Is pressure pulse from a PID detrimental to shot quality?

[russell]

I just don't have the tasting experience to judge this one myself. The flashing on/off of the heater at a 1 second interval under PID control is enough to affect the kitchen lights and you can see the needle on a pressure gauge during brew bobbing up and down by 1/2 bar. Somone posted that this doesn't matter, but Dan Kehn put a heater on/off switch on his PIDed Amica and turned it off during the brew (maybe not for this reason).

I'd like to hear from some of the experienced CGs on this one. It's not hard to turn off the heater during brew time if it might improve quality.

Russell

[russell]

Nobody has ideas on this? I went ahead and did a test that does show that the pulse makes a difference, if not to the taste, to the volume.

Running water at brew temp through the grouphead into a measuring cup, I got the following results:

Silvia:
30 seconds with heater under PID control: 12 oz.
30 seconds with heater turned off: 12.3 oz.
A difference, but not very large.

Elaine:
30 seconds with heater under PID control: 8.3 oz.
30 seconds with heater turned off: 9.3 oz.
That's a big difference. It must be related to the functioning of the gicleur and the rest of the complex E61 plumbing.

I ran two 27 second shots through the Eliane. They were really finely ground for the tamp pressure, the coffee (with mucho crema) just sort of oozed out, but the difference was still there, although less:
27 seconds with heater under PID control: 0.9 oz.
27 seconds with heater off: 1.1 oz.
Percentage-wise, that's a substantial difference.

Also I think the shot with the heater off was better, but I'm too inexperienced in tasting to have much confidence in that judgment.

Surely some on you have views on this.

Russell

[HB]

My motivation for cutting off the PID control to the heating element while pulling a shot was to decrease the temperature overshoot / recovery time for the next shot. I mounted the TC in the boiler well, which is not far from the water inlet. Once the pump started, the registered temperature would plummet. By clicking off the PID during this period, the boiler temperature had a few moments to stabilize and then the PID quickly returned to the target brew temperature.

I never tried tasting side-by-side a "pulsing PID" extraction versus one that wasn't. Your results make sense though, i.e., some loss of volume for the former case. It also seems intuitive that pressure drops at one second intervals wouldn't help the extraction, but only testing would tell...

[russell]

One comment on Dan's remark: I was talking with one of the tech reps at TTI Global yesterday about the return of a defective thermocouple and I asked him about when to autotune: when the PID has stabilized at the brew pressure or when the temp is much lower and climbing rapidly--e.g. after pulling a shot. He said autotuning should definitely be done at a much lower temp when the temp is climbing. (He appeared to be the most knowledge of the three consultants I've talked with there.) I reduced the temp on the Silvia to under 190F then started autotune. The overshoot on warm-up was reduced to 1.7F which was better than before. With this kind of temp stabilization, I don't thinking there is any reason to turn the switch off for temp control regardless of where the probe is located. But it's still an open question if it helps the brew. It seems unlikely it would make a discernible difference for the Silvia as the small volume difference indicates. But I think (or imagine?) I can actually see the thin coffee channel coming out of the PF pulsate on the E61 machine, and that hardly inspires confidence.

Russell

[russell]

I'm not sure anyone, other than Dan perhaps, is interested. but I just checked what happens to the temperature (i.e. PID displayed temp) during shots with the heater on and with the heater off. Here are the results:

Eliane:
Starting temp (both tests): 228.2 (set value was 228F)
Temp at end of 27 second shot: 214.4 (heater on) 211.1 (heater off)
Lowest post-pull temp/sec after start of shot: 214.2/30 (on) 208.3/42 (off)
Time after start of shot to recover to 228: 1:11 (on) 1:54 (off)
Overshoot, temp/time after start of shot: 228.4/1:30 (on) 228.7/2:20 (off)

Silvia:
Starting temp (both tests): 223.2 (set value was 223F)
Temp at end of 27 second shot: 215.5 (heater on) 215 (off)
Lowest post-pull temp/sec after start of shot: 212.8/50 (on) 210.3/57(off)
Time after start of shot to recover to 223: 2:10 (on) 2:18 (off)
Overshoot, temp/time after start of shot: 224.2/2:58 (on) 224.7/2:57(off)

In all four cases 17 grams of coffee ground at 7.5 over true zero in the Rocky were used. The only other significant difference is that the Silvia produced about 1.5 oz of coffee and the Eliane 2 oz. In each case, turning the heater off resulted in a slight (10%) increase in the amount of coffee produced. I have no idea why the Eliane produced more coffee than the Silvia. I used the stock double baskets and stock PFs in each case. Pressures were approximately equal, about 9.5 bar.

My conclusion: turning the heater off during the shot has no significant effect on temperature other than delaying the return to the set value (43 seconds for the Eliane, 8 seconds for the Silvia. So the decision whether to turn off PID control on the heater has to rest on the effect on the quality and volume of the coffee produced.

Russell

[tom_b]

For my PID Silvia, I turn the steam button on a few seconds prior to starting the shot, then turn it off ~25 sec into the shot, usually ending ~30 sec. I don't like the pulsing, associate it with early blonding, although I haven't quantified it. The practice is also intended to reduce temp drop during the shot, but my li'l bit of testing along those lines showed little if any difference - still 6-8F. It will cause a temp rise from the midpoint of the pour, so I cut it off before the end to stay below start temp.
tom:)

[AndyS]

One comment on Dan's remark: I was talking with one of the tech reps at TTI Global yesterday about the return of a defective thermocouple and I asked him about when to autotune: when the PID has stabilized at the brew pressure or when the temp is much lower and climbing rapidly--e.g. after pulling a shot. He said autotuning should definitely be done at a much lower temp when the temp is climbing. (He appeared to be the most knowledge of the three consultants I've talked with there.)

Can't figure out why he'd say that. Do you know his name? Dwight? Jason?

[another_jim]

If the lights are dimming when the PID pulses, you need to find a better wired outlet. If the other appliances are fine, and the pump pulses, replace the power cord and the wiring to the boiler's element with 12 or 14 gauge wire. Italian machines are built for 220, and use 18 gauge wire internally -- this is legal in the US, but bad practice, and a non-starter when one adds switching controls.

If rewiring the outlet (for the dimming bulb) is impossible, consider using an SCR and a 4 to 20 mA controller. This will not pulse, but works like a controllable dimmer switch.

Single boiler machines are piss poor for classic PID control, since the warmup time is a lot faster than the cool down time, and PID controllers are symmetric. I've played with the Silvia, and there are several possibilities:

1. Autotuning at any setpoint will produce the least overshoot, but very slow settling times (in essence, the autotune seems to use the cool down part of the cycle to set the parameters). The autotuners all use algorithms derived from the ultra-stable TLC tuning method

2. Classical tuning (e.g. ZN) will produce repeated overshoots but faster settling. You can get to this simply by narrowing the proportional band about 20% from the autotune, and raising the D parameter to 1/4 of the I setting

3. Unstable tuning by narrowing the proportional band to about 5F to 8F, setting the I to zero, and the D very high (to about 30 to 60 seconds, rather than the 10 to 20 the autotune will select). This is essentially an improved form of on/off control using the D parameter to cut the deadband in half from what simple on/off would produce. Greg Scace hit on this, and it's worth considering if you want to pull lots of shots in a row. Given the vaguaries inside the boiler, I doubt this form of control leads to any less repeatability in shots than classical control. However, I haven't tested this.