Over the past 3 weeks or so, I've been having extraction problems with my espresso setup. Originally I suspected my burrs, but finally it came down to my brew pressure being set to high (unbeknownst to me). As you can read in my concluding post, the Elektra appeared to be quite sensitive to excessively high brew pressures. When I asked him about the typical 8-10 bar "recommended" range, Jim clarified,

another_jim wrote:My guess is that Italian "custom" here is meaningful, but as usual, not part of anyone's conversation, not even theirs [...] in any case, rotary machines are set to 8 to 9 bar in Italy, and vibe pumps at around 10 to 12 bar. I was conflating the two when I wrote the guide, since I thought a properly OPVed vibe was identical to a rotary. But I think the initial attack of a rotary that Dan described may make a difference. Certainly on my Elektra, which runs at 11 to 12 bar, adding an OPV made no difference at all.

I don't have taste evidence. But Ken, who has identical vibe and rotary Cimbalis, needed a delay-on-make preinfusion relay on the rotary, but nothing on the vibe.

Jim was alluding to the (I believe later) well-documented fact that rotary pumps' high flow rate means that they infuse and come to full pressure dramatically faster than vibratory pumps (if they are not restricted by some other means). Ken found on his La Cimbali Junior rotary that the forgiveness factor was much lower than the vibratory pump version of the same machine, and that it needed some line pressure preinfusion to slow the ramp-up of pressure and make the machine as forgiving as its vibratory sibling.

On the other hand, John found on his La Spaziale that preinfusion was essentially a waste:

RapidCoffee wrote:My Spaz S1V1 has a 53mm grouphead, rotary pump, and no preinfusion. I added the mechanical preinfusion cylinder retrofit, then removed it after a few weeks. I saw no benefit from preinfusion on this machine. This might be due to any number of factors: smaller (53mm) grouphead and deeper (triple height) baskets, Robur grinder, decent barista skills, mediocre palate... But whatever the reason, my initial impressions were ho-hum. The Spaz certainly does not require preinfusion for consistently good pours.

Granted, John's machine's preinfusion method was wholly different from Ken's. The La Spaziale has a spring-loaded chamber that has to fill prior to coming to full pressure, giving it a progressive preinfusion similar (not identical, I believe) to the E61's. Ken's preinfusion method was to simply delay pump activation such that line pressure is delivered to the group for a few moments... similar to the way a paddle group La Marzocco or Synesso works. I've installed a mechanism to do this on my Elektra, and while I haven't studied it closely, I'd say it hasn't changed much of anything.

Finally, Tom (dsc) has had the opposite experience on his own setup, an Elektra T1 configured very similarly to my own:

dsc wrote:After various experiments with preinfusion I now abandoned the whole concept as I don't think it gives you anything. I'll say even more, I think it can cause more bad then good, especially when it soaks the puck only partially.

There's been a lot written about preinfusion, and indeed considerable discussion on this very forum. I'd say that the majority of observations fall in the center--that there's no real difference, though many people claim that preinfusion enhances the "forgiveness factor" of the machine. Certainly the average barista using a Synesso doesn't bother with preinfusion (in my experience). But there's no real consensus on this. Why is that? Perhaps it seems rather obvious that preinfusion's effectiveness depends on the machine's configuration: group geometry, installed gicleurs, pump flow rate, distance between pump and group, etc. Plausible if not quite probably, but woefully vague. Is there a rule of thumb here?

I'd love to hear further thoughts on this. If you have a machine on which you can control preinfusion length and/or toggle it on/off, have you ever studied the effect? I've surveyed a good bit of 'research' on this topic, but I'd welcome links to less-known studies, if they exist.

I don't want to give the impression that I'm anti-preinfusion. On the contrary: I believe that preinfusion accounts for a large part of the well-deserved reputation of the E61 as a forgiving grouphead. Since the Spaz S1 feels more finicky than my Vetrano (an E61 box), I was surprised that the preinfusion chamber did not improve my S1 pours. My experience was more along the lines of Tom (dsc) with his Elektra.

Anyway, still an interesting topic for investigation. I cannot flip a switch to turn preinfusion on and off. Perhaps someone with an S1V2 (or other programmable preinfusion machine) could be convinced to run some more tests?

For what it's worth, I don't think it's useless either, just trying to spark some interest. My hypothesis would be that its usefulness depends on other features of a machine's design.

For what it's worth, anyone thinking about doing some further experimentation may have the bar set high for themselves. This topic reminds me a bit of Adrian's thread on preinfusion and intrashot temperature stability from about a year ago. Specifically, I'm thinking of Ken's admonition:

Ken Fox wrote:There are two basic problems in studying anything like this, which are:

You have to define precisely what it is you are trying to study and how you will study it, and then after defining how you will study it you must evaluate the results of what you studying in a way that will stand up to some sort of "scientific scrutiny."

The first is possible, but doing only the first makes it likely that the response to these results will be "So What?

I'd be interested in getting an idea of the kind of resources that would need to be pooled to perform a test as narrow as answering the question, "what is the effect of preinfusion on machine x?" for a variety of machines x. Once some preliminary data based on an agreed testing scheme (blind taste test, etc.) is obtained for a number of machines, and assuming that the tests bear out the anecdotal evidence out there that preinfusion has different effects on different machines, we could begin to flesh out a hypothesis about what the relationship between preinfusion/type of preinfusion/type of machine.

Actually figuring out a hypothesis about that, not to mention a way to test it, will be substantially more difficult, but even if we never get to that point I'd claim that anyone who's willing to do some testing on this subject and provide complete data about their methodology and results would be doing service to the coffee community--in other words, if you're not up to following the scientific method as meticulously as Jim and Ken have in the, but willing to design some simple experiments that go beyond just making coffee in the morning and comparing your results to your vague memory of what happened yesterday and the day before, I think there's still a lot of useful information to be gathered.

For my part, among myself and Jon Rosenthal and Paul Jones, we've got a pair of Roburs, a Synesso, and soon enough a Speedster, both excellent machines with some degree of controllable preinfusion (not so much on the Speedster without physically removing the preinfusion chamber), and ideal for isolating the variable of shot-to-shot temperature stability with minimal additional effort. We might even be able to get our hands on a loaner Vivaldi with timer-based preinfusion as well... Maybe. I can't commit to testing without getting those guys on board, but perhaps we could all persuade them to loan out their equipment for the purpose and participate in a few afternoons' worth of blind taste tests over a few weekends.

no.

yes.

thanks for the time.

ShadowFax wrote:and soon enough a Speedster

You been talking to Kees Nicholas!?

shadowfax wrote:Perhaps it seems rather obvious that preinfusion's effectiveness depends on the machine's configuration: group geometry, installed gicleurs, pump flow rate, distance between pump and group, etc. Plausible if not quite probably, but woefully vague. Is there a rule of thumb here?

I don't use the 'variable' portion of the pre-infusion on the Speedster. It has a .6mm gicleur and a pre-infusion chamber so I cannot remove them from the equation, but after playing around with extended pre-infusion at line pressure, I don't bother. Just straight to pump pressure.

I tend to agree with the above quote, but I would also add that it probably depends on the grind quality, distribution, freshness of the coffee etc. If everything is pretty much spot on, the well talked about 'forgiveness' of pre-infusion is pretty much a moot point in my opinion.

Cheers, Chris

shadowfax wrote:(vast snippage) . . . .Jim was alluding to the (I believe later) well-documented fact that rotary pumps' high flow rate means that they infuse and come to full pressure dramatically faster than vibratory pumps (if they are not restricted by some other means). Ken found on his La Cimbali Junior rotary that the forgiveness factor was much lower than the vibratory pump version of the same machine, and that it needed some line pressure preinfusion to slow the ramp-up of pressure and make the machine as forgiving as its vibratory sibling.

The work that Jim and I did with my rotary machine, in the "preinfusion" threads we initiated, needs to be considered in the light of the dosing used. At that time I was dosing around 18 to 19g in a double basket, while using Cimbali Junior grinders, and there was no question that installing the delay timer in the rotary machine made a huge and obvious difference in the frequency of observed channeling. I now dose around 14g, and in addition all of my grinders are either Cimbali Max hybrids or a commercial conical (Compak K10 WBC). I have no idea at this point whether my current dosing pattern and grind quality would benefit from the delay-timer induced preinfusion that my rotary machine has.

I have left the delay timer and input line pressure reducer in place; I only disable the delay timer preinfusion setup when I do backflushing.

I could obviously test this, at least in an informal way, by simply turning off the preinfusion with the on-off switch I have installed on the machine.

My first reaction, however, is "why bother?" Any observations I might make would only be useful to the very small universe of people who own a modern Cimbali rotary machine, who dose as I do, who use predominently single origin coffee for espresso, and who have similar grinders.

Since I haven't been all that curious myself, given that the installation is already in place and causes zero problems, I just haven't bothered. My opinion is that with my current dosing and grinders, that the preinfusion mod I have installed would not yield enough benefit to bother with installing, but then, in reality, I just don't know.

ken

Ken Fox wrote:I could obviously test this, at least in an informal way, by simply turning off the preinfusion with the on-off switch I have installed on the machine.

That picture of the switch makes my head want to explode... you could at least label it "!pre-infusion"

gyro wrote:You been talking to Kees Nicholas!?

Alas, that level of capital outlay's not in the cards for me. That honor goes to one of my local friends.

I don't use the 'variable' portion of the pre-infusion on the Speedster. It has a .6mm gicleur and a pre-infusion chamber so I cannot remove them from the equation, but after playing around with extended pre-infusion at line pressure, I don't bother. Just straight to pump pressure.

I tend to agree with the above quote, but I would also add that it probably depends on the grind quality, distribution, freshness of the coffee etc. If everything is pretty much spot on, the well talked about 'forgiveness' of pre-infusion is pretty much a moot point in my opinion.

With the Speedster's fixed preinfusion, it's perhaps little surprise that variable line pressure preinfusion makes no discernible difference. In fact, I'd be interested to see a graph of the Speedster's pressure ramp-up. How long does it take for that preinfusion indicator to come out fully when you go straight to pump pressure? I imagine that might give you a smoother preinfusion than if you fill the chamber at line pressure then hit the pump switch after you've filled the buffer, making the ramp from line pressure to nine bars virtually instant. It doesn't seem to make a bit of difference for the Speedster either way, but could it on other machines?

Ken Fox wrote:My first reaction, however, is "why bother?" Any observations I might make would only be useful to the very small universe of people who own a modern Cimbali rotary machine, who dose as I do, who use predominently single origin coffee for espresso, and who have similar grinders.

Since I haven't been all that curious myself, given that the installation is already in place and causes zero problems, I just haven't bothered. My opinion is that with my current dosing and grinders, that the preinfusion mod I have installed would not yield enough benefit to bother with installing, but then, in reality, I just don't know.

I have to admit that this is the main reason why I haven't messed with my similar setup on the Elektra. Heck, I bought my relay for cheap ($10-20 shipped, IIRC) out of curiosity, tried it a bit, and surmised I wouldn't be able to tell the difference based on a little bit of fiddling, and I've also more or less left it at that. And I must say that my dosing habits are typically somewhat higher than the average Elektra-ite; I enjoy 17-17.5g often in my ridgeless double basket, leaving a screen imprint at the end pretty often.

All my exposure to preinfusion has led me to believe its effects are generally subtle at best, and certainly most claims about it center around an increased forgiveness factor, with claims now and then that "sweetness" or "smoothness" are enhanced. If it's true that preinfusion's effect tends to be something that won't show up in a taste test (either as a taste improvement or an increase in consistency), that will be fair enough grounds to cop out of experimentation. Who knows? If I can get my fellow Houstonians interested, I'll see how many machines we can do. Even if it's all moot, haven't you heard? Blind testing builds character: it makes you look like an idiot!

Hi guys,

well I still have my adjustable timer based preinfusion. It's 1bar for a duration of 0-9s, so I can probably run some tests. I do have an M4 Beswick fitting, a pressure transducer and some 1/16" copper tubing, but I haven't got anything to block the 4way junction after removing the existing PT line. I should get some parts probably next week, or the week after, so I might be able to see how the pressure changes on the group, as opposed to mid-installation.

I think there's plenty to test to be honest, low doses, high doses, double/triple baskets (due to Elektra being a bit picky when it comes to head space), long, short preinfusion, perhaps preinfusion pressure changes for those who are able to (I can't, I've got 1bar in the pipes and can't go any higher). I can do some simple testing today, a full and partial soak for example? If anyone comes up with a nice test procedure to follow I'm on board:)

Regards,
dsc.