Some thoughts on heat exchanger vs. double boiler espresso machines - Page 15

Need help with equipment usage or want to share your latest discovery?
User avatar
shadowfax

#141: Post by shadowfax »

On the leaching issue, isn't it a point in LM/Synesso's favor that their boilers (though not their interconnecting pipes) are stainless steel? I just compared my GS3 brew boiler water, which I cooled down to about room temp, with the cold water that feeds the boiler--just a simple side by side comparison (Full disclosure: technically, last night I drained both my GS3 boilers and refilled them). I can't even remotely tell the difference between the two. That's very different from, say, comparing my WBC-standard AB packet water with my softened, filtered tap water, where the formulated water tastes cleaner and certainly is easily distinguished from the tap water, which leads me (yes, I was already quite predisposed to the belief) to think that the inlet water quality is probably quite a lot more important than any leaching that happens in the HX/brew boiler. Probably your water makeup also determines how much leaching actually goes on in the first place, too. Anyway, that's my rambling bit of data and speculation.
Nicholas Lundgaard

User avatar
gyro

#142: Post by gyro »

shadowfax wrote:On the leaching issue, isn't it a point in LM/Synesso's favor that their boilers (though not their interconnecting pipes) are stainless steel?
Interestingly, it would seem that the hot water side of the plumbing on the Synesso (ie starting with the HX preheat lines) and continuing through to the brew boilers is stainless steel. This is not the case on the cold water supply side however.

I don't have a point to make myself as its all above my pay grade, but just an observation that for some reason they have chosen to do it this way.

Cheers, Chris

User avatar
cafeIKE

#143: Post by cafeIKE »

Just for grins, I measured the TDS* with water from the Vibiemme DB brew boiler, cooled to RT, and the same batch of water used to fill the machine Monday.
Boiler : ~103 over 3 samples
Bottle : ~110 over 3 samples

Jim, once the boiler surface is scaled, doesn't the metal transfer stop?

For a graph of rising short interval e61 HX PID brew temps, see Scale, HX Temperature Stability and a bit o' PID.

* cheep & cheerful HM TDS-EZ

User avatar
another_jim
Team HB

#144: Post by another_jim »

cafeIKE wrote:Jim, once the boiler surface is scaled, doesn't the metal transfer stop?
It would take a complete surface of scale to stop the transfer; never seen that in a working machine.

The TDS levels I've measured out of brass or copper brew boilers has always been higher than the water going in, but I'm not sure how the reading needs to be adjusted, since the TDS is calibrated to regular water minerals, i.e. calcium, magnesium, and carbonates, not metals. Moreover, some of that is minerals concentrating from steaming.

One would need to fill the boiler with waters of various hardness levels, not use the steam at all, and test the boiler water a week or two later to really get a feel for what is going on. I'm not inclined to waste my time on this, since I don't think there's any espresso gold to be had from a few copper ions.
Jim Schulman

User avatar
luca
Team HB

#145: Post by luca »

RapidCoffee wrote:The downside: changing the brew temp (especially reducing it) takes longer with large brew boiler volumes. Dunno if I agree about equilibrium; I'd guess thermal gradients are a bigger problem with larger boilers (but that's more dancing angels).
The amount of time that it takes to increase the brew temperature will depend in part on the wattage of the elements. You can't generalise and say that a bigger boiler will increase its temperature more slowly unless you also say that the elements of the boilers being compared are of the same wattage. FWIW, on the 3grp FB80 EE that I used to use, from memory the element would only go on for 30 seconds or a minute to ramp up the temperature something like 1C. IIRC, the Synesso was quicker to heat up, but tended to overshoot (this was with an earlier model; I hear that they have done some work on this now).

In terms of cooling, presumably we need to consider surface area to volume ratio and insulation. In the LMs and Synessos that I have seen, the boilers are relatively small in diameter (compared with a typical HX boiler) and are uninsulated. There has been discussion that part of the reason for this is to allow for effective cooling that the PID can take advantage of to maintain brew temperature. FWIW, I think that the FB80 certainly took more time to cool, but we aren't talking an eternity; we're talking a few minutes.

In terms of thermal gradients, I don't think that it's fair to say that bigger boiler = bigger problems. It might be fair to say that bigger boiler = bigger gradients, but whether or not this is a problem depends on the design of the machine and whether or not this has been taken into account. If you look at a saturated group, presumably hot water rises up to the top, then cools, then sinks back down into the boiler. Perhaps this movement keeps the group metal nice and toasty so that pulling multiple shots won't change the temp too much. FWIW, I remember scacing up the FB80 and getting great repeatability between shots, whereas something like a commercial e61 HX would heat up across a series of shots in a row. I also remember that the intrashot temperature would increase on the FB80, which might not be a good thing.

Anyway, as everyone has said, all of the above is pretty moot unless it gets you a more consistently tasty cup.

Cheers,
Luca
LMWDP #034 | 2011: Q Grader Exam, Brewer's Cup #3, Australian Cup Tasting #1

User avatar
RapidCoffee
Team HB

#146: Post by RapidCoffee »

luca wrote:The amount of time that it takes to increase the brew temperature will depend in part on the wattage of the elements...
In terms of cooling, presumably we need to consider surface area to volume ratio and insulation. In the LMs and Synessos that I have seen, the boilers are relatively small in diameter (compared with a typical HX boiler) and are uninsulated. There has been discussion that part of the reason for this is to allow for effective cooling that the PID can take advantage of to maintain brew temperature. FWIW, I think that the FB80 certainly took more time to cool, but we aren't talking an eternity; we're talking a few minutes.
Certainly you can postulate a more powerful heating element, oddly shaped brew boiler, insulation, etc., and this will change things. But all other things being equal, a large brew boiler will take longer to respond to changes in temperature than a smaller one. Perhaps I'm more impatient than most, but waiting more than a couple of minutes for the machine to stabilize at a new temperature is unacceptable. I'd like to taste the shot, make desired temperature adjustments, and pull the next shot at the adjusted setting as soon as possible.

A large boiler in HX machines makes sense, but I remain unconvinced that it's desirable in DB machines. A small brew boiler allows a DB machine to respond with reasonable agility to temperature changes, and should be perfectly adequate for the home environment or light commercial use.
John

User avatar
luca
Team HB

#147: Post by luca » replying to RapidCoffee »

I'm not trying to argue specifically for a larger brew boiler for a DB machine, but I am trying to argue that it's useless to consider that all things are equal, because they are not. Any espresso machine has a lot more going on than just the number of boilers that it has in it. Much better to consider how everything is interacting and what result is being delivered.

There are still more points to be considered on this topic. To take your example of larger brew boilers with everything else being held equal, one advantage is that changes to the system are proportionally smaller, so it's easier to maintain temperature. Specifically, the influx of, say, 100mL of water into the boiler when brewing is going to lower the temperature 5x as much in a 400mL boiler as it will in a 2L boiler. Similarly, if the elements are the same wattage in both boilers, the overshoot will be proportionally 5X greater in the 400mL boiler. But even that is an oversimplification! To give just two additional points, the water coming into the boiler probably won't mix evenly throughout it. This means that the point at which you draw the brew water from the boiler will be important. If you draw the brew water from a point near the cold water inlet and the PID maintains the temperature based on a probe that is far away from this point, the system is probably actually going to be pretty crappy. You can laugh about this, but I know a guy who had just this problem with a single boiler machine that he PIDded and he solved it by adding additional piping inside the boiler to draw brew water from a different point. Part of this also means that a larger brew boiler may have the advantage of separating the incoming water from the brew water to a larger extent than the smaller boiler and, so actually delivering brew temperatures that correlate better with the PID readout. That was the first point. The second point is that the inlet water might be different temperatures. For example, if the 400mL brew boiler is fed with water heated from a HX through the steam boiler (or a combination of steam boiler water and mains water mixed together) and the 2L boiler is fed with mains water, it becomes more difficult to predict which one will maintain more stable inter-shot temperatures. So making generalisations based on the size of the brew boiler and presuming that all else is kept even really doesn't get us very far. To this end, it's possibly a little ungenerous to describe the LM and Synesso brew boilers as "oddly shaped" rather than acknowledging the possibility that the shape is a result of some conscious design and engineering.

The group head is obviously another factor that needs to be considered. Generally speaking, if the group is basically a chunk of metal that needs to be heated by conduction of heat from the boiler and through the metal, presumably a bigger boiler with a bigger element will put out more heat, get the group up to temperature faster and, probably, maintain the temperature better so that it heats up less over successive shots. If the group has water circulating through it, presumably more brew water temperature circulating faster is going to do a better job of keeping it at temperature than less. To this end, I always found it odd that people market e61 groups based on how heavy they are, at least in the context of DB prosumer machines. If group A is lighter than group B, but it is lighter because it has more space for more water to circulate inside it, wouldn't group A perform better, notwithstanding that it is lighter than group B? I haven't used all of the different options, but the saturated group machines that I have used seem to have done a better job at putting out a consistent temperature than the others. Ironically, I would have thought that this would be more of an issue at home than at a cafe; at a cafe, constant use will keep the heads nice and toasty, whereas at home you are more likely to experience rising temperatures - if this is going to happen - because you will be pulling back to back shots from a cold start.

... and then after all of this, there's the issue of taste. If you are one of the many people who wants a machine that will give you a big margin of error to pull rich, full bodied, chocolaty shots from the same blend, you might find that a HX with a fairly large preinfusion time is your idea of heaven.

Cheers,
Luca
LMWDP #034 | 2011: Q Grader Exam, Brewer's Cup #3, Australian Cup Tasting #1

User avatar
RapidCoffee
Team HB

#148: Post by RapidCoffee »

luca wrote:I'm not trying to argue specifically for a larger brew boiler for a DB machine, but I am trying to argue that it's useless to consider that all things are equal, because they are not. Any espresso machine has a lot more going on than just the number of boilers that it has in it. Much better to consider how everything is interacting and what result is being delivered.
Lots of interesting issues, IMHO all debatable. For example, I don't agree that it's necessarily easier to maintain temperature in a larger boiler under dynamic conditions. A small boiler with a decent sized heating element comes back up to temperature very quickly after pulling a shot (30-40 seconds max in my S1), and responds more quickly to changes in temperature settings than a larger boiler. You're probably correct about the increased overshoot in a small boiler, but I'd predict greater mixing issues in a larger boiler, and increased problems with temperature probe placement. Note that single boiler, dual thermostat machines tend to have small boilers, typically under 1L. This is almost certainly a consequence of the need for rapid switching between brew and steam temperatures. Yet the brew temp stability of e.g. a PID'd Alexia (0.75L boiler) is reputedly excellent.
luca wrote:To this end, it's possibly a little ungenerous to describe the LM and Synesso brew boilers as "oddly shaped" rather than acknowledging the possibility that the shape is a result of some conscious design and engineering.
No idea who you're quoting, but it certainly wasn't me. I was trying to point out a simple consequence of boiler geometry and thermodynamics: when you increase the size of a standard cylindrical boiler, the volume increases more rapidly than the surface area. This means that a larger boiler will lose heat more slowly, and it will take longer to reach a new equilibrium at a lower temperature setting. To maintain the ratio of surface area to volume, you need a more elongated or irregular shape. Apparently LM and Synesso took this into account when designing their boilers. However, this is not true for all home DB offerings, especially those with twin steam and brew boilers.
John

User avatar
AndyS

#149: Post by AndyS »

RapidCoffee wrote:To maintain the ratio of surface area to volume, you need a more elongated or irregular shape. Apparently LM and Synesso took this into account when designing their boilers.
More likely that the long skinny boiler was a necessity for LM in order to create multigroup machines incorporating their "saturated group" design.
-AndyS
VST refractometer/filter basket beta tester, no financial interest in the company

User avatar
HB
Admin

#150: Post by HB »

Dan Kehn