will allow me to control/measure the piston pressure. Drawbacks compared to Alchemist's mod are errors due to friction, general coolness (always a factor), and the fact that I'll learn pressure with different angle/grip/hand position since I'll be using the torque wrench, not the normal lever. One advantage is that pressure meters are notoriously inaccurate, while a torque wrench can easily be 4% (mine) or 1% (for $$$). The simplest way to do this is to tap a 6M screw thread on the lever fork. Then I can put a screw in with a nice hex head to which I can attach my dial torque wrench. Any comments/suggestions?

I do not know by how much the torque wrench arm's contact with the fulcrum pin would distort the torque reading on the dial.

Regards
Timo

[matadero210]

Hi All,

The torque wrench idea initially failed for one reason: an M6 screw hasn't the strength to stand up to 200-400 cm-Kg. I spoke to a machinist friend who suggested that a larger screw should work, maybe M10 or bigger. But then he suggested that JB welding a small piece of tubing into a port drilled into the base of the cylinder would be his suggestion (ie: install a gauge). I'm going to try a 1/8" hole through the bell housing that should end at the place where the screen meets the bell, outside of the piston. Because the big O-ring sits just outside of that, if the hole is correctly placed (it will taper to 1/16 or less at this end) it should get cylinder pressure without compromising the basket-screen seal.

What I really need to get this right is a good cross section of the pavoni head, with dimensions. Anyone seen such a beast?

The new regulator arrived. I had gotten quite used to 20psi boiler pressure. I think 15psi is better, but it makes very little difference. The new reg also has a pressure gauge on it, so I can see how the larger expansion tank works (while the smaller one doesn't). The next step in plumbing is to try a back-pressure relief valve. I'll put this just after the regulator. This will do the job of the OPV and might allow discarding the expansion tank. Will let you know. As it is, the initial design is still the best (hence no new pictures).

r

[matadero210]

All,

I succeeded today in milling a 1/8" hole into the side of the Pavoni housing, diagonally, down to the place where the screen meets the housing (and the portafilter o-ring seals everything). I drilled a 0.040" hole straight up into the housing from the portafilter/screen side, so that the two holes meet. The plan is to JBWeld a 1/8" Copper tube into the hole and compression fit the other end to a gauge.

Will let you know how it comes out--should have photos of the complete system in the next few days.

raj

[hbuchtel]

Wow, you really get things done! I'm looking forward to the results!

Henry

[hbuchtel]

The next step in plumbing is to try a back-pressure relief valve. I'll put this just after the regulator. This will do the job of the OPV and might allow discarding the expansion tank.

You might want to check out this thread- variable-brew-pressure-via-steam-wand
if you have an extra steam valve I think you could put it before the boiler and use it to regulate the pressure during pre-infusion.

Henry

[matadero210]

Hi All,

Some quick snaps of the now working Pavoni, plumbed-in, with PID control, naked portafilter, regulated infusion pressure, and BREW PRESSURE GAUGE! (Forgive the shouting, but you can imagine my caffeine level).





The first shot shows the j-b welded tubing and gauge. Just to remind everyone, it connects to a 0.040" hole outside the cylinder, but inside the screen o-ring. The second shot shows the supply plumbing with infusion regulator and gauge (analagous to the normal gauge on the Pavoni Professional). Note the large expansion tank--I still like the feel of it better than the back-pressure relief valve (not shown). All parts come from mcmaster.com, which is quite addicting.

Must go now, can provide more detail over the weekend (do ask if you have specific questions). I did pull 2 shots on this at 8 bar--better than ever! Anyone in Silicon Valley is welcome to swing by for a pull or two.

raj

PS: would love to hear from anyone with ideas for how to improve on this. I can think of one major problem: poor insulation. This results in the heater coming on too often. Also, right now the controller uses its built-in mechanical relay. This weekend I'll switch to SSR, so maybe temperature will be more steady. I just wish I could think of a way of insulating the Pavoni without making it look terrible. In the absense of any new ideas, I may start working on a roaster or mill ....

[espressme]

"I just wish I could think of a way of insulating the Pavoni without making it look terrible. "

You could very easily make a sheet copper boiler wrapper and top. It could be held by insulation pressure with a a couple screws. In other words, make a sheet, curl it and leave room for the sight glass mounts where it comes together. put a machine screw top and bottom. Hammer out a cone that slips tightly over the top of the cover and fits close to the boiler neck. A few hours work total.

Good luck!
richard

[matadero210]

Hey Richard--

Good idea. I think all Pavoni's could benefit from boiler insulation, though without positive pressure you would never want to leave them on for long. I worked with Suyi at auberins.com and tweaked the controller to +-0.5C and <1min to recover between shots. Now I need to get that Robur grinder project worked out.....

raj

[mogogear]

This thread has been a smile from start to where it is now--------

FrankenPavoni meets the Tinman from the Wizard of Oz and after being pressured ( ;>)
........... steals his coat, bad * dap * dap!

[espressme]

This thread has been a smile from start to where it is now--------

FrankenPavoni meets the Tinman from the Wizard of Oz and after being pressured ( ;>)
........... steals his coat, bad * dap * dap!

With steam and a whistle we will have a hoot!
rp

[matadero210]

Hi all,

Just a quick update. The system is working fine. I did some tweaking to the PID to get +-0.4C and a resonably fast shot-recovery time (less than my grinding time = 1min). I lathed a 1.7kg stainless tamper that I'm experimenting with. Reg Barber's are too small. My tamper can't sit flat on the bottom of the basket: it stops 2-3 mm above the bottom (the puck is thicker, of course).

I'm curious about the fine-grind/light-tamp extremists. So I adjusted the Zass down and started tamping with just the weight of the tamper (1.7kg = 3.8#). It works fine and make an interesting cup, but very aggressive on young roasts. I wonder if 5-7 days is needed, because at 4 days past roast the cup is good, but wild.

Lastly, I still get a 'donut' during pre-infusion, although the pull is more uniform. I'll read up on it, but any suggestions are welcome. I distribute with WDT and leave the surface flat before tamping. Does Stockfleth's Move result in a concave distribution that would limit side channeling?

raj

[Stuggi]

Just a quick question. Could one just PID a M-model La Pav.? I mean without the water inlet. Does the La Pav work properly with just 93 degrees centigrade water in the boiler? And is it possible to attach the TC on the outside of the boiler?

[hbuchtel]

Hi Struggi, the problem is that the Pavoni requires a certain amount of pressure to force the water from the boiler up to the group, and that pressure comes from having the boiler at a higher temperature. (ie 93C would not be enough)

There is a good thread by cannonfodder about encountering this problem when putting a PID on his Gaggia Factory lever.

Regards, Henry

[Stuggi]

As I thought... Well oh well, you can't always luck out with these things... ^^