ciordia9 wrote:1. Greg Lepore advised me to add an additional T and just take the steam's temp and that was "good enough" for calibrating the pid. Which twomartini's told me that it wasn't such a good thought at a cupping.

I probably wouldn't PID an HX in the first place, but if I were going to do it, I would check if a pressure transducer would be a better choice. Sean let me borrow his portable pressure testing kit and it seems reasonable that if it can translate brew pressure into 20 readings per second graphs, controlling steam pressure would be a snap. Steam pressure is more readily measured and not subject to the vagaries of probe placement.

(The engineers / physics majors are welcome to support or shoot down my assertion.)

I had brought up the PID of a HX discussion a few months ago at Barry's gathering of the brains. I was contemplating using a PID in the rebuild of the Faema. I was told that PID and HX machines are not the best option.

It had to do with the PID attempting to correct for carryover. The PID will switch on and off quite rapidly while steaming in an attempt to compensate. What you end up with is not a smooth and stable temperature curve you would associate with a PID brew boiler, but a very ugly looking saw tooth temperature curve. The pressurestat or pressure transducer will give you a flatter temp curve on a HX machine (unless you never plan on steaming).

I did look at transducers, but an $80 Pstat is much cheaper than several hundred for a transducer.

I went forward; whats the worst that can happen?

I've learned what the overpressure value sounds like. I've learned what the boiler emergency pop sounds like. Neat experiences *wry chuckle*

No it's not all that bad. Just a note when you switch an auberin pid from 'f - 'c it uses the old numbers as the new numbers, literally and when you're dialing by a tenths of a degree you will not win the race of the boiler vs countdown. hehehe!

TC is an Omega 3"/K, sticks about 3/4" into the boiler.





Now I get to learn about how this thing is driven, and it's driving me crazy. I followed these rules:

at least 30m sitting
50ml flush
2m wait
pull for 22s
wait 2.7m
next..

I've got to think I can tighten this up. It feels high on the surf side now, each pull can average a 7'f change to the next shot. I'm going to lower the temp a bit more since most of this series has been out of my desired range of 93.8-94.4c. If Ken's right then I'm going to see these bands continue to slide around. I'm trying to visualize these different dynamic systems and I see the boiler being feathered at temp, shot causes new water to enter, temp flutters, should restabilize for another round.. if its not then maybe I need to hand tune the pid away from auto.. but I'm not sure what the shot temperatures are indicative of.. or it could be my time between shots should be longer.. or shorter.. or.. so I ramble.

Thoughts are welcome.

-a

Hi,

Sorry I have not had time to get into this thread beforehand, and have little time now, either. I'm in France for yet another month-long stint, mostly in language school, and have very limited internet access in an unpleasant venue (some of the Mcdonalds here have free WiFi, but since these places are jammed with people you don't want to spend time with most of the day, they are usable only in the morning before the lunch rush, which conflicts with my classes a lot the time).

Anyway, not responding to any particular posts or comments made before, I think that each HEX machine has its own natural temperature "behavior," and differing patterns of recovery. This is to say that my observations on my machines, although perhaps generalizable to a modest extent, are not going to work on other machines without a lot of trial and error. My 10 year old Jr. heats up its boiler water as quickly as does my newer rotary machine, but the response in the Heat exchanger as measured by shot temperatures is MUCH slower. As a result of this and perhaps other factors that I think are due to changes in materials used in the HEX by Cimbali over a 7 year period, I need both a higher boiler temperature (like 7degrees F, I think) in the older machine, combined by a 30 second longer inter-shot interval in order to get temperature stability at the same shot temperatures in both machines.

I have standardized on 50ml flushes in both machines and get good stability AS LONG AS ALL I DO IS TO PULL STRAIGHT SHOTS. If you introduce frothing into the mix, then all bets are off. I wouldn't even bother to check temps in that situation as I'm sure they are all over the map. My usage pattern is that cappas are never made after the first hour or so that the machine is used in the morning, so the machine restabilized after 10 minutes or so and it doesn't matter. For someone using a HEX machine to make a mixture of milk drinks and straight shots, who does not have a fairly long interval between shotmaking (say in excess of 10 minutes) I think that shot temperature stability is unlikely to be obtained.

As to what would work best in an Oscar, I haven't any idea. The PID does allow changes in boiler temps to be made easily, so play around with it, using various flush quantities until you find something that works. I'm assuming you will find something that "works," but of course I can't guarantee that!

Early on, don't bother to graph the results you get (unless you have software that allows that painlessly; I don't). You will become familiar with the raw data and whether it looks "good" without graphing it. Bad results can be erased from the Fluke's memory without bothering to type them out. Once you do find a regimen that "works," then you can start graphing it. Most things you try early on will NOT work; that I know from personal experience. As a general rule, the more you can reduce the boiler temperature, reducing flush size, the more you will be likely to attain stability, if my observations are any indication. How much this is machine-dependant I do not know.

Of course, if you are making a lot of milk drinks, reducing boiler temperature below the point where you can froth efficiently is a PITA, so you can decide that temp. stability isn't that important for milk drinks (my opinion) and you can bump up the temperature easily on a PID when you make milk drinks and reduce it when you make straight shots, allowing time for things to stabilize.

Good luck,

ken

Hi Andy,

I have recently bought an Oscar and am still very pleased with it. I watched your photos and measurements of the oscar before I decided to buy it. They were a great help in my decision, so thanks a lot.

I have one question about your PID-work. Why are you setting the boiler temperature to around 119°C? The steam-pressure of water is approximately 1.92 bar absolute pressure then. That would relate to a pstat-setting of just around 0.9 bar which is far too low for the oscar (you can use it on an isomac). I think 125°C which means a pstat-setting of 1.3 bar would be better for the oscar and should lead to a higher brew-temperature (and thus better steaming performance).

Greetings,
Andreas

DierdreDipstick wrote:Hi Andy,

I have recently bought an Oscar and am still very pleased with it. I watched your photos and measurements of the oscar before I decided to buy it. They were a great help in my decision, so thanks a lot.

Wow, I'm glad I could help and I'm glad you're enjoying the way of the Oscar!

I have one question about your PID-work. Why are you setting the boiler temperature to around 119°C? The steam-pressure of water is approximately 1.92 bar absolute pressure then. That would relate to a pstat-setting of just around 0.9 bar which is far too low for the oscar (you can use it on an isomac). I think 125°C which means a pstat-setting of 1.3 bar would be better for the oscar and should lead to a higher brew-temperature (and thus better steaming performance).

I may or may not be grasping your question, if I miss the mark I blame the morning The boiler temp at 119.3'c equates into brewhead temp of 201-202'f. I was concerned about brewhead temps for the dialing. Since I'm not on a pstat anymore and using the pid I don't set by pressure. I was worried about how steam pressure would be affected but truthfully I haven't noticed a difference at all. I can't quite step back into the map of the pstat as the regulator as it has been a while. I just remember that the outcoming pressure of the oscar at the brewhead was amazing and had to open the flow control nut to get it to come down to 9. The pstat is amazingly stable, just not granular for those quick adjustments.

I'm sure I've muddied that up.. finagle another question and I'll see if I can mash that as well

Hi Andy,

OK, you've had this setup for around 5 months now. The graphs you posted are from some time back and I assume you have generated some more. Obviously, you have left this PID installation in place and one has to assume that you find it on a par or better than your old pstat. Do you find that you frequently use the PID to set boiler temperature as you adjust for different types of drinks and beans? I certainly do.

The crux of my question is to try to figure out how generalizable were my own observations on my Cimbali Juniors, e.g. that one can use a PID and a consistent flushing routine to produce relatively stable shot temperatures in a Heat Exchanger machine. I don't know to what extent you have accomplished that at this point, but I'm sure that you do.

Since most people who read this who own HX machines don't own Cimbali Juniors, the generalizability of the observation might be helpful to others and certainly would be interesting to me.

Finally, reading one of your later posts in this thread, having to do with shot series, have you tried altering the intershot timing as opposed to boiler temperature and flush volume, to try to get relatively consistent shot temperatures in series, as if you had a bunch of people over for espressos? I've found considerable differences when shots are pulled at a rate of one every minute, one every 1.5 minutes, or one every 2 minutes, so standardizing on a production rate, it seems to me, could be a useful strategy in those uncommon home situations where one is pulling a lot of repetitive shots. At the same time, however, very few people (even coming to my house) request straight shots, so this observation may be of limited value in the real world, where one would have milk to soften any of the intershot temperature differences.

Best,

ken

Ken Fox wrote:OK, you've had this setup for around 5 months now. The graphs you posted are from some time back and I assume you have generated some more. Obviously, you have left this PID installation in place and one has to assume that you find it on a par or better than your old pstat. Do you find that you frequently use the PID to set boiler temperature as you adjust for different types of drinks and beans? I certainly do.

I've generated a few but I haven't been focused on it once I achieved results I thought were on par with expectations. As you know testing this stuff out pretty much takes your day and if you screw up a little something you just occupied another day of tests. I trip up and down the temp range only on new arrival of beans. I haven't gotten in to my own espresso blends so I don't get to take advantage of it on personal roasts. As I tour the myriad of other roastmasters though I do take advantage of listening to what they want their blends at and adjust accordingly, and if doing a review or some such will ladder up or down to see just what will occur with the blend.

Ken Fox wrote:The crux of my question is to try to figure out how generalizable were my own observations on my Cimbali Juniors, e.g. that one can use a PID and a consistent flushing routine to produce relatively stable shot temperatures in a Heat Exchanger machine. I don't know to what extent you have accomplished that at this point, but I'm sure that you do.

Since most people who read this who own HX machines don't own Cimbali Juniors, the generalizability of the observation might be helpful to others and certainly would be interesting to me.

I think that your observations and my own show that you can achieve at minimum singular shot stability. Where machines will differ is the amount of time till you can assure a followup shot will meet the same profile again. For the Oscar it takes about 6.5-7.5m before I would trust a shot to have the same profile. Anything less than that and the graphs are predictably out of alignment. Some weird observations were first shot clean, second shot naturally shallow in temp, third shot shows spike in temp with quick falloff. It's got some sort of rubberyness in the struggle to stabilize on shot after shot. As well walkups are so rarely different with the Oscar. I'm not sure if it's due to the way the heat bleeds off it, materials, or what, but I can have a shot in the morning and a shot in the afternoon with little (<1oz) or no flush and I'm where I should be. The only time walkup needs a flush is if you leave the machine running for a good day without pulling a shot, then you'll notice a burp in the line where I hypothesis that water has either boiled in the line and created the gap or perhaps its more simple and the system has just lost the pressure in the line due to inactivity. Regardless a quick flush after a day of nonuse and you're in line again. I don't find myself needing to flush for a duration and watch steaming water stream. I think this is machine/manufacturer dependent.

Ken Fox wrote:Finally, reading one of your later posts in this thread, having to do with shot series, have you tried altering the intershot timing as opposed to boiler temperature and flush volume, to try to get relatively consistent shot temperatures in series, as if you had a bunch of people over for espressos? I've found considerable differences when shots are pulled at a rate of one every minute, one every 1.5 minutes, or one every 2 minutes, so standardizing on a production rate, it seems to me, could be a useful strategy in those uncommon home situations where one is pulling a lot of repetitive shots. At the same time, however, very few people (even coming to my house) request straight shots, so this observation may be of limited value in the real world, where one would have milk to soften any of the intershot temperature differences.

By intershot timing, you mean like waiting the 7m ± for stability? That's the only thing I've found to really work. Pushing the temp up and flushing never showed a reliable multishot series. In effect it actually created more bounce in the series. Chris Keener wants to take some profiles of his machine so maybe I'll bring out the equipment and go back over this for show once I'm done with him. If I recall right though the predictability of it went out the window for me. I have much better luck just dealing with a latency than trying to force stability. These machines can't hope to compete on a professional level of one after the other shots (even then how many pro machines can?). There simply isn't a large enough boiler to take the dynamic shifts. Something is going to under or over compensate with these guys. Luckily between your use of high temp and flush and my learning what the realistic timing game is, it's about as close to perfection as we get until.......

In the real world I have one or two people who take their shots straight, and I put them in the time queue, after they have their drinks I then cycle in quick succession milk drink requests because it really takes a discerning palate to understand what is changing when you've got milk (&|| sugar).

I tell you what though, learning all of this spoiled something for me. Before doing all of this I'd take shots, make lots of drinks, and I was none the wiser. Now that I know I'm finicky on when I want to, and how I, use the machine. Even moreso knowing my grinder and how it likes to trap grinds requires me to clean it for each shot. The entire process/task has become more labor intensive because I understand what it is to achive quality. Kind of weird. Maybe I'll take a mortgage and get a versagrinder and G3! hahahah.

Thanks for the Q's Ken, you reactivated a part of my brain I haven't been using lately heheh.

-a

ciordia9 wrote:By intershot timing, you mean like waiting the 7m ± for stability? That's the only thing I've found to really work. Pushing the temp up and flushing never showed a reliable multishot series. In effect it actually created more bounce in the series. Chris Keener wants to take some profiles of his machine so maybe I'll bring out the equipment and go back over this for show once I'm done with him. If I recall right though the predictability of it went out the window for me. I have much better luck just dealing with a latency than trying to force stability. These machines can't hope to compete on a professional level of one after the other shots (even then how many pro machines can?). There simply isn't a large enough boiler to take the dynamic shifts. Something is going to under or over compensate with these guys. Luckily between your use of high temp and flush and my learning what the realistic timing game is, it's about as close to perfection as we get until.......

In the real world I have one or two people who take their shots straight, and I put them in the time queue, after they have their drinks I then cycle in quick succession milk drink requests because it really takes a discerning palate to understand what is changing when you've got milk (&|| sugar).

I tell you what though, learning all of this spoiled something for me. Before doing all of this I'd take shots, make lots of drinks, and I was none the wiser. Now that I know I'm finicky on when I want to, and how I, use the machine. Even moreso knowing my grinder and how it likes to trap grinds requires me to clean it for each shot. The entire process/task has become more labor intensive because I understand what it is to achieve quality. Kind of weird. Maybe I'll take a mortgage and get a versalab grinder and G3! hahahah.

Thanks for the Q's Ken, you reactivated a part of my brain I haven't been using lately heheh.

-a

By "intershot timing", I'm referring to repetitive shot series, where you would pull 5 or 6 or 7 or however many shots in succession with only enough of a pause to get ready for the next shot, plus perhaps an intentional gap to allow your boiler and HX and element to get back into swing. I have found that if I make shots every 60 seconds (e.g. a shot takes 30 or so seconds, then there is a 30 second gap for cleaning, grinding, and preparing for the next shot) that I get declining temperatures on successive shots. On the other hand, depending on which machine, if I make one shot every 90 seconds or every 120 seconds, I get much more repeatable shot temperatures. Granted, I would have difficulty pulling one shot every 60 seconds without risking channeling and other problems due to bad technique; I've never claimed to be an exceptional barista

In reality, I almost never use my machines like this; it is very unusual for me to pull more than 2 or perhaps 3 shots in a session, before there is at least a 10 minute gap in between them. Nonetheless, it is good to know at what pace you can get your machine to work, full out, while still maintaining more or less the same shot temperatures.

ken