A guide to managing HX brew temperatures

Beginner and pro baristas share tips and tricks.
Howardsmith

#1: Post by Howardsmith »

A guide to managing e61 Heat Exchnager machine brew temperatures

A link to a youtube video showing this process is below, I would suggest watching the video (may be best on a computer screen) & also reading.
To begin with I would like to say that managing HX (Heat Exchanger) brew temp is a minefield! Not only because of the technical & design difference between machines but also because there is a mass of information out there, some good, some not so good! Eric Svendson, creator of the "Erics Grouphead Thermometer" did a fantastic job in building the GH thermometer & has also posted a lot of useful info including two methods to manage brew temp named "Flush & Go" & "Flush & Wait". Before we continue with my method & the testing I think it would only be rite to give a quick overview of these two methods to help build the picture, once you understand what is going on here things become easier to take on board.



Flush & Go (F&G)

With the Flush & Go method we flush a sufficient amount of water out of the HX to cool the water inside the HX to below target brew temp. The GH (Grouphead) remains above target brew temp.

We then pull the shot immediately (or with a small delay). 'Too Cool' water exits the HX & is heated by the 'Too Hot' group. This results in some heat transfer & we have our brew water.

With the F&G method GH thermometer readouts will appear below actual brew temps during the shot. This is because the thermometer is an inch or so behind the puck & the 'Too Hot' group still has time to warm the water on its way past the thermometer to the puck.

An example of F&G could be as follows.


Remove PF (portafilter) & prep puck

Flush machine until temp starts dropping

Stop flush at 96°C Lock PF in & GO



During the shot the GH thermometer will read below actual brew water temp at the puck (this will vary machine to machine but somewhere around 1-3°C would be a ballpark figure)

The method above may give somewhere are a 93°C temp at the puck




Flush & Wait (F&W)

With the Flush & Wait method we flush a sufficient amount of water to cool the mass of the GH below target brew temp. The puck would them be prepped & the shot initiated at after 'waiting' until the GH thermometer hits a pre determined point.

With this method the GH is cooler than target brew temp. The water in the HX has likely recovered to full temp (there or there about, this is really not critical as you will see moving forward). On pulling the shot 'Too Hot' water exits the HX & is cooled by the 'Too cool' group. This results in some heat transfer & we have our brew water.

With the F&W method GH thermometer readouts will appear above actual brew temps during the shot. This is because the thermometer is an inch or so behind the puck & the 'cool' group still has time to cool the water on its way past the thermometer to the puck.

An example of F&W could be as follows.


Flush machine until GH thermometer drops to 93°C

Group will continue to drop

As soon as group begins to recover & temp starts rising remove PF & prep puck

Lock PF back & 'Wait'

Wait until GH thermometer hits 90.5°C & pull shot.



During the shot the GH thermometer will read above actual brew water temp at the puck (this will vary machine to machine but somewhere around 1-3°C would be a ballpark figure)



Both methods will get you where you wan to be. F&W is a little easier as we have nailed down a few of the variables. HX water is going to be about the same temp & GH is at whatever we decide. F&G is a little more tricky. Once the group starts dropping temp as we flush the thermometer is changing rapidly, a slight delay may mean a target flush to 96°C would quite quickly become 96.5 or 95. Not the end of the world but we can do better.


Now that I have given a general overview to the two most commonly used methods out there I want to say that both of these methods work. Both get you where you want to be but my biggest concern is the amount of water that is required to implement them. I know a lot of people have plumbed in machines however I would say a lot of people live in a hard water area & don't want the expense of filtrations systems or simply cannot be bothered with it. Also consider people like me who have a tank only machine. Once fully warmed up my Rocket Appartamento will take between a 150-200ml cooling flush. Follow that with 50ml for the shot & maybe a 30ml screen flush & I get less than 10 shots to a tank & multiple drip tray 'dumps'.

My frustration with misinformation, lack of support from retailers/manufacturers, long cooling flushes, constantly having to dump the drip tray & recycling bins full of empty water bottles forced me to look into how to get around these issues myself.

This is a guide that I have built for myself. It compromises of charts, videos & general discussion around what exactly is going on with a e61 HX machine but most importantly it includes an easily repeatable process to hit whatever brew temp you like at whatever boiler pressure you like, shot after shot without flushing. Please consider this is MY routine that works for me & with my machine. In general e61 HX machine will perform in a similar manner as the engineering is similar, however there are many factors that will change between machines so please see this guide as an example rather than the rule.



What will this guide show me?


Two sets of six 30 second shots pulled in a b2b (back to back) manner at a 'reasonable' workflow. Both sets took between 30 & 40 mins to record.

One set at 1.1 bar max & one set at 1.5 bar max.

You will see GH temp & 'at Puck' temp during all extractions (in the video) & data will be logged into graphs for reference & review.

You will see a method to produce "hot" "medium" & "cool" shots at both 1.1 & 1.5 bar with no flushing (other than screen flushing post shot) or changes to Pstat/PID during the sequence of extractions.

You will see some attempts to produce a "flat profile" shot where I am trying to eliminate the "HX Hump" at the beginning of the shot that this method produces.

You will also see a comparison of 1.1 & 1.5 bar shots pulled in the same manner producing almost identical brew temp profiles! A chart is included with 1.1 & 1.5 together.




The method

Before we talk about the method there are a few basics to understand. Understanding the basics will hep make sense of the process.

An e61 Grouphead is pretty much a solid block of brass & usually weighs around 4kg.

The amount of water required to produce a double shot of espresso would usually be between 30 & 100g. This is taking into account 'puck absorption' & filling any pre-infusion chambers & is somewhat of a 'stab in the dark' but the point I am trying to convey is that it is much less than the weight of the group.

Considering the water during a 'normal' extraction flows slowly there is plenty of contact time between the water & the GH. The GH does an excellent job of tempering the water & bringing it up or down as required.

This suggests that controlling GH temperature plays a large roll in the actual temperature at the puck. It is my belief that controlling GH temp is WAY more important than HX water temp. This will be demonstrated here & the results are shocking.

The method I have been using is to cool the GH with a small battery operated fan. This does not affect HX water temp much although the return will be a little cooler but not massively relevant to the outcome.



Position a small fan so that it rests on the top cap above the mushroom & blows directly down over the group

Cool the group to just above a pre-determined temp (group will continue to drop)

Pull the PF & prep the shot

Lock the PF in & wait for the GH to rebound to a pre-determined temp

Pull the shot at the pre-determined temp




With this method GH thermometer readouts will track above actual water temp at the puck. This is the same as with the Flush & Wait method as the GH is cooler than target brew temp & HX water is hotter than target brew temp.

The best part with this method is that we can fan the group to different temperatures & change actual brew temps shot by shot if we like. Boiler pressure/ temp appears to not drastically affect actual brew temp at the puck with this method.

This method has proven EXTREMELY accurate & repeatable shot to shot. I originally set out to try & produce a "hot", "medium" & "cool" shot but my results show we can target within 1°C fairly easily & even go between temps from shot to shot at any reasonable Pstat/PID setting.


Finally some numbers...


Hot Shot - Fan to 93°C - Pull Shot on rebound at 92.5°C - At puck temp 95°C

Medium Shot - Fan to 91°C - Pull Shot on rebound at 90.5°C - At puck temp 93°C

Hot Shot - Fan to 89°C - Pull Shot on rebound at 88.5°C - At puck temp 91°C




I have inserted the data referenced for the six shots at 1.1 bar & 1.5 bar. These shots were pulled over roughly a 40 min period over the course of two nights (one night at 1.1 & the other at 1.5 bar). A screen flush was performed after all shots & the PF was knocked out, dunked in hot water & put back in the group between shots unless otherwise stated.





Here are a few more graphs to demonstrate some of the findings.

Firstly we have a comparison between shot 1 and 4 at 1.1 bar.




Next up we have a comparison between shot 2 & 6 at 1.1 bar.



As you can see shot 2 (the blue line) had a higher 'hump' at the beginning of the shot. I suspect this is probably to do with the whole machine being just a little hotter.

Finally we have a comparison between 1.1 & 1.5 bar pressure.



As you can see although the 1.5 bar profile does take an upward climb towards the end of the shot the profile over a 30 second shot is similar. This shows just how much impact GH temp comes into play here.

As a bonus here is my best attempt at producing a "flat profile".



Please note this shot is at 1.3 bar which is where I have decided to now leave my machine after all of this testing. This is also a 40 second shot.

Machine was warming up from a cold start, PF was pulled at 90°C, puck was prepped, 5 second flush at 90.5°C, PF locking in & go.

This is an easy shot to pull, no 'fanning' required. This is now my go to morning shot every day.




Conclusions


I think from looking at the data it is clear to see that we can manage HX brew temperatures fairly easily utilising a small fan. We can also fairly accurately brew at a variety of temperatures without adjusting boiler pressure. The data also shows that boiler pressure/ temp does not directly have a significant impact on brew temp at the puck. Boiler temp will obviously influence the temp of the GH as the thermosyphon will run either hotter or cooler, however GH idle temp with this method has little bearing on the final result as we can regulate the temp with a fan anyway.



What do I recommend?

Well first off let me say, I am no coffee expert. I am a Telecoms Engineer. My job is problem solving, dealing with some simple & some complex faults on a daily basis. If I think something can be done better then to me that is the problem & I want to solve it. What I have done here is simply prove that a solution to my problem works in practice. I don't think this method will please everyone & that is fine. I am reluctant to tell anyone what to do, but rather share what I have found & what worked for me. What I will do however is share how I now approach managing HX temps after all of this learning.

My machine comes on at 06:45am every morning on a WiFi timer. I am down by 07:00am & flick the GH thermometer on. I keep an eye on it as I sign on for work, make the wife a cup of tea, make the kids some drinks, watch them beat each other up etc... Once the GH hits 90°C I pull the PF prep my shot, wait for it to hit 90.5°C, quick screen flush, lock it in & go. This produces that nice flat 93°C(ish) shot, easy!

If I want to make another shot any time soon I leave the machine on & fan down as the process above explains & carry on.

At the weekends the machine comes on as normal at 06:45am, same deal, I pull my first drink of the day as the machine is warming up. If we are going out I will turn the machine off, if we are staying in I will leave it on. The machine warms up fully & when I want another shot I will simply fan it down to the desired point & carry on as the process above explains.

For back to back shots I will generally move quickly & keep on top of the group temp (before it rises too high), sometimes hitting it with a bit of air. You get the idea. Keep the GH where it needs to be & you are good to go assuming you are not a super human barista who can knock shots out on the minute every minute I would assume HX water will be 'there or there about' where it needs to be. Please bear in mind all Rocket machines are fitted with a 3mm thermosyphon restrictor in the upper plumbing. This regulates (slows) the flow of the thermosyphon which in turn slows down the rebound of the group. If you have what Dan Ken of Home Barista would describe as a "Dragon" meaning a HX machine probably without a restrictor recovery time may be too fast to stay on top of. Fanning between shots pulled quickly will likely be still be required.




Considerations

The first consideration would be to discuss a point I just touched on. HX rebound time. This method relies on maintaining the temp of the GH & always initiating the shot at the same pre-determined GH readout. The other variable at play is HX water temp. I know I have said that HX water temp is almost irrelevant & I still believe this to be the case for the most part. However, large swings in HX water temps will show up in the results. How much? Hard to say but worth considering.


Flow..... Now flow is an interesting variable. Considering we are relying on cooling the "too hot" water by passing it through the group that is below target brew temperature the rate at which the water passes through the group make a difference. Too slow & the temp at the puck will cool below target brew temp. Too fast & the Hot HX water will not be attenuated enough. How much these results vary is another test in the making however I would like to acquire a Flow Control kit for my machine before perusing that avenue.

One thing worth mentioning is that errors in grind that result in a 'too slow' flow may be slightly buffered by the effect of a cooler brew temp at the puck. Meaning if you ground too fine & are likely going to be pushing into 'Over Extracted' tastes, these 'tastes' may be somewhat muted but the fact that the brew water ran a little cooler. Possibly making the e61 group even more forgiving? An additional note with regards to flow would be to consider the temperature profile throughout the shot as puck erosion occurs & flow increases.

Another consideration worth touching on is the commonly talked about 'advantage' of a HX machine vs a Dual Boiler (DB) & that is brew water 'freshness'. It goes without saying that generally HX machine have a quicker turn around of water being used for brewing when comparing to a DB machine. This is because the HX is significantly smaller than that of a conversional brew boiler on a DB machine. Not flushing a HX machine would obviously slow down the turn over of water (which is kind of why this whole project started for me) which may in turn show up in the taste. I couldn't tell, maybe you could, if you would like to 'freshen' things up a bit a 5 second screen flush prior to each shot will not negatively impact temp stability, it may in fact help produce a flatter shot if that is what your after, which conveniently leads me onto my next point.


'Flat' profiles vs 'Humped' profiles. Throughout this whole process here I have been mentioning the 'HX hump' & have also mentioned ways to avoid it. Please consider this is only for experimental purposes. There has been a lot of discussion around 'flat vs humped', 'inclining vs declining' temperature profiles & to be honest I think the 'jury is out' as to which is better. I'd say don't worry about it, play with both if you like, have some fun.


Further experimentation

Although I feel I have reached a conclusion & am now happy with how to manage my Rocket machine I simply cannot leave it there. This process needs automating. Unfortunately machine manufacturers generally want to focus on their higher end machines & simply slapping a PID on a HX & claiming it somehow magically produces normal brew temps is simply not the way forward.

Lelit are the only manufacturer I am aware of who have done anything that resembles a solution, the Mara X. They have however gone about it completely 'arse about face' if you ask me. I am not knocking them as I think innovation is important & it is great that they are trying, but with all the resources at their disposal to simply regulate brew temp by effectively ramping boiler pressure up & down (based on return thermosyphon water temp) is urrggghhhh, well it works, sort of but I just don't like it. Hard to really explain but I think they could have come up with a much simpler solution. I have one in the making that would actually work better, no boiler swings, no flushing (it is not a fan but I will not share until I have proven the concept).


I have automated the fan process myself with a small brushless DC fan linked up to a PID with a thermocouple mounted to the underside of the group. The PID can be programmed to turn the fan on/off at whatever temp you like. Maintaining GH temps this way results in a fluctuation of around 1°C at the group. Brew water at the puck would likely fall within the same range & brew profiles will possibly run 'flatter' due to the constant cooler idle that this process results in (return leg on thermosyphon will be cooler for extended periods may drop the HX temp a little, certainly stopping it 'superheating'). The set up for this is a little ugly, I am still at the proof of concept stage, playing around with a few options but I have a much better idea that has popped into my head that I will share if I ever get around to it.

Versalab: maker and supplier of finest espresso equipment
Sponsored by Versalab
User avatar
EddyQ

#2: Post by EddyQ »

Nice write up and info!

I went though a similar exercise with my KvdW Idro.
It is a off line HX, but a spring lever group. Initially I thought it would act like your e61. But no.

Jump to page 4. Time for an upgrade

I'm curious how you measured brew temperature. Now that I have both groups working well and pulling very nice shots, I need to go back and measure brew temps.
It is painful with the method I currently am doing.

Edit: sorry, I went back and watched the video. You have a thermometer on the group. Wish I had that kind of access to brew water.
LMWDP #671

Howardsmith (original poster)

#3: Post by Howardsmith (original poster) » replying to EddyQ »

For anyone wondering I will outline the 'at puck' thermometer set up.

I purchased a cheap meat thermometer from eBay, dismantled it and encapsulated the thermocouple and wiring in a few layers of foil tape. This contraption is then centred in a blind basket within a PF and formed to the shape of the basket/PF rim... It is then worked into final position slowly by pushing up the PF into the group a few times before finally being fully locked in with slightly firmer than normal force.

Removal of the PF now will normally result in the set up remaining in place but sometimes it needs to be repositioned. Removal is also easy.

The advantage of this set up vs a SCACE is that we are able to measure a 'realtime' extraction of a coffee puck where flow will vary unlike a SCACE where flow is a constant.

I have included a few photos photos to demonstrate the set up.





User avatar
EddyQ

#4: Post by EddyQ »

That method is similar to what I did and had trouble getting decent seal and my probe wire got damaged after a half dozen tests. But I was not using foil tape. That addition may help me a lot. Thanks!
LMWDP #671

dwightand1

#5: Post by dwightand1 »

Hello Howard;
This is great information! Thank you for taking the time and posting!
I have been doing something similar but not to this level of detail.

The following are graphs of an espresso shot. Taken after warm up. I use Eric's E61 group thermometer and record the process on video. Then afterwards, I use the play and pause in a video edit ap that displays time. I record the temp for each second in an excel spreadsheet. Finally, I use the data to create the graphs shown below.

My assumption is the error between the group measurement and the actual puck temperature is a systematic error or offset of three or 4 degrees.

I never thought to have airflow on the group, that is an interesting idea.

One interesting thing I have noticed deals with back-flushing. I recently back flushed with a cleaner and noted a significant amount of coffee colored water. After a good rinse, the temperature seemed to be a bit more stable. It could be my imagination. But it seemed like my initial temp went up a degree or two at the start of the pull.

Anyway it is encouraging to see this type of effort and work, just saying thanks!

Much Regards,
Dwight


Howardsmith (original poster)

#6: Post by Howardsmith (original poster) »

EddyQ wrote:That method is similar to what I did and had trouble getting decent seal and my probe wire got damaged after a half dozen tests. But I was not using foil tape. That addition may help me a lot. Thanks!
I had a few attempts at DIY scace device but in the end settled on this and now that I have this set up I don't think I would actually want to use a SCACE for testing instead.

Give the foil tape a try, it actually works if you pop the whole gasket and shower screen out and route it over the top of the gasket and then sandwich it between the gasket and the shower screen. I'd like to buy another thermometer and set it up like this as I think there will be much less ware and tear on the part. Only disadvantage being if you want to make something and actually drink it. I have been testing the PID controlled fan all day today (video and results coming soon) and the advantage of my current set up is that I can remove the 'probe' in 1 second and pull a non disrupted shot easily.

Set one up and post your results here.

Howardsmith (original poster)

#7: Post by Howardsmith (original poster) »

dwightand1 wrote:Hello Howard;
This is great information! Thank you for taking the time and posting!
I have been doing something similar but not to this level of detail.

The following are graphs of an espresso shot. Taken after warm up. I use Eric's E61 group thermometer and record the process on video. Then afterwards, I use the play and pause in a video edit ap that displays time. I record the temp for each second in an excel spreadsheet. Finally, I use the data to create the graphs shown below.

My assumption is the error between the group measurement and the actual puck temperature is a systematic error or offset of three or 4 degrees.

I never thought to have airflow on the group, that is an interesting idea.

One interesting thing I have noticed deals with back-flushing. I recently back flushed with a cleaner and noted a significant amount of coffee colored water. After a good rinse, the temperature seemed to be a bit more stable. It could be my imagination. But it seemed like my initial temp went up a degree or two at the start of the pull.

Anyway it is encouraging to see this type of effort and work, just saying thanks!

Much Regards,
Dwight

image
Hello Howard;
This is great information! Thank you for taking the time and posting!
I have been doing something similar but not to this level of detail.

The following are graphs of an espresso shot. Taken after warm up. I use Eric's E61 group thermometer and record the process on video. Then afterwards, I use the play and pause in a video edit ap that displays time. I record the temp for each second in an excel spreadsheet. Finally, I use the data to create the graphs shown below.

This is exactly how I am collating the data. To say that it is time consuming would be an understatement but I do find it interesting

My assumption is the error between the group measurement and the actual puck temperature is a systematic error or offset of three or 4 degrees.

Fahrenheit yes maybe, but it depends on the method, as Eric has said F&G and F&W are literally opposites, as mentioned towards the begining of my first post.

I never thought to have airflow on the group, that is an interesting idea.

Please try it... Follow the process as above and track how consistent your GH readouts become...

One interesting thing I have noticed deals with back-flushing. I recently back flushed with a cleaner and noted a significant amount of coffee colored water. After a good rinse, the temperature seemed to be a bit more stable. It could be my imagination. But it seemed like my initial temp went up a degree or two at the start of the pull.

Without data to back it up I suppose your findings here could be summed up as 'strange things happen'.... Maybe someone with more experience could chime in. My best guesses, although unlikley would be that your GH thermometer was 'gunked up' and the back flush cleared it out. Again, unlikley.


Anyway it is encouraging to see this type of effort and work, just saying thanks!

Much Regards,
Dwight

Thanks, like I said, I was on a mission to work my machine out, in the process I have learnt a bit about my machine and with all the data I have had to collate it made sense to share

Urnex: 100% dedicated focus on coffee and tea cleaning
Sponsored by Urnex
dwightand1

#8: Post by dwightand1 »

Fascinating! I will try it and see. With respect to warm up for the first shot. Could someone simulate this on an already warmed up machine, ie 2nd or 3rd shot of the day, by pulling water from the boiler and turn of the machine for a few minutes then turn it back on and wait till the group temp reaches 90.5...etc? You have given me many other things to try! Wow, what fun!!!

I like your thoughts or question on how would flow control provide yet another element to vary.

This is a very nicely done on the video. and I'm guessing some significant time to get it to this state.

My machine is a PID HX, as I like the ability to change the temp. But to your point, it is not a necessity.

Some time ago, I called into question the use of a PID due to the heat capacity and transport of water. Until you shared this post it had not occurred to me that the the group head itself creates a larger influence to heat transportation or transference of heat to the puck.

This is such fascinating work, thank you! I hope some of the manufacturers are taking notes! :)

Much Regards,
Dwight

Bluenoser
Supporter ♡

#9: Post by Bluenoser »

Current HX manufacturers need to re-think their design in an era where people are increasingly concerned with fine tuning the brew water temperature to adapt to lighter roasts; and to address the interest in flow profiling. They should cease the claim that a PID on the steam boiler will create an absolute and stable brew water. It won't. And instead, increase their education on how to get the best brew water control from the HX design. Certainly the MaraX is Lelit's attempt to address these issues with the HX design.

Having a temp Monitor for water inside the group, or at the screen allows a user to develop a useful workflow AND gives the user confidence on the absolute value of the brew water without having to invest significant additional cost on a group thermometer. It can also let you know if you've pulled back-to-back shots too quickly and your brew water has not reheated sufficiently.

Having a pressure gauge on the front measure the actual pressure in the group would not only help standard extractions, but would allow use of the E61 Flow mod without needing another pressure gauge. This keeps the aesthetics and makes the mod cheaper.

Throw in a bottomless PF instead of the single-spouted and you'd have an HX machine that would provide a significantly easier learning experience for those entering the espresso hobby.

Great work above to characterize the Rocket HX. And really great idea in looking at alternate ways to manage the brew water. I took that idea and now use a cold dish rag that I throw on my group If I notice it is above 201F. Instead of flushing, this brings my temp down to 195F, at which point I remove the rag, prep and pull. Waaayy easier than flushing and waiting ... and certainly saves on water.
★ Helpful

Howardsmith (original poster)

#10: Post by Howardsmith (original poster) »

dwightand1 wrote:Fascinating! I will try it and see. With respect to warm up for the first shot. Could someone simulate this on an already warmed up machine, ie 2nd or 3rd shot of the day, by pulling water from the boiler and turn of the machine for a few minutes then turn it back on and wait till the group temp reaches 90.5...etc? You have given me many other things to try! Wow, what fun!!!

I like your thoughts or question on how would flow control provide yet another element to vary.

This is a very nicely done on the video. and I'm guessing some significant time to get it to this state.

My machine is a PID HX, as I like the ability to change the temp. But to your point, it is not a necessity.

Some time ago, I called into question the use of a PID due to the heat capacity and transport of water. Until you shared this post it had not occurred to me that the the group head itself creates a larger influence to heat transportation or transference of heat to the puck.

This is such fascinating work, thank you! I hope some of the manufacturers are taking notes! :)

Much Regards,
Dwight
Fascinating! I will try it and see. With respect to warm up for the first shot. Could someone simulate this on an already warmed up machine, ie 2nd or 3rd shot of the day, by pulling water from the boiler and turn of the machine for a few minutes then turn it back on and wait till the group temp reaches 90.5...etc? You have given me many other things to try! Wow, what fun!!!

That could work but with the amount of water you would need to draw from the boiler maybe you could just do a F&W anyway.

I like your thoughts or question on how would flow control provide yet another element to vary.

This is a very nicely done on the video. and I'm guessing some significant time to get it to this state.

My machine is a PID HX, as I like the ability to change the temp. But to your point, it is not a necessity.

Some time ago, I called into question the use of a PID due to the heat capacity and transport of water. Until you shared this post it had not occurred to me that the the group head itself creates a larger influence to heat transportation or transference of heat to the puck.

This is such fascinating work, thank you! I hope some of the manufacturers are taking notes! :)

I doubt it....