Adding PID control to an HX espresso machine

#1: Post by **MpVpRb** » January 26th, 2015, 8:25 pm

I have an Isomac Tea (a heat exchanger machine with an E61 grouphead)

Although I am an engineer who could design a control system from scratch, I'm interested in the experience of the community

What is a good controller?

Should I drill a hole in the boiler for the the temp sensor, or use some sort of pipe fitting?

Should I use the temp controller instead of the pressurestat, or in combination with it?

I don't like milk in my coffee, so I want to optimize for brew quality at the expense of steam

But, my wife and guests sometimes like steamed milk, so I don't want to eliminate steam completely

Ideas??

#2: Post by **D'Laine** » January 26th, 2015, 9:16 pm

I have the PID version of the QuickMill Alexia. Always espresso, but can steam if needed.
I can send you the wiring schematic if you like.
The PID is a Watlow, but a Fuji would likely do the job as well.
Thermal sensor location is the question to follow up. It's very tricky. But simpler for reliability.
Or you could use a pressure transducer but more money. Then the logic is as designed by the manufacturer.

D.

#3: Post by **sweaner** » January 26th, 2015, 9:21 pm

Why do you want to PID an HX?

January 27th, 2015, 7:43 pm

I like to experiment, optimize and perfect stuff, even if it's already good enough (I'm a control systems engineer and tinkerer)

It always seemed odd to me that that the machine had no temperature control, only a pressure control

It seems that the control system was designed as a compromise between steam and brew water production

I want to optimize for precise brew water temperature control at the expense of steam

I also want to shorten startup time, since I only use the machine in the morning, and hate waiting

I am thinking of adding a circulating pump to the thermosiphon line

..and yeah, I realize I might make it worse in the process..but that's the life of an engineer who loves to tinker with stuff. I also realize that espresso machine design is a very mature engineering discipline, and improving it may not be as easy as I imagine..but hey..I'm optimistic, expert in my field, and fearless(and maybe a bit crazy)

#5: Post by **another_jim** » January 27th, 2015, 9:26 pm

MpVpRb wrote: What is a good controller?

Should I drill a hole in the boiler for the the temp sensor, or use some sort of pipe fitting?

Should I use the temp controller instead of the pressurestat, or in combination with it?

-- You can replace the pressure stat with a PID controller and use the tap for a well for the TC. This has been done, notably by Ken Fox, who lowered the boiler pressure to around 105C. He was dong it on a Cimbali Junior, so YMMV. Check what Swagelock catalog for the correct fitting. The steam boiler pressure is low enough to be easy to seal; so a simple fitting will work.

-- The Fujis are very good PID controllers and bullet proof enough to live inside an espresso machine. The Tea has an SSR, which it can drive. Their are cheaper ones available on Ebay, designed for ceramic ovens, that work well, but need to be mounted outside. A good solution might be to buy some feet for the Tea, and mount the PID under the chassis, where it is cool. While or at it, you can heatsink the SSR or remount it outside too -- it is prone to failure.

-- Buy one of Eric's E61 TCs that mount in the threaded tap of the group to monitor your temperatures real tine. The group's brew pressure is not trivial, so it pays to get gear that's been proven

http://users.rcn.com/erics/ · #6: Post by **erics** » January 27th, 2015, 10:26 pm

The Tea has an SSR, . . .

Jim - The SSR fitted to the Isomac has a control input of ~120 VAC whereas the typical PID control output for SSR's is ~12 VDC.

Mike - I do have the fittings to accomplish this and you can contact me at the email address below my sig. However, this appears to be your primary problem:

#7: Post by **another_jim** » January 27th, 2015, 10:52 pm

Oops. You can use the contact option on the Fuji, bit that would prevent the precise pulse mod control. A new SSR would be better. It's tru the tea might not be worth the trouble; chances are you'd just convert it to the equivalent of the old single boiler version.