Cafelat Robot Temperature Tests

A haven dedicated to manual espresso machine aficionados.

#1: Post by jpender »

I finally got around to installing a probe in my Robot, dangling below the piston so that it is immersed in the water once the PF is locked in. It's not the fastest responding probe so it still reads below the actual water temperature for the first 5 seconds or so of preinfusion. You have to sort of imagine the lines extending backwards in a quasi-linear fashion. But it helps to illustrate the effect of preheating strategies. The actual pulls started 10-15 seconds in.

I usually just heat up the PF (green line below) by placing it on the lid of a saucepan of actively boiling water. I compared this to doing nothing (black line), pouring just-off-the-boil water into the basket/PF so that it overflows for a short while (orange line), a preheat of only the piston in just-off-the-boil water (blue line), and a preheat of the piston, PF, and basket (red line).

Sort of interesting. The piston preheat really slowed down the rate of temperature decline.

★★ Quite Helpful

User avatar

#2: Post by Balthazar_B »

John, thanks for doing the experiment and posting the plotted results. Remarkably distinct outcomes!

Do you know what the ambient temperature was when you ran the tests? Did you commence every test with a stone-cold Robot?

If you have the time over the next few days, maybe it would be interesting to run each of them 3-5 times to see how much deviation you see from your first runs (i.e., degree of reproducibility). Yes, that's the memory of my high school physics teacher whispering in my ear. :)
- John

LMWDP # 577


#3: Post by max »

I love posts like these!

I will have to preheat my piston now. The only other Robot temperature experiments I've seen were with very low ambient temperatures, but, as you're in CA, I suspect this data reflects a normal situation better.

If you decide to do more data collection, know that it is appreciated :)


#4: Post by dak »

Great information and presentation of the data. Thanks for sharing! Really interesting how the Piston preheat helped with stability, but I guess it's not surprising given it's the biggest heat sink in the equation.

If you get a chance, I'd be curious to see how the Pour off method (fill to the top, wait 10 seconds, pour off, refill and pull shot) compares to the overflow method. I use the pour off method and I'm always amazed by how well I can pull some very light roasts without the overly sour/acidic notes I even sometime get with my Bianca.


#5: Post by jpender »

It's been kind of warm lately, which for here means 18-20°C in the morning. Yes, I made sure the Robot was cold before each test.

I ran at least three tests for each example except for the hottest one which I just did this morning for the first time. The runs were pretty tightly grouped except for the case of the so-called "double pour" or "pour off" technique, where you fill the basket with hot water, dump it, and then fill it again. That method was all over the map for me. I don't know why exactly. Perhaps the precise amount of water and the time were varying too much. I also only did the overflow method once because the first time I tried it it messed up the puck and I ended up dumping 18g of very nice coffee grounds.

I also had a piston-only preheat run that was 2°C colder because for some reason I couldn't get the PF to lock in for an extra 10-15 seconds. Things cool down fast!

Here are some of the runs to give you an idea:


There are some other things I want to try, mainly insulating parts of the Robot. Preheating the piston is kind of an annoyance that I'd rather not make part of my routine. And with the coffee I was using (Linea Espresso blend) the do-nothing shots were wonderful. So maybe this is all just an exercise in geekery unless/until I get interested in much lighter roasts.
★ Helpful

User avatar
Supporter ♡

#6: Post by yakster »

Great work, John, thanks for doing this and posting this.

LMWDP # 272


#7: Post by GregoryJ »

Hi John, can you post a picture of how you've mounted your probe?


#8: Post by Bluenoser »

How did you preheat the piston?. Huge difference with that. I was thinking of using the wand of my HX

Edit: Think I found it.. use double spouted.. I just tried and used tape on inside to block holes.. seemed to work pretty good.. will try out tomorrow morning.. seems to make big temp difference.

User avatar

#9: Post by Balthazar_B » replying to Bluenoser »

A mug, pyrex cup, or other receptacle with a diameter slightly wider than the piston/seal, nearly filled with boiling water, and set at the right height, would do the trick while the Robot's arms are down. Probably won't be long before some enterprising H-Ber rigs up a perfectly-sized reservoir equipped with a heating element to make it simple and repeatable...
- John

LMWDP # 577


#10: Post by Bluenoser » replying to Balthazar_B »

I just tried it with boiling water and the piston got pretty hot.. If the portafilter holder had a rubber-type insulated wrap all around below the handle, the PF would remain hotter..Maybe some closed foam tape would do the trick. This is what Flair has on its Pro2 PF. That is one heavy beast and it has the piston inside the PF as one unit along with a rubber wrap. That stays really hot and I'd say it could be pre-heated hotter a little easier than the robot.

Taking double spouted quickly off PF when hot is a little tricky and then to start pull within short time. But I think that would significantly keep heat in the water.

I added 3mm of foam to my PF and then wrapped in black tape.. should hold heat better.. will try tomorrow.