Flow rate of a rotary pump espresso machine - Page 2

[another_jim]

On the E61 groups, you have the additional preinfusion mechanism, so the water debit shouldn't matter. In any case, the "silent" Ulka has a low unrestricted flow. However, Micheal says the original E61 has a weakness, since the water feed to the dispersion screen isn't quite straight. He recommends a 0.6 or 0.5mm gicleur rather than the 1mm ones in the E61 boxes. Marshall reports he gets a more uniform "appearance of coffee" at the start of his shots on the Zaffiro since making this mod.

Al, the original poster on water debit on the alt.coffee thread I gave, works for Cimbali and is quoting their tech standards. Cimbali machines do not have preinfusion, and are relatively unforgiving even at 75 to 90 mL water debit. The post got a lot of echos, since the LM Lineas ESI was selling at that time were prepped for *$s and had massive debits in the 120mL range. Getting a properly levelled shot on them was a fairly heroic feat and most proper cafes started modding them with narrower gicleurs. 75 mL (2.5 ounces or so) became the informal standard for getting an easy to pack puck on LM groups.

Iirc, the non-stop/e66 family of groups you're working on do have some sort of preinfusion, so only the number of sink shots with a naked pf will tell you for certain if the puck is being gently treated, or whether you need a narrower gicleur.

[cannonfodder]

I think I have my mind wrapped around it now. On the plus side, having the high water debit that I am currently experiencing is forcing me to further refine my pack and tamp. I have also noticed that head space makes a big impact in this case.

On my Isomac, I could stuff the basket to the point I was getting almost a triple in a double basket with only minor issues from the machine. With the Faema, I am getting improved shots with more headspace and a lower dose. I have been going the opposite direction. Over the past week I have been progressively stuffing more and more coffee into the basket.

I do not normally weight my dose. I got out my scale, zeroed it with the empty PF and dosed it. To my surprise I have been progressively upping the dose and was now stuffing 20 grams into that double basket. I was not bottoming out the headspace yet. I would lock and remove the PF to make sure the screen was not tearing into the puck. I dropped back down to a 16-17 gram dose which gives me about 5/16-1/4 of an inch between the top of the puck and the top of the basket rim and was rewarded with a much better extraction.

I went the opposite way and kept reducing my dose, below 14-15 gram I went back to the uneven extraction and a wet and soupy puck. Using my 'normal' technique, I am averaging out at 16-17 gram per double and decent shots. I will have to see how my LM baskets respond to different doses in addition to the change in the jet.

[another_jim]

Sorry to change the subject; considering only the good shots, is the taste systematically different for the low dose versus high dose shots.

Does the shot blond more at the ristretto point for the higher dose shots? That would mean lowering the dose in the same basket gets you more espresso for less ground coffee.

[cannonfodder]

The larger dose shots were pretty much all bound for the sink. Regardless of the distribution method or tamp I got shots like this...






Severe blonding and uneven extraction on the front of the group. Now most of these are extreme examples, the last photo is what I would call a drinkable reject. I do not have any shots of the reduced volume to show in comparison. Each of those shots used different distribution techniques, NSEW, Weiss, modified Stockfleth, and the Chicago Chop. The latter of which tended to supply the best result. I gave the PF a couple of downward thumps on the forks of the Mazzer while dosing, which is how I jammed in so much coffee into the basket using each technique.

I would also tamp with the PF handle in different orientations. That way if it were me favoring one side of the tamper, the over extraction point would move, but it was always in the same position. I would even favor the front during the tamp and dose in an attempt to compensate but still got the same result or a completely choked shot.

Once I down dosed back to 17g and increased the head space, the extractions started to even out. Still not perfect, but better. One observation, once I engage the pump, I get a rapid beading on the bottom of the PF. In about two seconds I have a series of pearls across the bottom of the basket, then it pauses for another couple of seconds, then it starts to extract. The rapid beading on the bottom of the basket is part of why I started this thread and pointed to the high volume of water. The high flow rate is pounding the top of the puck instead of giving me a slower, more uniform preinfusion.

[another_jim]

This is the problem Marshall had, and his apparently had nothing to do with levelling/packing technique, but with uneven water entry. The smaller gicleur fixed it.

I'm trying to mentally picture how this could be; and the only thing I can imagine is an airpocket forms as the head space fills with water, so the puck doesn't fully soak during the preinfuse. A slower fill or a larger headspace prevents the pocket from forming.

[cannonfodder]

This is the problem Marshall had, and his apparently had nothing to do with levelling/packing technique, but with uneven water entry. The smaller gicleur fixed it.

I'm trying to mentally picture how this could be; and the only thing I can imagine is an airpocket forms as the head space fills with water, so the puck doesn't fully soak during the preinfuse. A slower fill or a larger headspace prevents the pocket from forming.

Good to know I am not going insane. I will pull a few shots when I get home with the lower dose and post the photo, maybe even try a vid with my photobucket account.

[cannonfodder]

Here are a 17g shot. The only thing that changed from the other blown shots is the amount of coffee in the basket and the amount of headspace. The extraction was still a little fast and the start is quite abrupt. The first photo is only about 4 seconds into the shot. The second is a little over half way and the shot has 'healed' itself, but the taste is still lacking, in part to the 22-second 2oz shot. A slight tweak on the mini and the timing is good.

As you can see, the extraction has evened out but the basket surface is still pitted with dead spots for the first few seconds.

[DavidMLewis]

That Faema I just finished rebuilding pumps a lot of water, and fast. I was wondering if it was flowing too much. After a flush, no spitting and sputtering, I get 4oz from the shower screen in 10 seconds. Does that seem reasonable or is my group jet too large? The pressure ramp is very fast but I am use to a little Isomac.

That would certainly be in the range. My Techno, unmodified, has a debit of 230 ml/10 s, as do LM Lineas if they don't have a gicleur. You have about half that, which is still a little high per Al's rule, but not so high that I'd worry about changing it.

Best,
David

[steffen]

Silvia (what a sweetie) just did somewhere between 3.33 and 3.50 ounces in 10 seconds. She has an orifice also and I seem to think it is about 1.0 mm.

Could somebody tell me where that orifice is build into the Silvia? Could it be changed with a 0.6mm jet? I would like to install a rotary pump into my Silvia and thus would be highly interested in that...

Thank you!

Steffen

[cannonfodder]

That would certainly be in the range. My Techno, unmodified, has a debit of 230 ml/10 s, as do LM Lineas if they don't have a gicleur. You have about half that, which is still a little high per Al's rule, but not so high that I'd worry about changing it.

Best,
David

The jets on my machine are incredibly easy to access. They screw into the top of the group, all you need is a socket and unscrew it. Jets are only $2, but for the sake of being through, I ordered the entire jet assembly, jet, screen, spring and body, whopping $11. It will be interesting to see how that changes the setup.

[erics]

The "orifice" in Silvia is located in the small 3-way valve adaptor that bolts to the grouphead. The reason I call it an orifice is that it receives the water from the, IIRC, 6.0-6.5 mm ID standpipe inside the boiler. Water from the "orifice" fills the small volume associated with this adaptor and 3-way valve, makes an abrupt U-turn, and heads towards the grouphead passageways.

IIRC, the size of this opening is about 1.0 mm (or less) and forms the seat for the 3-way valve in the de-energized position. I can say that there is no reasonably measurable resistance (pressure gage reads 0) under normal Silvia water debit flowrates (9.5 to 11.5 ml/sec) because of this "orifice."

The link to the lengthy AC posting which Jim Schulman provided before provides for some interesting reading. Not the least of which would be the posting from Barry J. that revealed he found NO difference "in the cup" on a pro machine with/without orifices and multiple operators.

Eric S.

[steffen]

Thank you, Eric!

Without having read the AC thread, I think that there is a difference, if the machine is operated with a vibration pump or a rotary pump. In the latter case, an gicleur is in my opinion needed because of the greater flow rate of this type of pump.

Steffen

[erics]

FWIW, here are the unregulated flowrates of the common Ulka vibe pump and the Procon rotary pump. I edited the Ulka graph from their website to lengthen the flowrate scale to accommodate the rotary pump. Data for the rotary pump is from the Procon website.

Eric S.