Buyer's Guide to the Quick Mill Alexia - Page 7

Behind the scenes of the site's upcoming equipment reviews.
Beavis

Postby Beavis » Jul 08, 2007, 8:41 pm

I hope & pray that after all of these reviews on ALEXIA are done, you EXPERTS can tell us if we need the PID kit or not to make consistently great shots.
A few of you have suggested this machine is a perfect upgrade from Silvia so I am watching this very closely. Thanks for the hard work!
Thanks, Beavis

User avatar
HB
Admin

Postby HB » Jul 08, 2007, 9:14 pm

As you may recall, I've shared my opinion about PID modifications before. From Thinking of adding a PID to Silvia:

Beavis wrote:Does it greatly enhance my shot quality?

HB wrote:No. It will eliminate temperature surfing, which greatly increases the convenience, but the "absolute" potential espresso quality will remain the same.


That said, Alexia does introduce a factor that a PID'd Silvia lacks: Nearly ruler flat intrashot temperature stability. Although I believe the benefits of hyper-precise temperature control are frequently overstated, if that's your thing, the PID'd Alexia beats the PID'd Silvia on intrashot brew temperature and its forgiveness of minor errors in barista technique, two important considerations for the espresso purist.

I see that the new prices for Rancilo Silvias are already up to $595. If the Alexia price holds, that's $300 extra for:
  • E61 group
  • larger boiler
  • more powerful heating element
  • all polished stainless steel
  • better quality steam (but you have to wait longer for it)
  • huge driptray
  • brew pressure gauge
  • magnetic water level sensor.
That's a nice list of extras, but only you can decide whether it's worth the added cost. I've used the PID'd Alexia for the Titan Grinder Project testing and its fast, accurate brew temperature recovery has made it much easier to compare back-to-back shots. If it were my money, I'd get the PID'd version of the Alexia over stock.
Dan Kehn

User avatar
jesawdy

Postby jesawdy » Jul 08, 2007, 9:54 pm

Beavis wrote:I hope & pray that after all of these reviews on ALEXIA are done, you EXPERTS can tell us if we need the PID kit or not to make CONSISTENTLY great shots.

Emphasis added to the above. If you want CONSISTENCY, I think the PID is a valuable option. That said, the PID does nothing to increase the machine's overall maximum potential. Let me explain why I think that is the case....

HB wrote:Note: This evaluation model is a slightly modified version of the PID kit offered by Jim Gallt / Chris' Coffee. When I did a few quick measurements of the original PID configuration, I noted temperatures would trend upwards. I noticed this same tendency with my PID'd Amica and added a manual cutoff of the heating element to reduce overshoot; I figured it might improve on the Alexia's already impressive performance. I contacted Jim about this and he sent me a cutoff relay for the heating element whenever the pump was engaged. While the difference in intrashot brew temperature is likely beyond the abilities of mere mortals to taste, the cutoff relay has the added benefit of eliminating the "aaaa-eeee-aaaa-eeee" sound of the pump as the PID cycles the heating element on and off during an extraction. At this price point, I believe the enhanced usability and elimination of the pump pressure zig-zap due to heating element cycling is worth it, but Chris/Jim haven't decided on the cost of this option.


OK, so on Dan's machine, when he lifts the levetta lever, the heater is cut out and the PID is doing NOTHING! It is not trying to compensate for incoming cold water, it's not trying to maintain the current set point, it is totally removed from the equation. That means that the intrashot flat profile is coming primarily from the machine's design; the boiler size, grouphead size and temp, water mixing, over (brew) temperature condition of boiler water, etc.

Now, sure I think it helps that the boiler has been sitting at a flat line temperature for some time prior to the lifting the lever. This likely helps control (or at least make consistent) stratification of temperature and eddies in the boiler, and perhaps most importantly control the grouphead temperature.

So on the stock machine, you have the same design, so you should have the same intershot potential. The trick is, you don't know the current boiler temp, and you don't know what the grouphead temperature is. You can only try to develop a routine that improves consistency, and that is developing some sort of temperature surfing routine. The downside to surfing is that you either wait and watch the heater light, or you doing something to force the heater on, use extra water and wait again. (Now Eric's adapter would take care of knowing the grouphead temperature, but would you also need to monitor the boiler?)

On a related note, I have been having a conversation offline with Gary. Gary has pointed me to a thread on CG, where he surfs the boiler and the grouphead temperature with a copper block on an Isomac Zaffiro. See Control of E61 group temperature using copper cooling block.
Jeff Sawdy

User avatar
Marshall

Postby Marshall » Jul 08, 2007, 10:21 pm

Beavis wrote:I hope & pray that after all of these reviews on ALEXIA are done, you EXPERTS can tell us if we need the PID kit or not to make consistently great shots.
A few of you have suggested this machine is a perfect upgrade from Silvia so I am watching this very closely. Thanks for the hard work!



I don't care much about intra-shot temperature stability, as long as the temperature stays within a reasonable range.

The most important advantage of the PID is easy and precise adjustablity. I adjust temperature for each blend (I keep a record of my preferences) and sometimes adjust while the coffee "ages."
Marshall
Los Angeles

User avatar
jesawdy

Postby jesawdy » Jul 10, 2007, 10:02 pm

Quick Mill Alexia - PID Use, Part 1

Image
Quick Mill Alexia - PID controller display; current temp (left), set value (SV) (right)

PID Kit Alexia Usage Instructions

The PID controller included in the Alexia kit is pre-tuned for the Alexia, and a set value (SV) of 230°F. No further tuning or programming* is required. The SV is indicated on the controller by the smaller green numbers on the right of the PID display. The current boiler temperature is indicated by the large red numbers to the left. The PID kit manual indicates that the SV of 230°F should correspond to a brew temperature of 203°F or 204°F for most machines. It also suggests lowering the SV by taste for lower brew temperatures and that actual brew temperature is typically 26°F lower than the SV.

(*- FYI, a full user's manual for the Watlow controller was included in the kit for those that wish to learn more about using and programming the PID controller.)

Here are the recommended PID machine usage instructions included in the kit:
  • Warm machine with portafilter installed for 35 minutes minimum,
  • Pull a warming flush of ~3 ounces,
  • Remove portafilter, wipe dry and build shot,
  • Pull your shot within 1 to 3 minutes of doing the warming flush,
  • Pull a warming flush if the machine has been idle for 5 minutes or more
These instructions are given with the caveat that following this procedure will give the most consistent results. It is also suggested that the warming flush can be skipped with the understanding that shots pulled in near succession will be 1°F to 2°F higher than the first shot.

PID Set Value (SV)

I played with the thermofilter all day on Sunday and pulled 40+ simulated shots (and probably only had 3 to drink :( ).

I spent a lot of time pulling shots at different set values (SV). The setpoint is simply changed by pushing the up/down arrows to the right side of the PID controller interface. My shot spacing varied, but it was never less than a 2 minute recovery between shots (interruptions from little ones sometimes changed my shot spacing :roll:).

I had started out with the idea of using a rather strict procedure:
  • Giving some time for the boiler to stabilize, 10 minutes or so,
  • Pull a 3 ounce flush,
  • Wait 2 minutes,
  • Pull all my shots equally spaced apart at 2 or 3 minutes (from end of previous shot),
  • Measure temp via Scace for a 30 second simulated shot.
As I got into doing this, I was seeing that it didn't really make much of a difference what I did, so I started to use this simpler procedure:
  • Wait 10 minutes between set point changes (sometimes longer),
  • Pull a warming flush of only 1-2 ounce flush between set points or if the machine sat idle for some long period,
  • Pull all my shots at fairly random intervals, never less than 2 minutes apart (from end of previous shot),
  • Before pulling a simulated shot, I dumped the Scace portafilter, ran the pump just long enough for a few drops to come out of the group's dispersion screen,
  • Measure temp via Scace for a 30 second simulated shot.
I did not datalog these shots (I have to be laptop tethered to do that), I just monitored the maximum temperature obtained intrashot. Just as a reminder, my review machine does not have the heater cutoff relay that Dan has installed. What I visually observe on most simulated shots is a rise to a peak temperature fairly early, a small fall of about 0.5-0.8°F, and then usually finish again at or near the same peak temperature at 30 seconds.

The Set Value and Brew Temp (°F) data is shown in the table below. The readings for a given setpoint (SV) are sequential as the shots were pulled, but I did not simply move up the SV in an ascending order. I bounced around a bit up and down for the SV. Glancing at the data I think you will see that any outliers are of ~1°F difference, I think that my slightly simpler procedures should suffice for most people.
    Image
And assuming the SV to brew temperature relationship is linear, I get the the trendline seen in the graph below. This means my PID to brew temperature offset is approximately 22°F
    Image
Temperature Profile

Here is a chart and table for two datalogged shots with a PID SV=223°F. I am using a Scace (v1) thermofilter device and an Extech meter. I did a series of 6 test shots at this set point, these are the only two that were logged. I watched the maximum temperature reached in the series of six shots, in this case:
    202°F, 201.4°F, 201.5°F, 201.5°F, 201°F, 201.3°F

    Image

    Image
In the next installment, I will post intershot performance with a more rigorous testing protocol, as well as some longer shots.
Jeff Sawdy

User avatar
Marshall

Postby Marshall » Jul 10, 2007, 10:29 pm

Thanks, Jeff. I'd be very interested in the effects of varying the warm-up flushes (say from 1/2 to 3 oz.) and then eliminating them altogether.
Marshall

Los Angeles

User avatar
jesawdy

Postby jesawdy » replying to Marshall » Jul 10, 2007, 11:15 pm

Marshall, this graph may show a little of what you ask. For this series, I had a 7 minute idle period before the first shot (recall the PID kit manual suggests a warming flush if idle for 5 minutes or more). IIRC, I did a ~1.5 ounce warming flush 2 minutes prior to the series. For each of these I did a "start dump" of the Scace, "spritzed" the group for just a few drops, locked in the Scace and pulled the simulated shot. The first shot tracks a bit cooler, but only by 0.5 - 0.7°F, the next two shots hug each other pretty nicely.
    Image
    And the data:
    Image
At a minimum, I will try to see what happens with no warming flush from idle.
Jeff Sawdy

Beavis

Postby Beavis » Jul 10, 2007, 11:29 pm

Curious if that 22-23 degree difference from the SV to the group head could be different on different Alexias with different PIDs? Are you all using the exact same Alexia for these tests?
Dan's machine was 24 degrees different...The Watlow kit says 26 degrees.
I know that the PID kit and the range (4 degrees (26-22=4)) probably is nothing to worry about but it goes right back to the same issue of trying to pull different blends at different precisely different temperatures. It's not a perfect world.
Maybe Chris could test each one (since they test the pid anyway) and measure the boiler temp vs. the brew group temp and tell us what our particular difference is on each machine ordered? :roll:
Sorry for the ramble...Thanks for the tedious work!! I think the PID is a no brainer though.
Thanks, Beavis

User avatar
HB
Admin

Postby HB » Jul 10, 2007, 11:42 pm

Beavis wrote:Dan's machine was 24 degrees different...The Watlow kit says 26 degrees.

Pshaw. The position of the thermocouple within the thermowell, amount of thermal grease used, variance among the thermocouples themselves, etc. could easily account for a degree or two. Ignore the number on the display and think in terms of "+2F" and and "-1F" for a given blend, not a raw number. I'll reoffer my advice from Is brew temperature the most difficult part of the equation to figure out?

Beavis wrote:I figure it's all about the temperature now?

HB wrote:I think the difficulty of temperature management is frequently overstated. Even the La Marzocco GS3, which is the current Holy Grail of brew temperature control, didn't magically transform my espresso (see One week with the La Marzocco GS3). In my opinion, getting the correct dose / distribution and diagnosing taste flaws are far far more difficult than temperature surfing whatever piece of espresso equipment you happen to have in front of you. My leading theory on why we obsess over it... because it's a lot easier for engineer types to measure brew temperature than the "quality" of an extraction.

...
I bracket temperatures and don't think about the actual number, other than in terms of cool / medium / medium hot / hot, roughly corresponding to 198, 200, 202, and 204F. In other words, repeatability is the key, not the "absolute" temperature.
Dan Kehn

User avatar
cafeIKE

Postby cafeIKE » Jul 10, 2007, 11:45 pm

It doesn't take many ±'s to get to ±2°F variation between machines.

You have the basic accuracy of the PID and boiler thermocouple, TC location in the boiler, the error due to temperature rise of PID, thermocouple and meter accuracy and calibration, position of the TC in the group, ambient temperature, aiflow and connector integrity all contributing.

What is important it is repeatable for any particular machine.