What is the purpose of long HX flushes?

[RapidCoffee]

A recent thread brought back a question that's been nagging me for some time. Why do HX machines require long flushes? That's one of the most common complaints about HX machines. You commonly hear of people flushing as much as 8oz/250ml. This doesn't make sense to me.

Assumptions:
1) Most home HX machines have a heat exchanger volume of about 100ml (3.5oz). There are exceptions, (e.g., Cimbali), but I believe 100ml is fairly common.
2) The primary purpose of the HX flush in an idling machine is to eliminate overheated water in the heat exchanger.

If these assumptions are correct, then why would a large flush be of any benefit? If the HX volume is 3oz, and you flush 3oz and replace it with fresh water, what does a larger flush accomplish? At some point you'll start to lower the boiler temp, but that's certainly not the purpose of the flush.

Perhaps an E61 machine with an overactive thermosyphon needs a larger flush to lower grouphead temp. But I thought the grouphead idled cooler than thermosyphon temp, not hotter, due to radiative heat loss.

FWIW, I do not find large flushes to be of benefit on my equipment. On my Vetrano, with the boiler pressure set to 1.3 bar (raised slightly to compensate for 4000' elevation - tip of the hat to Eric Svendson for pointing this out), a 3oz flush appears to be perfectly adequate.

I thought I understood heat exchangers pretty well, but apparently I'm still missing something. Maybe my guesstimate on HX volume is off. Any thoughts?

[jesawdy]

Perhaps an E61 machine with an overactive thermosyphon needs a larger flush to lower grouphead temp. But I thought the grouphead idled cooler than thermosyphon temp, not hotter, due to radiative heat loss.

John, what does your grouphead idle at? I thought you had a thermometer adapter now? If 3 ounce flushes are working for you, I would suspect that your grouphead idles very close to desired temp.

[Psyd]

I thought I understood heat exchangers pretty well, but apparently I'm still missing something. Maybe my guesstimate on HX volume is off. Any thoughts?

My HX volume is 250 ml each, or nearly eight and a half ounces. I pretty much fill my cup and then lock in, pull the shot after drying the cup, and end up at 196 - 200. If I want to go cooler, I flush a little longer and I can get a cup ready to go right after the flush, and I can get it somewhere in the range of 191 - 196, hotter is easier, just take three seconds and I'm in the 200 - 208 range. Yeah, I'm guessing that you're comparing oranges to apples a bit.

[jgriff]

Assumptions:
1) Most home HX machines have a heat exchanger volume of about 100ml (3.5oz). There are exceptions, (e.g., Cimbali), but I believe 100ml is fairly common.
2) The primary purpose of the HX flush in an idling machine is to eliminate overheated water in the heat exchanger.

If these assumptions are correct, then why would a large flush be of any benefit? If the HX volume is 3oz, and you flush 3oz and replace it with fresh water, what does a larger flush accomplish? At some point you'll start to lower the boiler temp, but that's certainly not the purpose of the flush.

Perhaps an E61 machine with an overactive thermosyphon needs a larger flush to lower grouphead temp. But I thought the grouphead idled cooler than thermosyphon temp, not hotter, due to radiative heat loss.

I have an Anita with Eric's thermometer adapter installed, and I definitely flush more than 3oz. Depending on how long the machine has been on, the GH temp at idle is usually around 212, but will sometimes creep as high as 216 if it's been sitting for a few hours.

I think you're right in assuming that the HX is only about 100ml of overheated water, but that's at the same temp as the rest of the boiler, 250-ish right? So I flush that out which is also adding heat to the brewpath and GH. Then I keep flushing to cool off the overheated GH so I don't end up with a shot around 212. Works for me, anyway. My normal technique is using flush and rebound, so obviously I'm giving the water in the HX some time to heat back up. I guess if I didn't do that the hotter GH temp wouldn't heat the water too much.

EDITED: So it seems you're right about the GH idling cooler than the thermosyphon, but the TS loop is close to the temp in the boiler, right?

[Psyd]

Weird, it's never done that before...

[TimEggers]

I've found that on my Anita I like a 5-ounce flush give or take a half-ounce (boiler @ 1.3 bar for steaming performance). I also like to flush lock and pull. The espresso is darn hot with good chocolate, caramel and raisin notes (this weeks home-roasted Liquid Amber from Sweet Maria's).

I wonder what the volume differences are between flush lock and pull, and the rebound methods are? Sorry if that has been discussed before.

[julioale]

My Levetta needs more than 8oz of cooling flush. The group temp is around 216 before 30 minutes of warm up (according to the Eric's digital adapter) and the boiler pressure set to .95.

3 oz of cooling flush, would be great.

Ale

[erics]

I am grateful that it is 11:45 PM EST because this is one of my favorite espresso subjects. Certainly there are several ways in which an hx espresso machine can be operated and even slight variations within those methods. On one end of the scale, it is entirely possible to pull a very good shot with absolutely no flushing whatsoever. This feat was first brought up by Ian Eales (CafeIke) using his PID'ed Vibiemme with his temperature controller set such that boiler pressure was around 0.70 bar. I duplicated this with Anita and provided a thermofilter shot with my pstat turned down to 0.70 bar (max reading). Steaming performance was merely "OK". Machine recovery after a shot is, well, not great. Machine recovery with the Vibiemme would be better than Anita.



The above graph has been previously posted but it serves to vividly demonstrate what different settings can do with respect to recovery time.

Now, here's the other end of the scale. Personally, I am a proponent of large flushes (~ 8 oz) to cool the grouphead down and then pull the shot as the grouphead is climbing back up to temp. I find that this provides for greater consistency in the espresso shot and I am able to easily adjust the temperatures if I happen to be using a bean with a known affinity for something either higher or lower than my norm (~200 F). I also start grinding and weighing and spooning in my WDT "clumpless " grind at about the same time that I start flushing. It takes me an inordinate amount of time to "prepare the puck" - especially at 7:00 AM but the timing always seems to work out pretty well.

This above chart is also available in the adaptor tidbits document on my "poor man's" website:

http://users.rcn.com/erics/

[RapidCoffee]

John, what does your grouphead idle at? I thought you had a thermometer adapter now? If 3 ounce flushes are working for you, I would suspect that your grouphead idles very close to desired temp.

Tonight my grouphead is idling at 209.5F. Boiling point of water is ~205F at my elevation (depends on barometric pressure). HX flush spikes at ~214F and typically settles down (quits flashing) at ~206F on the thermometer adaptor.

My HX volume is 250 ml each, or nearly eight and a half ounces.

This is your 2-group Astoria, right? I would expect a 250ml heat exchanger to require a 250ml flush on an idling machine.

EDITED: So it seems you're right about the GH idling cooler than the thermosyphon, but the TS loop is close to the temp in the boiler, right?

That's my understanding. When idling, the grouphead is continually radiating heat and cooling the thermosyphon. Thermosyphon temp should be hottest as it exits the boiler and coolest as it reenters the boiler.

Certainly there are several ways in which an hx espresso machine can be operated and even slight variations within those methods. On one end of the scale, it is entirely possible to pull a very good shot with absolutely no flushing whatsoever.
...
Now, here's the other end of the scale. Personally, I am a proponent of large flushes (~ 8 oz) to cool the grouphead down and then pull the shot as the grouphead is climbing back up to temp.

Hi Eric, and thanks for chiming in so late at night. I agree, there are several different ways to operate an HX machine effectively. But that doesn't answer my question: why are long cooling flushes seemingly required on some machines? (Expobar owners, in particular, complain that their machines run hot and require long cooling flushes.)

Let's ignore temp stability issues for the moment, and concentrate on a flush & go scenario. After flushing overheated water from an idling HX, shouldn't you be able to pull a shot immediately, without burning the grounds? On machines that seem to require a longer flush, does the initial flush overheat the grouphead, and the extended flush cool it down?

Still searching for enlightenment...

[RegulatorJohnson]

you answered your own question...

Perhaps an E61 machine with an overactive thermosyphon

(Expobar owners, in particular, complain that their machines run hot and require long cooling flushes.)

i probably am unique because i have actually owned both the pulser and the vetrano. in fact i have had both machines running together at the same time. (vibe vs rotary shootout anyone? ill pull the pulser from the basement and set it up before i sell it.)

erics TC has been in both group heads using the same probe and the same reader.

i have used the pulser with and without the thermosyphon restrictor. with it installed and with the pstats running 1.1-1.3 bar. i can flush ~3 ounces from idle and get within brewing temp. without the restrictor it would idle much higher than without the restrictor. i had it working really well was a "flush and go" type of system. i flushed very little water to get to the brewing temp of ~200° .

then i got the vetrano and everything changed. its seems to work well for the "flush and wait". i still only flush very little water. about 3 ounces even from idle. i am at about 4500 feet above sea level in SLC, UT. the vetrano runs between .9- 1.2 bar it idles at about 210.

i have noticed that with either machine that more than a few ounces will be too much especially if you flush until it stalls. i still dont get the stall. i try to avoid it.

it is late here so i don't remember the details of everything but i have posted on the subject on HB and EOG.

thanks for the time.
jon

[erics]

Assumptions:
1) Most home HX machines have a heat exchanger volume of about 100ml (3.5oz). There are exceptions, (e.g., Cimbali), but I believe 100ml is fairly common.
2) The primary purpose of the HX flush in an idling machine is to eliminate overheated water in the heat exchanger.

John -

I agree with your assumption of hx size and would add that an equally important purpose of a flush is to drive the average temperature of the grouphead towards the brew temperature range or value desired. This would be valid for pulling a shot after a long term idle.

I believe it is relatively easy to explain why one "Machine A" requires a longer cooling flush than another "Machine A". This can be attributed to different pstat settings or different locations (altitude) or both. The largest difference would be when the high altitude machine also has the lower pstat setting. In addition to this, I recently measured the boiler pressure on Anita with a 4.5" mechanical gage having a specified accuracy of 1/2% of full scale. There was a disparity between the gages of about 8% (the gage on the machine was reading ~8% higher than the purchased gage). All of the foregoing are quantifiable parameters whereas something like machine cleanliness is difficult to quantify and hence I make no attempt there. Imagine what the situation would be if all of these differences were acting in the same direction?

Now lets take the case of "Machine A" and "Machine B" (different manufacturers). Temperature measurement locations, although similar, are sufficiently different (geometrically) to report different temperatures when, in fact, the overall thermal energy of the two groups is not that far apart. Then you could look at my specific case wherein I am flushing through an empty PF which is always inserted into the machine and that adds a not so inconsequential thermal mass to the grouphead. In the case of the Expobar, design philosophy, in addition to grouphead geometry, plays a role. It APPEARS as though the Expobar is optimized for a speedy recovery (let's have a party ) and that most users pull shots at intervals greater than that originally contemplated by the mfg.

[RapidCoffee]

Some HX machines seem to require flushes that are significantly larger than the HX volume. The Expobar appears to be a prototypical example of this phenomenon:

I used a friend's Expobar Lever for a couple months; you aren't kidding - it requires an olympic-sized flush!

No joke here. With the top of my deadband somewhere around 1.14bar, I have to flush 8oz and immediately lock in (or within about 10 sec) for any beans darker than City roast. If I do a few pseudo-shots as I prep my basket, I can sometimes lower the flush volume to 6oz, but this really works better with the lighter coffees.

Could this be due to an overactive thermosyphon that overheats the group? Jon's results support this theory:

i have used the pulser with and without the thermosyphon restrictor. with it installed and with the pstats running 1.1-1.3 bar. i can flush ~3 ounces from idle and get within brewing temp. without the restrictor it would idle much higher than without the restrictor. i had it working really well was a "flush and go" type of system. i flushed very little water to get to the brewing temp of ~200° .

then i got the vetrano and everything changed. its seems to work well for the "flush and wait". i still only flush very little water. about 3 ounces even from idle. i am at about 4500 feet above sea level in SLC, UT. the vetrano runs between .9- 1.2 bar it idles at about 210.

Sounds reasonable, except for the following:

In all the HXs that I've measured, the temperature beneath the dispersion screen is in the 180-195F range. This is shown nicely in E61 Thermal Analysis Questions. I've excerpted my earlier reply below:

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).

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

According to this, the E61 grouphead runs cool, not hot. The HX flush warms the group, sure, but enough for a quick 3oz flush to overheat a 9# grouphead? I dunno, maybe it does - and the heck with E61 thermal stability.

Thanks for the responses. Still feeling a bit confused, but what else is new...

[RegulatorJohnson]

i have been shown this picture by a few people.

in my actual experience this is about 15° off.

if i dont use the machine for 4 hours then turn on the TC and read the temp 210-212° degrees. can i not assume that where the TC is placed (erics TC adaptor) that this picture is incorrect. the yellow band would need to be ~210°. so i still feel the same, considering my actual experience.

this graph could be replicated with a lower p-stat setting or a smaller restrictor.

i dont think that you can just point to a chart and say, this applies to any e-61 at any p-stat setting on any machine everywhere.

i also have a infra-red laser thermometer i can point it at the group at various places. it shows me that this graph is too low. at least for me, on my pulser and on my vetrano, in my kitchen, at my altitude, at whatever the barometer says, and whatever the moon phase was.

YMMV.

jon

[RapidCoffee]

i have been shown this picture by a few people.

in my actual experience this is about 15° off.

if i dont use the machine for 4 hours then turn on the TC and read the temp 210-212° degrees. can i not assume that where the TC is placed (erics TC adaptor) that this picture is incorrect. the yellow band would need to be ~210°. so i still feel the same, considering my actual experience.

this graph could be replicated with a lower p-stat setting or a smaller restrictor.

i dont think that you can just point to a chart and say, this applies to any e-61 at any p-stat setting on any machine everywhere.

i also have a infra-red laser thermometer i can point it at the group at various places. it shows me that this graph is too low. at least for me, on my pulser and on my vetrano, in my kitchen, at my altitude, at whatever the barometer says, and whatever the moon phase was.

YMMV.

jon

Well, that would help explain the long flushes. The thermocouple/thermometer is reading the thermosyphon temp, which I'd expect to be hotter than the idling grouphead temp in an E61. Remember, the grouphead is radiating heat and cooling the thermosyphon. But your IR thermometer reading is interesting. What sort of temps do you get? Any chance that someone with a stock Expobar (or other machine that runs "hot") could report back on this?

Thanks - John

[RegulatorJohnson]

i dont know if you can see this or not...

the infrared temp was switching between 211 & 212° and i had the infrared pointed just below erics TC.

the top number is the infrared reading.

the bottom number is the probe inside.

the machine has been idle for about 5 hours.



jon

[Psyd]

My HX volume is 250 ml each, or nearly eight and a half ounces.

This is your 2-group Astoria, right? I would expect a 250ml heat exchanger to require a 250ml flush on an idling machine.

At least, yeah. For quite a few blends the dance is rather complicated, flushing up to twelve ounces to get the thing back to where I want it by the time the cup is dry and the timer starts!

[jesawdy]

The thermocouple/thermometer is reading the thermosyphon temp, which I'd expect to be hotter than the idling grouphead temp in an E61.

No, at idle Eric's TC is reading the grouphead temperature. The thermosyphon is circulating through the lower part of the "mushroom" area.

Eric's annotated picture from E61 Group Espresso Machine: Detailed Interior Schematics, original image copyright Verna Design, Inc.

[jgriff]

I'm surprised that someone out there with access to nifty hi-tech tools and better math skills than most hasn't already analyzed this. How hard would it be to measure the temp of water in the thermosyphon as well as the GH temp and come up with some sort of calculation regarding heat loss? Then couldn't you also use some of that data and some known facts about specific heat and water flow, etc. to figure out just how the physics of the whole thing works?

Personally, I suspect that most unrestricted e61 groups idle higher than desired brew temp and that water from the HX first adds a little heat to the group and subsequently cools it back down as the overheated water is exhausted while flushing. Between the cooler water that runs through the group while the flush is taking place and the stalled thermosyphon action (when the HX water is cooler than the group temp) you get the combined effect of bringing the group to proper temperature. It's interesting to note that (at least on my machine) the temperature I see on Eric's thermometer adapter starts out showing the idle GH temp, then jumps up to a higher number when I start the flush, comes back down to around the idle temp and stays there for a bit, then starts rapidly declining until I stop the flush, it then bounces back up to some lower (than idle) number for a second, then gradually comes back down for the next 45 sec - 1 min and bottoms out, holding there for at least 30 seconds or so before it starts climbing again.

Another thing I find interesting is that I assumed the Quickmill machines would all behave the same, but maybe with the Vetrano or machines newer than mine they've modified a bit so it doesn't run as hot(?). Of course, all of this is just for fun as I'm happy with the shots I get using my flush regimen and I'm sure John, Jon, Jeff, Eric, Dan and all the rest are as well, no matter how dissimilar our techniques might be.

(I have to give a lot of credit to Eric for his education via e-mail about the temps the thermometer reads and GH temperature behavior.)

Justin

[erics]

The ORIGINAL intent of the thermocouple adaptor (and also digital thermometer adaptor) was to measure the temperature of the water as it approached the puck.

As knowledge was gained via this forum's contributors and thinking caps properly adorned, it became (to me) obvious. A temperature measuring device at or slightly below the recommended depth of the thermocouple is measuring a temperature (at idle) that is VERY REPRESENTATIVE of mean grouphead temperature.

[RapidCoffee]

No, at idle Eric's TC is reading the grouphead temperature.

A temperature measuring device at or slightly below the recommended depth of the thermocouple is measuring a temperature (at idle) that is VERY REPRESENTATIVE of mean grouphead temperature.

Oops, of course you're correct. Pardon the brain f@rt.

Jon's temp measurements are in good agreement (but are at odds with the diagram, and what Dan reported). It would be interesting to get temp measurements from a "long flush" E61 machine (like a stock Expobar). But this goes a long way towards explaining things. Thanks!

Personally, I suspect that most unrestricted e61 groups idle higher than desired brew temp and that water from the HX first adds a little heat to the group and subsequently cools it back down as the overheated water is exhausted while flushing. Between the cooler water that runs through the group while the flush is taking place and the stalled thermosyphon action (when the HX water is cooler than the group temp) you get the combined effect of bringing the group to proper temperature.
...
Another thing I find interesting is that I assumed the Quickmill machines would all behave the same, but maybe with the Vetrano or machines newer than mine they've modified a bit so it doesn't run as hot(?). Of course, all of this is just for fun as I'm happy with the shots I get using my flush regimen and I'm sure John, Jon, Jeff, Eric, Dan and all the rest are as well, no matter how dissimilar our techniques might be.

I'm becoming convinced this is the correct explanation. Some speculation: flush differences in otherwise-similar QuickMill machines could be due to the higher flow rate of the Vetrano's rotary pump. For a given volume flush, there would be less time for heat transfer to the grouphead.