Theory behind flushing an E61 HX machine

Beginner and pro baristas share tips and tricks.
User avatar
kenja

#1: Post by kenja »

I recently purchased an E61 machine and have read all the articles about getting the temperature right by flushing the grouphead. I still have a couple of theoretical questions about the process, though. Let me try to restate what I've read first.

By flushing you are running water out of the boiler through the heavy, thermal mass of the E61 grouphead to try to get it to the correct temperature for extraction. Once you finish flushing, you give it 15-30 seconds to "bounce back" to the desired brew temperature. At that point you are ready to pull your shot. Because the thermal mass of the head is so big, it does a good job at maintaining the temperature you set through the flushing process while the brewing is taking place.

If I didn't get that right, feel free to correct me. So here are my questions:

The boiler contains water that is above the temperature you want to brew at, right? It is at 1-1.2 bar which means it has a head of steam and is super-heated water, right? If that is true, why does flushing cool the head down instead of heating it up? Are you relying on cold water in the reservoir being pumped into the boiler to cool it down?

Once that is cleared up, what is the process that causes the grouphead to heat back up once it is flushed? The water in the boiler must start to reheat, but there must be some heat transfer from the boiler to the grouphead. Is the water from the heat exchanger cycling through the system even when the pump is off? How closely do the grouphead and the boiler temperatures track one another during this process?

After the 30 second wait between flushing and pulling the shot, the grouphead temperature seems to pop back up by 15 degrees or so. At that point you start running brewing water out of the boiler and through the head for pulling the shot. If the grouphead temperature rises and falls by 15-25 degrees during the flushing and waiting process, how can you expect the temperature to be maintained to 1-2 degrees during a 25-30 second shot?

It doesn't add up for me, so hopefully someone can shed some light on this. If I understood the process a little better, I feel like I may be able to manage it better.

I just bought Eric's thermometer attachment, so maybe that will help me figure it out, too.

Thanks,
Kenja

User avatar
Compass Coffee
Sponsor

#2: Post by Compass Coffee »

Brew water doesn't come from the boiler on an HX machine. The HX (heat exchanger) is a separate water line which runs through and is heated by the boiler. Hence while idle the HX line in the boiler will rise to over heated temperature and running incoming water through the HX line brings temp down to production operating temperatures.
Mike McGinness, Head Bean (Owner/Roast Master)
http://www.CompassCoffeeRoasting.com

User avatar
jesawdy

#3: Post by jesawdy »

Some recent discussion about this in What is the purpose of long HX flushes?
Jeff Sawdy

User avatar
erics
Supporter ◈

#4: Post by erics »

kenja wrote:I recently purchased an E61 machine and have read all the articles about getting the temperature right by flushing the grouphead.
The E-61 machine you have is, AFAIK, essentially identical to the Vibiemme Super Domobar so a little additional reading on that machine's fine review may shed new light. One exception is that the Giotto Premium may not have a thermosyphon restrictor.
By flushing you are running water out of the boiler through the heavy, thermal mass of the E61 grouphead to try to get it to the correct temperature for extraction. Once you finish flushing, you give it 15-30 seconds to "bounce back" to the desired brew temperature. At that point you are ready to pull your shot.
As Mike said, when you flush, you are "flushing" water out of the heat exchanger. See the diagram of the Quickmill machines below. Water flows out of the injector tube and out through both ends of the hx towards the grouphead. The water at the bottom of the hx is ABOUT 15 degrees cooler than the water at the top of the hx prior to any flushing. The percentage of "bottom" water that gets fed to the group compared to the amount of "top" water is fixed by the length of the injector tube and is typically fine-tuned by trial & error after a lengthy session with a computational fluid dynamics guru. I would regard this process as very complex as it pretty much sets the tone of the particular hx machine.

Image

The AVERAGE temperature of the water in the thermosyphon loop (which certainly includes the hx) is about 215 F but would surely vary with the specific machine, its internal cleanliness, and the pstat setting. Flushing easily lowers this temperature but, of course, the group takes a little more effort.

After a, say, 8 ounce flush, the water in the hx is much cooler than the grouphead and so a reverse thermosyphon action action takes place whereby heat is transferred from the group to the thermosyphon flow. This goes on until the temperature of the water in the hx, being heated by the water & condensing steam in the boiler, rises above that of the grouphead.
I just bought Eric's thermometer attachment, so maybe that will help me figure it out, too.
So much depends upon your modus operandi. The thermometer is intended to provide a greater degree of consistency. There's probably more info on the FTP site in the adaptor tidbits document - http://users.rcn.com/erics/
Skål,

Eric S.
http://users.rcn.com/erics/
E-mail: erics at rcn dot com

User avatar
cannonfodder
Team HB

#5: Post by cannonfodder »

Wonderful description Eric, and nice diagram.

I use the group thermometer more as a reference point than an absolute. Do your flush, observe at what temperature you stop the flush. Make note of your rebound time (the time between when you stop the flush and start the extraction) and the taste of the espresso. If it is bitter, flush another degree or rebound a few seconds less, if the shot is sour stop the flush at a higher temperature or extend the rebound. Personally, I would work with 30 second intervals for the rebound. Keep in mind that the group thermometer will read around 5 degrees F higher than the water hitting the coffee puck.

Image
Dave Stephens

berndb

#6: Post by berndb »

erics wrote:The water at the bottom of the hx is ABOUT 15 degrees cooler than the water at the top of the hx prior to any flushing. The percentage of "bottom" water that gets fed to the group compared to the amount of "top" water is fixed by the length of the injector tube and is typically fine-tuned by trial & error after a lengthy session with a computational fluid dynamics guru. I would regard this process as very complex as it pretty much sets the tone of the particular hx machine.
Eric, here I come with some questions:

Do you definitely know, that the water at the bottom of the H/X is 15 (F or C?) colder than on the top?

Is it really true, that the design of the thermosyphon is based on any results of computational fluid dynamics. Given the E61 Legend's simple adjustable flow restrictor (which does not just change the idle temperature of the group but the temperature of the brewing water as well) there might be a lot trial and error but I can't see an indication for serious science there.

Why does the length of the injector changes the percentage of top and bottom water that gets to the group. Up to now I did believe that it changes the temperatures of top and bottom water differently, but the mix of top and bottom water in the brew head should just depend on the (dynamical) flow characteristics of the top and bottom part of the H/X.

Regards Bernd

User avatar
erics
Supporter ◈

#7: Post by erics »

Good Morning Bernd -
Do you definitely know, that the water at the bottom of the H/X ist 15 (F or C?) colder than on the top?
I don't definitely know it, i.e. even though I have stick on Type T thermocouples still in the box. I was taking my data from a posting by CafeIke (Ian) here: Array/espresso-machines/what-is-purpose-of-long-hx-flushes-t4737-20.html#51817. In addition, if you do a google search on alt.coffee using "thermosyphon" as the key word, you will run into his write-up on a unique thermosyphon stall problem he had with his Vibiemme and he provided temp graphs at the upper and lower tubes immediately where they enter/exit the grouphead. At his point of measurement, you could, maybe add a touch to his upper temp and subtract a little from his lower temp to get the delta T across the hx. I trust his data.
Is it really true, that the design of the thermosiphon is based on any results of computantional fluid dynamics. Given the E61 Legend's simple adjustable flow restrictor (which does not just changes the idle temperature of the group but the the temperature of the brewing water as well) there might be a lot trial and error but I can't see an indication for serious science there)
My most truthful answer is "I don't know." However, given the profileration of the E-61 grouphead, it would surprise me if there did not exist some pretty hard core studies of its heat transfer and fluid dynamics characteristics. I can also surmise that these studies are not in the public domain. I don't consider the injector as being a part of the thermosyphon system (because it is not active when the thermosyphon system is active and vice-versa) although I could see the argument that it is a part of the thermosyphon system. I don't personally know how well the E-61 Legend's adjustable flow restrictor works - there has been, AFAIK, no published info on thermal performance. I bought one to install in Anita but I'm waiting on a rainy day because Anita has some rather tight quarters in that area. A calculation I once did showed me that the thermosyphon flowrate was "in the same ballpark" as the AVERAGE brew flowrate and, based on that, I agree with you that the restrictor also changes the temperature of the water mix entering the grouphead.
Why does the lenght of the injector changes the percentage of top and bottom water that gets to the group. Up to now I did belive that it changes the temperatures of top an bottom water differently, but the mix of top and bottom water in the brew head should just depend on the (dynamical) flow characteristics of the top and bottom part of the H/X.
Let's say the injector has a 3" length, just for discussion purposes. If you changed that to 2.875" or 3.125", I think you might be pressed to detect some brew water temp change. However, if you extended it to, say, 6" or reduced it to zero, I believe you could easily measure the difference. Here the size of the injector (inside diameter) would affect the fluid velocity and thus the dispersion within the hx. The relative amount of water (inertia), above & below the injector tip would govern the mix to the grouphead, in addition to the diameter and effective length of the two tubes. There are many factors, as I believe you know, and that is all the more reason why a model of the heat/mass transfer within these machines is constructed.
Skål,

Eric S.
http://users.rcn.com/erics/
E-mail: erics at rcn dot com

User avatar
cafeIKE

#8: Post by cafeIKE »

Group Inlet and Outlet Temperatures on Vibiemme Domobar Super w pstat
K Thermocouples wedged into junction of group and thermosyphon tube
Image

User avatar
jesawdy

#9: Post by jesawdy »

cafeIKE wrote:Group Inlet and Outlet Temperatures on Vibiemme Domobar Super w pstat
K Thermocouples wedged into junction of group and thermosyphon tube
Thanks, Ian. That's a pretty graph... we are looking at power on for 1 hour 2 minutes on the x-axis there?
Jeff Sawdy

User avatar
cafeIKE

#10: Post by cafeIKE » replying to jesawdy »

Logging started when the upper thermosyphon reached 200°F, continue for 01:03:50 @ 10s intervals.