samuellaw178 wrote:That's totally valid here. That's the highest temperature reasonably before it may char or change the dissolved solid in the coffee (since 90-95C is the brew temp). Also, I did some research here before on HB, and I think I see a reference to using that temperature. I would like to get even hotter for faster drying, but am afraid it will have unintended effect.

Sample size is 5g(thereabout). It depends on the accuracy & resolution of the scale. The more sensitive it is, the less sample you can get by. I thought of using my own 0.1g scale, but the resolution is insufficient and I have to use at least 100g sample for that. The miligram scale is able to resolve 0.1mg, and can do so with confidence in +-1mg. So I want my dried coffee weight to be at least 10 times more than that to be measuring something significant. The dried coffee ended up in the range of 60-70mg, which is good enough to get me in the range of fairly good estimate.

It took me about 2 hours total for drying and to confirm it's dried. It's in a tube so takes longer to dry.

I have verified it through taking duplicate sample (twice). Both trials seems to be very very close (1.435% vs 1.427%, and another one I need to check back my notes if that's important).

Accuracy wise, I would think dry and weight method is the gold standard for getting proper TDS (no brix meter or refractometer etc), before VST refractometer takes over. VST is essentially a more sensitive/accurate refractometer that takes refraction information. And that refraction data is correlated to the established dry-and-weigh TDS so that users don't have to go through the drying process every time.

Thanks for the response. I can't justify the cost of a refractometer when my primary brewing device costs $25, so I've done what you're doing, albeit with less expensive equipment.

I tried to figure out what the optimal temperature for dehydration of coffee is and couldn't access a standard (you have to pay for that) but got the sense that it was something pretty close to 100°C. One document I found suggested just a hair over that value, and I've been drying my samples at 102°C. But I really don't know what is best.

You can't really do statistics with just two samples. You're probably getting good numbers though.

The reason I ask about accuracy is because, well, how do you know you're not introducing some weird systemic error? For example, you are concerned about the temperature being too high but aren't really sure what it should be. Or: How are you managing the hygroscopic nature of the dried coffee between when you finish drying it and when you weigh it? You really need something reliable and proven to compare your results to or else you could be getting consistent numbers that are always off for some unanticipated reason.

This is largely academic since for coffee if you're getting 1.435% TDS and it's really 1.527%, who cares? The fish you're trying to fry are bigger. But it is still a question worth asking, if you have the means to answer it.

My own efforts require 2-3 times your sample size, roughly the same amount of time to dry but likely less precision (~±0.02% TDS).

samuellaw178 wrote:Thanks so much for your comments. You're spot on with the muted bitterness. That's what my dumb palate tasted but doesn't know how to describe.

As for the coffee dosage, if I bump that up with the others held constant, I would expect to see a bit more than 1.7% in the undiluted, and >1.43% in the final cup with extraction yield closer to 20%. Should be a stronger cup but less the bitterness. But from practical point of view, I think I will coarsen the grind size first, then reduce the steeping time, to achieve lower undiluted TDS (will aim for 1.5% range), that gets me closer to 20% yield.

Just curious, did the VST software come up with the 1.7% (for 41.5g dosage calculation) or it was a user-input value?

I got it from you!

It was your estimate for TDS of the undiluted brew, based on TDS of the diluted brew adjusted by the dilution percentage. In fact, all of the parameters came from you: 1.70% YDS, 35g coffee and 460g of brew water. With those inputs, the VST software calculates 23.73% extraction yield and 385g of beverage.

But looking back at your chart, I see that you have two estimates for beverage weight, which differ depending on the weight of the retained water. Estimating the pre-dilution and post-dilution beverage weight doesn't make sense when you can easily weigh the final beverage to get the post-dilution weight and simply subtract 100g to get the pre-dilution weight.

I couldn't find anywhere in this thread where you state the weight of the diluted beverage, so I'm assuming you haven't weighed it. If not, please do.

Armed with that number, you can calculate the pre-dilution TDS using this formula given to me by a knowledgeable friend:

(Final %TDS x Final Weight of combined Beverage) / Initial Weight of Beverage

For example, if the final (post-dilution) weight of the beverage really is 490g, then the initial (pre-dilution) weight of the beverage is 390g (490-100). If you you measure 1.43% TDS, then the undiluted TDS would be 1.7966667%. Let's call that 1.8%, which is higher than your 1.7% estimate. The VST software calculates over 25% extraction yield for those parameters.

My point here is that there's no need to estimate the pre-dilution TDS. If you weigh the final beverage, and your measurement of the post-dilution TDS is reasonably accurate, you can calculate the post-dilution TDS without making guesses about the weight of the liquid retained in the grounds.

The missing piece of data is the actual measured post-dilution weight of the beverage. If you give me that I can re-run the parameters with the correct undiluted TDS and get you the updose weight that will give you 20% extraction yield. Again, you have to keep the grind and steep time constant for this to work.

samuellaw178 wrote:As for the coffee dosage, if I bump that up with the others held constant, I would expect to see a bit more than 1.7% in the undiluted, and >1.43% in the final cup with extraction yield closer to 20%.

I forgot to address this in my last post.

No, the TDS will remain at 1.70%. If you updose without changing any other parameters, the pre-dilution TDS doesn't change because you're extracting less weight from the grounds. The beverage weight will decrease slightly as well. Less coffee extracted, more water retained. Extraction yield and beverage weight decrease, but TDS stays the same.

Note that the VST software doesn't calculate the %TDS. It's always a target or a measured value. I used it as a target because that's what you estimated the pre-dilution TDS to be, and we want it to be the same so that the strength will be near the middle of the range after dilution. I also entered your net brew water weight (460g) as an input. I set the extraction yield target to 20%, middle of the range. The VST software calculated a dose of 41.5g and a beverage weight of 370g.

Again, given the actual beverage weight, I can give you a more accurate updose weight.

Sorry for missing in action for a while! Been busy working hard at work over the weekend.

jpender wrote:The reason I ask about accuracy is because, well, how do you know you're not introducing some weird systemic error?

I got your question's context wrong previously. In this case, you're right. I can't be sure that I am not introducing some systemic error. But what I am sure is that I can't eliminate the systemic error, simply due to the different experiment condition than what they're doing. Even with the same procedure as CBI, I may still be introducing some unintended systemic error into my result.

But what I can do is, to make sure that the systemic error is consistent throughout, and to eliminate any obvious error that is more likely to be significant. Luckily enough, the range for ideal tds/extraction is quite broad. Due to that, the systemic error is very likely to be insignificant. And most importantly, like you, I am only using the number as a guideline. I don't need to know that my prefered 1.21% is actually their 1.24%. They all fall pretty close into the range anyway, unless there is some obvious error.

But if we want to be on the same page as everyone that's using a VST(that our 1.200% is really their 1.200%), buying one VST is the only way. For me I don't need that and can't justify that either. The VST price has climbed up significantly I think, 1.1K AUD here. I could've bought a much better roaster or another espresso machine to play with for that money.

Peppersass wrote:I got it from you!

Suspected so. I thought for a moment the VST was able to predict that, which was unexpected.

Peppersass wrote:I couldn't find anywhere in this thread where you state the weight of the diluted beverage, so I'm assuming you haven't weighed it. If not, please do.

Yes, you're right, I did not measure that yet. With that value, I can know how much water exactly is retained in the coffee ground as well.

Peppersass wrote:No, the TDS will remain at 1.70%. If you updose without changing any other parameters, the pre-dilution TDS doesn't change because you're extracting less weight from the grounds.

Hmm, I would disagree with that actually. Imagine if you brew French Press(it is immersion method like Vac Pot without dilution), one with 35g coffee to 500ml water, another with 41.5g coffee to 500ml water, I would argue the 41.5g brew is stronger (because TDS is higher). The increase in TDS is not proportional to the updose which makes it complicated (ie: 50% updose doesn't mean 50% increasse in TDS).

Since you have the VST refractometer, do you think it's possible for you to confirm a small trial on that (with maybe 10g vs 13g coffee instead of the large amount quoted)?

Peppersass wrote:Again, you have to keep the grind and steep time constant for this to work.

You're asking me to delibrately running a bad overextracted brew, aren't you?

Do give me a few days(a week I think), I will run a few Vac Pots and come back with proper TDS & weight measurement. My beans have run out and it's probably meaningless to try drawing conclusion from comparing different beans.

samuellaw178 wrote:You're asking me to delibrately running a bad overextracted brew, aren't you?

Hmm, I would disagree with that actually. Imagine if you brew French Press(it is immersion method like Vac Pot without dilution), one with 35g coffee to 500ml water, another with 41.5g coffee to 500ml water, I would argue the 41.5g brew is stronger (because TDS is higher). The increase in TDS is not proportional to the updose which makes it complicated (ie: 50% updose doesn't mean 50% increasse in TDS).

Uh, yeah that's probably right, and for vac pot you have the added issue of the draw down time increasing when you add more coffee, and that'll increase the extraction as well.

You'll have to grind coarser, too. You could grind coarser with the 35g dose, but my guess is that the TDS will drop enough that the diluted beverage will be too weak. So updosing and grinding coarser should get you a lower extraction yield but will also put the strength in the middle of the range.

I'll have a better idea of the right dose to use for a 20% extraction yield after you measure the beverage. Then all you'll have to do is adjust the grind to hit something like 1.35% TDS in the diluted beverage.

I'll try to run some experiments this week with my vac pot. I can't test the dilution, but I can compare TDS, extraction yield and draw-down times for different doses while holding the grind constant.

samuellaw178 wrote:But if we want to be on the same page as everyone that's using a VST... buying one VST is the only way.

It's not the only way. But I agree with you that consistency is more important than an absolute number, as long as you're not way off.

Hmm, time for some update. I did more measurements, and things are getting more complicated instead.

How much water does the coffee retain in a Vac Pot?
I measured the ground water retention by measuring the syphon top globe with wet coffee ground(immediately), and compare it to cleaned and dried top globe. The ground seems to retain about 1.47-1.5g water/ground. Mark (MWJB) is on the spot with the 1.5g/g. Of course, the dissolved solids is ignored for the calculation, otherwise it's too complicated.

Water evaporation
The extent of water evaporation depends on how long is the steep time (and the brew process). I am getting evaporation loss in between 14g-23g. That's about 2-4% of total brew water (560g).

Grind vs steep time
For my run, coarsening the grind size by 0.2mm on the Rosco Mini did not seem to have much impact. The difference is from 1.43%(previous reading) down to 1.42%(new data), but I wouldn't bet on it because they're of different coffees and I don't think my measurement is confident within the 0.01%.

Reducing steep time (including draw down) seems to have more influence. I got 1.365% final TDS by reducing 1 minute steeping time (4 minute down to 3 minute).

With some rough assumptions made, my prediluted TDS is about 1.58% and extraction yield is about 21.9%. Just a bit higher, but I think I can optimize it further by steeping 30 sec less. (Again, don't bank on these numbers, they're just estimation range based on assumptions made)

===

After all these measurements, I don't think the dilution component of Vac Pot is as significant as I thought it to be. The Vac Pot seems to compensate that by extracting more efficiently than the French Press (with the same grind/steep time).


p/s: Just for reference, I am using 560g starting boiled water and 35g coffee for all my vac pot above.

samuellaw178 wrote:After all these measurements, I don't think the dilution component of Vac Pot is as significant as I thought it to be. The Vac Pot seems to compensate that by extracting more efficiently than the French Press (with the same grind/steep time).

I suspect you're right regarding the dilution aspect. Between the full pot (70g dose, 1120g brew water) and the half capacity pot, there should only be significant difference in the strength of the final result, if they hit the same %TDS in the initial immersion phase in the top chamber - however at same temp, grind & time, they should hit a reasonably comparable yield (but at a different %TDS), then after going through draw down & dilution in the bottom chamber, both brews should pretty much normalise & come out the same, or close to.

samuellaw178 wrote:Grind vs steep time
For my run, coarsening the grind size by 0.2mm on the Rosco Mini did not seem to have much impact. The difference is from 1.43%(previous reading) down to 1.42%(new data), but I wouldn't bet on it because they're of different coffees and I don't think my measurement is confident within the 0.01%.

Reducing steep time (including draw down) seems to have more influence. I got 1.365% final TDS by reducing 1 minute steeping time (4 minute down to 3 minute).

Initially you complained that your vac pot brew was "delicate & weak". If you want a stronger brew why coarsen the grind & shorten your steep time? A finer grind (drip range) would make more sense.

Did that 4 minute steep time include the drawdown? With my Hario 5C tabletop model typical steep time is 45-48 seconds with an average total contact time time of 2:25-2:40 which includes steep time & drawdown.

MWJB wrote: Between the full pot (70g dose, 1120g brew water) and the half capacity pot, there should only be significant difference in the strength of the final result, if they hit the same %TDS in the initial immersion phase in the top chamber - however at same temp, grind & time, they should hit a reasonably comparable yield (but at a different %TDS), then after going through draw down & dilution in the bottom chamber, both brews should pretty much normalise & come out the same, or close to.

Thanks Mark, you raised a good exploration point. I will do a full pot vs half pot as my last one. The reason I don't do full pot now is because 70g itself is almost half my roast batch (from 200g green). Using a small roaster here. I agree about your point and think they should be more or less the same final TDS. Let's see.

JohnB. wrote:Initially you complained that your vac pot brew was "delicate & weak". If you want a stronger brew why coarsen the grind & shorten your steep time? A finer grind (drip range) would make more sense.

Actually my motivation for measuring the TDS is to understand more about Vac Pot method and to see if I can improve it, not exactly because it's weak. But yes, initially I thought my Vac Pot is extracting less, not more. Because...it seems to be less 'heavy' (for lack of better word) than the French Press, and it feels more delicate. Now after the measurement, I think the delicateness in fact is the (desirable) nature of the Vac Pot, not much I can change there. And the way to improve it is to reduce the extraction that I've been doing towards a better flavor extraction. It seem counterintuitive, but does make sense once you think of it this way.

May be the sludge is contributing to the increased 'perceived strength' on the Espro Press. My TDS reading for Press is 1.2-1.25% so the 'heaviness' definitely does not come from strength (alone).

JohnB. wrote:Did that 4 minute steep time include the drawdown? With my Hario tabletop model typical steep time is 45 seconds with an average drawdown time of 2:25.

4 minute includes the drawdrown. What grind size do you use for that 45 sec steep time? I reckon I could do drip grind too, just by adjusting the steep time. Another thing to do.