Tamper bottoms: Flat or convex?

[HB]

(Images courtesy of Home-Espresso.com)

[ThaRiddla]

I'm a flat kinda guy.

I haven't ventured too much into the world of espresso tampers, but I'm doing some testing on new and experimental tampers now, so I might be changing my mind.




Whatever produces the best results.

[richard_bone]

Even though I don't have one at home and have almost no experience using them, I can't believe that a flat tamper would be better than a slightly convex one.

My reasoning (and it may be flawed), is that you'd have less of a tendency to channel on the edge of the puck because it would be pushed a bit against the wall of the basket...

[malachi]

Honestly, with good technique I've seen no real difference (either observed with the naked portafilter or tasted in the cup).

[PheasantCreek]

A lot of the channeling on the side I have seen and discussed with patrons have come from the tap method of tamping. People tend to tap the side of the portafilter too hard and it shocks the puck from the sides of the portafilter. You can see the shock waves from hitting a half-filled water bottle.

[HB]

I guess that I'm a flip-flopper. My first quality tamper was a flat Reg Barber (tall). The next was a Reg Barber convex (short). Then an Espro (flat). And now Ken Nye's convex:


(Image courtesy of Espressocraft.com)

It may be hairspitting, it may be my imagination, but my preference leans towards convex because side-channeling does seem less common (note that Reg Barber and Espressocraft's are very gently domed). To be fair, the sort of channeling I'm referring to would be visually undetectable without a naked portafilter, i.e., nothing like this disaster:



I'm inclined to attribute a little of this occasional weepness at the edges to distribution faults, since I've noticed it occurs in pretty much the same location -- directly above handle. When I turn the portafilter 90 degrees while dosing, the weepy location changes. I won't swear to it (yet), but the lightly domed tamper bottom seems to reduce the edge-weepiness.

By the way, being a flip-flopper can't be all bad. David Schomer switched sides from 1999 to 2004.

[shadowfax]

I'm thinking about getting a convex Espressocraft tamper to replace my rosewood Al tamper that I need to give to my sister. Are they still not selling them, or do I need to email them to get one ordered?

I'm also a little lost on the sizes (heights). I guess I should go measure my current one...

[framey]

I've got 4 flat and one convex tamper so I guess that puts me in the flat favouring camp

When I asked one of the more respected coffee houses here in Australia (Campos) why they chose convex tampers I was told that your tamper shape should match your basket shape, which in their case (Linea) was slightly convex. I'm not sure I've ever seen a convex basket?
Do baskets change shape with use ie the constant pressure of commercial coffee making?

I'm sure there was a very long and scientifically proven point on CG or Alt.C that a convex tamper will inevitabley place a great deal more pressure in the center of coffee and less toward the edges therefore giving technically an uneven tamp. The convex tamper I have is the Lava Deluxe and it seems to be quite rounded ie strongly convex. Anecdotally I haven't noticed any really difference between the Lava convex and my other flat tampers.

Stephen

[Ken Nye]

I just received an e-mail from Dan regarding this thread....so here are my two cents.

Regarding the flat vs convex issue, I still do not feel that one is truly better than the other. My own personal preference is mild convex. After speaking with many pro-baristi I beleive that the majority prefer convex as well, especially in Europe. If I had to put a number to it, I would say approx. 70% convex/30% flat.

I definitely notice better results with convex when working with triple baskets. They are more prone to channeling at the perimeter than doubles, and I feel that the convex provides a bit of reinforcement in that area. I have experiment with various degrees of convex pistons and feel strongly that a very mild convex with a smooth radius is best. No high points, etc. Currently we offer both flat and mild convex pistons at EC, but if I were to offer only one it would be mild convex.

There are many factors that play a bigger role in the tamping process than the shape of the piston. A level tamp, an even and full distribution of coffee in the filter basket, the distance between the puck and the group screen are all crucial to shot quality. Another often overlooked problem is the fit between tamper piston and filter basket. The majority of baskets that I have used, OEM or aftermarket, are slighty out of round and variances of as much as .5 mm (inside diameter) can be present. I generally like to see .5 - .7mm gap between piston and basket, depending on the amount of taper in the basket sidewall. Needless to say, if the fit is poor, the risk of channeling at the perimeter increases. This FB issue has become a pet peeve of mine, so much so that I have designed my own line of filter baskets with very specific tolerances (they will be out this summer).

I am unaware of any FBs that have a convex bottom when new. In a commercial environment baskets tend to stretch, and may appear convex. When our baskets stretch we replace them. Distortion in the bottom of the FB can cause irregularities in the filter holes which leads to uneven extractions.

Last but not least is handle length, shape, weight and balance. This mostly depends on hand size and tamping technique. We offer four "stock" lengths of handle, all evenly balanced and weighted. The shape that we choose took us over a year to design. Our goal was to design a line of tampers that could accomodate as many hand sizes and accepted tamping techniques as possible...not an easy task. Generally speaking, I find that tampers between 3.25"-3.50" in length suit the majority of people. We keep two sizes on the counter at our espresso bar and this seems to keep all of the baristi happy. It is also nice to go back and forth between sizes, this helps us practice different techniques and avoid too much repetition.

Hope this helps....
Feel free to post or e-mail with questions or input.

Thanks
Ken Nye
Espressocraft, Inc.
Ninth Street Espresso

[Abe Carmeli]

I have not fully tested it but I tend to think that if you use Staub, a flat tamper may be preferable. A convex in such case may produce an uneven "crater" in the middle of the puck

[PeterG]

I prefer flat.

I've always felt...the bottom of my pf basket is flat, so the tamper should be too. I don't want the center of my puck to be thinner than the edges, even slightly.

I also staub tamp, so I have always felt that the convex shape would do some weirdness in the middle, as Abe describes.

Finally, I can't see the convexness creating any lateral pressure, "sealing" the coffee to the sides of the pf. If anything, the opposite is more likely. Think of this scenario: you distribute and level the coffee perfectly in the pf. If you use a convex tamper, the middle would be slightly more compacted than the perimeter, leading to increased channeling there. To be sure, this is pure conjecture. But that's what bulletin boards are for, eh?

Ken's post does have me thinking, though, about engine piston design. All the pistons I have ever seen are slightly convex. And if it's good for my Hemi, it might be good for my Faema.

Peter G

[Ken Nye]

I've got the crew at the shop working with flat for the next week or so, I will report any interesting feedback.

[espressobsessed]

I've noted some seriously flawed thinking within this thread. There are many ways to dose distribute and tamp to achieve one end: a technically flawless shot.

1. First flaw: not choosing the piston to match the showerscreen. If you're runnning a curved showerscreen, you want your tamper to match that curve, so you can maintain that 2mm gap across the top.

2. Some baskets will become curved, some are less prone. The 18g LM basket seems to be the best, the 16g Rancilio - not so much. Curving in the basket will cause problems, but you can compensate for them with your distribution.

3. A curved piston WILL NOT reduce channeling on the perimeter - that is, depending on your dist. technique. If your distribution is perfectly level, then a curve will pack coffee in the centre more densely. That dense coffee will reach it's maximum tamp faster than the outside.

My last swipe in distributing creates a slight divot in the coffee bed. My tamper is the Reg European curve: I tamp to create a bias in the centre: when I watch my extraction at work, a canadian dollar sized brown spot appears in the centre of the bottomless PF, then continues outward. I find this technique lets me run shots longer before blonding. Since switching to the euro curve, I've stopped using the staub.

I should also add, I get the most consistent results when using a hydraulic Rancilio. I've used procon machines - I find them to be less forgiving: flow pulsates, channeling occurs (generally) more often. (yay hydraulic)

Moral of my story: tailor your tamper to your showerscreen (and whatever yields the longer shots and most consistency)

[espressobsessed]

I also highly recommend BURNING your clicker tampers.

If you want them as a learning tool - they're FAR different than regular tampers. The violent action of the click at the end is not natural (and also causes a pressure spike). Use a bathroom scale instead, then you or your baristi can see the pressure progression during the duration of the tamp.

The only reason why clickers are popular: everyone loves a gimmick (usually).

-j

[shadowfax]

when I watch my extraction at work, a canadian dollar sized brown spot appears in the centre of the bottomless PF, then continues outward. I find this technique lets me run shots longer before blonding. Since switching to the euro curve, I've stopped using the staub.

speaking of flawed reasoning....

when your shots start in the center, they almost always blond first in the center. The reason your shots "take longer to blond" is that the outside starts late. then when the center starts to blond and overextract, the outside of the flow is still brown and hides it. This is why your shots appear to blond later.

as for tamping, there's so much variance that I hardly buy any of this. you will buy a tamper, and then adjust your method to it. the convex of the tampers I have seen is minimal anyway... when you distribute, it's not based on getting absolute evenness in terms of density across the puck. You distribute to compensate for how your shots run. in other words, if I notice that my shots are starting on the outside, I will attempt to get more coffee out to the sides as i distribute using stockfleth's. It doesn't matter if you are using a convex or a flat tamp, you will end up compensating with your method, unless you just don't care. a tamper is supposed to provide a consistent surface to pack the coffee. if its curved, then you will need to have more coffee around the edges. And, based on feedback from the bottomless, you will do this whether you know it or not.

I don't know about the shower screen thing. I've never seen a convex shower head, although lots of groupheads have screws, and it's probably good to leave a little extra room in the center for that. but ultimately, I don't think it matters at all. equipped with the feedback of a bottomless portafilter, again, you will come up with a distributing/tamping method that matches your tamping surface to produce even extractions.

Or you just won't improve because you're lazy, and complain about how you wish you had a different type of tamper, because of course, if your espresso sucks, it's your equipment, not your technique, right?

[espressobsessed]

Most e61 hx machines used a curved showerscreen. Maintaining an even distance is based on schomer's methods.

[terryz]

I also highly recommend BURNING your clicker tampers.

If you want them as a learning tool - they're FAR different than regular tampers. The violent action of the click at the end is not natural (and also causes a pressure spike). Use a bathroom scale instead, then you or your baristi can see the pressure progression during the duration of the tamp.

The only reason why clickers are popular: everyone loves a gimmick (usually).

-j

I just set a pile of them on fire! Yikes what a mistake

In real world training, take five people and ask them to tamp at 30-40 pounds. What you get is a wide spectrum of different tamping weight. Click tamps provide initial training to a set weight and that all.

Using one of these devices in a production enviroment is not what they were produced for ( speaking for my own product)

The violent action you refer to must be theory only. There was allot of work behind creating that little bit of kinetic energy and how to stop it before it travels. To be honest this was a concern but after 9 months or so, I have a different theory, and one that will be produced into another "Gimmick" tamp.

A bathroom scale is what brought about the click type tamp. A bathroom scale is intended for wieght measurement of 80+ pounds and tend to be inaccurate. The large spring in a bath scale is unstable until it reaches outside of its free space.

Anyhow, burning clicker tamps is one idea, but it is also flawed as they would be easier to simply melt. Then you could use the raw materials to design and build something better, ,,,.......Well if there was something better Until that time bath scales belong in the bath and click type tamps belong in the hands of those who need them.

Gotta go build the next gimmick, CYA......

[buzzword]

3. A curved piston WILL NOT reduce channeling on the perimeter - that is, depending on your dist. technique. If your distribution is perfectly level, then a curve will pack coffee in the centre more densely. That dense coffee will reach it's maximum tamp faster than the outside.

My last swipe in distributing creates a slight divot in the coffee bed. My tamper is the Reg European curve: I tamp to create a bias in the centre: when I watch my extraction at work, a canadian dollar sized brown spot appears in the centre of the bottomless PF, then continues outward. I find this technique lets me run shots longer before blonding. Since switching to the euro curve, I've stopped using the staub.

For a rookie, for a guy who's tasted good coffee and wants to make it for his family and (lucky) friends, that is very helpful. I posted on CG yesterday, asking advice on choosing the Euro or US curve in a Reg Barber tamper. No answers on CG yet, but this is the kind of word I was hoping for!

Also when obsessed wrote concerning 'clicker' tampers:

The violent action of the click at the end is not natural (and also causes a pressure spike).

it was a useful consideration for me, cause that's all I've used since I moved from my Capresso C1000. I confessed on CG that I thought I might be getting better results with my plastic convex (the last day or two) that came with my Oscar. Not that I'm competent to judge the clicker-tamper myself, but obsessed provides good food for (my) thought.

The Espro tamper I've been using moves pretty smoothly though about 3/16", but then it does kind of pop through a final 1/16 or so.

[malachi]

the clicker is a useful training tool. seriously.
would i use it to actually make shots? no.
but i used it in every training session i've done since it came out.

as for the flat vs convex thing... seriously, it doesn't make any difference that i can discern. i've cupped a whole lot of coffee from all sorts of tampers trying to taste minute changes. i've evaluated tons of shots through a naked portafilter trying to identify altered extraction. so far, there is a lot of theory - but at the end of the day it's what's in the cup that matters and i see no difference.
use what makes you feel good.

personally, i think handle length, balance and weight are all FAR more important that the shape of the bottom of the piston.

[cannonfodder]

I have one of each. Depending on the shape of the shower screen determines which one I use. My Factory lever has a flat screen so I use a flat tamper, my Millennium has a domed shower screen so I use a convex tamper. It just made more sense to me to use a tamper that matches the screen. That way you have an even height between the puck and screen from center to edge.