shawndo wrote:I've also been trying the WDT, since it's really difficult to distribute with such low doses without a needle.

shawndo wrote:A couple of Robur E shots. One with WDT and the other without.

Your technique looks good. The first video shows clear signs of a donut extraction, so you might try downdosing slightly. The second video looks much better. Both shots appear to blond early, but perhaps that's the lighting.

Are you are using a flat tamper? In my experience, convex tampers make it easier to get a "pretty" bottomless pour.

Note: In the "WDT" video, you appear to be leveling the top of the puck with a brush. The WDT uses a dissecting needle (or related implement) to stir the grinds. I doubt this will make much difference with your high end grinders, but it's worth a try.

It's a little hard to tell from the video, but the needle end of the tool you're using to WDT looks really short. I think you want the needle end to be long enough to hit the bottom of the basket. Something like what the person in this youtube video is demonstrating: http://www.youtube.com/watch?v=lICr7GZo ... ata_player

I asked for the Robur videos to make sure the problem isn't with your Versalab grinder. Although the flow looks slightly better with the Robur, and I didn't see any spritzes, the pull isn't all that different from the Versalab.

Shawn, there are two fundamental problems with your shots: 1) the flow rate is way too fast: it goes blond only about halfway through the pull, and 2) you're getting major halo-extraction/side-channeling. Although multiple streams are OK, the center area of the screen remaining uncovered through a significant portion of the pour isn't OK.

Significant blonding occurs at the edges about 15 seconds into the pour, and the shot is going full blond in about 20 seconds. You're letting it run almost white-blond for a very long time. I would think the shot tastes bitter, weak or both. Although I'm a proponent of pulling a target beverage weight, and feel that the "blonding point" can be very subjective, the fact remains that color and surface tension of the stream are very important indicators that you're nearing the ideal extraction point. The target weight must be achieved before the shot goes appreciably blond, or at least within a few seconds either way depending on the coffee and your preferences. Unless John is right about the lighting, all of your videos show shots that run 10-15 seconds past the point when they should be cut.

I don't believe either of these problems has anything to do with your machine, lack of head space, variable puck expansion, etc. It could be, however, that the coarseness of the grind and/or the way you handle the distribution and tamping are causing the puck to expand more than it normally would, but I think the expansion is a symptom rather than a cause.

In my previous post I recommended that you grind finer, but after reviewing the Robur videos it's not clear whether the grind is too coarse or your distribution/tamping is flawed. It could be either or both.

I think it's better if you start by trying to solve the halo/side-channeling problem first, then fix the flow rate. That's because the extraction problem may be partially responsible for the fast flow-rate (though it could be vice-versa, as I point out below.)

While your technique looks pretty good, I saw several things that give me pause. Chief among them is the fact that you tamp and weigh with the basket outside the PF. Dosing directly into the basket is OK because it lets you weigh the grounds. My concern is that your distribution and tamping efforts could be compromised by the slight shock of putting the basket on the scale after tamping and then back into the PF. This would be exacerbated if the grind is coarse. I know it's a ridgeless basket, but my philosophy is to be as gently as possible after the tamp.

Others have already commented that your WDT tool looks odd. I agree with John that it probably isn't a factor, given your Titan grinders, but it's worth using the recommended tool if you aren't already.

Finally, your nutating tamp is way too light to do any good and your final tamps look rather light as well. I can't tell if you're using any pressure when you polish the puck at the end, but you should not. The step is unnecessary, but if you're going to do it use only the weight of the tamper and make sure you don't disturb the puck in the process.

Try this:

1. Dose into basket
2. Weigh basket (now, not after the tamp)
3. Put basket back in PF
4. WDT (can be eliminated with Robur grind after fixing problems)
5. Distribute (gently tap PF on counter to settle grounds if necessary to avoid losing grounds)
6. Tap PF on counter 2-3 times to settle fines (this is a harder tap than in #5)
7. NEWS with about 10 lbs pressure
8. Nutate with about 10 lbs pressure (your nutate pressure in the video is too light)
9. Tamp straight down with 30 lbs pressure (looks light to me in the video)
10. Lock and load, pull

Rapping the PF on the counter just before tamping was recommended to my by Vince Fedele of VST when I was getting under-extractions. This step may not be necessary with WDT, though I believe there has been some dispute over whether WDT results in fines migration to the bottom of the basket. Either way, my belief is that fines migration will give you a slower flow rate and a higher extraction yield.

A variation on this is to dose into a small cup or ramekin, weigh the grounds, then tap the grounds into the basket (which is in the PF). This allows you to evenly distribute the grounds around the edges and in the middle. I use an OE plastic dosing tray for this, but they don't seem to sell it anymore. They do sell a rubber version that should work. Anyway, this is what I do and I believe it will eliminate any need to WDT.

You may have to practice a bit with the NEWS, nutate and final tamp to get it right. Make sure the puck is level when you're done. Be very gentle when you lock the PF into the machine. The goal is to not disturb the puck.

I recommend that you shoot for a 30-35 second extraction. The flow should remain dark or well tiger-striped to at least 25 seconds or more. You can let it run a few seconds past the first sign of blonding, but not more. The flow should never get as blond as it does in your videos.

If you get a good-looking pour, but the flow rate is so slow that you can't get your .6 ratio in 35 seconds or less, I would simply cut the shot at the higher brew ratio and drink it ristretto. IMHO, most specialty roasts, especially blends like Hairbender, are designed to be pulled ristretto and will taste better that way. Personally, I shoot for about .7 or so. A typical recipe for me would be 19g-20g in a 20g VST basket pulling 28g in 30 seconds (I cut the shot at about 25g because there's some dripping after cutoff.)

If the flow rate is still too fast, or if you still get a major halo extraction, try grinding finer. Alternatively, keep the coarser grind and increase the dose. I would think 20g-23g for Hairbender pulled in 30-35 seconds would be reasonable. If you use VST baskets, be sure to use the basket rated for the dose (i.e., 18g basket for 18g dose, 20g basket for 20g dose, etc.) As I said before, do not under-dose or over-dose VST baskets.

Though some on this board dispute it, tamping pressure can definitely make a difference when it comes to side-channeling. It can also slow down the flow rate, though usually only a little. I've read a lot of posts about success with 5 lb tamps (mostly from lever-heads), but they always cause horrible extractions on my GS/3. You may be able to back off of the full 30 lbs, but I wouldn't try that until you solve the extraction problems.

Generally I would agree with John that convex tampers make for prettier pours, but I use a flat-bottom tamper with my VST baskets, as recommended by VST. If the above steps don't quite solve the halo problem, it's certainly worth experimenting with both.

You may have answered this before, but are you using a 58mm tamper or a 58.3mm-58.4mm tamper as recommended by VST?

FYI, I made a lot of edits to the above post after submitting it, so if you read it before 3:30 PM EST Monday January 7, 2013, you may want to take another look.

Peppersass wrote: You may have answered this before, but are you using a 58mm tamper or a 58.3mm-58.4mm tamper as recommended by VST?

- Yes 58.3-4, I'm actually using the tamper that VST sells on their store.

-I'm not sure I know what "NEWS with about 10 lbs pressure" means Thats distributing with a finger, north/east/south/west, right? you're saying to push down on the puck while distributing?

-I think it was pulling a fast because I remember reading that as suggested from the Bench posts on Hairbender. I was going for 18g, .6 yield (30g) in about 24-25sec.

-As far as blonding, i double checked the dates on my beans last night and they were pushing 13 days. Just got a fresh batch in today. I didn't remember it blonding as quickly before, so I'm hoping that's all it was.

-One thing I am remembering is that LM "Strada" baskets are marked 1g lower than their VST counterparts. My 22g VST is identical to the 21g "strada" basket. Also, I was thinking that the nickel test should override the rated VST marking. That's the impression I got from going through one of the VST threads.
In any case, I will be trying your suggestions over the next few days.

If I don't figure it all out this week, I'm going to register for a counter culture class. It seems they have them twice a month here and I never realized it

Peppersass wrote: 1. Dose into basket
2. Weigh basket (now, not after the tamp)
3. Put basket back in PF
4. WDT (can be eliminated with Robur grind after fixing problems)
5. Distribute (gently tap PF on counter to settle grounds if necessary to avoid losing grounds)
6. Tap PF on counter 2-3 times to settle fines (this is a harder tap than in #5)
7. NEWS with about 10 lbs pressure
8. Nutate with about 10 lbs pressure (your nutate pressure in the video is too light)
9. Tamp straight down with 30 lbs pressure (looks light to me in the video)
10. Lock and load, pull

I am going to gently disagree with Dick on a number of minor points:

1. The above protocol is unnecessarily complicated. Grind, dose, distribute, tamp - get it right, and you can eliminate both the tapping (steps 5 and 6) and multiple tamps (Staub tamp in step 7).

2. Like you, I prefer to prep the puck with the filter basket outside of the PF. I have been doing this for years on a relatively unforgiving machine (Spaz S1, rotary pump, no preinfusion), and it works just fine. Inserting the basket into the PF does not disturb the puck, as long as you remove the spring click or use a ridgeless basket. Placing the basket on the scale is a nonissue if you handle it gently.

3. Dose is important, so keep weighing your dose for consistency. Regardless of the basket manufacturer, some adjustment of dose is reasonable. Donut extractions can often be cured by downdosing slightly.

4. Because nutation is a distribution technique (not a tamping technique), nutation after tamping doesn't make sense. 10# is on the heavy side for nutation, and I'm guessing Dick recommends this because he tamps before nutating. I only nutate before tamping, using the weight of the tamper.

5. Tamp pressure is relatively unimportant, and tamper piston fit to the basket does not have to be precise. My relatively loose fitting convex Bumper tamper gives me prettier pours than my flat 30# Espro tamper.

6. Aging coffee blonds faster, and might be the primary cause of early blonding in your videos.

7. Fines migration is an unproven theory. Even if it turns out to be true, the idea that stirring the grinds somehow replicates espresso brewing conditions (near boiling water under 9 bars pressure) is ludicrous. (Sorry, but I'm getting really tired of hearing that the WDT causes fines migration!)

8. When it comes to espresso, everything matters. Dick's espresso machine is similar to yours, and mine is not. So when in doubt, you should probably listen to him and not me.

Good luck!

Shawn, if you want you are welcome to come here and try to pull shots on my gear.

I wanted to offer this to you a long time ago but there was a gap in your posting and I thought maybe you got it right.

chapelizod wrote:It's a little hard to tell from the video, but the needle end of the tool you're using to WDT looks really short. I think you want the needle end to be long enough to hit the bottom of the basket. Something like what the person in this youtube video is demonstrating: http://www.youtube.com/watch?v=lICr7GZo ... ata_player

A stainless steel bicycle spoke is also a nice tool for WDT. About 20-cents at any decent bike shop.

RapidCoffee wrote:8. When it comes to espresso, everything matters. Dick's espresso machine is similar to yours, and mine is not. So when in doubt, you should probably listen to him and not me.

I really appreciate John pointing this out after disagreeing with almost everything I recommended .

If you're still game to try some of my suggestions, it will be very interesting to see if any of them work -- i.e., whether your issues and my solutions are specific to the GS/3 design. I've suspected as much, since a few other GS/3 owners who use VST baskets have followed my procedures or a subset with some success.

By way of gentle rebuttal, I would say that what I do isn't overly complicated or time consuming. It just looks that way because the steps are broken out in detail, but most of them take virtually no time at all and can be done in a series of smooth motions.

"Grind, dose, distribute, tamp and go" sounds simple until you break down the steps in detail. Then it starts to sound as complicated as my technique. For example, John's WDT requires a whole series of extra steps that you're doing. I don't do WDT, so I don't have to do all those steps. [As I said, I don't think you'll have to do WDT when you solve your extraction problem, at least not with the Robur (all bets are off with the Versalab, IMHO.)]

Even if you're not doing WDT, you may have to tap the basket to settle the grounds enough to distribute without losing coffee. Well, that would be another step in my list, but it takes less than a second and can hardly be considered added complication (and pros who updose without a scale do it all the time.)

My point is, when you magnify technique to each required movement, it looks more complicated than it really is. My advice is to do what works, not what takes the fewest steps or what you think you should be doing. After all, you're a home barista, not a pro. Time is on your side.

Let me clarify a few other things:

1. In recommending that you try dosing with the basket in the PF, I'm only trying to eliminate one possible cause of the halo extraction. John may very well be right that this is a non-issue, but it will only cost you one or two pulls to find out.

2. John and I agree on weighing the dose for consistency. I also agree that tweaking the dose might help with the halo extraction, though this may be more true for the La Spaz S1 than a GS/3-style machine. However, I don't think adjusting dose "regardless of the basket manufacturer" is the right advice for someone using VST baskets (and, I believe, John does not use VST baskets because they aren't available for his machine.) VST recommends a deviation of +/- 1g from the nominal dose. Further, Jim Schulman's study showed that VST baskets are particularly sensitive to significant variations in dose, so I think that for the sake of consistency you should either stay close to the nominal dose with VST baskets or switch to baskets from another manufacturer.

3. Sorry for the confusion about NEWS. There are two kinds of NEWS distribution techniques: one you do with your finger to swipe the grounds around the basket; the other you do with the tamper to move grounds below the surface to the edge of the basket. I actually do both. After filling the basket, I use a combination of NEWS swiping (N-E-W-S with my forefinger, not the tamper) and modified Sockfleth, with the goal being to distribute the coffee as evenly as possible across the surface and fill in the edges. Your technique in this regard looks similar. The only difference is that when I do the NEWS and modified Sockfleth I try not to lose any coffee because I weigh before distributing. The second kind of NEWS that I described in my list uses the tamper. This is a distribution technique whereby you lower the tamper straight down on the puck, without pressure, then rock the tamper towards the edges in a North-East-West-South pattern, pressing lightly or using only the weight of the tamper. The goal is to ensure that the grounds are pushed all the way to edges, even below the surface. You can find videos of this technique on this site and others. Again, I do both kinds of NEWS distribution.

4. By nutating, I mean rocking/rolling the tamper around the edges, either with just the weight of the tamper or with very light pressure.

5. It's probably not necessary to do both NEWS and nutation, but I always do -- belt and suspenders, I guess. I can't remember when I started doing this (probably when I had a less capable grinder), and haven't tried just one or the other in a long time. I think I added nutation after NEWS because it looked like the tamper wasn't contacting all the grounds at the edges. All I know is that no matter how carefully I distribute, I'll likely get a halo extraction if I don't do something to distribute the coffee at the edges below the surface. This is particularly true with VST baskets.

6. One thing I forgot to mention is that I level the tamper, and hence the puck, after the NEWS and nutate. I saw you do something similar. Again, this is for distribution.

7. I do not do NEWS (with the tamper) or nutate after tamping. I don't know where John got that idea, but possibly because I overestimated the pressure I use. 10 lbs was just a guess on my part, and likely a bad one. I've never measured the tamper pressure when I do these steps. I tend to add a little more pressure than the weight of the tamper itself, but it's probably less than 5 lbs. Might even be 1 lb or less. I don't consider this a tamp because light pressure is applied only at the edges, not in the center. Yes, these are distribution techniques, not the tamp, which comes after.

8. I tend to agree with John that tamper fit is not a big deal. But I figured since you're getting halo extractions, and are using VST baskets, that the slightly larger 58.4mm tamper might help. That said, it's just as likely that the fit is too tight and that you're pulling coffee away from the edges when you lift the tamper. For that reason, I'd try a standard 58mm. I'd also try a convex tamper. For both, you can used the silly plastic tamper that probably came with your machine. Why not? A one-pull experiment might reveal something.

9. I disagree about tamp pressure not mattering. Long-time readers will remember that Jon Rosenthal did, too (rather vociferously, as I recall), and he was using a GS/3 at the time. Maybe this is strictly a GS/3 phenomenon, but I find that the flow rate is affected by the tamping pressure. Usually it's not a huge effect, but going from, say, 5 lbs to 30 lbs with my machine can make a 5-second difference in pull time. I know others have disputed this effect, but I'm simply reporting what I see. My sense is that those who tamp very lightly are either using machines that are insensitive to it or are simply grinding finer than I do.

10. I didn't realize you were pulling old coffee, which certainly can cause early blonding, but I thought you reported earlier that your pulls were bad with fresh coffee, too. I wonder if you're flushing the stale grounds out of your Robur before grinding for the test shots. If not, the coffee you're pulling is really stale and will blond really fast. Can't say that about the Versalab, however, but that grinder scares me anyway .

11. I can't comment one way or the other about the theory of fines migration, but I do know that my extraction yield went up about 1% following Vince's recommendation that I rap the PF on the counter a couple of times before tamping. However, that could have been for any number of reasons, including the coffee I was pulling at the time, and I haven't done systematic experiments to determine whether it really has any effect at all. I recommended it to you because if fines migration to the bottom of the basket really does slow the flow rate, it might help cure the early blonding.

In summary, don't worry about the number of steps, at least not yet. Now is not the time to streamline your technique. You need to solve your extraction problem. It will only cost you a few pulls to try my suggestions -- so why not? This is how you solve problems -- by careful experimentation. Once you nail down the cause, then you can work on streamlining. Clearly, I need to do some experimentation myself to see if my technique can be streamlined!

Peppersass wrote:This is how you solve problems -- by careful experimentation. Once you nail down the cause, then you can work on streamlining. Clearly, I need to do some experimentation myself to see if my technique can be streamlined!

Just to add some unneeded support to this (from someone with much less expertise than peppersass): I recently switched machines, pump to lever, and while I wrapped my head around what parts of my routine were useful and what weren't, I ended up creating spreadsheet for every shot across two nights (~4 hrs total) of dedicated tuning, around 45 shots. I tracked bean, roast, roast age, basket type, dose weight, grinder time, distribution style (WDT,stockfleth,news,etc), tamp style (nutate, pressure,etc), preinfusion time, extraction time, post-extraction time, puckology (wet, dry, sneeze, wait-time), taste, other notes.

This really helped me get back to the basics (Espresso 101!) and avoid flailing by changing too many variables at once. It left me with a very streamlined routine [grind, tap a few times, light tamp, pull*] that yields great results (YMMV) as well as a better understanding of my grinder and espresso machine. It also helped make it apparent when the VST basket made sense versus the OEM basket for my machine.

All in all, it sounds rudimentary and tedious (and the beans aren't cheap either!), but the results have been worth it.

(*) preinfusion + shot