The impact of filtering on refractometer TDS and extraction yield measurements

Postby **namelessone** » Feb 14, 2019, 4:46 am

catalinH wrote:The coffee used was Ethiopia Uraga from da Matteo roasted for filter.
My Ek is OCD alligned,also I've replaced the carrier with a turkish one from Frank.
Although your conicals may produce less boulders than my Robur,you will still not hit a 25%EY using a conical grinder because of the widw particle distribution.Another obstacle in achieving high extractions is the obsession of producing ratios
so my questions remains: What is there to love about 1:2 ratio?

You need to use a syringe filter and/or centrifuge before you use your refractometer for espresso shots. There is no way that reading is accurate.

...split from Top 5 Grinders for Light Roast Espresso Application by moderator...

Postby **JoeSventek** » Feb 14, 2019, 5:40 am

namelessone wrote:You need to use a syringe filter and/or centrifuge before you use your refractometer for espresso shots. There is no way that reading is accurate.

Are you sure filtering the samples makes the readings more accurate?

Here's an interesting comment from https://socraticcoffee.com/2016/02/exam ... ssessment/

Jeremy wrote:[...] Our work suggests the VST syringe filters do significantly lower TDS. Looking at it more closely, we see that the filters retain some dissolved coffee solids as well as release something (not coffee) that can register as a dissolved solid (unluckily, this is not a 1-for-1, as the filters retain more than they release). We have asked Atago why they do not recommend filtering coffee and here is their response, "By using those filters and syringes the espresso is no longer the same as what people actually drink. Then we thought, what does the Brix (TDS) value after filtration mean?" This is a valid point and something to consider. Atago and MISCO (the VST device manufacturer), and lots of other manufacturers, make refractometers with a variety of industry scales. Interestingly, we have not found any of them selling sample filters. VST suggests the syringe filters' use is to remove noise in the sample and create a more stable reading. But one might ask, "A more stable reading of what?"

Postby **CwD** » Feb 14, 2019, 5:53 am

JoeSventek wrote:Are you sure filtering the samples makes the readings more accurate?

Here's an interesting comment from https://socraticcoffee.com/2016/02/exam ... ssessment/

More than sure. http://awasteof.coffee/science/espresso-filtration/

Planning to make a post for this specifically after I make sure a post about something on my own site is kosher, so let's not derail into this topic too far.

I also found that the VST readings are not significantly lower than centrifuged ones, so the idea they could remove tds and make it less representative of the drink is bs at best. They're not lowering the tds, just debris that register as tds. Same as the centrifuge, which has nowhere for truly dissolved solids to possibly be removed to.

Postby **JoeSventek** » Feb 14, 2019, 6:20 am

Unfortunately I don't have the time to read your article just yet but I added it to my reading list to not forget about it.

Thank you for doing the work and sharing it, Mitch!

Postby **catalinH** » Feb 14, 2019, 6:33 am

namelessone wrote:You need to use a syringe filter and/or centrifuge before you use your refractometer for espresso shots. There is no way that reading is accurate.

I don't filter,but I take 3 measurements of TDS from 3 samples and the variations are in the limits of what my colegues described.I pic the lowest reading as a reference.I also allow the samples to cool down and reach equilibrium with the refractometer.
I'm aware there are errors but not so that when you estimate the EY it will reduce it by more than 0.2-0.3%

Postby **CwD** » replying to catalinH » Feb 14, 2019, 7:15 am

More like .8% in terms of EY error from my experiments. Even the very lowest unfiltered readings were multiple tenths of a percent higher in tds, let alone extraction yield.

Postby **catalinH** » Feb 14, 2019, 8:00 am

All I can say is that my estimated TDSs shown by my refractometer don't vary more than 0,05%. When this happens I check my calibration with distilled water and retake the measurements. I never allow that my readings go all over the place. I only pick the lowest value. This means that I'm consistently right or consistently wrong

.
Furthermore 25%EY estimate is not an isolated event. So are my readings so far off

?

Postby **namelessone** » Feb 14, 2019, 12:37 pm

According to the data posted by CwD there was difference of for example 8.39 (filtered) and 9.26 (unfiltered), assuming 1:2.5 ratio this can correspond to difference of more than 2% EY (20.98% vs 23.15%).

It's also likely the finer you grind -> the more undissolved solids -> the higher difference between filtered/unfiltered. Even on my drip brews there's been some difference in TDS filtered/unfiltered when I'm grinding very fine, due to more silt in the cup.

Centrifuge seems good enough for espresso, perhaps not for cupping QC or similar activity (0.05% TDS is still significant at immersion ranges).

Postby **Denis** » replying to namelessone » Feb 14, 2019, 12:47 pm

In our small geek group we obtain way over 25 EY% out of light roasts. So please stop trying to minimize the gap (with errors and whatsoever).

There are people on other forums that pull 27% EY out of ristrettos (18 TDS out of 1:1 or 1:1.5 shots). We are not talking about 1 coffee out of 10 we are talking about consistent shots.

Postby **catalinH** » Feb 14, 2019, 1:57 pm

namelessone wrote:According to the data posted by CwD there was difference of for example 8.39 (filtered) and 9.26 (unfiltered), assuming 1:2.5 ratio this can correspond to difference of more than 2% EY (20.98% vs 23.15%).

It's also likely the finer you grind -> the more undissolved solids -> the higher difference between filtered/unfiltered. Even on my drip brews there's been some difference in TDS filtered/unfiltered when I'm grinding very fine, due to more silt in the cup.

Centrifuge seems good enough for espresso, perhaps not for cupping QC or similar activity (0.05% TDS is still significant at immersion ranges).

What are the chances when I take 3 readings of TDS and they all vary less than .05%(eg 9.72,9.73 and 9.76) and I pick the lowest one each time that all of them are wrong every time?
This means that all my samples are murky, filled with fines that artificially increases TDS seriously overestimating EY.

Home-Barista.com

The impact of filtering on refractometer TDS and extraction yield measurements