Reducing bitterness w/ high-extracting grinders (SSP HU etc.)
-
NoStream
Traditionally, it's been pretty easy to to dial in to a reasonable starting point - say something like 1:2 ratio in 27 seconds. With higher-extracting burrsets, especially SSP HU, it's gotten more complicated. If you blindly dial in to traditional parameters, your shot will be heavy, bitter, overwhelming, etc. So the new orthodoxy would be something like 1:2.5 in 20 seconds. But I think this is missing out on a little complexity that can get us to a better shot. And it's become more apparent now that we have higher-extracting baskets; you can easily get bitter cups if you blindly push grind ultrafine but great ones if you find the right sweet spot. So here are some dimensions worth playing with and dialing in:
(0. Flow rate) - higher flow rates produce cleaner and less bitter but thinner and sometimes less sweet. Past maybe 2 ish mL/s (very approximate, not scientific), EY will decline a bit, which is not necessarily a bad thing. (With the coffees I prefer and HU burrs) from 3-5 mL/s peak flow can be ok sometimes ok. I've increased the flow limit on extractamundo dos. You'd probably pick out a flow rate ahead of time and then fine tune from coffee to coffee.
1. Grind - closely related to flow rate, of course, but not necessarily if you use flow caps, pressure caps, etc. Going coarser can help a lot. I typically hold dose constant and dial in grind first to get desired flavor balance - this will end up affecting flow rate, then adjust dose to tweak the peak pressure and flow to fine-tune the flavor balance. Since the profile I'm using puts a soft limit on flow rate, this means peak pressure might be slightly lower at a coarser grind. IME overly low pressures sacrifice complexity and body and at extremes EY as well; overly high pressure produce shots that "just taste like espresso" and are bitter. But since grind size has a bigger effect, where too-fine shots hit a wall of aspirin-y bitterness, I start with grind.
2. Dose - back in the day, you could just use the finest grind before EY dropped (channeling). But with HU, I actually am not trying to maximize EY - 23-25% at 1:2.5 is very possible, but I typically prefer 21-23%, depending on the coffee. So if you want to grind coarser, increase the dose. The DE 1.0's massive headspace helps here, coupled with spacers or puck screens as needed (I have a variety of 1.7mm, 1.0, and SWorks ultra-thin ones). Stephane Ribes showed that higher doses were typically associated with higher EY, no need to be afraid of dosing 19 or 19.5 g in an 18g basket if you have enough headspace.
(3. Temp) - most people use lower temps with HU nowadays. This is good overall but some of the really low temp profiles seem to sacrifice a bit of sweetness, aroma, and complexity? Need to do more testing here, but it would be super cool to see more data here. Overly low temperatures are a bit like overly-low pressures - it will still taste pretty good, and it's safer, but you'll reduce the top quality possible from your shots. (speculation)
(4. Basket, bottom filter) - with VST baskets missing holes around the edge, I got profound improvements from using ICM to mound grounds away from the edges. With newer baskets (Wafo etc.) this isn't necessary and shots can be cleaner. They also have higher flow though, which means you can grind finer (higher EY in e.g. Hendrick's tests, sweeter, more bitter) or updose (similar flavor, cleaner). Closely related, in my informal back-and-forth tests, bottom paper filters can affect taste, and I _think_ that less clog-prone papers like Chemex and Cafec medium dark v60 papers produce a cleaner and less bitter taste than robot filters (less clogging? or actually filtering out bitter compounds? will possibly start a thread on this at some point)
(5. coffee choice) - above assume you're using a very clean, light roast from an excellent roaster like Wendelboe, Sey, Nomad, etc. I'm not 100% sold that Nordic-style as light as possible without underdevelopment is optimal for this type of shot since it can mean you have to grind so much finer than you get channeling. Something a hair darker like Onyx or George Howell might work better(?) But I haven't really explored this and mostly just buy the roasts I prefer on filter.
(6. ratio) - imo pulling shots longer doesn't decrease bitterness, just dilutes it. And pulling shorter shots to decrease EY results in a super concentrated style that is not my preference, though I know some people like 12 TDS HU shots.
(Why bother? Well, a well-dialed-in HU shot tastes amazing if you like the style: very transparent, aromatic, sweet, and juicy. Obviously, this isn't an attempt to figure out what style of espresso is "best.")
I don't think the process of dialing in every variable above is necessary for every shot. FWIW on my Decent, using a modified version of Extractamundo Dos, I'm updosing a bit (19-20g in a WAFO), using a flow rate around 3.5-4 mL/s / total shot time around 17s, and using Chemex bottom filters. From there, I just tweak grind, then dose if needed, then possibly ratio if the shot is too muddy and concentrated or too thin.
Would love to hear people's strategies and thoughts on making SSP HU and other high-extracting burrs taste clean and sweet and not muddy or bitter. What has or hasn't worked for you? I realize some of this might be Decent-specific but wanted to open this up to a broader audience. And sorry for the length of this post, but I prefer to be overly thorough than omit detail.
(0. Flow rate) - higher flow rates produce cleaner and less bitter but thinner and sometimes less sweet. Past maybe 2 ish mL/s (very approximate, not scientific), EY will decline a bit, which is not necessarily a bad thing. (With the coffees I prefer and HU burrs) from 3-5 mL/s peak flow can be ok sometimes ok. I've increased the flow limit on extractamundo dos. You'd probably pick out a flow rate ahead of time and then fine tune from coffee to coffee.
1. Grind - closely related to flow rate, of course, but not necessarily if you use flow caps, pressure caps, etc. Going coarser can help a lot. I typically hold dose constant and dial in grind first to get desired flavor balance - this will end up affecting flow rate, then adjust dose to tweak the peak pressure and flow to fine-tune the flavor balance. Since the profile I'm using puts a soft limit on flow rate, this means peak pressure might be slightly lower at a coarser grind. IME overly low pressures sacrifice complexity and body and at extremes EY as well; overly high pressure produce shots that "just taste like espresso" and are bitter. But since grind size has a bigger effect, where too-fine shots hit a wall of aspirin-y bitterness, I start with grind.
2. Dose - back in the day, you could just use the finest grind before EY dropped (channeling). But with HU, I actually am not trying to maximize EY - 23-25% at 1:2.5 is very possible, but I typically prefer 21-23%, depending on the coffee. So if you want to grind coarser, increase the dose. The DE 1.0's massive headspace helps here, coupled with spacers or puck screens as needed (I have a variety of 1.7mm, 1.0, and SWorks ultra-thin ones). Stephane Ribes showed that higher doses were typically associated with higher EY, no need to be afraid of dosing 19 or 19.5 g in an 18g basket if you have enough headspace.
(3. Temp) - most people use lower temps with HU nowadays. This is good overall but some of the really low temp profiles seem to sacrifice a bit of sweetness, aroma, and complexity? Need to do more testing here, but it would be super cool to see more data here. Overly low temperatures are a bit like overly-low pressures - it will still taste pretty good, and it's safer, but you'll reduce the top quality possible from your shots. (speculation)
(4. Basket, bottom filter) - with VST baskets missing holes around the edge, I got profound improvements from using ICM to mound grounds away from the edges. With newer baskets (Wafo etc.) this isn't necessary and shots can be cleaner. They also have higher flow though, which means you can grind finer (higher EY in e.g. Hendrick's tests, sweeter, more bitter) or updose (similar flavor, cleaner). Closely related, in my informal back-and-forth tests, bottom paper filters can affect taste, and I _think_ that less clog-prone papers like Chemex and Cafec medium dark v60 papers produce a cleaner and less bitter taste than robot filters (less clogging? or actually filtering out bitter compounds? will possibly start a thread on this at some point)
(5. coffee choice) - above assume you're using a very clean, light roast from an excellent roaster like Wendelboe, Sey, Nomad, etc. I'm not 100% sold that Nordic-style as light as possible without underdevelopment is optimal for this type of shot since it can mean you have to grind so much finer than you get channeling. Something a hair darker like Onyx or George Howell might work better(?) But I haven't really explored this and mostly just buy the roasts I prefer on filter.
(6. ratio) - imo pulling shots longer doesn't decrease bitterness, just dilutes it. And pulling shorter shots to decrease EY results in a super concentrated style that is not my preference, though I know some people like 12 TDS HU shots.
(Why bother? Well, a well-dialed-in HU shot tastes amazing if you like the style: very transparent, aromatic, sweet, and juicy. Obviously, this isn't an attempt to figure out what style of espresso is "best.")
I don't think the process of dialing in every variable above is necessary for every shot. FWIW on my Decent, using a modified version of Extractamundo Dos, I'm updosing a bit (19-20g in a WAFO), using a flow rate around 3.5-4 mL/s / total shot time around 17s, and using Chemex bottom filters. From there, I just tweak grind, then dose if needed, then possibly ratio if the shot is too muddy and concentrated or too thin.
Would love to hear people's strategies and thoughts on making SSP HU and other high-extracting burrs taste clean and sweet and not muddy or bitter. What has or hasn't worked for you? I realize some of this might be Decent-specific but wanted to open this up to a broader audience. And sorry for the length of this post, but I prefer to be overly thorough than omit detail.
-
coyote-1
The burrs in mine are not SSP, but they are titanium coated and look incredibly similar. Clearly modeled after the SSP burrs (which are themselves a copy of an earlier European burr).
My strategy prior to my current one was simple: 25-28 seconds after beginning the shot I would end it. That works well; the shots were consistently enjoyable. My current method of emulating spring lever machines via flow control:
- 6 seconds or so preinfusion at 4-5 bar.
- 15 seconds or so 'rest' (no pressure, blooming).
- ramp up to full pressure
- after 10 seconds or so, begin gradually backing off pressure (depending on flow)
- continue until blonding begins. The moment I begin to see blonding, I stop the shot.
My temps are around 192F at the beginning of a typical shot, higher if using lighter roasts. I am not weighing anything. It's my impression that up until blonding begins I'm extracting the good stuff, and that blonding means that's been exhausted and from that point I'm pulling the bitter compounds. The results in the cup tell me that's an accurate assessment.
My strategy prior to my current one was simple: 25-28 seconds after beginning the shot I would end it. That works well; the shots were consistently enjoyable. My current method of emulating spring lever machines via flow control:
- 6 seconds or so preinfusion at 4-5 bar.
- 15 seconds or so 'rest' (no pressure, blooming).
- ramp up to full pressure
- after 10 seconds or so, begin gradually backing off pressure (depending on flow)
- continue until blonding begins. The moment I begin to see blonding, I stop the shot.
My temps are around 192F at the beginning of a typical shot, higher if using lighter roasts. I am not weighing anything. It's my impression that up until blonding begins I'm extracting the good stuff, and that blonding means that's been exhausted and from that point I'm pulling the bitter compounds. The results in the cup tell me that's an accurate assessment.
-
NoStream (original poster)
I've heard of people doing this! I think it makes sense, but it seems like the shots would be overwhelmingly strong.
-
NYPeter007
- Supporter ♡
Is it possible the SSP burrs need to be seasoned? I have my SSP MP in a Super Jolly that I use for filter and over time the coffees have been tasting cleaner and brighter. Same for the DF64 with the HU burrs.
-
NoStream (original poster)
I think there's a mix of actual seasoning and "brain burn-in" especially for filter with SSP HU since the style of drip coffee is so different. (Still not 100% my favorite, prefer pre-15 EK or 64 MP.) I've had my SSP HU since 2018 though, so plenty of seasoning. And this problem wasn't so much of one in 2018 when my technique was worse (worse WDT, no filter below, VST rather than wafo/sworks basket, etc.)
-
Radio.YYZ
I have ssp burrs on p64, i use light roast single origin coffee. Few things you can do (super easy on your de)
- try the londinium style shot without changing anything else.
- try (as someone mentioned) using 1.25x and 1.5x ratio instead of 2x.
I know you mentioned you liked high yielding shots but the high yield would include bitters/astringent flavours as well. High pressure throughout the shot starts to increase bitterness as time progresses in my experience, it is more pronounced with some coffees vs others.
Failing all that you can try crafting water yourself to overcome the unpleasant flavours, the late rplavis' water profile is great for it.
- try the londinium style shot without changing anything else.
- try (as someone mentioned) using 1.25x and 1.5x ratio instead of 2x.
I know you mentioned you liked high yielding shots but the high yield would include bitters/astringent flavours as well. High pressure throughout the shot starts to increase bitterness as time progresses in my experience, it is more pronounced with some coffees vs others.
Failing all that you can try crafting water yourself to overcome the unpleasant flavours, the late rplavis' water profile is great for it.
Good Coffee: Technique/Knowledge > Grinder > Beans > Water > Machine
-
TomC
- Team HB
NoStream wrote:Traditionally, it's been pretty easy to to dial in to a reasonable starting point ...
Glad you're posting again, whatever the topic. Folks might not always agree with you, but you've posted insightful or thought provoking content over the years.
Join us and support Artisan Roasting Software=https://artisan-scope.org/donate/
-
PIXIllate
- Supporter ♡
In my experience with multiple sets of coated burrs (Shuiken Kafatek in my case) there was a persistent bitterness when using higher extraction techniques until I hit the 30-40lb mark. At which point the bitterness decreases allowing for a return of sweetness and complexity. Things improved further for the next ~30lb and then finally plateaued as I got to ~70lb of beans through.
Seasoning is a real thing, which is why I don't put much stock in reviews that aren't done long term and with burrs that have been used a significant amount.
Seasoning is a real thing, which is why I don't put much stock in reviews that aren't done long term and with burrs that have been used a significant amount.
-
Jeff
- Team HB
Great to hear of your observations and approaches!
Let's narrow things down a bit, as I see comments about things that seem to be pretty tangential.
High-extraction burrs are to me things like 98 HUs, 64 MPs and marginally Bentwood, Lagom Mini Moonshine, 64 Cast v2. I can't comment on the Mizen burrs as they haven't arrived yet. I can't comment personally on the Kafatek burrs, but I would guess that the Flat Max burrs intended for light roasts are probably high-extraction burrs. 64 HUs are not in this category, nor are pretty much any "stock" burr, even if it looks like one of those.
Lighter roasts -- If you're pulling anything from darker than Tim Wendelboe's "espresso" roast, you're in a different category of extraction. Tim Wendelboe's espresso roast level is lighter than many American filter roasts. From what I've seen from US roasters, if they label it as "for espresso", it isn't the kind of lighter roasts I'm talking about, no matter how they describe the roast level. If it tastes roasty, it isn't a lighter roast level to me.
I would put (5. coffee choice) much higher on the list. The more revealing your burr set is, the more it reveals greens and roast defects.
I concur that while you can push a light roast with a Niche Zero to 21-23% EY in a high-flow basket, there is a cliff where you abruptly hit astringency, bitterness or both. Without a high-flow basket, the shots are generally lacking in complexity, you're fighting the sour/bitter/astringent balance, or both.
With high-extracting burrs, I find that the shot is "done" faster, no matter if I'm shooting for a high-flow ("turbo") shot or a low-flow one. I can pull 60-second shots with my 98 HUs that are very tasty. However, the stream seems to blonde much faster at a given flow rate with the "faster" burrs compared to the "slower" ones (that are still fast, compared to traditional burrs).
My flow rates, when dialed in, tend to be in the 3.5-4 g/s range at the end of the shot for the 98 HUs and more in the 3-3.5 g/s range for the Bentwood and Lagom Mini. I don't pull much with the P64 since I consider the 98 HUs to be slightly better than the 64 MPs for the coffees and styles I prefer. My overall shot times tend to be in the 15-20 second range, from pump on, including a brief hold. My ratios for coffees such as Coffee Collective, Apollon's Gold, Manhattan, Hydrangea, Flower Child, Mood Trap ("Prady") tend to be 17:30 through 17:38. Ratio (mass) is my target, time is the output. I'll occasionally pull a 30- or even 60-second shot, either because I completely guessed wrong on the grind, or to try a very different dial-in for a specific bean.
Damian's profiles (including LRv2) are excellent for what I believe they are trying to achieve, a smooth cup with minimal astringency from medium-roast coffees. I find that the extended hold profiles of that general style tend to emphasize the generic chocolate and nuts notes at the expense of lighter and brighter notes.
I believe that going much over 6 bars isn't adding much in the cup. I believe that going much over 8 bars can result in bland or even cardboard-like notes, even with coffees that don't reveal those notes when cupping or as filter.
The short hold with decaying pressure, such as in Extractamento Dos! does seem to be advantageous. I ended up at a similar place after working with profiles inspired by Stephane and by Luca.
I am mixed as to if not reaching 5-6 bars at least briefly is detrimental to the overall quality.
Water makes a big difference. I am using 20/80 GH/KH most of the time for espresso.
I haven't decided if I have a preference between Ahlstrom 237 or 909 papers on the bottom. I'm using a VST basket and a puck screen (primarily for cleanliness). Low-flow baskets, precision or otherwise, as well as ones with reduced hole coverage may be great for classic espresso, but I don't find them appropriate here. I have not convinced myself to spend $50 on shipping for a WAFO basket yet, or if I want to try Sheldon's first. I might try to borrow one locally.
I haven't dipped as low as 80°C, but have been looking at the mid 80°s now for a couple years. How this number translates to a conventional machine's number, like an E61 double-boiler or a La Marzocco, I don't know.
I agree that the highest EY is not always the best coffee. I find this to be true with both espresso and filter. Some just don't hold up well when pushed much past 20-22%. Others just get better and better.
Let's narrow things down a bit, as I see comments about things that seem to be pretty tangential.
High-extraction burrs are to me things like 98 HUs, 64 MPs and marginally Bentwood, Lagom Mini Moonshine, 64 Cast v2. I can't comment on the Mizen burrs as they haven't arrived yet. I can't comment personally on the Kafatek burrs, but I would guess that the Flat Max burrs intended for light roasts are probably high-extraction burrs. 64 HUs are not in this category, nor are pretty much any "stock" burr, even if it looks like one of those.
Lighter roasts -- If you're pulling anything from darker than Tim Wendelboe's "espresso" roast, you're in a different category of extraction. Tim Wendelboe's espresso roast level is lighter than many American filter roasts. From what I've seen from US roasters, if they label it as "for espresso", it isn't the kind of lighter roasts I'm talking about, no matter how they describe the roast level. If it tastes roasty, it isn't a lighter roast level to me.
I would put (5. coffee choice) much higher on the list. The more revealing your burr set is, the more it reveals greens and roast defects.
I concur that while you can push a light roast with a Niche Zero to 21-23% EY in a high-flow basket, there is a cliff where you abruptly hit astringency, bitterness or both. Without a high-flow basket, the shots are generally lacking in complexity, you're fighting the sour/bitter/astringent balance, or both.
With high-extracting burrs, I find that the shot is "done" faster, no matter if I'm shooting for a high-flow ("turbo") shot or a low-flow one. I can pull 60-second shots with my 98 HUs that are very tasty. However, the stream seems to blonde much faster at a given flow rate with the "faster" burrs compared to the "slower" ones (that are still fast, compared to traditional burrs).
My flow rates, when dialed in, tend to be in the 3.5-4 g/s range at the end of the shot for the 98 HUs and more in the 3-3.5 g/s range for the Bentwood and Lagom Mini. I don't pull much with the P64 since I consider the 98 HUs to be slightly better than the 64 MPs for the coffees and styles I prefer. My overall shot times tend to be in the 15-20 second range, from pump on, including a brief hold. My ratios for coffees such as Coffee Collective, Apollon's Gold, Manhattan, Hydrangea, Flower Child, Mood Trap ("Prady") tend to be 17:30 through 17:38. Ratio (mass) is my target, time is the output. I'll occasionally pull a 30- or even 60-second shot, either because I completely guessed wrong on the grind, or to try a very different dial-in for a specific bean.
Damian's profiles (including LRv2) are excellent for what I believe they are trying to achieve, a smooth cup with minimal astringency from medium-roast coffees. I find that the extended hold profiles of that general style tend to emphasize the generic chocolate and nuts notes at the expense of lighter and brighter notes.
I believe that going much over 6 bars isn't adding much in the cup. I believe that going much over 8 bars can result in bland or even cardboard-like notes, even with coffees that don't reveal those notes when cupping or as filter.
The short hold with decaying pressure, such as in Extractamento Dos! does seem to be advantageous. I ended up at a similar place after working with profiles inspired by Stephane and by Luca.
I am mixed as to if not reaching 5-6 bars at least briefly is detrimental to the overall quality.
Water makes a big difference. I am using 20/80 GH/KH most of the time for espresso.
I haven't decided if I have a preference between Ahlstrom 237 or 909 papers on the bottom. I'm using a VST basket and a puck screen (primarily for cleanliness). Low-flow baskets, precision or otherwise, as well as ones with reduced hole coverage may be great for classic espresso, but I don't find them appropriate here. I have not convinced myself to spend $50 on shipping for a WAFO basket yet, or if I want to try Sheldon's first. I might try to borrow one locally.
I haven't dipped as low as 80°C, but have been looking at the mid 80°s now for a couple years. How this number translates to a conventional machine's number, like an E61 double-boiler or a La Marzocco, I don't know.
I agree that the highest EY is not always the best coffee. I find this to be true with both espresso and filter. Some just don't hold up well when pushed much past 20-22%. Others just get better and better.
Sponsored by La Marzocco · Home
