Bump for subscription to this mad project! I love it.

I don't mean to be arguementative, but I'm not sure I agree with Jacob and Everman. I guess a measurement of the lateral play of the outer burr in a real machine would shut me up, but I think Zassenhaus designs -- all conical burr -- are free to wiggle laterally. That's because when grist is present it only takes a half turn of the axle to self-center the inner burr. So, the only precision bearing is in the axle direction.

Regardless, I'm planning to machine a brass housing along the lines of a zassenhaus design and see how it works (I have access to a mill and lathe). The design will leave open the possibility of a second bearing to lock everything down, as suggested. One question I haven't resolved: how do we judge grinder quality? The only reproducible way is to use a series of test sieves, but that's going to cost $150-ish. So, I'll start with microscope pictures of the grist and see how that looks. But it sure would be great to set up test sieves and get samples from everybody: we could assemble a database of grind distributions! Maybe my wife will buy me test sieves for my birthday.

r

Hi all,

I disappeared the last month to finish my home-made lever machine http://www.home-barista.com/levers/robo-pavoni-learning-to-make-espresso-machine-from-scratch-t4141.html. Now that it's finished, I'm ready to take another look at the grinder project. Thus far, I have some rough drawings. Please forgive the roughness, I'm just learning SolidWorks, but I wanted to share what I have thus far.







The basic innovation is to use a lever with the adjuster. The adjuster itself is 100tpi screw, which is 3 or more times finer than any existing mill, I think. With the lever on the top, the force on the screw is reduced by 3x, while the tpi is effectively 300. This means that 1 turn of the screw should give .003" displacement, about a fourth of the optimum espresso particle diameter. Thus, my grinder will be 4 full rotations of the screw from choking. The screw, of which only the bushing is shown, will have an allen key to give a 4" radius. The fulcrum is the thinned material to the left of the shaft.

The shaft needs further refinement (the Robur burrs are 14mm ID, not quite .500", and the various clip rings and bosses need to be added), and the bottom bearing needs a rework. I'm not sure about the bench: my tests at home suggest that the mill should fit into a housing (fine hardwood?) that clamps to the bench or my espresso machine so that it only takes 1 hand to crank and none to hold the mill. The burr as shown (in the middle of it all) is assumed to have flat faces that top and bottom brackets can reference.

raj

Hi All,

We'll, I finished my Rajjenhaus Robur! Based on consultation with many mechanical engineers, I decided to omit the bottom bearing for the time being (Versalab's omission of a bottom bearing was decisive). The other change relative to the drawings above was to make the flexure out of a second piece of brass, pinned to the body. This was primarily for ease of manufacture. I'm hoping to do a video, but here are stills (the color is off: the body is brass and the burr, at bottom, is steel):



You'll notice that I've used the crank from my Zassenhaus Turkish mill. There are some stainless shims to establish the basic grind, and the 100tpi screw for day-to-day adjustments. My first shots give 1 complete rotation of the screw = 10-15 sec of shot time. The knob is nice for general adjustments, but there's a socket for an allen key which allows maybe 30-50 positions per rotation. Now that's adjustability.

I machined a 1.5 kg stainless tamper a while ago I wanted to see if I could use only the weight of the tamper, with NO additional force. That proved impossible before, but is easy as pie with the new grinder. Now I just rest the tamper in the basket, give a 90 degree turn, and repeat once. Seems to be quite reproducible. You'll notice the robo-pavoni in the background. It now sports an insulated boiler and is controlled to 0.1C. For more on that, see http://www.home-barista.com/levers/robo-pavoni-learning-to-make-espresso-machine-from-scratch-t4141.html



I'll work on money shots, but double ristretto gives 8-10mm crema and an improved flavor over my previous grinder (Zass grain mill). Normal doubles give 6-8mm crema. Flavor is definitely richer than before, more transparent.

I'm thinking about a base so that I don't have to hold the grinder while cranking and a receiver that would allow grind-into-basket with a guide to ensure normality of tamping (ie, the mill would lift out of the base/receiver and the tamp would be dropped in. The funnel would guide the tamper to be vertical). But so far its fine as-is. A grind takes about 100 turns (80-120 is the range thus far).

raj

ps: I got a set of super-jolly burrs a few weeks ago. It would be nice to make a grinder for those too, but I've used up my free time for a while.

You sir, are a wild man- I love your ability to make proto-types-you have to hold it to know what you like and don't like. I did this also when developing my stove and stand designs for Mo-Go-Gear... not to mention the tooling and jigs for production. There just weren't any stores to go buy off the shelf stuff.... Thanks for living outside the box ..

Hi Raj,
That do mean king! Thanks for stretching the limits! on the cheap!
richard

It looks like you have a nice adjuster solution, but perhaps limited in range. I just read this thread today, so this suggestion is a little late, but maybe for the next one?

Some planes (for planing wood) use a differential screw-thread for fine adjustment. Basically, you have a fixed inner screw with an internally threaded shaft from the adjusting knob running on it. This shaft also has external threads of a slightly different pitch which mate with internal threads on the thing-to-be-adjusted.

So the movement is the difference between the thread pitchess. For example, with 3 tpi and 4 tpi you would get 1/3 minus 1/4 or 4/12-3/12=1/12 inch/turn. Clearly, you can get movement about as fine as you could want. Without flimsey super-fine threads.

Hope that's understandable.

Hi Dick,

A good idea. I have a differential screw with 0.5um markings--I thought that might be a bit too extreme for this! It's sitting in my bag of tricks for some other project ....

raj

ps: you are right about the limited range: its about 6 turns = 20mils vertically, enough for espresso, but not the full range of coffee grinds. One mill per brewing method .... (insert LOTR humor here)

Hi Dick,

I was thinking about your differential screw idea. To me, a differential screw is a $400 item that you buy (as it is in optics, in which I was trained). But reading through the lit, your idea is a great one. If I threaded my axle with M4 x 0.7 and had a keyed bushing with M5x0.8, then a simple coupling with M4 and M5 at either end would give 100um per turn (or 1/4-27 and 1/4-28, which gives 1/27"-1/28" ~ 35um/turn). My present design has 250um/turn on the end of the lever (and <100um / turn actual adjustment)--with your design I could achieve the same performance without the lever and with greater total adjustability.

I may have to buy more axle material and try a retrofit. Its a good thing I pinned the lever--it comes right off. Now that you mention it, why doesn't Mazzer do this themselves? It would be pretty easy given how their adjustment mechanism works. Patent it and then sell them the idea!

raj

matadero210 wrote:I'm hoping to do a video, but here are stills <snip>

I'd love to see how you use and adjust it. (video posting help)

matadero210 wrote:Hi Dick,

I was thinking about your differential screw idea. .... If I threaded my axle with M4 x 0.7 and had a keyed bushing with M5x0.8, then a simple coupling with M4 and M5 at either end would give 100um per turn (or 1/4-27 and 1/4-28, which gives 1/27"-1/28" ~ 35um/turn). My present design has 250um/turn on the end of the lever (and <100um / turn actual adjustment)--with your design I could achieve the same performance without the lever and with greater total adjustability.

I may have to buy more axle material and try a retrofit. .....
raj

Actually, it hadn't occurred to me you could use a different arrangement of internal/external threads and moving/locked elements. Cool. (Obvious once pointed out.) This is sort of the planetary gear-set or threads, or maybe the chain-fall of threads.

I'm not sure I fully understand, from your drawings, just what turns within what in your grinder. (I'm somewhat new to espresso and the equipment.) Of course, one burr turns against the other, but which moves and which is held and how?. And how are the burrs held together? Does the upper case latch to lower, etc.? I haven't seen one of the German hand-grinders or taken apart another grinder, so maybe that's what I'm missing. Or perhaps I just haven't read this thread closely enough.

Great stuff!

matadero210 wrote:I ordered the burrs today.
raj

Which ones did you order (71 or 83 mm)?

mathias wrote:Which ones did you order (71 or 83 mm)?

I bought the 71mm, because they were cheap(er) and available. The bigger diameter works against a hand grinder, which is already free of over-heating problems. Increased diameter would increase the torque needed--and its a fairly firm crank at 71. It would be a lot easier if my hand was larger or the mill smaller (71 for the burrs, 74mm is my OD for the mill).

Dick, regarding your question: the outer burr is screwed into the base: stationary. The inner burr is attached to the axle with a 3mm machine key. There is a hand crank on the top to turn the axle. The single plane bearing holds the axle in the center of the housing, and the lever/screw have a thrust bearing to pull up on the center burr.

In other news: I'm thinking about a grinder that uses differential screws. This would be two plates, with flat burrs mounted on each, and an axle and crank. The cool thing is that I can make several different augur/cones to allow use of any sized flat burrs--this would be a universal flat mill. I'm thinking of starting with Mazzer Major burrs, but would also like to get some Macap or others. A different set of screw holes could be used for each set of burrs. The axle would be fixed to the lower plate and lower burr--all that would be stationary. The top plate would have a crank, the top burr, and a single-shot hopper (2cu in), as I did with the conical. A cone would be used on the axle to make sure all the beans are fed to the burrs. The adjustor would be 3/4-27 and 1/2-28 threads

3/4-27 and 1/2-28 = 37.0mils - 35.7mils = 1.3mils/turn = 34um/turn

As you can see, the adjustability will be excellent. This will almost certainly be less work that the conical mill, and I really like the idea of being able to use any flat burrs without change (I have been deciding whether to 'stock up' on Robur burrs, as stupid as that sounds).

Hi all,

I've been using the new mill for 3-4 shots per day for more than a week. Once I got everything dialed in, the unit is very reproducible. Typically from day-to-day, with beans aging and temperature and all that, the first shot of the day may be 5-10 seconds off the ideal. A half turn will get me back to 20-25 seconds. At that length pull, my machine recovers temperature during the shot, so its ready to go immediately. Do others typically need a 5 second adjustment from day-to-day as the roast ages?

The mill has settled into 60-70 turns for 11gms (the Pavoni 45ml 'double'), and I've gotten good at pouring measured beans into the 0.5" hole in the body (small enough that grinding chunks hardly ever fly out). That's about 30seconds of grinding. I use a frothing pitcher that just fits the burrs as receiver, which gets dumped into a dry funnel that just fits the basket. There is no distribution--the dumping of the grinds and a slight shake produce a level surface, and the 1.5kg stainless tamper does the rest (I turn it 2-3 full rotations to settle everything rotationally).

Crema is 6-8mm over several roasts, and >1cm for 14gm ristrettos. I expected more variation from roast-to-roast, but clearly the grinder is just way better than the Zassenhaus grain mill and Turkish mill that I was using. I normally roast in a cast-iron-pot, which gives good mouthfeel and nice, 12 minute profiles to the start of 2nd snap. With a popcorn roast to Vienna (which I normally hate), the body was less, but the taste was way better than the same roast with my old mills. Not that I intend to roast that dark, but it's an indication of much sweeter the shots are.

Stay turned for more on the universal flat mill and videos (my webcam keeps crashing as I record the whole grind/tamp/infuse/press cycle, but I'll figure something out).

raj

You mean like this?



It actually worked, but was so slow I had to put a drill on top.

mteahan wrote:You mean like this?

It actually worked, but was so slow I had to put a drill on top.

Hilarious. Yes, I was afraid this whole thread might turn to a joke if it took too long to grind (or if the torque was too high). 30 seconds does not seem very long--and it never takes more than 90 strokes max when properly adjusted. I think that's the advantage of very large burrs for handmills--more throughput per turn. I'm hoping the Mazzer Major works out similarly.

raj

Kidding aside (though its hard), an electric pepper grinder could actually work. They use conical mills, are adjustable and would grind right into the basket.

I'm almost serious here:




Could be a new thing, like the pop corn popper coffee roaster.

mteahan wrote:Kidding aside (though its hard), an electric pepper grinder could actually work. They use conical mills, are adjustable and would grind right into the basket.

I'm almost serious here:

<image>


Could be a new thing, like the pop corn popper coffee roaster.

Michael,

I think it would only need better burrs--you could take a smaller set of conicals (Baratza?) and sub them in. Espresso is not for millionaires, its for the millions.

On another note: of all the burrs you've handled, which seemed the sharpest and best made? I've been a bit disapponted in the Robur, which has some metal shavings and tool marks on it. The Mazzer mini burrs are very clean and nice--perhaps even sharper than the Robur. Are there any flat or conical burrs that stand out for their workmanship or sharpness--those would be great candidates for me.

thanks,
raj

I even have a pepper grinder with ceramic mills.

Grinds like crap, though.

If the end of the mill is sized right, it could even self tamp like the swift; grind from the bottom and work your way up. Could be interesting. . . .

One of the interesting observations I've made after three years of using second-hand hand grinders is that for three espressos a day the burrs appear to be good for years. Case in point, the grinder I use at work is 50+ years old, and I've been using it for three years. The grind is very even, and its capable of thoroughly choking my lever machine. I'm almost convinced that the grind has become more even over time, and that as the burrs wear they become capable of finer grind. In theory, if one had conical burrs that were just a polished smooth cone and matching outer burr, the grind would be ultra-fine wouldn't it? Think of the sawdust generated by coarse-grit sandpaper, versus that from fine emory cloth.

Unless Robur has manufacturing practices seriously inferior to the German "Stahlmalwherk" burrs from the 1950's, I doubt you'll be needing to stock up on burrs.

I hardly consider 30 s of effort, while enjoying the aroma of freshly-ground beans, to be a hardship. Of course, a longer handle will make the grinding easier.


PC