Hello:

Update ...

erics wrote: .... opt for a copper sealing washer (or teflon washer) where the screw meets the grouphead ...

You were right on the dot there.
It did not leak (not even a wee bit) for the longest time.
But one day, while doing a back flush I looked and noticed a very small drop of water starting to form.
Teflon used as per your suggestion and problem solved.

erics wrote: Measuring the temperature at a location closer to the point of dispersion as Ken suggested ...

Right again ...

Ken Fox wrote: .... interested to see how useful the different measurement points end up being as regards shot temperature management, which is the only thing that really matters ...
... can see a real potential benefit to the shorter probe, but not having tested it in actual use this remains merely my unproven, hopefully educated, guess.

Quite so ...
And a very well educated guess it was. =-)
Both you and Erics were right about the location of the probe being better right above the dispersion point.

Unfortunately, I am unable to find my notes with the hard data (ie: multiple scrawls on various pieces of paper while under the influence of caffeine) but what I can do is attest to the usefulness of a probe inside a well ending above the centre of the shower screen.



With the probe inside a deeper well (ie: where it intercepted the water coming from the HX) the readings fluctuated sharply, both up and down as the brew water passed through it and as a result, the temperature of the brew head itself (ie: what we really need to know) was something that could not be monitored reliably.

Now, with the probe located above the centre of the shower screen, this did not happen. Of course, the readings did fluctuate as water was pumped through the brew-head but this fluctuation was smooth and not sharp or abrupt. It is to note that it's magnitude decreased (but did not cease) till it reached a point around the value set by the pressurestat as the Cimbali's 'body temperature' stabilised.

I have the idea that the use of melamine foam (or any other thermal insulation) to cover the top of the brew head should keep heat loss to a minimum and help this process become smoother still. Some sort of heat loss protection to the boiler should also help, but there more than anything by keeping pstat cycling down, and saving energy, of course.

Cheers,

CIV

civ wrote:Hello:

Update ...


Quite so ...
And a very well educated guess it was. =-)
Both you and Erics were right about the location of the probe being better right above the dispersion point.

Cheers,

CIV

Thanks for the update. What would be hugely interesting would be to use a dual channel datalogger with inputs from Eric's adapter and from a Scace device. One could then determine the best strategy for using the information from Eric's adapter.

I can think of 3 possible strategies; I'm sure there are more:

(1) It might turn out that the idle temperature of the TC probe in the group head, just above the group screw, gives enough information by itself to devise a flushing strategy (e.g. to tell you how much water to flush before making a shot, in order to get a desired shot curve temperature result).

(2) Another possibility is that you could use the real-time readings from Eric's adapter to enable reaching a "flush to" temperature, following which you would know what your resulting shot temperature would be. This would be more similar to what most of Eric's "customers" have used their adapters for, in other machines.

(3) If the boiler was PID'd (and hence could have its boiler temperature easily changed), one could take account of the boiler temperature AND the group head temperature, at idle, which could tell one what the shot temperature would be with a standard volume flush. If this temperature was not what was wanted, one could then adjust the boiler temperature up or down to get the desired shot temperature.

I'm not assuming that you have a dual channel datalogger, or a Scace Device, but rather musing about how cool it would be if you had one, and if you were inclined to do this sort of study.

Perhaps I should contact Eric about this and get another adapter probe from him, to retrofit into one of my Juniors. IF I could find the time to test it, I do have a datalogger and Scace so I could test the theory. The problem is the time, and since I'm not using either of my Juniors at the moment, it would be a real time sink to do so.

ken

Thanks for the update. Is the PID integrated with the grouphead probe?

Hello:

Ken Fox wrote: Thanks for the update.

You're welcome.
Noblesse oblige.

Ken Fox wrote: ... to use a dual channel datalogger with inputs from Eric's adapter and from a Scace device.

Quite so ...

Ken Fox wrote: (1) ... the idle temperature of the TC probe in the group head ... ... gives enough information by itself to devise a flushing strategy (e.g. to tell you how much water to flush before making a shot ...

With this alternative we may be able to do without a data logger and a Scace device. The existing digital thermometer probe with another (naked) identical used in an 'over the lip' fashion or even inside a properly insulated styrofoam cup will get us close. The thing is 'how' close and that is probably something for the palate to decide to then adjust the temperature parameters accordingly.

Ken Fox wrote: (2) ... use the real-time readings from Eric's adapter to enable reaching a "flush to" temperature, following which you would know what your resulting shot temperature would be ...

This is more or less how I am using it now.
I have to add a stop watch and come up with a way to hack the brewing circuit to be able to start/stop on demand. Eventually, I guess.

Ken Fox wrote: (3) If the boiler was PID'd (and hence could have its boiler temperature easily changed), one could take account of the boiler temperature AND the group head temperature, at idle, which could tell one what the shot temperature would be with a standard volume flush. If this temperature was not what was wanted, one could then adjust the boiler temperature up or down to get the desired shot temperature.

It is the one I like best.
The best and most accurate would be using a Scace device and the datalogger but could also be done, albeit with less accuracy, using the PID's output and the thermometer I am now using in the brew head.

Ken Fox wrote: ... that you have a dual channel datalogger, or a Scace Device ...

Unfortunately not.
Yes, it would be way cool [ 8-D! ] to have at least a single channel datalogger, but I what I have seen locally is quite expensive. For the time being I'll have to make do with the digital thermometer and my (now) malfunctioning Sanyou TA-4, provided I can manage to get it re-calibrated properly. =-/

Ken Fox wrote: ... get another adapter probe from him, to retrofit ... IF I could find the time to test it, I do have a datalogger and Scace so I could test the theory.

Yes, it is a time sink, no doubt. I may be able to (eventually) manage to do something using what I have at hand and a stopwatch but it is a 'may' with no deadline.

In any case, methinks that there are a number of factors that conspire against coming up with a 'universal' protocol for brewing with a Cimbali Junior. Water temperature (cold plumbed in water is not the same as warm water in the reservoir), ambient temperature and even pump type and pressure are a few I can name and I'm sure there are a few more.

I think that the best any of us can do is observe the variables closely and come up with a way to prepare the machine to behave as predictably as possible in the face of varying external conditions. eg: I am planning to use melamine foam to insulate the top of the brew-head up to where it meets the boiler and add some thick fibreglass weave to the underside of the cover that goes over it. That's a big hunk of brass that will stop losing heat as it is doing now.

The same will be done to the plate that shields the back of the boiler and I am still studying a way to insulate the boiler without taking the machine apart. I'm relatively confident that this will help in keeping brew head and boiler temperatures stable, once the machine is, of course.

Cheers,

CIV

Junior wrote:Thanks for the update. Is the PID integrated with the grouphead probe?

I assume you are asking about my machines, being as CIV does not appear to have a PID installed on his.

The PID in my machines is controlled by a probe located in the boiler. I believe there are people who have modified their espresso machines to control the temperature in the grouphead, although generally have also PID'd the boiler. In order to do this in the grouphead you would need a separate heat source in the grouphead, and either a 2 channel PID or two separate PIDs.

To try to control the shot temperature in a HX machine, by using a probe in the grouphead, to control the boiler element, is probably not a good idea. In that case one would be using the temperature probe "information" to control another process, e.g. the boiler temperature. Second-order control would be likely to have bizarre results that are not correlated with what you really want to control in this situation. As an example using a Cimbali Junior, the group head temperature results from factors such as convection heat from the contiguous boiler and heat exchanger, plus heat loss to the ambient surroundings; it does not have its own heat source. During the start up phase, the grouphead temperature is going to significantly lag behind the temperature rise in the boiler until equilibration sets in. The result would be that you would likely blow the pop valve in the boiler during the start up phase, because the head won't get hot enough, fast enough, to shut off the boiler element before this occurs.

Granted, someone could probably either program a PID to allow for this or add in other complexity, but in the end you would be unlikely to get better shot temperature control without a whole lot of experimentation and time expended. I think you would want to do things as simply and as straightforwardly as possible, in order to get the desired result (tight shot temperature control). We have some resident geniuses here and it would not be beyond their capabilities to design a system that could do these things, but I doubt they would consider the use of time necessary to be worth the possible benefits.

ken

Thanks Ken for the well thought-out response. I guess I was too quick to wonder about the use of group head data for automatically adjusting temperature; however, of course that would be impractical in adjusting the boiler temperature on the fly as the sole data point.

I would be curious about seeing the data collected from the tc probe on CIV's unit and conclusions for flush regimes. CIV's machine is 220v correct? Would this impact any element of temperature management (pressure is pressure in the boiler regardless of how it got there) other than recovery time?

Hello:

Junior wrote: Thanks for the update.

You're welcome ...

Junior wrote: Is the PID integrated with the grouphead probe?

No ...
Everything is ready (see the photos of the port and SSR mount on pg.1) to monitor boiler temperature.
But the Pt100 probe went south and as a result the Sanyou TA-4 (inexpensive, but avoid it) now needs recalibration. As a result of this, there's no functioning PID installed at this time. =^/

Ken Fox wrote: Second-order control would be likely to have bizarre results that are not correlated with what you really want to control in this situation.

Very well explained.
Far too many variables and quite complex to get working.
I think it could even be dangerous to try.

Junior wrote: ... data collected from the tc probe on CIV's unit and conclusions for flush regimes.

Eventually I may come up with some data, once I was able to say that it was something 'repeatable'.
But then, I'd be talking about 'this' Cimbali in 'a' particular situation and 'a' specific set of additional variables. I can't but wonder about how useful the hard data could be to anyone else.

Junior wrote: CIV's machine is 220v correct?

Yes.
A 220V, 1800W D/1, plumbed in, with vibe pump and doser.

Junior wrote: Would this impact any element ... ... other than recovery time?

No. As you say, pressure is pressure but then the same applies to wattage. ie: 1800W are 1800W here in AR and in Iceland. 8^D

The difference is in the construction of the heating element and the (probably) dedicated 20A circuit you will need to use if everything in your house is 110v.

1800W @ 220v = 8.18 A
1800W @ 110v = 16.36 A

Basically recovery time will have to do with a set of environmental variables that will affect it such as ambient temperature and mains water temperature, among them.

Cheers,

CIV