E61 Thermal Analysis Questions

[bruce]

I completed a simple steady state thermal analysis of the E61 , which I posted on Lino's schematic thread.

The results show about a 10 degree F delta from the end of the assembly to the thermo siphon cavity. It also shows a small temperature difference from one side of the shower screen to the other. Does his mean that the water will be hotter on one side of the puck? Would it help to add insulation around the group assembly?
I was going to take the analysis one more step- by considering the transient case and the effects of the water flowing through the assembly.
I had some questions which haven't been answered, and would be helpful in completing the model:
What is the brass alloy used for the part?
Has anyone taken a thermocouple and mapped the surface to verify my steady state results? I would like to know if I am way off.
What is the flow rate through the thermosyphon cavity?
What is the diameter of the small duct which delivers the water to the screen?

[rfc]

I'm not an expert in thermodynamics, and I didn't even stay at a Holiday Inn Express, but I'm always happy to share my opinions.

First, with regard to the asymmetric distribution of the temperature left to right across the screen, my guess would be that is because the holes drilled to communicate the water from the "mushroom" down to the screen are drilled asymmetrically, presumably to get the path over to the "infusion device chamber" on the right side of the machine.

I would question the modelling however. From what I have read, the best shots come after the machine is fully heated and you're pulling shots one after another at a practical rate.

In that case, the temp. delta would probably be lower, but whether or not it is, might be beside the point, because it is the water that ultimately hits the coffee, not the E61 housing, so the water temp drop would probably not be as much and (given that it is starting out life at somewhat higher than 203 degrees), it will probably be at or near that (in ideal conditions).

Now, with regard to your other questions:

"Would it help to add insulation around the group assembly?"

It wouldn't help the look of the machine. There is something very tactile about the whole process of making coffee and it's not just about touching the machine (which admittedly is quite hot). But you do have to clean it, and if whatever insulation you chose was not cleanable, forget it. It can't delaminate either--that would be a big problem. Look at the space shuttle.

"Alloy of the brass"

Don't know, but it has a lot of copper in it. I doubt it would affect the model much. Try a few of the most common alloys and you'd probably be fine.

"Flow rate through the Thermosyphon loop"

I've wondered about this myself. I think it is fairly significant. If you remove the mushroom while the machine is on (not while pumping, obviously), you can see it flowing quite spiritedly. If you timed the fluid flowing the first 1/4" out the top tube, you could do the math on the I.D. of the tubing and come up with what might be a pretty good number.

"Diameter of the small duct delivering fluid to the screen"

My guess: 3 millimeters, but it intersects with the larger one above and behind that is more like 4 or 5.

One last piece of advice:

Before you keep crunching these numbers, I hope you've pulled yourself a couple of nice doubles to keep the brain going!

[HB]

Has anyone taken a thermocouple and mapped the surface to verify my steady state results?

Although you hear lots of discussion of "overheated" groupheads, even the hottest ones I've measured idle a couple degrees below brew temperature at the dispersion screen (e.g., 195F), which seems reasonably consistent with your diagram below. That is, I assume your model only takes into account the solid brass, so my delta temperature between the back "red" of the group and at the screen would be a couple degrees more (yours is 197.5-191.8 = delta of 5.7F; mine would be 7-8F, yielding an effective brew temperature of ~202F).

It's interesting that you ask about insulation. Please excuse my ignorance of temperature models, but intuitively I would think that an insulated group's temperature would be easier to control. Would a form-fitted insulation jacket around the group result in a more even distribution of temperatures in your static model, or less?

From E61 Group Espresso Machine: Detailed Interior Schematics and isotherm.blogspot.com

[bruce]

Rfc, HB,

Thanks for the reply. Yes, I really do need a couple of doubles to get going in the morning.
Some comments/clarification: I used a water temperature of 200F within the thermosyphon chamber, with the water being stagnant. This gives a 2.5 F difference between the water and the metal surface. Since the water isn't really stagnant, adding some flow could reduce the 2.5 F significantly. RFC: it sounds like you have taken your machine apart: can you venture a guess at what the fluid velocity was?

Regarding the brass: don't ask me why (I am not a metallurgist), but adding a small amount of zinc to copper reduces the thermal conductivity to a fraction of pure copper. Within the various types of brass, the thermal conductivity can vary plus or minus 50%. A pure copper part could reduce the delta of 7.5 F by 2/3 or more. Unfortunately pure copper is so soft it would probably bend out of shape the first time you torqued the filter into place.

Dan- thanks for confirming the numbers, this is just the sort if 'sanity check' I was asking for- if you had measured 180F, for example, it would tell me the model is way off.

We are on the same wavelength with the insulation. I was thinking that some type of cover over the cylindrical portion would reduce the temperature difference to a negligible amount. This section of the assembly is really acting like a thermal fin, by increasing the surface area of the group head exposed to the surrounding (cool) air. The machines that have the group head under a "hood" do not allow air to flow around the assembly, and are therefore a much better thermal design. Of course they are not as pretty as the E61. Maybe aesthetics has as much to do with its longevity as its function.

I think we could design a cover for the E61 that would be functional and aesthetically pleasing. A sculpted metal piece could be created to cover the existing assembly. A fabric glove between them would provide the required insulation.

[gscace]

I've embedded a bunch of stuff in your quoted text to keep it next to the relevant passages. Hope it's not too much trouble.

I completed a simple steady state thermal analysis of the E61 , which I posted on Lino's schematic thread. http://www.home-barista.com/forum...hematics-t397.html
The results show about a 10 degree F delta from the end of the assembly to the thermo siphon cavity. It also shows a small temperature difference from one side of the shower screen to the other. Does his mean that the water will be hotter on one side of the puck?

Not necessarily because the volume above the dispersion screen on the groups I've looked at provides enough conductance that the off-center water delivery hole has a negligible negative effect on the water dispersion.

Would it help to add insulation around the group assembly?

Definitely not. the group should be thought of as a heat exchanger. If the group is insulated then there won't be enough temperature difference to drive the convection loop and the thermosyphon won't work. Your model needs to consider rejecting the heat to the surrounding air. If you insulate the group, the heat transfer to the environment will be less and that means less water cooling. Since water density difference is what drives the thermosyphon loop, you can imagine that the convection cycle will become very sluggish if the density difference is small. In addition to heating the group, you are also cooling the water in the hx, which would otherwise rise in temperature to that of the surrounding water in the boiler. One of the nice things about the thermosyphon idea is that it provides some cooling for otherwise waaaaay overheated hx water when the machine is idling.

I was going to take the analysis one more step- by considering the transient case and the effects of the water flowing through the assembly.

It's an interesting exercise. You should consider cases where the group is hotter than the water in the hx. In this case the thermosyphon runs backward. You may also consider cases where the group and water in the hx are nearly the same temperature. In this case, the thermosyphon doesn't run at all.

I had some questions which haven't been answered, and would be helpful in completing the model:
What is the brass alloy used for the part?

Marine bronze. The stuff is pretty easy to machine, but grabs sharp tools with a vengeance.

Has anyone taken a thermocouple and mapped the surface to verify my steady state results?

My group has a thermocouple probe inserted into the water passageway right where the water exits into the dispersion screen. I've also measured temperature inside the large cavity, and while brewing, using a prototype thermofilter.

I would like to know if I am way off.

I think the 10 degree number from the inlet to the group and the water passage exit to the coffee is reasonably close - within a factor of 2 anyway. When my group is idling, I see about 199 degrees at my probe. This is considerably lower than the hx temp, of course. When I do my cooling flush, the water heats up to 206 degrees, which means that the water had to have been hotter, since it's losing heat to a relatively cooler group on the way to my probe. So I think the 10 degree delta is within reason.

What is the flow rate through the thermosyphon cavity?

Depends on the length of the tube between the hx and the group, the diameter of the tube and its wall thickness, the ID of the inlet fittings, and any shape factors required to deal with direction changes in the plumbing.



What is the diameter of the small duct which delivers the water to the screen?

On my Astra, the diameter is 3mm.


My Astra, from the factory ran too hot for boiler pressures that provided good steam (at or above 1 atm steam pressure when the valve was open). I lengthened the tube between the top of the hx to the group, which cut down the flow to the group. This dropped the temp of the group, which dropped my brew temps once the hx is adequately flushed. I reckon that the group temp when idling should be close to the desired brew temp since it's such a good temperature attenuator. I published some measurement results and my interpretation of them on alt.coffee last year. Do a Google groups search for "thermosyphon behavior" or something similar to find these. It's a pretty interesting topic. Dive into it a bit, and you'll realize that while the e-61 group has a pretty revered status, it's the wrong group to use in situations where you have the group temp and boiler temp close together, such as in the Brewtus or Chris's Coffee's new double boiler machine. But it sells machines.

-Greg

[malachi]

Do a Google groups search for "thermosyphon behavior" or something similar to find these. It's a pretty interesting topic. Dive into it a bit, and you'll realize that while the e-61 group has a pretty revered status, it's the wrong group to use in situations where you have the group temp and boiler temp close together, such as in the Brewtus or Chris's Coffee's new double boiler machine. But it sells machines.

Thank you.

[rfc]

I might be missing something, so correct me if I am wrong, but the group head needs to dissipate heat under normal circumstances, not get hotter (like it would if you insulated it).

Normally, you need to flush to get the temperature DOWN to the correct brew temperature. The temperature certainly does come down because as you flush, cooler water is coming into the middle of the HX, but at least part of it would be absorbed by the grouphead's temperature differential.

If the entire grouphead were 203.5 degrees, my guess is that flushing to get to the proper temp would take longer than it does with it in the high 190's.

[gscace]

I might be missing something, so correct me if I am wrong, but the group head needs to dissipate heat under normal circumstances, not get hotter (like it would if you insulated it).

Normally, you need to flush to get the temperature DOWN to the correct brew temperature. The temperature certainly does come down because as you flush, cooler water is coming into the middle of the HX, but at least part of it would be absorbed by the grouphead's temperature differential.

If the entire grouphead were 203.5 degrees, my guess is that flushing to get to the proper temp would take longer than it does with it in the high 190's.

This is correct. Also, the idling group temperature can affect the temperature profile during brewing and also the temperature from shot to shot during continuous duty conditions.

-Greg

[bruce]

It took me awhile to figure out what you guys have been talking about. I have a single boiler machine (Zaffiro). I was thinking that temperature of the group should be as close as possible to the water temperature in the boiler. It seems obvious that you would want to add insulation in this case. A completely different problem than the HX machine, where you want the group head to be cooler than the boiler. For an HX machine, adding cooling fins is the solution. Size them correctly and you will not have to flush so often. Now that would be ugly.

[bruce]

Was the E61 designed originally for an HX machine? Makes more sense than a double boiler machine.

[HB]

Was the E61 designed originally for an HX machine? Makes more sense than a double boiler machine.

The E61 patent ("ALTERNATELY SEATING VALVES") addresses the question of preinfusion and makes no specific mention of the boiler design. It's a safe bet though that it was an HX, as that was pretty much the norm those days as best represented by the Faema E61.

PS: Below is a couple pictures of Ken Nye's Faema E61 Legend from his shop at Ninth Street Espresso:

[Marshall]

Not necessarily because the volume above the dispersion screen on the groups I've looked at provides enough conductance that the off-center water delivery hole has a negligible negative effect on the water dispersion.

-Greg

This has recently become a subject of great interest to me, because I found a very different effect in my Zaffiro. Alt.coffee regulars may recall I had an ongoing problem with "split" pours. The front half would start first, beginning at the very front edge, and turn blonde well before the back half. I tried all sorts of cleanings, adjustments and tests, none of which made any difference.

In frustration, I took it over to Michael Teahan's shop. With some effort, Michael removed the dispersion screw, which I had thought might have a back-end blockage. It was clear. But I discovered for the first time that the E-61 water inlet was at the back of the chamber, not directly above, as I had assumed. We both surmised that the water was jetting forward and exiting the chamber with more force at its front.

Michael opening up the top of the head and found that Isomac was using a 1.0 mm gicleur, instead of the Faema 0.5 mm standard. After trying an intermediate size (0.7), we went with the 0.5. This was over a month ago.

It's hard to measure the improvement that one tweak makes, when my machine has had many others. But, this may rank with the PID installation. I am getting beautiful, even pours with a regularity I had not previously thought possible. This means I can fill my demitasse, and the cup greatly quality exceeds my previous premature short shots.

So, in sum, I think the off-center delivery hole can have a serious negative effect, if the water flow is not adjusted to allow for that fact.

Marshall

[gscace]

Not necessarily because the volume above the dispersion screen on the groups I've looked at provides enough conductance that the off-center water delivery hole has a negligible negative effect on the water dispersion.

-Greg

This has recently become a subject of great interest to me, because I found a very different effect in my Zaffiro. Alt.coffee regulars may recall I had an ongoing problem with "split" pours. The front half would start first, beginning at the very front edge, and turn blonde well before the back half. I tried all sorts of cleanings, adjustments and tests, none of which made any difference.

In frustration, I took it over to Michael Teahan's shop. With some effort, Michael removed the dispersion screw, which I had thought might have a back-end blockage. It was clear. But I discovered for the first time that the E-61 water inlet was at the back of the chamber, not directly above, as I had assumed. We both surmised that the water was jetting forward and exiting the chamber with more force at its front.

Michael opening up the top of the head and found that Isomac was using a 1.0 mm gicleur, instead of the Faema 0.5 mm standard. After trying an intermediate size (0.7), we went with the 0.5. This was over a month ago.

It's hard to measure the improvement that one tweak makes, when my machine has had many others. But, this may rank with the PID installation. I am getting beautiful, even pours with a regularity I had not previously thought possible. This means I can fill my demitasse, and the cup greatly quality exceeds my previous premature short shots.

So, in sum, I think the off-center delivery hole can have a serious negative effect, if the water flow is not adjusted to allow for that fact.

Marshall

The reduced gicleur size has the effect of increasing the conductance of the volume above the screen. This is a good solution provided that the water debit remains big enough.

-Greg

[Marshall]

The reduced gicleur size has the effect of increasing the conductance of the volume above the screen. This is a good solution provided that the water debit remains big enough.

-Greg

Just what I thought! What's "conductance?"

My water debit is a bit more than 60 ml.

Marshall

[gscace]

Just what I thought! What's "conductance?"

My water debit is a bit more than 60 ml.

Marshall

Conductance in this case is borrowed from vacuum system technology. Imagine that you have a small diameter pipe through which you are trying to cram lots of material (presume a liquid). In order for you to get this stuff through the pipe you have to put a lot of pressure on it upstream. As the liquid goes down the pipe, the pressure drops. If it comes out into the room, then at the end of the pipe the pressure is atmospheric, although to cram lots of liquid through, the pressure will be very high upstream of the exit. This sort of a system is qualitatively "low conductance." A high conductance system is just the opposite. In this case, very low pressure differential pushes a large amount of material. In an espresso machine group, low conductance areas include the tiny holes above the shower screen in the e-61 style groups I'm familiar with (the Astra has a brass diffuser block just above the screen). In order to flow lots of liquid through, the pressure drop along these small holes is relatively high. By contrast, the disc shaped chamber above the diffuser block has a lot of volume and flows the same amount of liquid through it at very low pressure drop and velocity. There's a quantitative value for conductance ( that I'd have to go look up since I don't use it daily). If the ratio of high conductance to low conductance in the group passages I just mentioned is sufficiently big, then the pressure across the entire high conductance volume will be constant. This means that the flow rate through the little holes in the diffuser will be constant throughout the diffuser, and the coffee will get wetted equally. Provided that the conductance ratio is favorable enough, any negative effect of the angled water passage above the diffuser block on water dispersion will be eliminated.

-Greg