My commercial, non E61, rotary, 90ml HX, 4L boiler, machine (Rancilio S26) seems to have a very narrow espresso flow rate requirement to delivers its flattest brew temperature.

I drew a representative (but not pretty) illustration to communicate this:



Unless the shot flows relatively fast (a bigger normale), brew temp will climb up during the second half of an extraction. It's so bad that for a true ristretto, brew temp graph will look like half a circle (top line in illustration).

I haven't seen many questions on this before so I'm assuming this isn't common HX behavior. Besides sticking to a faster-flowing extraction, my attempts at flattening brew temp included cooling my grouphead (don't ask how!) during an extraction. While cooling the GH works, it isn't predictable nor repeatable nor convenient nor pretty.

I did try different boiler pressure settings (from 0.5 up to 1.5 bar) but couldn't notice a change in behavior. For a long time, I kept my p-stat at around 1.0bar for a flush'n'go routine. I also tried keeping my p-stat at around 0.5bar or so for some time. But what I settled on for a long time now is keeping my p-stat at 1.4bar and 2-3 minutes before I brew I turn off the heating element and quickly bring boiler pressure down to 0.5bar(via manual boiler fill and releasing steam). I settled on this because it keeps the group head hot enough and ready so I don't have to do long warming flushes. I also settled on this because it was the most predictable way for me to manage my machine; of all other things I tried over the years. I did read a lot about managing temp on a HX over the years, but every time I try to apply what I read, my machine doesn't seem to match what's expected. Ideally, if a pro spends enough time on this machine, he'd crack its code, but, alas, I'm not pro and I couldn't crack this machine after years of trying.

Any one sees possible ways for me to have a flat-ish line temp ristretto with this machine?

The Quickmills have the OPV on the hot side of the HX to alleviate this effect afaik. Either way, you could simulate that by putting a small orifice in the hot side of the HX to simulate that.

Or you could put a flow restrictor in the top of the HX to take more water from the cold side of the HX in the group.

Bluecold wrote:Either way, you could simulate that by putting a small orifice in the hot side of the HX to simulate that.

My machine doesn't have a thermosiphon so I assume you mean I should restrict flow of hot HX water to the group head by an orifice in the path from the HX exit to GH? But what would that accomplish after the initial preinfusion, once the brew circuit is at full pressure?

Bluecold wrote:Or you could put a flow restrictor in the top of the HX to take more water from the cold side of the HX in the group.

This sounds like modding the HX or possibly the HX injector to achieve a certain HX output mix. If so, this sounds right on, but I'm not sure I have the skill to go there -- although I mod like mad, the HX isn't something I've touched before...I'm afraid I'll make serious damage if I go there. As a last resort, maybe.

Edit: here's a diagram showing the HX (#34) with the injector inside it. It looks like the whole boiler has to be disassembled to mod the length of the injector to tweak HX mix output.

My machine doesn't have a thermosiphon so I assume you mean I should restrict flow of hot HX water to the group head by an orifice in the path from the HX exit to GH? But what would that accomplish after the initial preinfusion, once the brew circuit is at full pressure?

There'd be more flow through the HX, similar to that of a lungo.

Basically, your brew temperature is determined by the boiler pressure, the flow through the HX and the amount the cold water travels through the HX. If the injector is very long, you'll only push out very hot water out of the HX, if the injector is very short, a larger percentage of cold water will exit the HX.

Everything I know from HXs and tuning them I've learned from this dutch video by Kees van der Westen explaining his Mirages

Everything I know from HXs and tuning them I've learned from this dutch video by Kees van der Westen explaining his Mirages

Same, I don't speak Dutch though so we had to tear one to bits to understand that video

The Mirage is very stable, mainly due to an oversized system that reacts very little to the input of cold fresh water and a carefully tuned(length and size) brew water path that is quickly flushed. All very clever.

Bluecold wrote:There'd be more flow through the HX, similar to that of a lungo.

Basically, your brew temperature is determined by the boiler pressure, the flow through the HX and the amount the cold water travels through the HX. If the injector is very long, you'll only push out very hot water out of the HX, if the injector is very short, a larger percentage of cold water will exit the HX.

Sure, no question about that but it requires modding the HX or injector, which is major surgery as I explained above. I thought you had an easier to implement idea for achieving the same effect by restricting flow to the grouphead, did I misunderstand the following?

Bluecold wrote:you could simulate that by putting a small orifice in the hot side of the HX to simulate that.

Isn't that to restrict flow out of the HX to the grouphead? That's what's not clear to me.

Hmm, it's a bit unclear indeed. I was talking about a small bleed valve that would let a small amount of hot water bypass the group and get dumped in the driptray to simulate a normale-lungoish shot from the HX's standpoint.
With an orifice I meant a small hole in the pipe to use extra water.
Like this:


The orifice is shown in red.

Ahh, now we're talking . That looks like a much more possible mod. There are two places in my machine's system where this mod is implementable (just in theory!) like you explain above:

1. Through the brew solenoid setup
2. Through the screw at the top of the grouphead

Solenoid: doubt it's possible to mod that to bleed water.

GH Screw: looks more promising. That screw at the top of my GH is M6. GH is enclosed inside the machine (unlike those exposed E61s) with little space around it to add "stuff". This "stuff" would need to be a tiny M6 "T" fitting & a tiny adjustable flow valve. Anyone knows where to find these online?

Hmm, interesting idea! Thanks Roeland.

coffee.me wrote:Solenoid: doubt it's possible to mod that to bleed water.

On second thought: no I don't!

Maybe the solenoid can be modded to achieve our objective of bleeding water during an extraction. If I block the solenoid from fully "closing" the exhaust hole when it is engaged during an extraction, it will bleed some water that was meant to to go the the coffee. I would probably also need to make the exhaust hole smaller. Bingo! This won't even affect its function as a backflushing mechanism. Bingo!

I'm thinking loud here, very open to other ideas.

coffee.me wrote:Unless the shot flows relatively fast (a bigger normale), brew temp will climb up during the second half of an extraction. It's so bad that for a true ristretto, brew temp graph will look like half a circle (top line in illustration).

I don't know the particulars of your espresso machine or if your brew profile plots are accurate, but what they depict is typical when the grouphead is too hot. Lowering the pressurestat setting, flushing 2-3 seconds longer, and allowing a slightly longer rebound might help.

coffee.me wrote:Any one sees possible ways for me to have a flat-ish line temp ristretto with this machine?

Why the concern about the shape of the brew profile? Is there a taste defect you're trying to correct?