From a few months back (for those who don't frequent Home-Barista) Interesting discussion on H-B here

 

First Post of discussion thread :
It's holy writ: espresso grinders need to be adjusted frequently and precisely. Holy writ is wrong; instead, you only need to adjust the grinder when you change the dose, the blend, or the style of shot (e.g. from ristretto to normale). All other grinder adjustments are a sign of sub-par technique.

How can this be? We all know that a grinder with too few grind settings will produce inconsistent shots, and we all know any cafe where grind adjustments are not allowed will have espresso that sucks. It turns out that frequent grinder adjustment is what economists and engineers call a 2nd best solution, something you do when the best solution is unavailable:


As the weather changes, as the coffee ages, as the static charges on the grinder wax and wane, the ground coffee becomes fluffier or less fluffy, and also more or less compressible.
These changes affect the density of the prepared puck.
So, if you dose by volume, it will vary the weight of coffee you use from shot to shot.
But the flow rate depends almost entirely on the weight of coffee.
So if you dose by volume, you will see frequent changes in the shot's flow, and have to make frequent grind adjustments.
Since these grind adjustments are retrospective and cannot anticipate how the ground coffee characteristics will change for the next shot; volume dosing will always be jittery, both in shot by shot flow rates, and in the compensating grind adjustments.

If you dose by weight, the jitters go away. Instead, the same dose from the same blend will always get you the same flow. Call this the principle of grind-weight invariance.

How do I know this? I first noticed it when I was doing the TGP. I could pencil in the grind settings on the grinders, come back to them four to five days later, and still be perfectly dialed in. You have to be more precise on the weight and the grind adjustment for smaller burrs or flats than for larger burrs or conical ones; but the principle of grind-weight invariance holds. Since then, for the past two years, I've been weighing doses. I've found, without exception, that when dosing by weight, once a blend is dialed in, it stays dialed in.

Consequently, I believe that the overwhelming reason for inconsistent shots in high end cafes is dosing by volume or grind time. Once they find a way to routinely dose by weight, the consistency of the shots will improve tremendously.
Jim Schulman

Views: 703

Reply to This

Replies to This Discussion

Basically I agree with what you're saying and didn't mean to imply it would be easy. Going to the moon didn't even seem possible let alone easy yet once the mind of man was set to make it happen it happened in short time. (Obviously at great expense:)

But I disagree that most current commercial grinder designs are good candidates for modification. While Mazzers are great work horse durable grinders, every model including Robur E have the same big basic flaw, huge grind throat hang. This alone makes them poor candidates for any type of real grind dose accuracy, if you want fresh grinds accuracy. Have a lull in production and you've got to purge and waste damn near a whole double shot to clear the throat hang. Every Mazzer grinder is based on the stone age grinder usage method of filling the hopper so the horizontal grind path was mute, grinds hanging around in the chute path didn't matter since they were hanging around in the doser anyway. The E models did nothing to eliminate throat hang, just eliminated the doser, the basic grinder design remained the same.

The only major production grinder that appears to deal with the grinds throat hang issue is the Nuova Simonelli Mythos. I say appears because I haven't had the opportunity to use one, going by their parts illustration showing near horizontal versus vertical mounted burr motor which would move the grinds downward from the burrs instead of horizontally through a chute like most other grinders, except the Versalab of course.

IF a dose by weight grinder does ever go into serious R&D whether to weigh the beans pre or post grind could be a consideration. Just as many Home Baristi only put enough beans in the grinder for one pull at a time, enabling back to back pulls of different coffees (something I routinely have done at home for close to a decade), a possible design might be the beans fed from the hopper to weigh chamber, then to the burrs. Any scenario actively weighing be it pre or post grind would add time, but me thinks could be done adding maybe only a second or two, making grinding with a Robur (or Major) take around 6 or 7 instead of 5 seconds for a double. The benefit of highly accurate dosage would out weigh the very small added time to me. In my world 1g dose variance can be huge in the cup, especially dealing with dynamic coffees be they SO or blend.

Would I be interested? Absolutely. Would I be able to afford a bunch of them or even one is another question! And of course that is the question any manufacturer cares about.

Brady said:
Mike McGinness said:
Some say it can't be done. Others find a way and make it happen.
For those that like to poo-poo Home-Barista.com and some of it's regulars you'd best think again. Who do you think invented the Scace Thermofilter? And invented the espresso machine torture test used to certify competition machines including up to the WBC? Wasn't one of the "big boys", nope, a lowly "home barista". Who do you think had WORKING variable pre-infusion and full variable profile pressure control espresso machines BEFORE Slayer or Strada even prototyped? Again a Home Barista. The technology readily exists to build a grinder that doses by weight. A grinder with virtually zero "throat grind hang". Coffee (and countless other comodities) packaging machines do much the same task everyday. It's a matter of either one of the "stuck in the grinder stone ages" manufacturers investing the R&D to make it happen or a new comer to the market. But it won't happen if there's not a perceived demand. Out of hand discounting the concept most certainly won't help make it happen.

Yes Brady, I've mentioned within .5g dose accuracy training. Timed grinding helps, but is a 2nd best solution. It is not as good (accurate) as a dose by weight grinder "could" be.

Maybe its going to take a home barista to make this happen. Heck, maybe that thread will make it happen.

Not to defend the lack of this product, but this is not an insignificant engineering challenge for a grinder. Time-based dosing is probably seen as "close enough" and is a simple plug-and-play solution using off-the-shelf components that does not negatively impact durability, reliability, or (for the most part) price. Remember too that we are a tiny portion of the market, and that concerns like ease of use and price sensitivity are real. They probably already have a dusty one sitting on the R&D shelf but don't feel like we'd buy it.

I've thought about how to do this quite a bit. To achieve a weight-dispensing grinder that performs better than a timer, you need a sensitive scale that is isolated from the vibration of the motor and burrs. This works against the "grinder on a scale" approach. It has to be accurate, which works against the "weigh the beans" approach. The arrangement needs to be rugged enough to hold up to regular bar use, which works against the "weigh the doser" or "weigh the portafilter" approach. It has to be as fast as current technology, which works against the "weigh the basket" approach. It also has to preserve the grind distribution, which works against "weigh into a container" approach. There'd have to be an easy way to calibrate too. Not so easy.

I think the best approach would be one that can be added on to an existing grinder - like the plug-in timers that the uber-geeks were using in the days before the Mazzer E. One of the tinkerers out their could make them, then once people were raving about it online the big boys will pay attention. Maybe a freestanding grounds hopper like the "e", only with a gate at the bottom. The support structure for that unit could have a load cell built in, and it could be easily isolated from the grinder. The gate could be opened over a portafilter to dispense the ground espresso... hmm too bad I don't have much free time... The tricky part is the sparky-bit - the part that reads the load-cell signal, compares to a setpoint, and triggers a relay to shut off the grinder. The rest is just parts...

So I guess I'm not saying that it can't be done - clever engineers and tinkerers can do just about anything. All I'm saying is that we shouldn't underestimate the amount of engineering effort that will be required OR overestimate the net improvement over existing technology that will be realized.

Let's ask - what sort of weight accuracy would you guys be satisfied with? If a "dispense 17.8g every single time" grinder existed, would you be interested?

Good discussion.
Shadow-
Not wanting to be more accurate with your espresso dosing is just not a smart move. A shot with a 17gram load and a shot with an 18 gram load are going to taste quite different to a discerning palate. Also, if you are swiping 1 gram off of the top of your PF every time you distribute and you produce 500 espresso drinks in a day you are looking at more than a pound of coffee just swept straight into the knockbox at the end of everyday.

Is .5g close enough? Probably. But we shouldn't be focused on close enough or good enough. 15 years ago, if we would have stuck with "good enough" we wouldn't have PID temperature stability and for all we know Starbucks could still be "good enough."

Brady-
I'm with you on the thinking here. I thought about it a little more on my drive home after I posted my initial response. Things that passed through my mind were if we can do it on our packaging equipment we should be able to do it on our grinding equipment. A doserless-like grinder that dispensed into some sort of detached scale should work fine. We have packaging equipment that can vibrate like hell to move beans down a tray and then shut off the tray or decrease the dispensing speed as you get closer to the set goal weight. Why can't we do this with grinding equipment on a smaller scale? (no pun intended)

When we press "go" it would start the motor. Once it got within, say... 1 gram of the set point it could slow to just pulsing stop/start until it landed at the set point.

I'm sure there are reasons why this can't work, but my brain is going, so I'm typing... haha.

-bry
Shadow said:
Interested? Not really as I don't think weighing to a .1 gram resolution is all that important. Being within .5-1 gram is plenty close for me.
Well.... I try not to waste any grounds and swipe any excess into the basket and not into waste. Simply do a light tamp to settle everything, then swipe grounds from basket rim onto the top of the puck then do a slight nutation to ensure a good puck/basket seal.

Personally I like what I do to be an artform, in that it isn't identical every single time. Sure I want consistency once I lock onto what is working at that particular time. But I don't want to be robotic in what I do. If that was the case I might as well go super auto and that sure as hell isn't happening.

I'm simply not going to nitpick about .1 grams shot to shot. Rely on what skills you have instead of using a piece of equipment to do that for you. I find my eyeballing to be more than adequate.
I was thinking about this problem while working on my shop's Fetco timed grinders (blogged about it). I'm not much of a machinist, but it seems like there might be a few different ways you could get a machine that grinds by mass.

Does anyone know if there is an advantage to using disc or conical burrs, as opposed to "roller" burrs? It seems like that would allow better vertical feeding.

If you could find a way to get the grounds that are lost on the burrs/throat (technical term? burrstick?) to an absolute minimum, or at an acceptable and consistent volume, you could weigh the beans before grinding/dosing, which would eliminate the vibration problem.
Mike McGinness said:
But I disagree that most current commercial grinder designs are good candidates for modification. While Mazzers are great work horse durable grinders, every model including Robur E have the same big basic flaw, huge grind throat hang. This alone makes them poor candidates for any type of real grind dose accuracy, if you want fresh grinds accuracy. Have a lull in production and you've got to purge and waste damn near a whole double shot to clear the throat hang. Every Mazzer grinder is based on the stone age grinder usage method of filling the hopper so the horizontal grind path was mute, grinds hanging around in the chute path didn't matter since they were hanging around in the doser anyway. The E models did nothing to eliminate throat hang, just eliminated the doser, the basic grinder design remained the same. The only major production grinder that appears to deal with the grinds throat hang issue is the Nuova Simonelli Mythos. I say appears because I haven't had the opportunity to use one, going by their parts illustration showing near horizontal versus vertical mounted burr motor which would move the grinds downward from the burrs instead of horizontally through a chute like most other grinders, except the Versalab of course.


I agree with all of this. All I was suggesting is that a DIY mod that could be applied to a "stone-age" grinder might advance the technology to the point that the manufacturers might pay attention. Your point earlier in the discussion about the innovations advanced by home baristas is well made - many are engineers, inventors, researchers, and tinkerers - just the kind of people that might head down to the workshop and make something like this happen. This kind of grinder advancement will probably happen in someone's basement first.

Your observation is right-on about the Mythos. I thought it was a big step in the right direction, despite the fact that it looks kinda like a slushy machine and comes with an auto-tamper option. The burrs are actually at an angle (around 45 degrees, iirc) from horizontal, which allows beans and grounds to flow nicely. It also has one of the shortest exit chutes I've seen. They did their homework.

Good discussion.
I just replied to this, it was good. It was also deleted... anyway. You have a low profile grinder, around 5 inches tall from counter to bottom of hopper, in front sits a scale. scale is connected to grinder via usb or ir or telepathy or something. weighs portafilter, resets, sends signal to grinder. Starts grinding. Compensates for grounds still in free fall, tells grinder to stop. Done. There will be an iphone/blackberry/droid app which will allow you to program the weight you desire. Gonna keep this one short because I've got to get to the patent office. This will be out in 3 years and cost $1,499 and will be called "TIMED GRIND".
Shadow said:
Interested? Not really as I don't think weighing to a .1 gram resolution is all that important. Being within .5-1 gram is plenty close for me.

As one of those HOME baristi (but one who is familiar with a production barista's environment and difficulties), lemme tell you that half a gram can equal up to seven or eight seconds of extraction, depending on the bean and the grind. Four or five, most commonly, but even that will have a discernible impact on the resulting shot. I tend to measure out the morning's dose right onto a scale, just because I have on right there (and no line waiting for me, thank you very much) and I'm measuring out withing .2-.3 just by eye more often than not.
.5 is a pretty big change-up, and I'm *not* as finicky as some of those H-B'ers.


Brady said:
The burrs are actually at an angle (around 45 degrees, iirc) from horizontal, which allows beans and grounds to flow nicely. It also has one of the shortest exit chutes I've seen. They did their homework.

Ehm, Krups actually had the idea long ago, and put it into one of their first lil burr grinders. I still have one,a and I got it before I ever knew about a coffee community of any sort. It's the Krups Coffina 223-A. Check out the Mr. Fusion in 'Back to the Future".
Yeah of course half a gram CAN make a difference (never seen 7-8 seconds difference regardless of the bean/grind/etc), but in a true commercial setting eyeballing works just fine. No luxury of weighing the doses, so you just do the best you can to keep things consistent in order to achieve a certain standard and simply go with it.

If the barista is in tune with the equipment you just get a feel for what the dose should be. Remember there should be somewhat of an art to espresso (or I believe so anyway) and if every single extraction were flawless the human element would be removed from the experience.
Minor quibbles:

Shadow said:
never seen 7-8 seconds difference regardless of the bean/grind/etc), but in a true commercial setting eyeballing works just fine. No luxury of weighing the doses, so you just do the best you can to keep things consistent in order to achieve a certain standard and simply go with it.

"Can be". I used those terms because It's happened to me, and then suggested that four or five are far more common. In a true commercial setting, whatever technique you use should get you closer than half a gram, and four or five seconds away from your target.


Shadow said:
... there should be... ... an art to espresso ... if every single extraction were flawless the human element would be removed from the experience.

Art is a communicative form. You meant 'craft' Art is when you use your craft to tell people something.
And the human element is what makes a decent lever pull so much more 'flawless' than a super-auto. If every single pull were flawless, we'd spend a whole lot less time in these coffee community fora! ; >

Reply to Discussion

RSS

Barista Exchange Partners

Barista Exchange Friends

Keep Barista Exchange Free

Are you enjoying Barista Exchange? Is it helping you promote your business and helping you network in this great industry? Donate today to keep it free to all members. Supporters can join the "Supporters Group" with a donation. Thanks!

Clicky Web Analytics

© 2024   Created by Matt Milletto.   Powered by

Badges  |  Report an Issue  |  Terms of Service