All,

I'd like to share my first Pavoni mod. I started with a pre-M europiccola and drilled the heater/base to include ports for TC and water inlet (1/8" npt). The whole is mounted on an oak base with PID controller, power switch, and steam override switch.



From the back you can see the stainless flex tube that connects to my kitchen faucet.



I'm just learning about how this works, but the first shots were very nice. The water supply is regulated to 20psi, so I can leave the water connected and pull shot after shot. It takes <1min to stabilize temp after each shot (and the cold water admission). 20psi is on the edge of too rough for pre-infusion, so I can also close the water inlet valve and do a more gentle pre-infusion (I monitor this with a bottomless PF). The TC is t-type (mcmaster.com) and the PID (which comes with internal 10A relay) is from auberins.com. More details to follow (once I stop shaking from caffeinosis) after I master use of the machine.

share and enjoy,
raj

Awesome. Did you come up with some kind of expansion valve for when the cold water comes up to boiler temp? Wouldn't want the thing to explode.

Yup,

I also relieve excess pressure while flushing/heating the head.



I use a commercial anti-hammer assy first, but a simple expansion tank seemed better.

r

Raj,

Oh man, that is impressive. Very ingenious.

The water supply is regulated to 20psi, so I can leave the water connected and pull shot after shot. It takes <1min to stabilize temp after each shot (and the cold water admission). 20psi is on the edge of too rough for pre-infusion, so I can also close the water inlet valve and do a more gentle pre-infusion (I monitor this with a bottomless PF).

Before actually pulling the shot, do you get resistance at the lever from the uppermost position (i.e., right after pre-infusion)?

Javier

Javier,

There are two ways of dealing with pre-infusion. The problem arises because Pavoni's are not really air-tight--only partially so. They leak slightly. One method is to open the inlet valve and jiggle the lever until the downward pressure feels right and then close the inlet valve. After 10-20 seconds the temperature stabilizes and I can pull the shot--it takes about 5# of upward force to raise the lever.

The second method is to leave the inlet valve open for the entire time. In this case it takes about 15# to raise the lever and the pre-infusion is a bit fast (about 5 seconds to visible liquid on the bottom of the basket).

Both of these methods work fine with a warm machine. Coming from cold, the second method has too much false pressure, which I relieve with a warming blank shot.

After the pre-infusion, I feel some pressure at the top of the stroke. I think your question is getting at the issue of whether the cylinder is full of water or water+air. I haven't worked this out, but with the higher pre-infusion pressure (second method) it seems that I get almost all water. I don't really understand how the Pavoni can take advantage of a 45ml stroke--I tend to get ~35ml at best for a single pull. I'm looking at ways of changing the pre-infusion to start earlier (before the lever is fully raised), but I'm not sure if I'll be able to incorporate that mod.

raj

Hello Raj,
Now you've done it! That is something I've wanted to do to mine in the course of a rework.
How do you prevent the boiler from overflowing when the boiler is cold and the water is on. Or, in other words, how do you monitor the water level? Do you fill to a point on the sight glass, or does the 20PSI balance the steam pressure and the whole thing becomes moot?
Again WOW!
sincerely
richard / espressme

Do I understand correctly that, for brewing, you temporarily "convert" the machine into an open unpressurized kettle (albeit with a boiler cap) by keeping the boiler temperature below boiling? Afterwards you can reset the PID so that the temperature will rise to above boiling and the machine can be used to produce steam?

I realize you might not have measured these things yet, but questions about brew temperature spring to mind.

If the temperature of the water inside the boiler is ~198°F, what is the temperature of the water hitting the puck? How much heat does the group sink away? With the Peppina's relatively lightweight group, as a point of reference, there is a drop of only a couple of degrees between kettle water temperature and brew temperature. But with the heavy brass group of the Pavoni, which is designed to handle 212°F water and bring it down into brewing range, one would expect a drop of more than a couple of degrees.

You mention above that it takes 1 minute to stabilize after a shot. After an influx of ~50ml of cold mains water, how many degrees does the water temperature in the boiler drop? When pulling multiple shots, is there an upward brew-temperature creep? With shots at 2-minute intervals, say, does the group sink away less and less heat with each shot, so that at some point, the boiler temperature and the brew temperature are quite close to one another, only a couple of degrees apart? Or does the group also recover in 1 minute?

Regards
Timo

All,

Thanks for all the great questions. I'll do my best, but there is still a bunch of metrology to do (I love that word). For those who care, a peek up her skirt (or his kilt):



NB: I had the two-element brass heater. I drilled out and removed the 200W heater and removed the heat shield. This improved circulation around the element (if that helps I have no idea why). The 200W heater was in the way of my TC, a 1/16" x 8" stainless grounded probe.

Operation goes like:

1. turn on gate valve to water. This lets cold water into the 20psi regulator (fixed, but I've got a better one on order).
2. power up the pid/heater and let heat.
3. because I live in fear of electrocuting myself, place measuring cup to catch group+opv drips.
4. at 50C, wiggle lever to check boiler pressure. if too high, close inlet water and drain by raising lever.
5. at temperature (95c at present), open inlet valve (bringing boiler to 20psi), and let 100ml run through head to pre-heat.
6. wait 1 minute (boiler may cool with inlet water, but comes back). pull shot.

5. (alternate) at temperature, open and close inlet valve, and let enough water run through the heat to pre-heat without exhausing the boiler pressure. keep inlet closed and use reduced pressure for shot.

I haven't been able to measure the puck temp, but taste is great and consistent. water in the measuring cup is 180f. I had plans to insulate the head, but I don't want to uglify too much. Maybe a wooden box around the head followed by spray-in foam. That would eliminate the heat-sink capability of the head. I also have a heater band that I might use on the head, but I think I'll wait. Pre-heating the head with a blank shot is sooo easy when plumbed-in.

Timo,

yes, right now I set the pid to 105c for steam. there are two switches on the panel, but not yet connected. One for pid power, and the other for pid override. That will be the steam switch.

Regarding the temperature drop. With the basic procedure, it varies, because the pressure regulator is a bit sticky. Sometimes a shot results in almost no inlet water, but with full inlet about 9C drop which takes ~30sec to recover (the pid is pretty good once auto-tuned). In the alternate procedure (with no inlet during the shot), the 9c happens when I toggle the inlet. In 4 tries the variation was 8-10c.

As for your last question, my reasoning is that if I always flush with 95C water 1 minute before the shot, I'll always have the same group head thermal loss (whatever it might be). unless I rig up something better for measuring puck temp, I'm not sure how to test this. The copious flushing may not really fully pre-heat the head, but it seems to come close.

Measuring in the cup at 1 minute intervals, drawing 50-75ml from the shower screen to cup, with boiler set to 95C:
183, 187, 187, 186, 188F.

I think most of that variation was in the volume drawn (since its pre-infusion water, not real shots), and pre-infusion water is less affected by head temp that real shot water. I'll try to repeat the experiment with real shots shortly.

r

So you allow water into the unit and then turn the water off to operate the unit and then between shots allow more water into the system? Otherwise you would be over pressured during operation and forcing water out the OPV ...Right or Wrong?

TUS172 wrote:So you allow water into the unit and then turn the water off to operate the unit and then between shots allow more water into the system? Otherwise you would be over pressured during operation and forcing water out the OPV ...Right or Wrong?

I have tried both ways. With the 20psi regulator, the pressure is firm but OK. The problem has to do with the size of my expansion chamber. Its about 20cc, which is too small to allow the lever to be raised easily. So, when I first fill the pavoni, the sealed boiler is filled with air at 0bar. As the water comes in, I end up with about half a boiler at 1bar of air. In this circumstance (400+20ml of air in the system), I can leave the inlet open and pull shots just fine. Over the course of 2-3 hours, however, the air gradually escapes (the Pavoni is not an air-tight design), and I am left with only 20ml of air (just like operating the pavoni with the boiler filled to the very top).

Tonight I'm installing a 500ml expansion chamber into the cold-water plumbing. Then I should be able to leave the inlet open all the time. The basic problem is the 45ml that the piston displaces when rising. With the new expansion tank, the pressure in the system should stay 20psi regardless of 100ml changes to the system.

So to your question: right when the system has leaked the air out (after a few hours) and in incompressible, wrong when there is plenty of air in the system.

I'll post a picture of the new, larger expansion tank tonight. It's overkill (.5L of air), but the stabilized pressure will make steaming easier, I think. Let you know tonight.



raj

matadero210 wrote:Over the course of 2-3 hours, however, the air gradually escapes (the Pavoni is not an air-tight design), and I am left with only 20ml of air (just like operating the pavoni with the boiler filled to the very top).

Does the air escape if you idle the machine for 2 hours, or only if you use it periodically?

I ask because I was kibbitzing when we had a new boiler put in the house where I used to live. The heating man said it was okay if the radiators had some air in them, or even if they had a lot of air in them, because the air would dissolve in the boiler water, then get released in a special valve as the water cooled, and that after a few days *all* the air would be gone from the system.

(I remembered this conversation when I was thinking about saving electricity costs on my upcoming 2-group thermosyphon leva machine by not priming one of the groups.)

My Comocafe (Hi Tim!) has two rubber flapper valves, and both are water-tight, but only one is air-tight. I assumed the Pavonis used more robust valves, but maybe I'm wrong.

Edit: does anyone else get a close gvideo tag if they click submit while the edit window has the cursor?

Paul the Roaster asks about how long air lasts in my system.

Today I fired it up and the pressure release valved leaked the air out in 15 minutes (at 20psi). I glopped some silicone grease on it, and retried: it took 30 minutes this time. So I think its much less subtle than air dissolving in water. at 95C, I don't think it would happen, anyway. Not much solubility. But with the new expansion tank, air isn't needed in the boiler anyway.

Today I enlarged the expansion tank did 10 shots in a row. First, the new tank:



Then, I switched to Celsius scale, autotuned, and moved the TC wiring away from the AC. The result was improved TC stability. I also hooked up the power switch and steam switch.



The steaming works fine: its fun to heat to 120C and open the steam valve and watch the heater try to keep up (it stabilizes around 110C with the steam wand fully open after 2 minutes). Needless to say, to get to steam takes much longer (95-120c in about 2 min) because the boiler starts off cooler.

The expansion tank has resulted in much more stable pressure, so that I left the inlet valve open for 2 hours and 10 shots without issue. When I figure out how to post video, I'll put up the money shot. Its not perfect (a little blonding at the end), but a FULL 45 ml shot with 1 lever pull through 11g of 2-day-old roast ground fine, tamped very lightly:



raj

matadero210 wrote:Needless to say, to get to steam takes much longer (95-120c in about 2 min) because the boiler starts off cooler.

Suppose you steam the milk first: how long (and what volume of water) does it take to bring the temperature down to brew range by flushing with the fresh-water inlet valve open?
Regards
Timo

All,

I learned a new thing from the Pavoni this morning. As you know from other discussions, the pre-M piston is backed with steam (normally), while the M piston is backed with water. My system, although pre-M, is so full of water that the piston gets backed with water. This morning I noticed that if I left the machine (at temp) for 5 minutes and then dry-pumped the machine--lifting the lever almost but not quite to pre-infusion and back down--the boiler registered a 5C temperature drop. I think this is water that has cooled in the head, behind the piston, returning to the boiler. This is the reason why my machine at 95C has a burnt taste, but at 93.5 is much better and brighter (I'm still working my way down). A few dry-pumps 20 sec before pulling a shot probably heats the head to boiler temp, or very close.

So, no need to insulate the head, just dry-pump it! (ok, now you can giggle!)

raj

matadero210 wrote:All,
....snip....
So, no need to insulate the head, just dry-pump it! (ok, now you can giggle!)
raj

Hi Raj,
I would agree that a lower temp in the boiler should help.
Let me offer another idea:
Why not get a LED beam level indicator and use it to monitor your liquid level in the sight glass. a reflective type would probably not even show from the front. Or a through type could be small enough to be almost un-noticable. Use it to control a simple solenoid valve in your supply hose.
I love the project!
sincerely
richard / espressme

matadero210 wrote:All,

I learned a new thing from the Pavoni this morning. As you know from other discussions, the pre-M piston is backed with steam (normally), while the M piston is backed with water. My system, although pre-M, is so full of water that the piston gets backed with water. This morning I noticed that if I left the machine (at temp) for 5 minutes and then dry-pumped the machine--lifting the lever almost but not quite to pre-infusion and back down--the boiler registered a 5C temperature drop. I think this is water that has cooled in the head, behind the piston, returning to the boiler. This is the reason why my machine at 95C has a burnt taste, but at 93.5 is much better and brighter (I'm still working my way down). A few dry-pumps 20 sec before pulling a shot probably heats the head to boiler temp, or very close.

So, no need to insulate the head, just dry-pump it! (ok, now you can giggle!)

You might want to recalibrate your setup using briskly boiling water to see if it registers 100°C. 203°F boiler water temperature is not egregiously hot, and it is probably lower when it hits the puck because the group will have sinked some heat away.

On the dry-pump technique ... it is your fun toy to play with, and you can dry-pump it to your heart's content but if you have the machine hooked to a fancy PID, why not try to dial in the temp and flush time/volume before you introduce flying-by-the-seat-of-the-pants techniques such as "a few dry pumps" -- that path leads to the enchanted land of Oxymoron.

Regards
Timo

Timo--

Thanks for the very sage advice. My TC registers 101.6C at vigorous boil, so I was almost certainly burning the first few shots. I had no idea how inaccurate TCs could be: should I be using pt100 or something else? My most recent shots are at 93 (nominal), but I'll try lower. As for the pumping, I'm thinking that 'enough pumping that the temperature remains stable after the last one, then immediately pull the shot' is the best policy. The real issue is that vigorous pumping causes such a fast drop in temp that the PID overdoes it and overheats the boiler by 1C, which take a minute or two to dissipate. Must learn to pump gently!

raj

matadero210 wrote:Timo--

Thanks for the very sage advice. My TC registers 101.6C at vigorous boil, so I was almost certainly burning the first few shots. I had no idea how inaccurate TCs could be: should I be using pt100 or something else? My most recent shots are at 93 (nominal), but I'll try lower. As for the pumping, I'm thinking that 'enough pumping that the temperature remains stable after the last one, then immediately pull the shot' is the best policy. The real issue is that vigorous pumping causes such a fast drop in temp that the PID overdoes it and overheats the boiler by 1C, which take a minute or two to dissipate. Must learn to pump gently!

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.

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?

Regards
Timo

timo888 wrote:Suppose you steam the milk first: how long (and what volume of water) does it take to bring the temperature down to brew range by flushing with the fresh-water inlet valve open?
Regards
Timo

Its on the order of 100ml, but I haven't measured the volume precisely. The trick is to limit the pre-infusion flow so that the boiler doesn't cool off too fast--so I suppose the answer is 'as fast as you want'. I think you are absolutely right that this is the way to go: steam the milk first, open the inlet, dry-pump the head, and pull the shot. Thanks for the idea,

r