Coffee side flow restriction

by **Paul_Pratt** on Fri Dec 11, 2009 2:08 am

Whilst hard at work yesterday at the Cafelat offices (read messing about with machines) I was testing the upcoming knockboxes with an older La Marzocco that I had just rebuilt. However on this machine the flow rate is monstrous, 500ml in around 15 seconds. Remember back in the GS3 threads Bill mentioned something around 400-500ml a minute is ok.

The reason for this is because my machine does not have any kind of restriction, the pick up tube in the boiler has a 1mm hole. If I was on the ball I should have done something about this before I assembled the group, now I would have to take it apart to braze over the hole and drill a smaller one. This is not an option as I would most probably wreck all the new group seals I put in a few days ago.

So I dug out a diffusor screw that I modified a few months back when I was messing about with pressure profiling. I was trying to smooth out the ramp up and down of the pressure of pre-infusion and post-infusion. Here is the dispersion screw that I used on my Linea machine. Stock screw on the right, modified on the left. All you do is tap an M4 thread into the screw and screw the stock LM gigleur into it.*

The screw is the last thing the water touches before the coffee. What I did was restrict the opening of the screw to 0.6mm. Works a treat.

So here is the older Marzocco diffusor screw (similar to an E61 screw) with the restriction.

With this in place my flow rate is down to 500ml in around 50 seconds and I have much more control over the pre and post infusion rates. By having the screw in this place the post-infusion rate is very much like a lever machine. Rather than a one-off pressure release you get a very gradual release.

What I have been doing on my manual (paddle machines) is when brewing as normal you back off the paddle and the pump disengages. You then get a lovely gradual decline in pressure and I love the coffee this way. On the solenoid machine I used with the Linea screw above, you still get a slow pressure decline.

The only drawback I could see with this is the potential for blockages with coffee. However I have been using a screw now for 2 months and not had any problems. Even if it did it is a 10 second fix as you just unscrew it and clean it.

The upside of this mod is that it is cheap and if you don't like it you just take it out in a few seconds.

As I said I like this as it has allowed me to control the flow on a machine where installing a restrictor in the usual place is not an option. Better coffee? well I like it this way.

Paul

* Actually I used a lathe for this and also turned down the gigleur head so it was more low profile.

by **Paul_Pratt** on Fri Dec 11, 2009 2:11 am

Here's the machine I installed it in.

Yeah I know, too many machines....

by **eastpresso** on Fri Dec 11, 2009 9:09 pm

Neat!

The only draw back I can think of right now is that you probably have to remove it when backflushing (edit:) as well?

(If you have too many machines I know a solution :mrgreen:

)

by **Paul_Pratt** on Sat Dec 12, 2009 5:02 am

I keep it in there when backflushing.

I just threw this out there as I was keen to see if anybody had any thoughts on other drawbacks to this. I have measured the pressure at the head and it is all good. I just figured that restriction anywhere before the coffee was acceptable.

Paul

by HB on Sat Dec 12, 2009 9:37 am

Paul_Pratt wrote:I just figured that restriction anywhere before the coffee was acceptable.

I agree that a slower pressure rampup will make an espresso machine more barista-friendly, but ever since writing Pressure profiles, preinfusion and the forgiveness factor, I've wondered if it compensates for a group's poor diffusion. That is, a group that forcefully squirts water here and there will behave more civilized when its incoming water pressure rises more slowly.

As part of the Mypressi TWIST - Second Look, I made this video to demonstrate how even the initial wetting of the puck is. Spring powered lever espresso machines have a similar full-on shower spray look to them:

Since you appear to be asking for other avenues to pursue and you're a talented machinist, I thought to pose this question to you: How well does the La Marzocco design initially diffuse water across the puck compared to other espresso machines you've used? I've guessed at this contributor to the "forgiveness factor" by observing the water spraying forth from the group, but Ken proposed interrupting the extraction and doing some puckology:

From How to Preinfuse; Extraction Pressure Redux

It would be interesting to see if a pre and post pressure restricted pucks show differences in initial wetting, and by extension, whether improvements to the initial diffusion would carry the same benefit.

by **Jepy** on Sat Dec 12, 2009 1:05 pm

When I was building my temp profiler I made a series of these type gicleurs from .5mm to .9mm, the main reason was that the plumbing in the machine didn't allow an easy way to mount the standard gicleur, so I opted for the screw method. Never had a problem with these, even in a standard Synesso. As Paul said, you just unscrew it if you want to clean it. I would soak the screen and screw in the cleaning solution periodically anyway.

by **eastpresso** on Sat Dec 12, 2009 9:22 pm

How long does it take for the group to depressurize? Do you experience any 'portafilter sneeze' that necessitates a short wait before you pull the next shot?

It would seem odd that manufacturers do not adopt this method if there are no drawbacks?

by HB on Sat Dec 12, 2009 9:33 pm

eastpresso wrote:Do you experience any 'portafilter sneeze' that necessitates a short wait before you pull the next shot?

That's handled by the 3-way solenoid, i.e., it opens an exit pathway to the driptray when the group solenoid closes the brew pathway (and vice versa).

eastpresso wrote:It would seem odd that manufacturers do not adopt this method if there are no drawbacks?

But there is a clear drawback -- added maintenance because narrow jets clog more easily with scale buildup.

by **eastpresso** on Sat Dec 12, 2009 9:53 pm

The gicleur sits directly above the puck and before the solenoid - when the solenoid opens the gicleur works in the other direction which results in a gradual pressure release. I was just wondering how long it takes before you can safely remove the pf?

by **shadowfax** on Wed Dec 16, 2009 2:46 pm

Paul,

As we discussed here the probable issue is the increased propensity to clogging from coffee particles in the water path. On the other hand, the gicleur experiences a periodically bidirectional flow--that is, the water goes back and forth as you brew and then depressurize the chamber. This may help, and of course most of us (I home? I sure do) take the screw out regularly for cleaning anyway.

I did want to ask, did you use the stock gicleur? It looks like there may be more machining challenges to modding a stock La Marzocco to do this than your post suggests. The standard LM ruby gicleur, near as I can tell (I have 2 next to me right now, measured with a set of digital calipers), is M5. it Looks like the dispersion screw is M6. The inside hole of the dispersion screw (again, on my 2 spares from EPNW) is ~3 mm. From what I'm reading you should drill a 3.3 mm hole before tapping to M4 x .7, suggesting that the screw ought to be widened somewhat before tapping it. The LM gicleur at M5 also is way too big, and it looks like you'd want to source some M4 gicleurs (like these, as Eric S. suggested). I don't recognize the ones you've used in the images above... Are these from older LM machines or something?


Contact Us	Legal disclaimer and copyright: This content and all its parts are copyrighted, all rights reserved. Content including all text and images in part or in whole may not be reproduced in any electronic or printed medium without prior permission. Home-Barista, Home-Barista.com and related logos are trademarks of Home-Barista.com, LLC. Other trademarks are the property of their respective owners. By accessing this site, you agree to the terms of service. Overall speed: , current page: s, avg: s, max: s