Recently I have been using the Abid dripper, scale, TDS meter, Hario Skerton, and correlate the flavor with the Coffee Brewing Control Chart.

This is the formula I deduced from reading the Coffee Brewing Handbook:

Strength (in TDS or %) = CoffeeWeight*Extraction / (WaterWeight - CoffeeWeight*2.086)

I use the metric system, as it is a bit tedious to convert oz and gallons. I also find it is easier to just place the Abid dripper on a scale, zero it, add 11-12g of coffee, zero it again, and add 200g of water. It is actually more convenient to measure the water by weight, rather than volume. The original Coffee Brewing Control Chart is geared toward commercial setting, therefore to measure by weight is probably not possible.

One interesting observation. When I leave the whole apparatus on the scale, I can see the weight constantly decline. Presumably this is due to the aromatic molecules, water vapor, and carbon dioxide loss. With 200g of water, after 3-4 minutes, approximately 8-10g will be lost. The original Coffee Brewing Control Chart probably ignored this.

I also find, for my personal taste, the ideal is more like the Norwegian control table.

The Hario Skerton is quite ideal for single serving brewed coffee. The manual grinding likely produces very light heat. There is practically no waste. The glass and whole grinder can be rinsed and cleaned without residual oil. The shape of the glass container is also conducive to evaluate the dry fragrance. I did some unscientific side by side "smell test" of the same beans and roast, with the Hario Skerton and Versalab M3. The dry fragrance of the Hario grind is stronger.

I am trying to find the original research by Dr E E Lockhart to no avail. Does anybody here know where it is published?

Hi Henry,

Some thoughts:

- I'm not sure how accurate your results will be using just a TDS meter (by which I'm assuming you mean a conductivity meter, like those used to test water hardness). Prior to the advent of the reasonably affordable refractometers designed specifically for coffee, I know people did use TDS meters, and more often brix refractometers, to help measure extraction, but apparently there is a lot of work involved in measurement and calculation to give you even close to an accurate reflection of the extraction %. See this thread for more info. The cheapest but most PITA method is baking the coffee grounds after brewing to evaporate all water, then weighing the grinds to see how much was extracted from them.

- with one of the coffee refractometers I linked to above, you can easily get an accurate TDS% reading, and then you plug it into this formula (or just use the fantastic software also designed by Vince Fedele):

Extraction % = (final weight of brewed beverage * TDS%) / weight of ground coffee dose * 100

- I'm pretty sure most of the weight loss you're seeing over brew time is due to evaporation. I'd strongly suggest covering the Abid during the brew, even if only to prevent heat loss during extraction. In fact, you may even want to grind finer and steep for a shorter time to avoid the heat loss that occurs over 4 minutes. (I'm not quite sure why everyone uses 3-4 minute steep times with the Clever/Abid -- probably because those are typical French Press brew times, and people see the methods as similar because they involve full immersion. I typically only steep my Abid for 45 secs to a minute, plus drawdown, and adjust my grind until I get the extraction profile I'm looking for.)

- the SCAA lists 15 works by Ernest Lockhart in their library, but none of them appear to be available online through the SCAA. I've also found a few PDF versions through Google, but none free.

- finally, make sure you're using the correct updated control chart -- the SCAA updated the chart in 2009 to correct for a long-standing error, but there are plenty of the old versions still around in print and online.

Chris,

Thank you for your reply. My original goal was to actually review the original research by Dr Lockhart from the 1960's. Most of the SCAA references were Coffee Brewing Center pamphlets. The Coffee Brewing Handbook by Lingle does not contain any references. The research on water, freshness, chemistry, etc, can be found in American Chemical Society publications.

The oven drying method on page 6 can be duplicated. However, for example, I would like to know under what conditions 1g of coffee will bind to 2.086g of water, it is time dependent, grind dependent, etc. Another example, water density is 1027.845 kg/m3 at 100C, vs 990.685 kg/m3 at 20C. Because of this, the volume is significantly different and the control chart does not account for this temperature change.

The formula I deduced was purely from reading the book. I use it to gauge percolation time or grinding. For example, if I want to extract 20% with 11g of coffee in 200g water:

TDS = 11*0.2/(200-11*2.086) = 1242 (1.24%)

I'll just adjust my percolation time or grind to arrive at 1242 on my TDS conductive meter reading. It will be interesting to compare side by side a TDS conductive meter vs refractometer.

Yeah, I'm highly suspicious of the 2.086g figure being a constant across all conditions and coffees, which is why I prefer to measure the weight of the brewed coffee, not the brew water beforehand, therefore taking how much was absorbed by the grounds out of the equation entirely. I would suggest you do the same, it's easy and will immediately give you more accuracy (i.e., why calculate something you can measure?). For instance, in your example, replace (200-11*2.086) with the weight of the beverage after brewing.

The water density issue is one of the things that Vince Fedele has corrected for in his software (I believe you either enter weight, or you enter volume, but it corrects for temp in in its calculations) . I think the SCAA chart might be heading in this direction now, but I can't say for sure, not having seen the updated chart. Certainly sticking with weight is the way to go, in my opinion.

For a fascinating review of the kind of issues these errors have been causing for coffee brewing, see the appendix of Scott Rao's book 'Everything but Espresso.' Pretty much all the things you're bringing up here are covered there. In fact, if I understand the issue with the coffee brewing chart correctly, you should not be looking for 1,242 TDS, you should be looking for 12,420 (the error in this case being the SCAA's mislabelling of the ordinate axis of the chart back in the 90s -- 1.25% brew strength is 12,500 ppm, not 1250).

Chris,

Thanks for the post again.

Just to clarify for people who have not read the Coffee Brewing Handbook. A typical conductive meter can be calibrated to read sodium chloride or calcium carbonate equivalent of TDS. The TDS either by sodium chloride or calcium carbonate correlate to coffee "TDS" by a factor of 10. This is the reason the SCAA or Norwegian brew control charts are set to be off by about 10X on the TDS ordinate, ie, Strength expressed as percentage. This was actually not a mislabeling of the chart, as the chart is an approximation of the dissolved solid and not an absolute measurement, which requires the bake oven analysis and not practical in daily use. The reason for the conversion factor of 10 was discussed on page 10 of the book.

I thought about the water binding factor of 2.086g water per gram of coffee, and don't think it could be omitted in the calculation. If you measure the weight of the brewed coffee only, it will be useful to use the 2.086g conversion factor. If you measure the weight of the brew water before brewing, you also need to use this factor to approximate extraction rate.

I have not read Scott Rao's book, and maybe I will obtain one in the near future to see his take. For now, my goal is to numerate my range of preference for a brewing process that is reproducible.

The refractometer also suffers the similar measurement error, as it is also an indirectly measurement of coffee TDS. Index of refraction is set to correlate coffee TDS, instead of electrical conductivity. My guess for the commercial refractometer or conductive meter advertised as coffee usage probably just use a proprietary index to correlate numbers from the bake analysis.

As for the conductive meter conversion of water TDS, coffee TDS, and NaCl and CaCO3, the table is at page 10 of the Coffee Brewing Handbook.

Hopefully we can stimulate more discussion regarding the Brewing Control Chart.

Hmmm. I don't have the full CBH, just the 'Basics of Coffee Brewing' shorter version, which I got the other day (published in 1996). It doesn't seem to say anything about the factor of 10 issue you mention. In fact, it has a chart labeled as 'Relationship between conductivity and dissolved solids' (which I believe is the same chart you're referencing) that states, for example, if the conductivity meter reads conductivity at 3000 michromhos, that indicates a coffee TDS of 1680.4, a NaCl-ppm of 1550, and a CaCO3-ppm of 1300. So there's a set relationship, which is how the meters calculate coffee TDS, but it's not a factor of 10, and therefore would not account for the error in the chart. That said, perhaps you have a more recent version that corrects all of this. I'll see if I can get my hands on one, for sure.

As for how refractometers work compared to conductivity meters, I don't believe it's simply indexing info from a bake analysis. This provides lots of info, and from the sounds of it, it's measuring the refractive index of the sample, compensating for temp, and then looking that up in a table that correlates refractive index to TDS for coffee specifically.

I'm not sure why you feel you can't omit the calculation for water retained, and just substitute the measurement of the final brewed coffee. The portion of your sample calculation I quoted is in fact *trying to figure out* what the weight of the final coffee is -- it's taking the 200g of brew water you used, then, *assuming* that each gram of coffee absorbs 2.086g of water, multiplying the coffee grams x 2.086 and subtracting that from 200 to 'figure out' what the final brewed coffee weight is. Why not just weigh the brewed coffee, and thereby eliminate any error the (I believe estimated) factor of 2.086 might be causing? Because the core of the calculation is the amount of coffee solids (by weight) that makes it from the grounds to the final beverage, the weight of the final beverage is most important, not the weight of the brew water. For instance, how much of the brew water is evaporating through brewing, which is also not accounted for? Just weight final beverage, I say.

I do think that you're using the chart and your meter for the exact right reason, though, to help give you something objective to which you can correlate your taste buds, to help you figure out what you like, and then how to get it more often. For some reason many coffee people are resistant to trying these things, so it's great that you're getting how valuable they are. Oh, and Rao's book is really, really worth it.

My goal and deduced formula was used to estimate a range of palatable TDS at a set extraction rate, for example 20%, and adjust grind and percolation. Your formula is to estimate extraction, while keeping TDS constant. One observation I have is that the drip rate of the Abid is not fast enough, thereby altering the final brewed coffee weight. For example, the first 30 seconds will have different TDS than the last 30 seconds. Now flip the logic around, and assume the first and last 30 seconds will have the same TDS, because the coffee all ends in the same cup, then it does not matter what the final weight of the brew coffee is.

The 10X factor is necessary because a conductive meter measures the electrical current across the solution, which is affected by ions. In brewed coffee, significant amount of TDS is contributed by polysaccharides and proteins, which amount to at least 30% of the TDS, which are NOT ions, and therefore will not be measured by the conductive meter. For this reason, the coffee TDS is approximately 10% of conductively measured ionic TDS, therefore the control chart appears to be off by factor of 10 as you observed.

The refractometer has a similar problem. In the laboratory, pure/deionized water is used as reference, and a known concentration aliquot is used to "zero" the machine. A common molecules, like glucose, reflects and deflects light. In the laboratory, I'll just change the wavelength of the incident light to obtain optimal reading and therefore the concentration for different molecules. Coffee contains more than just one type of molecules. In coffee refractometer measurement, the light will be significantly affected by other solutes, ie polysaccharide, protein, etc, therefore the "strength" reading again will be an approximate, based on a conversion factor, probably derived from data points obtained with bake oven numbers.

I suspect conductive coffee TDS measurement will also be even more helpful in siphon, where the "travel south" rate is relatively rapid, unlike the Abid dripper. I'll conduct some test when I have time.

Ok, I think I get it, you're using the formula to try to get you a TDS to shoot for, so you need to do the calculation before you make the coffee, therefore you can only enter brew water weight. I would encourage you after brewing and measuring the TDS to run it back through the formula using the actual brewed coffee weight though, to see if the 2.086 factor is throwing things off at all.

I used to worry about the fact that TDS in certain parts of an extraction from the Abid (or any drip-type brewer) would be higher than others until I realized that that's kind of always the case -- even in a French Press for instance, water that's a little closer to a particular coffee molecule will contain more solubles than one that's farther. There's also the different particle sizes to consider -- water near smaller particles with greater exposed surface area will absorb more solubles faster than water near bigger particles. And then I thought about the fact that these are solubles -- they're dissolved in amongst the water, and can be redistributed by a simple stir. So really, it's about avoiding extraction flaws (loss of temp, uneven extraction of grounds, bad quality grind, etc.), and then just ensuring that your grind setting/contact time/temp/agitation, etc., result in an 18-22% extraction *overall*. I think the Abid handles the first part better, and allows me to hit the second part more reliably, than any other paper filter method, including the syphon.

I believe that the index of deflection angle to coffee soluble concentration that the refractometer is using was probably created taking into account the other solutes you mention, but I must admit that I'm really out of my depth when it comes to talking about the workings of refractometers, conductivity meters, etc., so I'm just going to stop talking about them now.

To calculate TDS (a.k.a. brew strength):
TDS = mass of coffee × extraction yield ÷ ((mass of water + (mass of coffee × extraction yield)) - (mass of coffee × water absorption factor))

To calculate Extraction Yield
Extraction Yield = TDS × (((mass of water × TDS) + mass of water) - (mass of coffee × water absorption factor)) ÷ mass of coffee

You could use 2.086 for the water absorption factor, but 2.0 is generally sufficiently accurate. Remember significant digits?

longer form:

Given extraction yield, mass of coffee, mass of water, and water absorption factor, calculate TDS:
TDS = mass of solubles ÷ mass of solution
mass of solubles = extraction yield × mass of coffee
mass of solution = mass of water + mass of solubles - absorbed water
absorbed water = mass of coffee × water absorption factor
Therefore, TDS = mass of coffee × extraction yield ÷ ((mass of water + (mass of coffee × extraction yield)) - (mass of coffee × water absorption factor))


Given TDS, mass of coffee, mass of water, and water absorption factor, calculate Extraction Yield:
Extraction Yield = mass of solubles ÷ mass of coffee
mass of solubles = TDS × mass of solution
mass of solution = mass of water + mass of solubles - absorbed water
absorbed water = mass of coffee × water absorption factor
Therefore: Extraction Yield = TDS × (((mass of water × TDS) + mass of water) - (mass of coffee × water absorption factor)) ÷ mass of coffee