When I first got a Vibiemme Double Domobar (version 3) to replace my Isomac Tea, I got generally very good results, but I only occasionally got the great shots I had come to expect from the Tea. I thought it might have something to do with the temperature vs. time profile. The Tea (HX with vibe pump) had a very flat temperature profile during the shot, about a 2 F peak-to-peak variation. For example:

The main problem with the Tea was that changing coffee and temperature, for my wife's latte say, took a while to dial in. I figured a double boiler would fix this, and it has for the most part.

The Tea profile was taken with coffee in the basket and a thermocouple draped over the basket rim. See Jim Schulman at Seeking low tech way to measure brew temperature.
A similar plot for the VBM double boiler with rotary pump shows more than double the temperature excursion:

(The shot ended at 45 seconds.) The 4.5 F p-p variation in this temperature profile is typical of my VBM, and it was a surprise and a disappointment. I was expecting something closer to the profile from my Tea, and I thought the temperature variation might have something to do with my problem that there seemed to be no temperature setting that gave a really excellent shot. This raised three questions: Why is there such a large temperature excursion during the shot, why was it a surprise, and what can be done?

To answer the second question first, most temperature profiles are measured with a
Scace Thermofilter Temperature Device, which has very low flow from the beginning. When I measure the temperature profile of my VBM DD with a backflush disc that has an 8-mil hole in it, a poor-man's Scace thermofilter that allows water flow of about an ounce in 30 seconds, the temperature profile looks very flat, as shown here:

The shot ended before the last 2 F rise, so the basket temperature peak-to-peak variation appears to be about 2 F. This is what most people show, hence my surprise.

Now for the reason: A basket of dry coffee grounds will allow very rapid flow until the puck saturates. And, the rotary pump on my VBM DD provides much more rapid flow than the vibe pump on my Tea. I am pretty sure this rapid flow of hotter boiler water into the group and puck is the origin of the temperature spike in the group head and in the basket (at 20 s in the plot). As the puck saturates and the flow slows, the water temperature drops closer to the group temperature, then toward the end of the shot both temperatures rise as the group heats from the flow of warmer boiler water.

If this is the correct explanation, then preinfusion should mitigate the problem. There is an enlightening preinfusion thread from Ken Fox and Jim Schulman here:The Impact of Preinfusion on the Taste of Espresso Shots.

The basic idea with a plumbed-in rotary machine is to lift the brew lever enough to let in water at the plumbing pressure, 2.5 bars max in my case, until the puck is saturated but before any drips land in the cup, then to turn on the pump. I moved the pump switch away from the lever cam slightly to be sure the pump did not turn on unintentionally. I also repeated Ken and Jim's experiments to see how much preinfusion time was required vs. pressure. I did not find a dry cone of coffee in the center as they did even down to 1.5 bar, but I did find that when drops began to form on the outer rim of the basket there was a circle of unwetted coffee a few grains thick in the center of the puck. I was using a 15-g VST basket, which is slightly convex, with 15 grams of coffee. (I used a flat tamper for these tests, but I have ordered a European curve in hopes of getting totally uniform wetting, though I doubt those few grains are going to make any difference.)

After several days of practice, my procedure now is to use a bottomless portafilter so I can watch the first stages of extraction (via its reflection in the shiny stainless of the drip tray). To begin, I preinfuse with mains water. In 8-12 seconds, depending on whether I'll get a lungo or a ristretto, the outer rim of the basket bottom will begin to color. At this point I immediately raise the lever so the pump comes on, then finish the shot as usual. The temperature profile generally has a p-p variation of 2 F and looks like this:

This technique produces awesome shots; this particular one had slight almond and a citrus flavor I've seldom tasted from these beans. These beans, BTW, are Roman Espresso from Blue Bottle. They seem to be unique in preferring the lower temperature you see here.

The advantage of watching the extraction is that the preinfusion time can be adjusted depending on the flow rate. If experience shows that, say, an 8-second preinfusion works for most shots, I may consider putting in a delay-on-make timer, though the stainless sawing required to easily bypass it for flushing and cleaning shots may stay my hand.

With so much written on this forum and others about all aspects of espresso, in spite of the heroic efforts of the forum moderators and the awesome search engines available today, it is sometimes difficult to know whether one is repeating the obvious or advancing the field. I did not find any discussion of the relationship between temperature profile and preinfusion, so I think this is a new look at the topic. If you know of earlier discussions, please let us all know. I did find a reference to very short E61 preinfusion with a rotary pump,
Pressure profiles, preinfusion and the forgiveness factor, which referred to preinfusion times of 2-3 seconds. My own measurements of the VBM are closer to 1.5 seconds. This short time is likely the root cause of the temperature excursions without additional preinfusion. Incidentally, I don't think the issue will be as much of a problem for flooded groups, which are essentially at the boiler temperature.

Hope this helps all you rotary pump owners with E61 groups out there. I'd love to see any corroborating or contradictory evidence.

I'm taking votes: Will the initial temperature spike be reduced to none at all if I lower the preinfusion pressure from its current 2.5 bar? Will the taste deteriorate with lower preinfusion pressure? Vote today!

Gerry,

Sorry I have not responded sooner. Somehow I missed the post here, and your E-mail to me got snagged by my ISP's E-mail spam filter at the server side, so I just found it this morning.

I have not taken the trouble to quantify and graph my results in terms of preinfusion temperature variations. What I have found by taste is that a preinfusion time that is right around when the first drops fall into the cup seem to work very well. It is to the point now that I preinfuse every extraction that way- until a full ring of drops form around the perimeter of the basket (I always use a bottomless PF) or a drop falls into the cup.

You had asked me about the "detent" of the lever at the preinfusion point. Technically speaking, preinfusion begins with just a little more movement than when the "detent" is reached. The feel in the lever when an increase of force is needed to mover the lever further is the internal cam hitting the brew valve, and at that point a bit more force is needed to overcome the brew valve's spring.

This morning I updosed just a tiny bit. Maybe about one gram. The PF was full and I had just a bit more left in the doser. I piled it on top of the leveled PF, and I used the edge of my leveling tool to distribute it and press it in as evenly as I could. I preinfused (of course) and the pull went at least three ounces before beginning to lighten. I was steaming the milk at the time and kept looking over at the pull wondering if the cup would hold it all. When I poured the milk there was only room for about 1½ ounces! It was the best cappa I have ever made... Heck. I think it was the best cappa I ever tasted, and it was my home roast and blend as well!.

It's been ten weeks and I am being impressed every day by the performance of this machine.

In regards to the new DD owner's manual covering V.3 (the current version) of the machine: Stefano recently received the pre-final (post-Beta, if you will) version which he said he was going to post to his website. I have not heard when that will be. I assume in the next week, but you know how that goes. I will go so far as to say that I feel it is my best work to date and it has a section that is the most detailed explanation, fully illustrated, on how an E-61 group operates that I have ever seen (for what that is worth, but I did search before writing it). I am excited for owners and non-owners of this machine alike to get their hands on all 70+ pages of it!

Thanks for your comments, Randy. At the risk of continuing what is mostly a monologue, here are the results from lowering the preinfusion pressure.

I haven't received any votes, but I'm sure all you VBM DD owners out there would like to know that I've lowered the supply pressure to 1.5 bar with good results. The preinfusion time is 20 s compared to the 10 s it took for color to form at 2.5 bar. The temperature spike in the basket is indeed reduced, though the one in the group head is still there. The resulting shots are exquisite.

Bottom line is that it I have a slight preference for lower preinfusion pressure, partly because of my limited taste tests and partly because it's easier to start the pump before any drops fall that might influence the taste by contamination with a low-pressure extraction. This is a different conclusion from The Impact of Preinfusion on the Taste of Espresso Shots. The main difference is that my basket is not up-dosed; it's 15g in a 15g VST basket, which leaves plenty of headroom to get equal flow over the entire puck.

Ideally, the E61 group would take care of all the preinfusion needed. I'll investigate reducing the flow rate with a smaller gicleur (jet) to see if that helps.

Preinfusion (and flushing) Update:
Now this really is becoming a monologue, but I'll continue in the hopes of helping any VBM double boiler owners out there. Since the earlier posts I've replace the stock 0.070" gicleur (jet) with a 0.055" diameter one. This about doubled the preinfusion time, so that now with 2.2 bar preinfusion takes about 20s. I haven't noticed any taste change with this longer preinfusion time, but I still get better-tasting shots with mains preinfusion than the E61 provides on its own. A comparison of the pressure profile under different conditions, including mains preinfusion and a profile of EricS' Anita that he kindly sent me, is shown here:

You can see that even with the smaller gicleur the flow limited only by the E61 preinfusion is faster than the Anita.

Another important improvement to my shot quality with the VBM DB V.3 has been a flush routine originally suggested to me by RandyG, who wrote the manual for this machine (so he knows what he's talking about). As shown in earlier posts, with a 2-oz. flush the group is still enough colder than the boiler water that it causes temperature variation during a shot. When the steam boiler is on, I've found that a 4-5-oz. flush is just about right; shorter and there is a temperature rise at the beginning of a shot, longer and there is a temperature drop in the middle, followed by a rise at the end. To see how this is working out, four nominally identical shots taken yesterday and today are shown here:

The times have been aligned so that the preinfusion starts at about the same time. It's comforting that all these shots are within a degree of so of the set point. One is about a degree hotter than that; often the first shot of the day, after several hours of sitting, is a degree or two warmer than the later ones, in spite of the flush. Nevertheless, you can see that the group TC just before preinfusion is a good indicator of the shot temperature. I expect that waiting till the TC cools a bit will collect these temperatures even closer.

THEN I'll be ready to give Vivace Dolce another try. It is a demanding mistress .

It's me again, with an update on getting even flatter temperature profiles from my VBM DD v.3, with the additional result that with this sequence one can know the average shot temperature by measuring only the group temperature with the kit supplied by EricS.

First, I've standardized on a 15-second pre-shot flush, which results in 3.5-4 oz. of water through the group. This seemed to minimize the variation in group head temperature during a shot.

Second, you might have noticed (though I didn't right away) in the second plot of my previous post, that the flattest profile has a wait after flushing significantly longer than the usual 20-25 seconds. Quite by accident I noticed that an even longer wait produces a yet flatter profile. Here is a sequence of four shots where the ~20s preinfusion was started about 60 seconds after the end of a 15-s flush:

I find the 0.6F standard deviation over all four shots, spanning 3 days, to be quite remarkable. These curves were taken while pulling excellent chocolaty shots of Vivace Dolce, a roast that is known to be demanding of temperature control.

I've since noticed that a similar profile results for 45- and 90-second waits, and what's more surprising is that the shot temperature is nearly the same, independent of wait time over this range. This suggests that as the group temperature cools after the flush, the boiler temperature rises just enough to compensate. Good machine design, VBM! Since the group temperature is decreasing with time while the shot is staying the same temperature, the difference between average shot temperature and the group temperature just before the shot does depend on the wait. I find that for a 45s wait the shot is about 1F cooler than the group just before the shot, and for a 90s wait they are about equal. With this information I can dispense with the basket thermocouple, except for the occasional check!

I'd be interested to hear from any other VBM DD owners out there, just to know that I'm not talking wholly to myself.

That is quite a nice study you have made of this machine. While I am not set up to gather and assimilate such data, I do find that in the cup this machine is remarkable. The day-to-day consistency is a wonderful thing on which one can depend. Drop the setpoint one degree and you can be fairly well assured that this one degree drop is what you will get at the coffee. My usage certainly verifies what you have seen.

Jim told me that he had been doing quite a bit of testing using a multi-channel, NASA surplus (or military- I don't remember) digital temperature recorder and I think a lot of this stability comes from his research as well as VBMs efforts. Stefano also assisted. To me? Don't really care who takes credit. The coffee I get from the DD is more consistent and more controllable than I have ever experienced, and with less effort.

This morning after I pulled the first double into the cappuccino cup, I placed it on the table and handed my wife, who normally is not prone to enjoy straight espresso, a demitasse spoon. I turned my back and began to prepare the second extraction. From over my shoulder I heard the click of the spoon on the cup, and then, "Mmmmmm." Need I say more?

Hello,
I only found your thread today, but this is really amazing research. I am expecting my DD to arrive some time in the near future and I expect your precise information to be of great use when I start to find my modus operandi with the VBM. I'm afraid to be of little use in adding to the discussion though, since my expertise lays not in such precise technical measuring and experimenting. But let me at least thank you for your research.

Somewhat troubling to me seems to be that I don't expect plumbing in to be viable option for the coming months, which of course means I wont be able to use the preinfusion to stabilise shot temperature. Any other strategies to dampen the temperature swings? Will the flush and wait still help somewhat without preinfusion?

Pieter and Randy, thanks for your notes. I haven't tried a shot without preinfusion for some while now, but I'll try a few in the next few days and see what happens. It's possible that enough else has changed in my routine that the preinfusion does not change the flavor that much now, though earlier both Randy and I noticed a definite improvement with preinfusion. More later...

I've tried several shots without mains preinfusion recently, but I have not been able to keep the peak-to-peak temperature excursion at the top of the basket less than 2-3 F. Typical examples of preinfusion and no preinfusion are shown here:

It's clear that the preinfused shot has a much flatter basket temperature profile. The not-preinfused temperature profile is similar to the one I posted at the beginning of this thread, so not much has changed in the intervening three months, in spite of my changing the gicleur and optimizing the flush/wait cycle. In the early days I tried several different flushing schemes but could not get the profile reliably flatter without preinfusion.

As far as taste goes, this few degrees variation from a not-preinfused shot is not important as long as the temperature doesn't wander into regions were the bean is particularly sensitive. It makes a big difference with Vivace Dolce, which gets bitter if a little too hot and sour if a little too cold. The shots shown here were of Compass Delirium, which is very forgiving at the lower temperature ranges, 192 to 196 F. I got some very tasty not-preinfused shots of Delirium and a couple of Dolce, so flat temperature is not a necessity, but it does seem (to me) to give a shot with more clarity than when the profile wanders more than 1F or so, so plumbing in the VBM is worth the effort.

I continue to believe that this initial temperature spike without preinfusion is caused by the rapid inrush of water from the boiler to saturate the puck. Once the puck is wet, the coffee grounds swell, slowing the flow so that the hotter boiler water is cooled more by the group, which causes the drop. At mains pressure, it takes 20s to saturate the puck; at pump pressure it takes 10s, so clearly the flow is faster with the pump. As someone on this forum said, all machines are heat exchange machines, it's just a matter of degree.