Volumetrics, how accurate are they?

Need help with equipment usage or want to share your latest discovery?

Postby ShelbiRyan » Jan 05, 2019, 6:13 pm

Years ago when I started with my little Breville, I had used volumetrics. But of course being a beginner, I wasn't weighing out my shots. So I can't completely remember how accurate or consistent that machine was. I then moved to a double boiler E61, single dosed in and perfectly weighed my shots out, manually of course. So my accuracy was pretty damn good.

I've now moved to a pressure profiling machine with volumetrics. I've noticed after setting a certain amount of ml output, with all other parameters the same, my output weight can vary as much as 3 grams above or below my original setting. I have a hunch that the pressure profiling may be causing inconsistencies.

My question is, how accurate are other non profiling and profiling machines? The article linked below states that the accuracy of volumetrics using precise basket dosing is pretty bang on....but not using any kind of profiling.

So far my experience is far from those results. I'm trying to narrow down the culprit...

http://www.lamarzoccousa.com/blog/ben-k ... lumetrics/


Postby Znekcihc » Jan 05, 2019, 7:38 pm

I'm curious about this as well. My next dream machine is the LMLM or GS3 and this is the main feature I'm looking forward to.

User avatar

Postby AssafL » Jan 06, 2019, 5:02 am

Flowmeters push out a pulse (or two) per revolution. So the resolution is the volume per pulse. However, and this is where inaccuracies come from, flow isn't linear. So that at the extremes, while the paddle rotates, the volume isn't as the flowmeter claims. So flowmeters have orifices that linearize some section of the pulse-per-flow curve.

What does this mean? Well - there are two main use cases for flow meters:
1. Inter shot repeatability (same dose, same grooming, same packing) - same water and same bev weight.
2. Measuring water volume.

For 1 - as long as you are close to the linear zone, it should be okay. On the GS/3 AV it seems the choice was for a rather high flow flowmeter so it work reasonably well - but not perfectly well.

For 2 - Same GS/3AV Gicar flowmeter was useless for evaluating flow volume for extremely long air purge and Scott Rao Fingering Pause (what we used to call PreInfusion back in the days...). The Gicar was very much non-linear in those ranges and even when using a 2 or 3 step calibration curve proved to be inadequate for the task (my measurements were posted here in other threads).

The Digmesa I use now (kind gift of member Piezo) send order of magnitude more pulses per flow and I can get accuracies down to a single ml over very low flows (30ml/min or so). My PI is now fully based on volume.

Much more on this on Jakes PI thread, on the Chimera thread and on the Flowmeter accuracy threads:
Gicar flowmeter accuracy
My long and rambling path to preinfusion/pressure profiling

Of all the mods I did on the GS/3 AV "Chimera" the flow meter ranks probably second in terms of usefulness (I actually use the Volume and Flow Rate readouts more than anything else). In order of preference:
1. Obviously, using a water debit starved gear pump is by far the best mod... For any number of reasons. I'll never use a rotary vane ever again. An Ulka would be an upgrade...
2. Accurate flowmeter - PI progression metering when combined with manometer.
3. Slayer style bypass comes in a distant third. Over the last month my bypass wiring became loose and I ended up using the metering valve and gear RPM as my sole control and it was sweet. I may bypass the bypass altogether (if it weren't for the effort put into building the damn thing; but it works well and extends the dynamic range of the flow paddle).
4. PP - sucks. pressure profiling sucks since the gear pump is volumetric. Manometer as feedback is cool - but pressure sensor is rather overkill.
Scraping away (slowly) at the tyranny of biases and dogma.


Postby ShelbiRyan » replying to AssafL » Jan 06, 2019, 2:27 pm

Thank you Assaf for your reply. Seeing as my R Nine One is only about a month old, I don't really feel like modding it :wink:

But the longer I play around with this machine the more I realize that measuring water volume with a flow meter is not a great fit for repeatability when pairing them with all sorts of different pressure profiles. Especially long slow pre infusions and basing these shots on output weight. I was hoping someone with a DE1 would chime in on their experiences. As I'm not sure if this inaccuracy is only with this machine or if it's a system or component issues.

You have mentioned that the volumetrics on the GS3 AV are fairly accurate with high flow scenarios, which I'm assuming is a traditional ramp to 9bar shot? If that's the case, what would have been the average margin of error?

Correct me if I'm wrong, but I'm starting to think that gravimetric's may be the only way to accurately repeat pressure and or flow profiled shots? I do understand that there are many variables to this (dose, groom, tamp and even the age of the bean) but let's just say all those variables are the same....

I've read in your attached threads that most people have some variance in their volumetrics from day to day as things change, my issues are shot to shot. Pull one shot and the output is 40 grams, repeat that exact shot and the result is 44 grams. And this is is with a traditional 9bar shot, no fancy profile.

User avatar

Postby AssafL » Jan 06, 2019, 7:06 pm

I didn't have that much variability inter shot wise on the Gicar flowmeter. Repeated pulls measured very similar to each other.

But I was unable to correctly deduce volume of water by counting pulses. With the Digmesa I am able to rather accurately.

Gravimetric is nice but in cup. It tells you little of the PI (which happens mostly prior to first drops).
Scraping away (slowly) at the tyranny of biases and dogma.


Postby ShelbiRyan » Jan 06, 2019, 7:37 pm

AssafL wrote:Gravimetric is nice but in cup. It tells you little of the PI (which happens mostly prior to first drops).

Yes of course, I should of clarified what I was talking about a little better. Gravimetrics would be ideal for shot output targets.

User avatar

Postby AssafL » Jan 07, 2019, 3:19 am

The best is both.

For an inter shot consistency the Gicar should be fine. If the Gravimetric output differs shot-to-shot for the same
Coffee and same prep (dose, groom, tamp), I'd be suspicious of the prep.

In fact I'd say that 99% of the time I second guess myself rather than blame a machine/ component/burr. It takes me a month to come up with the conviction that it isn't me... which it usually is...

As long as the flow rate is sufficient to rotate the propeller, it doesn't matter if it is linear or not. For the same flow pressure/flow profile, same puck density distribution, you'll get the same number of 'ticks' from the flowmeters open collector Hall effect sensor.

If the flow is marginal, meaning the propeller starts and stops, all bets are off. But that is unlikely. I have to get substantially under 20ml/min for the Gicar to stop 'ticking'.

One point though is that coffee is not the same coffee day in and out. In fact, midday it may change (e.g. humidity). So as coffee dries out, the inter shot consistency suffers greatly. It should be accurate for an hour long session at a coffee bar, but all my bets are off if you compare coffees over 24 hours.

For that reason I hooked up an Arduino and counted pulses to try to assess volume of water and that turned out to be difficult. This is after reading the Gicar catalog which wasn't very prescriptive about how they measured (or what they meant) by accuracy.

BTW- it may just be that the only problem with the Gicar as far as linearity is the nozzle size. The espresso machine makers all seem to use the large nozzles (probably avoids clogging) which doesn't help linearity. I spent a few hours trying to find smaller nozzle bodies but wasn't successful.
Scraping away (slowly) at the tyranny of biases and dogma.


Postby ShelbiRyan » Jan 07, 2019, 11:37 pm

AssafL wrote:The best is both.

Agreed, I'd like to see more machines incorporating this. AFAIK Decent seems to be the only home friendly machine doing this.

In fact I'd say that 99% of the time I second guess myself rather than blame a machine/ component/burr. It takes me a month to come up with the conviction that it isn't me... which it usually is...

I would also have to agree with you on this. Like you 99% of the time my inconsistencies are my own doing. Which is why I rarely make posts like this, because shortly after I hit submit I'll get that "duh" moment. But I feel after 563 shots with the machine (I checked the counter) and countless hours of trial and error.....I think I might be in the 1% here.

This morning I went about my daily 2 shot back to back routine. I went a step further and even weighted my EG-1 output, even though I know I didn't have to. So 19.6g in and 19.5g out. The regular blind shaker shake, went a little extra step with some more WDT. ONA level, Kafatek tamp and selected my pre-made "traditional" profile D. Which is set to stop at 72ml giving me an output of 40g. First pull went well. Nice even extraction which resulted in 39.1g in the cup. Second shot was another nice even extraction which yielded 45.6g in the cup.....

So here I am scratching my head again. I've gone as far as reaching out to Rocket about this issue. I don't feel that other users with other volumetric machines have this much variance in shot output. Or maybe I'm wrong.....?

User avatar

Postby AssafL » Jan 08, 2019, 4:22 am

Do you flush before the first pull? It may be that there is some extra volume in the pipe that fill up?

If it sometimes happens in reverse (first being 45 and the second being 39) then obviously this isn't it. But if it always goes up - it may be that.

BTW. This is a >10% error. It should be within a gram (even less). So there is something wrong.

You can connect to the flow meter with a scope or an Arduino and count the pulses. See if during the pull pulses are missing and at what point in the pull.

You can also pull the cap off the flowmeter (if it is a gicar) and see if the is debris or something clogging the nozzle. If you do don't bend the tabs - if they break it is an expensive part.
Scraping away (slowly) at the tyranny of biases and dogma.


Postby michael » Jan 08, 2019, 2:22 pm

I've noticed this as well with the benefit of the acia lunar so I did a few tests, water only no coffee

My water only shots with 8 seconds preinfusion to start are sometimes right on but can vary 2gs, with the occasional 4gs, still not that much, but then my machine is getting old and I'm thinking flow meters have improved

I'm curious what the results would be with a new gs3 8)