The next time that I have the espresso machine apart for descaling & checking gaskets, I'd like to add active cooling for the group head. I'm already getting pretty good shots from this machine, so the improvement may be modest. Or possibly even non-existent. But at least it will make a fun weekend project. I'll outline a few thoughts about this below & if anyone has any comments or wants to point out the error of my ways, feel free to jump in.

The Cimbali Junior pulls great shots but having a big HX, it runs hot. To get the brew temperature into the 200F-ish range for most coffees means adjusting the pressurestat down to about 0.8 - 0.9 bar. Amazingly enough steaming is still adequate at 0.8 bar, but recovery is slow when you are trying to make a ten or twelve cappuccinos at once when entertaining. Cranking the boiler up to 1.2+ bar yields endless steaming power but also nets unpalatable shots even if you run very long cooling flushes. So I've been mulling the idea for a while of running the boiler at say 1.4 bar and actively cooling the group head to get it back down to an appropriate temperature range for shots & hopefully get the best of both worlds in one package.

Many of the Cimbali machines use the same group head (PDF file). On the larger, multi-group M30 machines, Cimbali added a passive group head cooling arrangement by routing the HX feedwater through the group head twice to cool the group. This creates a water jacket (see pictures in this thread) (or this one) in the group head that is plumbed with several copper tubes. By shortening or lengthening the copper tubes you could bump the group temperature up or down. The Junior model uses the same group head, but the two bosses are not drilled out or used, perhaps because the Junior has a smaller boiler and less recovery capacity than it's bigger brothers.

Anyway, I was hoping to drill out & tap the Junior group bosses to form the group "water jacket" like the M30 has. But rather than running the HX feedwater through the group, I was thinking of using a small pump to circulate water in a closed loop with a five gallon reservoir. I can drill the group like this to mount a thermocouple above the group dispersion block & use a PID controller to regulate the cooling water flow. Sounds good so far, but snag number one appears to be that the Jr. HX water feed bisects the rearmost "cooling boss", rather than entering from the top as in the M28/29/30. So I'd only be able to drill & tap for one water path, rather than two on the Junior. I haven't actually pulled the group head out yet to confirm this, but it does look likely based on all of the Cimbali photos that others have already posted. It seems like a lot of work to drill, tap & plumb everything & perhaps find that one cooling path is insufficient. So I will probably need to source a group head from a M28/29/30 machine to implement this. It would surely be cost prohibitive to buy a replacement group head from Cimbali, but I'm hoping that I can find someone with a M28/29/30 machine in bad shape that is willing to part out the group head for a reasonable price.

I don't suppose anyone here has a Cimbali M28, M29 or M30 group head for sale?

I solved a similar problem :
HX Heaven or 1½ Boiler

Thanks for the suggestion. There's always a variety of solutions to any problem and you have hit on another one. I currently have the boiler PID set with an inflated derivative value so that the heating element goes full on as soon as I start steaming, which helps. But it still makes the brew temperature oscillate more than I would like.

Since the water jacket possibility is more or less built into my machine already, I'm still inclined to pursue active cooling. It seems that this will give me the most stable group temperature while remaining largely independent of the boiler or ambient temperatures. Or at least that's how I hope it will work.