SB E61+PID, Episode 2: Tested properly this time, with much better results

Need help with equipment usage or want to share your latest discovery?
boren

#1: Post by boren »

After the previous disappointing results (see Episode 1) I decided to retest my Quick Mill Alexia Evo temperature stability. This time I tried implementing the (excellent) advise I received in that thread and tested temperature during actual extraction, including at the coffee puck. It turns out it was well worth "wasting" 150 gram of coffee beans for this test, because the results are very different from what flushing water would imply.

You can find a spreadsheet with the full results here. Note that the first 17 rows are hidden. If you unhide them you'll see details about boiler temperature, grouphead temperature and brew pressure gauge (phantom readings). I also measured the cup heater temperature (nothing exciting there, it quickly gets to ~40c and stays there). The more interesting things though start at row #18, with this first shot pulled (at the 20 minute mark).

Some info about testing and methodology:

I used a K-Type thermometer (TASI-8620) with two probes - a flexible thermocouple probe with an exposed wire for the puck, and a 10cm steel probe to measure temperature in a paper cup.

For the puck temperature I initially put the probe over the portafilter lip and laying on a 58mm puck screen. I later came to the conclusion that the probe might not be sitting flat, affecting measurement. I added another 49mm puck screen (which I use for my Elektra MCaL) on top of the probe. When the probe was sandwitched between the two screen measurement was more stable and the reported temperatures were higher. I was surprised to find that sealing wasn't compromised much by this setup and that only a few drops of coffee managed to escape from the side.


For the cup temperature I placed the cup into a bottomless portafilter, and then inserted the probe from the side of the cup. I held its tip to the bottom of the cup with a zip tie. Later on I punctured the cup to allow the coffee to drip out a bit. I'm not sure if this was needed, but I wanted to avoid pressure buildup.


Before each shot I flushed the machine for 4 seconds, taking note of the temperature before and during the flush (2 and 4 seconds in).

In each shot I noted the temperatures 2, 10 and 20 seconds into the shot, as well as at the end of it.

Observations:
  • The grouphead temperature seems to stabilize after 40 minutes. The boiler needed 3:40 minutes to reach its the target temperature of the PID (95c).
  • Below is a summary of the results. It is limited for the grouphead from shots #3 to #13, at the puck for shots #8 to #13 (where the probe was sandwiched between puck screens) and at the cup for shots #10 to #13 (small sample size, I know).
  • Throughout each shot, the temperature at the grouphead changes by about 2.5c. When measured at the puck from 2 seconds and until the end of the shot, the temperature changed by as little as 2.9c in one shot, and up to 5.5c in another. For these shots, if averaged from 10 seconds into the shot and until the end, the changes were much smaller - between 0.7c and 1.2c.
  • Grouphead temperature before flushing was relatively stable, ranging between 84.4c and 86.2c
  • The brew pressure gauge showed very odd readings. During machine warm up it went, for no obvious reason, from 0 bar to 7 bar after 2 minutes, then back to 0 bar after 7 minutes, then to 1 bar after 11 minutes. After pulling a shot it usually stayed high at around 5 or 6 bar, and then dropped to about 1 to 2 bar after a few minutes.
Takeaways and conclusions:
  • Temperature stability when pulling actual shots (with coffee in the basket) is much better than my original conclusion, after just just measuring temperature by flushing water. It is good enough most of the time to get tasty espresso, but a range of about 3.5c between shots (at the grouphead) and 4.5c (at the puck) is not great. My expectation from a machine with a PID was to have better temperature stability.
  • The brew pressure gauge can just be ignored. It's only useful during extraction.
  • A 4 second flush is not sufficient to get the temperature in the beginning of the shot (2 seconds in) close enough to the rest of it. It's probably worthwhile to experiment with longer flushes.
  • After I'll nail down a workflow that gets temperature to be as stable as possible I'll finally install the flow control system I bought a while back. If it improves my coffee enough I'll keep my machine. If not, then it's probably time to move on to a saturated group Lelit Elizabeth.
One last question - if the temperature at the puck is about 85c throughout most extractions, does it mean I need to bump up my PID temperature by 8 degrees (from 95c to 103c) to get to the frequently recommended brewing temperature of 93c?

_Ryan_

#2: Post by _Ryan_ »

New to your journey, sorry if you've answer this elsewhere, but I have a couple of questions:
1. Why do you flush your single boiler machine before a shot? I thought this was something only done by HX operators to get their temps down. (I have a single boiler myself and haven't done this in seven years, especially not now that I have a PID installed.)
2. How did you calibrate the offset for your PID?

boren (original poster)

#3: Post by boren (original poster) »

I flush to increase temperature before the actual shot. You can see in the linked spreadsheet how temperature increases at different stages (during the flush and during extraction). It also helps clean up the shower screen from any coffee that might be left from the previous shot.

I didn't change the configuration of my PID and that also applies to offset.

Bluenoser
Supporter ♡

#4: Post by Bluenoser »

Interesting data! I remember a poster criticizing the need to measure the brew water precisely, but only by these types of investigations do you really see what is going on in E61 machines. If your PID is set to 95 and you are getting 84.1 as your average puck temp (I'd ignore the cup temp), then many PID systems allow you to add an "offset" to the PID reading, which you alluded to .. which can make the display reflect the final brew water temp. Now the problem with some vendors is that they don't calibrate this offset before shipping to a customer. And sometimes, customers get machines with the offset unset (set to zero) and so their brew water is wildly different than they think.

Your probe in the cup experiment might be too 'slow' and have too much thermal mass to accurately give you a reading. I found a foam cup better.. and I used a K type thermocouple which had very little mass and reacted quickly. I also only found it was accurate for about 3-6 seconds and after that time I couldn't get a good seal against the PF.

The Breville DB was designed using a SCACE (from what I read) so its brew water is one of the more accurate and stable on the market; and shows the benefit of a small/integrated brew boiler right at or in the grouphead. Your results are reasonably consistent, but the data will show that the E61's thermal state will have a significant impact on the final brew water temp. You can also see that the spike the grouphead thermometer shows (from your original post) is 'eaten' up by the E61 and this is one reason the E61 is still used.. to even out thermal changes in the brew water. But without this type of data, one can't ever find out the relationship between the group thermometer and the final brew water.. I find the relationship between my HX group thermometer and the final brew water temp is very consistent from about 10-20 seconds into the extraction; which allows me to now just use the group thermometer as my guide.

If you consider your numbers from about 5 seconds to about 30 seconds it looks like your brew water is probably within 2C which is as good as SB/DB can do with an E61, I think.

For HX machines, the PID is even less relevant as the Thermosiphon design also significantly impacts brew water temp. If manufacturers added a single temperature measurement near the screen of all their E61 machines, users would be able to make much better use of their machines. I suspect many E61 users without any thermometers are suffering from brew water temps that are sub-optimum.

boren (original poster)

#5: Post by boren (original poster) »

Bluenoser wrote:If your PID is set to 95 and you are getting 84.1 as your average puck temp (I'd ignore the cup temp), then many PID systems allow you to add an "offset" to the PID reading, which you alluded to .. which can make the display reflect the final brew water temp. Now the problem with some vendors is that they don't calibrate this offset before shipping to a customer. And sometimes, customers get machines with the offset unset (set to zero) and so their brew water is wildly different than they think.
These are my PID settings:

Celsius mode
P=3
I=0.05
D=2
F.04=12 (offset between boiler and group)

Does it mean I should change the last one from 12 to 23 to compensate for the actual offset?

Edit: The maximum offset value available is 16.5c. I set it accordingly and will increase the PID temperature setting to more than 95c.

Pressino

#6: Post by Pressino »

Very clear and practical presentation of how to measure temperature stability. I'm sure this will be extremely useful for users of e61 (and actually any other type of) espresso machines.

Thanks! 8)

User avatar
Jeff
Team HB

#7: Post by Jeff »

For me, brew temperature is "a number", not an absolute. It's important to me that "94" (for example) repeatedly represents close to the same temperature. I don't really care if that is 92.6 or 96.5, because I can get to the point where "normal" is 94 on my scale and 92 is "on the cooler side" and 96 is "on the warmer side". I don't see a lot of bragging rights with it being exact. As long as I can understand another's suggestion that "this coffee tastes better to me pulled a couple degrees warmer than my normal", that is actionable for me.

User avatar
cafeIKE
Supporter ★

#8: Post by cafeIKE »

If you do much of this, and your TC wire is not molded at the plug, be aware that coffee will be forced down the sheath and may leak into the datalogger. Learned this lesson over 15 years ago :oops:

For this reason, either use an extension or carefully cut an exit gap in the sheath at a convenient location.

It's normal for the brew gauge to oscillate if there is a good seal. If pressure drops to zero on a hot boiler, there is likely a small leak somewhere.
boren wrote:One last question - if the temperature at the puck is about 85c throughout most extractions, does it mean I need to bump up my PID temperature by 8 degrees (from 95c to 103c) to get to the frequently recommended brewing temperature of 93c?
My DB is set to 103°C for 91°C shots - no PID offset. IMO, offsets are wishful thinking.

It doesn't make sense that the shot temperature is so low with a 12°C offset. It's like the offset is not being processed or the sign is wrong. I'd check the meter and probes by inserting in a kettle spout. Should read very close to 100°C, slightly less at altitude. If you want to be really precise, get local atmospheric temperature and pressure to adjust if not standard.

I'd set offset to 0, boiler to 103°C and re-run.

Bluenoser
Supporter ♡

#9: Post by Bluenoser »

boren wrote:These are my PID settings:

Celsius mode
P=3
I=0.05
D=2
F.04=12 (offset between boiler and group)

Does it mean I should change the last one from 12 to 23 to compensate for the actual offset?
Great info about the water creeping up the probe!!.. never thought of that..

I would not expect more than about a 10-15C drop on a heated group.. if you are getting 23, wonder if the puck temp is giving you a lower reading than it should.. How much water is exiting when you do your shot.. are you getting about 40-50ml over 30 seconds?.. the nice thing about a SCACE is that there is a simulated puck and a valve that drops about 60ml(g) of water over 30-40 seconds. If you are near an espresso service store you might try to borrow a SCACE for a day.. I was never confident of my home-made readings enough .. but when I used a SCACE it confirmed whether I could trust my own man-made experiments.

Edit: As I reread your original post, sounds like you used actual coffee during the shots.. which should be a gold standard for testing the temp.. So I don't have an explanation why you are finding about a 20C drop..

As Ian said, it is not necessary to put an offset.. just to know what it is for your machine. And finally when you are confident you are close to 93C on a shot, adjust the PID up and down a few degrees on different shots to find what is the optimum for taste. You'll eventually get a baseline and can move up/down based on the roast level and your own taste preference.

boren (original poster)

#10: Post by boren (original poster) »

Is the recommended temperature (e.g. 93c) supposed to be at the boiler, the grouphead, or the puck (coffee bed)? I'd assume the latter, but want to be certain.