Thanks for running the test and sharing.

cafeIKE wrote: My DB is set to 103°C for 91°C shots - no PID offset. IMO, offsets are wishful thinking.

What do you mean? The way I understand, PID offset is merely to offset the temp displayed. In your case, if you set it to 12C, it will display 91C but actually target 103C in the boiler and brewing temp will match the display.

@Bluenoser, I used singles, not doubles. 11.7 gram ground coffee, 23 gram in the cup, at about 25 seconds. You can see the specific timings of each shot in the spreadsheet I linked in the OP. The first few shots were longer, as I adjusted the grinder for new beans I didn't previously use.

The amount is water that managed to escape from the side (where the cable of the probe passed) was negligible. Just a few drops.

K7 wrote:What do you mean? The way I understand, PID offset is merely to offset the temp displayed. In your case, if you set it to 12C, it will display 91C but actually target 103C in the boiler and brewing temp will match the display.

The display cycles ±1°C at idle. The group remains constant. When I start to brew, the display can drop 5°C, depending on the reservoir temperature, where the machine is in its heat cycle, etc.

If I bump the idle temperature up 4°C for a hotter brew temp, the shot temperature does not rise 4°C.

When the machine is coming due for a descale, shot temperature drops. The display temperature remains the same.

When you want to know the shot temperature, monitor it at the group.

Thanks for the clarification. I now understand where you were coming from.

cafeIKE wrote:If I bump the idle temperature up 4°C for a hotter brew temp, the shot temperature does not rise 4°C.

In that case, how does it rise in your experience? 2C? 3C? 5C? Just trying to understand how much deviation one can expect from these machines.

When the machine is coming due for a descale, shot temperature drops. The display temperature remains the same.

Interesting point. Maybe the OP's machine is suffering from this?

Yes, scale might be an issue. In the last couple of years I'm using Rpavlis water and previously used TWW, but during the first months after buying the machine I used bottled water that I know wasn't soft. It could be that there's some scale left from that period. I was planning to descale it back then but this warning in the extended manual (PDF file from chriscoffee.com) gave me pause:

"Often times descaling can cause more problems than it solves. It can react to the
minerals and foam over ruining electrical components. If the solution is too strong it can
cause the chrome plating inside the group to flake off and get in the coffee or if it's too
weak it can dislodge minerals and cause a blockage. For liability reasons we strongly
discourage descaling and will not provide any instructions on the process."

How problematic is it in practice in a machine like the Alexia Evo? I never had any issues descaling my previous espresso machines (using Durgol Swiss Espresso Descaler).

boren wrote: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.

I personally flush after each shot to clean up the shower screen, more experienced members may have thoughts on this.
I'm questioning the flush as each time you flush you're purging stable, hot water and introducing cool water into the boiler and entering into the same recovery cycle you'd be going into between each shot. again, more experienced members can correct me here.

Assuming I'm reading this thread correctly, as your offset = 0 and your boiler temperature is 95 degrees, I'm not surprised that your brew temp is far below what you want.

boren wrote:These are my PID settings:

Celsius mode
<snip>

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

Our machines are somewhat similar (750ml boiler, single boiler e61 etc.), for 93.5deg brew temp my boiler target is ~ 104deg. (My offset is -11deg C from memory)
I've accepted with an e61 this will end up being +/- 2deg with consecutive shots. I just ensure the first shot is mine.

I've measured the first ~4-6sec of a shot with a temp probe held to the water coming out of the screen (yes, burns occurred, worth it), water in an insulated cup, and done the flash boil method to calibrate my offset. The last part was key.
(I was going to make a thermofilter and pressurefilter but waiting to see what comes out of a local vendor in the next month instead)

To me, if the water coming out is at the target temperature for the initial shot, that's as good as it's going to get for an e61, the remaining questions are "does it taste good enough?" (the answer is a resounding "yes!") and "does it recover quickly enough between shots?" (quickly enough that I haven't bothered to tweak my PID knowing I could improve recovery further).

If I've read your threads correctly you've set your boiler temp. as your target brew temp and are getting the expected fall off. Bump it up and hopefully that will address your primary concern - try boiler temp of 102-104 deg Celsius based on our machines being very similar.

cafeIKE wrote: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.

Good idea. I'll check my thermometers and probe with a stove-top pot of boiling water. This should be stable enough at 100c.

_Ryan_ wrote:I personally flush after each shot to clean up the shower screen, more experienced members may have thoughts on this.
I'm questioning the flush as each time you flush you're purging stable, hot water and introducing cool water into the boiler and entering into the same recovery cycle you'd be going into between each shot. again, more experienced members can correct me here.

I flush for 4 seconds before each shot and for a second or two after each shot. I think the amount of flush water is negligible considering the size of the boiler, but I plan to test different timings to fine tune what provides the most stable shot - no flush, 1, 2, 4, 6 and 8 seconds.

Assuming I'm reading this thread correctly, as your offset = 0 and your boiler temperature is 95 degrees, I'm not surprised that your brew temp is far below what you want.

My offset was actually the default for my machine 12 degrees (you snipped that part ) . However, I increased it yesterday to 16.5 and changed the PID temperature to 99 and just pulled a shot. Temperature at the puck was 96.3c (after 2 seconds), 98.3c (between 5-10 seconds), dropped to 97.8c (after 15 seconds), and then 97.3c (20 seconds until shot end at 25 seconds). This is pretty stable and about 4 degrees higher than the target. I reduced the PID to 95c and will test again later. I think I'm getting closer to a good setup.

I also sent an email to Quick Mill asking them about the best practices for descaling my machine. I'd like to cover all possibilities.

I tested a couple of shots without flushing, about 20 minutes apart.



I guess flushing isn't needed, as the temperature of both shots was quite stable after about 5 seconds. Still, I'll repeat the same test with different lengths of flushes.

The actual offset between boiler and puck is about 18.5c. Unfortunately the PID can't be set beyond +16.5, but I can achieve the real target temperature by adding 2c to the PID target. I just need to remember to do this.

Edit: Scratch that. With a 4" flush and offset of +16.5 the result is even better:



Hopefully this is repeatable...

boren wrote: My offset was actually the default for my machine 12 degrees (you snipped that part ) .

Firstly, apologies - i saw the 0.4!

boren wrote: I flush for 4 seconds before each shot and for a second or two after each shot. I think the amount of flush water is negligible considering the size of the boiler, but I plan to test different timings to fine tune what provides the most stable shot - no flush, 1, 2, 4, 6 and 8 seconds.

Fair enough, interested in your test results for "no flush".
I'm mildly curious what the impact of the 4s flush is, I think that would be a 30-40ml flush(did a quick google of flow rate of the pump not under pressure), so you're right, in the context of a 750ml boiler it's next to nothing. But I'm not knowledgeable enough with this stuff to know where the brew water is drawn from and if there is any stratification in the boiler. I'm probably over thinking it (or underthinking it considering the role of the grouphead hehe).

boren wrote: My offset was actually the default for my machine 12 degrees (you snipped that part ) . However, I increased it yesterday to 16.5 and changed the PID temperature to 99 and just pulled a shot. Temperature at the puck was 96.3c (after 2 seconds), 98.3c (between 5-10 seconds), dropped to 97.8c (after 15 seconds), and then 97.3c (20 seconds until shot end at 25 seconds). This is pretty stable and about 4 degrees higher than the target. I reduced the PID to 95c and will test again later. I think I'm getting closer to a good setup.

I also sent an email to Quick Mill asking them about the best practices for descaling my machine. I'd like to cover all possibilities.

So the Offset was increased by 4.5degrees and the PID SV/brew temp was increased by 4 (from 95 to 99) resulting in a change of 8.5 degree change in target temp? Glad it's closer to what you want.
then a reduction of PID SV back to 95 (-4) leaving a net change in target brew temp of 4.5 degrees?

I ask as I quickly threw together a comparison of some of your data and compared to my machine as our designs are so similar - my PID offset is set so that my PID SV is pretty much the water coming out of the screen, so the flash boil method is relevant (see Getting accurate shot temperature displays on PIDed double boilers without a thermometer ) I also did testing with a cup but can't find my notes. I didn't bother drilling a basket and fitting a probe as a device is due to market shortly that I'll likely buy and taste is guiding me at the moment and saying I'm in a good place.

If you're interested in comparing you could repeat the flash boil test on your machine and calculate the delta between boiler temp and screen. My delta was 11.5 deg C. I could potentially try and find my remaining unbroken probe for a few puck tests if further base-lining was deemed relevant.

boren wrote:I tested a couple of shots without flushing, about 20 minutes apart.


I guess flushing isn't needed, as the temperature of both shots was quite stable after about 5 seconds. Still, I'll repeat the same test with different lengths of flushes.

The actual offset between boiler and puck is about 18.5c. Unfortunately the PID can't be set beyond +16.5, but I can achieve the real target temperature by adding 2c to the PID target. I just need to remember to do this.

Edit: Scratch that. With a 4" flush and offset of +16.5 the result is even better:



Hopefully this is repeatable...

Nice, much better. Although I can't help but notice in your spreadsheet two things:
- "boiler temp" as 96 when offset is 16.5, with PID SV as 95. 95 +16.5 = 111.5; therefore boiler temp would be 111.5 (PID target + Offset = Boiler temp)
- "PID real offset" as 17.2, when the offset configured in the PID is 16.5; this may confuse readers Might I suggest naming it something else? if boiler temp is corrected, simply "boiler temp - target temp" - Sounds like this is some compensation you're having to add to your SV to get the target you want, see waffle below...

Sounds like you have a decision to make, sucks that you can't configure the offset to reflect the delta to the puck. I'd personally move the calibration point up one 'link' in the chain for a logical representation, but you have options.
1. Do as (I think it was Ike?) suggested and set Offset to 0 and know what boiler temp will get you what puck temp
2- Do as I and many others do(although mine will be changed to puck temp in time), calibrate offset for water coming out of the screen, aligned to boiling point, and do your compensation based on this figure. At least this should be consistent(set and forget).
3- Do as you are and compensate at the puck level.

_Ryan_ wrote:(or underthinking it considering the role of the grouphead hehe)

My guess is that. The back of the group head is heated by a thermosiphon, driven by the cooling of water in the head. The front of the group head is losing heat to the atmosphere, so is at a lower temperature than the back of the group head. A brief flush will raise the internal temperature of the passages to the brew chamber.

See, for example, E61 Thermal Analysis Questions