Now that I have a temperature stable machine, I've been finding that small changes in temperature can have a dramatic effect on flow. Nothing new in that observation, although I was finding the magnitude of the effect difficult to predict. However, looking back at a sequence of shots of yesterday it all became a bit clearer - increasing temperature was decreasing the eventual steady-state flow rate (15secs onwards) but having little effect on the initial flow (8-15secs). In fact, the final flow seemed to depend only on the temperature and the initial flow only on grind, within the narrow band I was looking at.

I routinely log flow, and three representative shots are shown below (this is flow into the boiler). Initial debit is ~60ml/10sec, set by needle valve. Black shot was a bit flat, so I reduced brew temperature by 1 deg C to get the blue, and flow sped up by about 4 secs (stopped it early, and didn't taste), I then ground finer to get the red - which was very tasty, and 19% extraction yield (hadn't measured the others). Comparing the red to black, the initial flow is slower, but the final flow is faster, and matches that of the more coarsely ground shot at the same temperature. That is, the red and blue end up parallel, with the same slope. I did run a couple more shots to confirm the effect (and finish the bag). All shots dosed at 18g in a VST basket, 90ml corresponds to about 36g in the cup. Coffee was a Nicaraguan SO from Hasbean, 7 days post roast, and machine a heavily modified Gaggia with two boilers (one for preheating) a cartridge heater in the group and three pid loops. Grinder is a Nuova Simonelli 75mm flat. Anyway, search didn't throw much up - so I was just wondering if others have made similar, or conflicting, observations.

During temperature testing in the Favorite Espresso Blends study, I have noticed that minor grind adjustments are needed to maintain consistent flow. But that's over a much wider temperature range (typically 88-94C). Your flow rate is almost twice as high as mine (80-90ml in 25s); perhaps that magnifies the effect. You also mentioned "finishing the bag" - hopefully you maintained a consistent bean load in the hopper, otherwise that could easily account for the faster flow rate.

So yes, I've noticed an effect of temperature on flow, but not a dramatic one. I'll try to keep better notes next time I do a round of temperature testing. Thanks for pointing this out.

RapidCoffee wrote:During temperature testing in the Favorite Espresso Blends study

Thanks for the reply, I did have that study in mind.

RapidCoffee wrote:Your flow rate is almost twice as high as mine (80-90ml in 25s); perhaps that magnifies the effect.

That's 80-90ml of water into the boiler - of which only a little over 30ml ends up in the cup. Most ends up in the drip tray or stays in the puck. The graphs give a rough approximation of flow into the cup from around 8-10 secs onwards (i.e. starting slowly then building up).

RapidCoffee wrote:You also mentioned "finishing the bag" - hopefully you maintained a consistent bean load in the hopper, otherwise that could easily account for the faster flow rate.

This was single dosing. I did originally think that this might have been the problem, so repeated a couple of these shots to confirm. I have also previously compared single dosing vs loaded hopper (at one temperature) with this equipment and found that with the grind adjusted to compensate for the change in flow then everything else stays the same (i.e. flow vs time, extraction yield - and taste).

I'm sure the effect isn't always this dramatic, but what I found striking was that changing the temperature and then the grind to get back to the same average flow resulted in a markedly different flow characteristic.

GlennV wrote:...That's 80-90ml of water into the boiler - of which only a little over 30ml ends up in the cup...

That is very interesting data. How do you correlate flow rate through the puck with flow rate into the boiler, with presumably varying resistance in the puck?

-Ed

decaf_Ed wrote:That is very interesting data. How do you correlate flow rate through the puck with flow rate into the boiler, with presumably varying resistance in the puck?

-Ed

Water is incompressible, so once the air in the system is compressed then flow into the boiler equals flow into the group which pretty much equals flow into the puck, this is the case from around 8-9 seconds on in these graphs. There's about 50ml of trapped air here, which gets compressed to 5ml or less so at 9 bar and increased temperature. Even if the pressure then drops to 7 bar at the end of the extraction this only increases by 1ml or so, an insignificant error. What would be trickier (but both interesting and, I think, doable) is estimating the amount of water that's already gone into the puck before the system is fully pressurised.

The 30ml+ number is simply from weighing the cup - 36g of brew at ~10% tds contains ~32.5g of water which had a volume of ~32.5ml before it was heated.

That's too bad. I thought you were graphing actual flow rates through the puck, which is IMHO much more interesting than the system behavior of your one-off Gaggia.

If you pursue this, I suggest that you simply measure the extracted beverage weight after a fixed amount of time, changing only brew temperature (not dose, grind, or anything else). That should remove most confounding factors.

Since the taste changes with temperature, the extraction must change as well. Whether this amounts to a measurable difference in flow rate (and beverage weight) is still unclear.