G'day Guys, I'm trying to sort out a problem with a Bezzera BZ-40 which is being particularly fussy about the dosage. I'm finding channeling to be a major issue and the channels seem to coincide with the holes on the dispersion block (Bezzera call it a shower holder). The machine is fitted with a Procon pump and it's brewing at around 8.5 Bar, 60ml, 30 secs, double basket. I find I have to up-dose considerably and using a Mazzer Mini E the grind is somewhat finer than the "start here" point.

The free flow straight from the shower is 550ml in 30 secs which seems to me that there's no restriction in the line at all, and that's roughly what a Procon pump will flow at 0psi. When I got the machine the gicleur (Bezzera = jet) was missing completely so I machined one up from some available stainless 5mm thread however it wasn't quite concentric in the lathe when I did it and I reckon it's gone over-size (1mm drills and stainless are fun...NOT).

Just wondering on some thoughts from those who may be familiar with this or similar machines if I could be on the right track in suspecting the gicleur. Unfortunately Bezzera put their "jet" on the outlet of the HX and I had all sorts of grief getting that particular fitting to seal properly and since it's under constant line pressure I'm very reluctant to crack it again unless I'm pretty sure it's a likely candidate.

Incidentally, before somebody says it, yes I do know how to dose and tamp; it's the machine.

Cheers
Pete

I agree you're on the right track in suspecting the overenthusiastic water flow. IIRC, gicleurs range from 0.6 to 1.0mm. I don't know if this applies to your Bezzera, but a friend of mine had a Livia 90 and complained about how fussy it was. We thought it might be the 6 hole dispersion block wasn't doing the job, so he had the holes filled and then redrilled the block with 10 smaller ones. It helped quite a bit.

Thanks Dan, I had considered two options to modify the machine. One was to fit a second gicleur just above the "coupling ring" as Bezzera call it. The brass block it attaches to has a fitting which is threaded all the way through so I think I could be able to get one in there and it's often where they're fitted on other makes. The other is to completely machine up a new dispersion block, it's a fairly straight forward lathe job, with more/finer holes as you suggested. However the overwhelming thought in my mind is that it shouldn't be behaving as it is now, and before I begin modifying things I would like to get it working precisely as the manufacturer intended. I generally always try to go back to this base line, with pretty much everything I work on, as generally the manufacturer put quite a lot of R&D into their products and I figure they probably know a thing or two about the product

PeteF wrote:The free flow straight from the shower is 550ml in 30 secs

If you can get the free flow rate (what Al on alt.coffee used to call "water debit") down to 65-90ml in 10 seconds, your extractions will be more forgiving (IMO).

I recently purchased a similar machine. Are you aware of any electronic manual that is available for this machine which is a 1995 model. I cannot find one anywhere (see Need manual for 1995 Bezzera).
Turned machine on and pump runs but there is no filling of the boiler.
Thanks
JM

Thanks Andy, yeah that would seem about right just roughly eyeballing it, it seems to run at about twice the rate I would expect. I will try replacing my DIY gicleur with the Bezzera stock one, and if it's still flowing fast machine up one still smaller.

I'm not sure on any electronic "owners manual" however there are parts diagrams if that's what you may be after?

Pete

I machined up a new gicleur, this time 0.6mm. I use Loctite 567 I think it is, and find it works best if the machine is heated then turned off before the pressure builds as the loctite goes off faster/stronger. In the process I found the OPV was jammed closed which hopefully would explain why I kept blowing a joint seal on the HX inlet as the water expanded when idling!

I have still to pull any shots but did quickly do a flow check and it's now 300ml/30 sec which seems a little more reasonable but I must confess still higher than I was expecting for such a small gicleur. I would imagine most single group HX machines using the Procon would flow at around the same rate.

Look forward to cranking it up in anger come morning.

Pete

PeteF wrote:I have still to pull any shots but did quickly do a flow check and it's now 300ml/30 sec

Sounds better, still a touch on the high side.

The reasons that water debit is generally measured per 10 sec (eg, 75 ml/ 10 sec) are:
1. It puts more emphasis on the initial ramp to pressure, which is a significant factor
2. It saves 20 sec of your time per measurement!

FWIW, your actual brew pressure at the group will be a little less than before, because you'll have a greater pressure drop through the plumbing. Not a big deal, just something to be aware of. It may not show on your gauge, depending on where it senses pressure.

Interested to hear how it affects your espresso making. If you're up for more machining, it would be interesting to compare the results of the new 0.6 and a newer 0.5 mm. My experience is the smaller the gicleur, the finer the required grind. Up to a point, smaller gicleurs may improve the espresso by decreasing the degree of microchanneling. Smaller gicleurs also increase maintenance because they have a tendency to scale up faster.

That last item is NOT a positive.

BTW here's a picture of my gicleur: .

AndyS wrote:If you can get the free flow rate (what Al on alt.coffee used to call "water debit") down to 65-90ml in 10 seconds, your extractions will be more forgiving (IMO).

Given the machine configuration, it might be easier to install a delay-on-make relay and use line-pressure preinfusion. In that case, the larger water debit is what he'd want. This assumes that the line pressure is regulated someplace in the 2-3 bar range, of course.

Hi Pete -

Are you sure you have the "right" Procon pump installed? What does the brew pressure gage go to when you are doing this free-flow test?

BTW, a good friend who has a Quickmill Vetrano (original owner, excellent shape) flows 497 ml/min in a free-flow test.

erics wrote:a good friend who has a Quickmill Vetrano (original owner, excellent shape) flows 497 ml/min in a free-flow test.

Hi Eric:

JOOC, what does your friend think measuring a full minute of flow will tell him that a 10 sec measurement won't?

And Happy Holidays to you.

I know that you are an advocate of the "10 second - water debit" parameter and, way back, I researched the alt.coffee archives to understand the background info. I certainly have a lot of respect for someone who takes the time to contribute to the various coffee forums (Al Critzler(?)) but my interest dimmed a little when I read "As has been explained to me, acceptable water debit through . . ."

http://groups.google.com/group/al...4/fb226db4ea9f8deb

Having observed several espresso machines in both the home and cafe environment, I saw almost instantaneous flow when the switch was depressed or the lever actuated and that tilted me in the direction of measuring flow over a longer period of time for the sake of accuracy. I positively agree that there probably exists instances and/or machines where one would want to take into account the ramp-up as a trouble-shooting tool but in the posts I've seen here, the problem centered around flow from a tired pump or an incorrectly adjusted valve and I was always comparing their numbers to those published by Ulka (now CEME), Fluid-O-Tec, or Procon.

As regards my friend with the Vetrano, he sorta split the difference - he made several 30 second flow runs over a period of maybe an hour and consistently hit 7 ounces on the nose. I just did the conversion.

Thanks guys, ok I've pulled a couple of shots now. First went down the drain, a gusher. Second was ok but still having dreadful channeling issues. Haven't used the naked PF on it much, no need to, channeling is as plain as day fairly early into pour. Grrrr! Not sure what I'll try now, I've gradually worked my way up to considerably updosing so maybe try going down in dose/finer grind? I generally find if I can go back to "Italian Standards" on a machine it generally behaves ok, not surprising given those times/dosage/volume are what they're designed for, but I must admit I haven't come across a machine as temperamental as this little bugger!

Yep, got the right Procon in it, well as far as I can tell anyway. It's a series 1, which would be as installed. I was under the impression that the series 1/2/3 had identical specs anyway? The reason I use 30 secs rather than 10 to measure the flow is that, firstly it's what's on my shot timer More the point, as Eric mentioned, in this design of machine there is no chamber etc to fill when the brew button is pushed. The circuit is charged up to the 3 way solenoid so when the pump comes on/solenoid opens the water simply goes down the final 10 cm (6") to the dispersion block. I just find it easier to accurately measure the larger volume also, but if you divided by 3 it would be pretty much identical to a 10 sec measurement. Eventually I will probably modify to delay the pump activation but I feel that type of thing should be further down the track and my emphasis at the moment is getting the machine/Barista working as designed i think the pump delay will be a very worthwhile mod on this machine however.

Andy, yes the design of the machine puts the pressure gauge on the inlet to the HX so it will most certainly affect the indicated pressure. I was brewing at 8.5 Bar previously and adjusted the pump to 9.0 to compensate. I did see some graphs somewhere (Illy??) that showed the affect of different sized gicleurs on pressure drop and while I can't recall the specifics I guesstimated 0.5 as covering it. A 0.5 hey, hmm, well now we're pushing it 0.6 was the smallest drill bits I had but could certainly get smaller, however just want to get everything working as spec for the moment and experiment from there.

Pete

Bezzera shows the BZ40 as having a 100 litre/hr Procon which I equate to their 25 gph build. The numbers 1322 should be on your pump as per this link:

http://www.proconpumps.com/PDFs/E...20Series%20123.pdf

Yes, the series 1,2, & 3 have same specs but there exists nine different specs within each series.

Thanks Eric, I didn't know about that link, champion! I contacted Procon both here (in Oz) and in the USA a few weeks back about different pumps in their range but neither bothered to reply

The pump reads CB1305BLF and G2244. What is presumably the build date on the reverse side is 09/96 which coincides with the only other thing I can date, the start capacitor on the motor, also coincidentally 09/96. I would say it's the motor that came with the machine. It could well be that the pump is getting to the end of its life but I can't see that causing channeling issues or high flow. I thought you may have been on to something there though, oh well

erics wrote:Having observed several espresso machines in both the home and cafe environment, I saw almost instantaneous flow when the switch was depressed or the lever actuated and that tilted me in the direction of measuring flow over a longer period of time for the sake of accuracy. I positively agree that there probably exists instances and/or machines where one would want to take into account the ramp-up as a trouble-shooting tool but in the posts I've seen here, the problem centered around flow from a tired pump or an incorrectly adjusted valve and I was always comparing their numbers to those published by Ulka (now CEME), Fluid-O-Tec, or Procon.

Perhaps we're talking about two different issues. The 10 sec water debit is mainly concerned with quality of extraction. Your 1 min flow test is about whether the pump is operating properly. Seems to me the OP's is with the former, not the latter.

Andy, can you please elaborate in what you mean by, "The 10 sec water debit is mainly concerned with quality of extraction." By this do you mean the time of pre-infusion before full water flow is achieved? In other words, you would really have to look at both the 10 sec and 30/1 min figures for a full account of what the machine is doing. Having said that, in a machine such as the BZ40, there is little proportional difference between the two ie 30 sec volume is pretty much exactly 3x 10 sec volume.

Incidentally I do notice on the pressure gauge a longer time for full pressure to be applied, though it's still not a huge period of time. Overall, I feel the 0.6mm gicleur will be a positive move. Also, I have been out and have critically low coffee levels (again), but did pull a shot as promised much closer to the 14g "standard" obviously with a finer grind. Certainly a better "quality" shot and showed promise. ... mind you I was looking at the amount of head-space on the puck and was thinking OMG I could do laps in there, but it did seem to show potential Guess the Italians know a thing or two about coffee after all!

Pete

PeteF wrote:Andy, can you please elaborate in what you mean by, "The 10 sec water debit is mainly concerned with quality of extraction." By this do you mean the time of pre-infusion before full water flow is achieved?

Yup, the 65-90ml/10 sec is about saturating the puck before you blast it with full pressure. But such a machine will require a little more careful management (water treatment, descaling, etc) because the small gicleur will tend to clog or scale up more easily. The payoff is more even extractions, less channeling.

Consider that the old Starbucks Lineas were set up with minimal flow restriction (200-350ml/10 sec water debit) so that maintenance issues were minimized. Since Starbucks had little interest in serving quality espresso, a little channeling (or a lot) was OK. Newer LM models (GS/3, GB/5), sold to more discerning customers, have low water debits to maintain puck integrity. On the old machines, liquid espresso appeared at the portafilter within a couple seconds. It was all about minimal maintenance and maximal shots per hour. On the new machines, it normally takes 5+ seconds for liquid to appear.

PeteF wrote:In other words, you would really have to look at both the 10 sec and 30/1 min figures for a full account of what the machine is doing. Having said that, in a machine such as the BZ40, there is little proportional difference between the two ie 30 sec volume is pretty much exactly 3x 10 sec volume.

From my point of view, if the pump isn't shrieking (from cavitation, bad bearing, busted vane, etc), and you can adjust the bypass so the pressure gauge reads 8-9 bars while you're pulling a shot, then the 10 sec measurement is all you need worry about.

PeteF wrote:Incidentally I do notice on the pressure gauge a longer time for full pressure to be applied, though it's still not a huge period of time. Overall, I feel the 0.6mm gicleur will be a positive move.

As others have suggested, if you're still dissatisfied you can try a pump delay preinfusion. And if that isn't enough, re-drilling the dispersion block might help a lot. You can always buy a GS/3.

G'day Andy, off the shelf the BZ-40E wasn't (indeed isn't as they're still sold) exactly an entry level machine (well unless you're talking commercial machines of course) and is capable of producing very good quality espresso. Clearly there are better machines available however it shouldn't just channel as a "norm". There will definitely be things I will be able to do to this machine to make it even better however. Nevertheless that's for the future. I've had a look at the pucks coming out and there are no obvious holes in the puck as there were before so hopefully I'm getting somewhere.

I'm pretty sure mechanically the machine is now running 100% but that still leaves normal machine adjustment variables (brew pressure and water temperature), Dosing variables (amount/grind), and Barista technique. Unfortunately with so many variables it's difficult not to begin chasing one's tail, however I want to eliminate as many of the objective variables as I can before starting to change more obscure factors.

Unfortunately, as with any new install, many things are changing in a short space of time and it can be very difficult to track down when something starts running off the rails. One of the things was a change of grinder, from a Rocky to the Mini E, and there is no doubt I'm now getting more channeling with the Mazzer which is more than a little frustrating/disappointing to chase down why. At least I'm a bit happier with the rate of water flow now, as it did look very excessive previously. Whether it was contributing to the channeling issues, well ...???

Pete

PeteF wrote:A 0.5 hey, hmm, well now we're pushing it 0.6 was the smallest drill bits I had but could certainly get smaller.

I'm old enough to remember carburetor jets. I'm fairly certain that they weren't drilled to size. I believe the final sizing was done by testing against a pressure/flow meter. I'd be very surprised if gicleurs aren't manufactured the same way. What I'm saying here is that something you made with an 0.6 mm drill bit might very well flow significantly more water than a pukka piece. The commercial ones seem to have either 4x0.5 threads or 6x0.7 ones; why not build an adapter for a commercial gicleur that would be much easier to machine?

Best,
David