Many brewers wonder what difference
bottle conditioning makes in brewing. One aspect of bottle
conditioning is the presence of live yeast its effects on the aging
beer.

Based on an on-line discussion, which I
had with fellow brewers, I designed an experiment where I added a
small amount of live yeast to my Doppelbock when I bottled the
carbonated beer after cold conditioning.

The recipe, which I brewed about 7
month before this sampling, was very similar to the recipe posted on
braukaiser.com. It was brewed with an enhanced double decoction. The
first week of primary fermentation was done at 8 C (46 F) and the 2^nd
week was done at 10 C (50 F). This was followed by a 1 month
maturation at 12-13 C (54-56 F) during which time the beer was racked
off the primary yeast into a Cornelius keg with shortened dip tube
where it was allowed to reach its final gravity of 4.8 Plato. To help
this maturation the beer was kraeusened with WLP830 (German Lager) 2
weeks into this maturation rest even though the primary fermentation
itself was done with WLP833 (Ayinger Lager). The attenuation going
into the following cold conditioning was 73% while the attenuation
limit of the wort was 76.6%. During the maturation rest the beer also
built up natural carbonation.

The beer was cold conditioned for about
1 month and 2 weeks and then bottled straight from cold conditioning
at 1 C (34 F) into chilled bottles. The bottles were not purged with
CO2 prior to filling. Oxygen scavenging caps were used because these
were the only ones I had at hand. 3 of the bottles received about 300
Million cells of an active WLP830 (German Lager) culture. This
yielded about 1 Million cells per ml in these bottles. No sugar or
other fermentables were added since the beer was already carbonated.

After bottling the bottles were stored
in the basement. The ambient temperature started at about 13 C (56 F)
at bottling time and rose to about 17 C (64 F) over the following 3
months. 2 days before sampling the bottles were cooled to 10 C (50
C).

3 month after bottling and about 7
month after brewing I sampled a bottle that was bottles without
additional yeast (non-yeasted sample) and a bottle that was bottled
with additional yeast (yeasted sample) side-by-side. At the time of
sampling I knew which was which but didn't know what to expect from
the yeasted sample.

Aroma:

The non-yeasted sample showed the
typical dark fruit (including black currant) and malt aroma of a
German Doppelbock with a hint of alcohol while the yeasted sample
showed a much more restraint aroma. The malt notes and dark fruit
notes were rather faint. The aroma was more that of the young beer.
There was also a hint of alcohol in its aroma.

Appearance:

Both beers were clear at serving
temperature. The yeasted sample formed a thin, yet dense, yeast layer
on the bottom of the bottle.

Foam stability:

My standard foam stability test is to
take a Koelsch glass, pour the beer down the middle to let it foam up
until the foam reaches the top of the glass. Then the time it takes
for the foam to collapse and show the beer surface is taken. For both
samples it took more than 13 min for that to happen.

Taste:

The taste experience was similar to the
aroma experience. The non-yeasted beer showed a much stronger and
more complex taste while the yeasted sample was more subdued. Both
samples did not exhibit any off-flavors. In both cases the bitterness
was low and did not linger into the finish. The non-yeasted sample
was a bit sweeter which was also reflected in its lower attenuation

Mouthfeel:

Both samples exhibited the same
mouthfeel.

Stats:

The yeasted beer did ferment a little
further since at bottling time a small amount of residual fermentable
sugar was available as can be seen from the attenuation to
attenuation limit difference of 2.6% for the non-yeasted beer.

Conclusion:

The result of this tasting did surprise
me yet supports my thinking that the hallmark flavor of German Bocks
and Doppelbocks is in fact the product of oxidation and staling of
the beer. It is assumed that the yeasted beer sample did not exhibit
that flavor as strongly since the yeast was able to scavenge the
oxygen that had been introduced during the bottling process.

The added yeast did not affect the head
retention negatively in this case. One way it can do this is by
releasing excessive amounts of Proteinase A into the beer which can
break down foam proteins.

The pH was not negatively affected
either which is a sign that there was not an excessive amount of
yeast autolysis or not enough yeast to make a difference.

The idea that big dark beers benefit
from small amounts of post fermentation oxygenation has also been
brought up by fellow home brewer Fred Bonjour and warrants further
investigation into optimal oxygenation rates ad well as aging times.

When reading up on brewing Weissbier (also known as Bavarian Wheat) one of the suggestions is a ferulic acid rest. This rest around 43 C
(110 F) works best at a pH > 5.7 and liberates ferulic acid into the
wort. This ferulic acid is the precursor to 4-Vinyl-Guajakol which is
responsible for the the clove flavor produced by Weissbier yeats. The
more ferulic acid there is in the wort the more 4VG should be produced
by the yeast and the more clove character the beer should have.

This
is what I wanted to test. So I brewed a Weissbier recipe twice. Once
with a simple Hochkurz mash and another one with an additional 30 min
43C rest at a pH > 5.70. For the second beer acid malt was added at
61C. This is above the optimal range for protoelytic activitry since I
also wanted to limit the protein degradation during the time the mash
spent in the 45-55C range.

The following table lists the process steps taken for the 2 beers:

Note that the fermentation for the 2nd
batch slowed down signficantly after it reached a gravity of 6 Plato.
At this point I decided to pitch a lager yeast and I cooled the beer
for the time it took to propagate that yeast. This was to drop out most
of the original yeat and limit autolysis. This was unplanned and I hope
it is not the reason why the results of the experiment are like they
are.

Tonight I tasted the two beers:

Conclusion: For the chosen yeast holding the ferulic acid rest
didn't make any noticeable difference in the clove flavor that was
produced during fermentation. While additional experiments should be
made to confirm these findings it is very much possible that this rest
is not worth the additional work.

(comments are disabled b/c
of problems with spammers. Send your questions and comments to kai at
braukaiser dot com)

Many articles about foam stability,
like this BYO article for example, mention that foaming during beer production can hurt the head
retention of beer. The explanation is that foaming consumes the
foaming compounds in the beer. I wanted to see if I can demonstrate
this with a simple experiment.

I took two 500 ml PET soda bottles and
filled each with about 300 ml of 1 C carbonated beer using a beer gun
like device. The beer was batch A from the Kraeusen Experiment. Both
bottles where then purged of air by squeezing and closing them once
all air was squeezed out. To fill the head space with CO2 and provide
the same storage conditions for both beers both bottles where shaken
up once. After that they were stored in a 8 C fridge. The control was
not shaken anymore but the “shaken” beer was shaken 3-4 times a
day over the next 3 days. Each time the foam was allowed to fall back
into the beer before it was shaken up again. Including the initial
shaking, which was also done for the control, the “shaken” beer
foamed up 10 times. This is 9 more than the control.

Shortly after shaking up the "shaken" sample.

On the 4^th day, after the
foam has settled completely I filled the beer into 4 Koelsh glasses
(2 for the control and 2 for the “shaken” beer) using a funnel
that was held about 12 inches above the bottom of the glass. Each
glass was filled until the foam reached the upper rim of the glasses.
A timer was started when the first glass was filled. “shaken 1”
was done filling at about 5s, control 1 was done at about 12s,
control 2 was done at about 20s and “shaken” 2 was done at about
25s.

Here are images I took while the foam
was collapsing.

Shortly after the glasses were filled. The two glasses on the left are the "shaken" beer and the 2 glasses on the right are the control. The order of filling them was 2, 3, 4, 1 with 1 being the left most sample and 4 the right most.

2 minutes into the experiment

After 8 min. All samples appear to have the same amount of head left. The numbers in the lower section of the picture indicate the order in which the glasses were filled. 

 Looking onto the top of the beer. In all samples the surface of the beer starts to show.

 After 9 minutes the foam in all samples receded to a point where beer is visible.

For both beers it took about 8-9 min
until some beer showed through the foam when looking into the glass
from the top. There was no significant difference in foam stability
between the two beers. The same was true for lacing. While this
method of evaluating foam stability is not as precise and repeatable
as the ones performed by beer analysis labs, it is good enough to
provide an objective assessment of the foam stability of a beer
using.

Conclusion

I'm not saying that foaming does not
affect foam stability at all but I was not able to demonstrate this
using this simple experiment. It is possible that other factors play
a role or that the effect is too small to be detected with such a
crude approach. But for now, I would not worry about heat retention
being reduced noticeably when the beer happens to foam during the
brewing process.

Common brewing advice in American home
brewing is to let the Kraeusen fall back into the beer after primary
fermentation finishes. Very few brewers question this advice.
However, all the German books I have read about brewing and some
American home brewing books state that the bitter gunk on top of the
Kraeusen should be removed. If it is allowed to fall back in the beer
it will impart a harsh bitterness. As a result I have always
fermented in 5 gal carboys and removed the Kraeusen through blow-off.

In order to find out how much taste
difference that makes I set out to conduct an experiment. I brewed
two batches of my Altbier. Both were fermented in buckets. On the
first batch I allowed the Kraeusen to fall back into the beer. In
fact I helped it a little towards the end since I needed to rack the
beer before it was completely fermented. This however should not
invalidate the results since most brewers leave the beer in their
carboys well past the end of primary fermentation and until all
Kraeusen has fallen back. For the second batch I skimmed the brown
gunk off the Kraeusen regularly.

Both batches finished fermentation
during a maturation phase in a corny keg before they were moved to 4
C to settle the yeast and precipitate haze. Since they were still
cloudy after 2 weeks I added 3.5 g (½ pack) dissolved gelatin to
each keg which helped clearing the beer.

The following table outlines the
brewing process used for both beers:

 *) that pH value doesn't make sense to me. It should not be higher than the pre-boil pH and not that much different from the pH for batch B. But I did take the pH measurement on a sample of wort that had been standing unpitched for 24 hrs as opposed to the batch B post boil pH which had been taken the same day.

Both batches were fermented in buckets. A clear pot lid kept contamination out and provided easy access to the Kraeusen. The picture shows the batch A. The opening in the blue bucket lid was later enlarged to allow regular Kraeusen skimming for batch B.

The beer was partially naturally
carbonated during maturation and then force carbonated during cold
conditioning. The carbonated beer was bottled straight from the cold
conditioning kegs.

I presented the beer as a double blind
(participants didn't know the difference ) triangle test to 7 club
members. Only 3 were successfully able to tell the difference. Those
who were able to separate the beers correctly reported the following:

I was very surprised how few were able
to tell a difference which appears so clearly to me. So I poured
myself 3 triangle tests and have to admit that I only got 2 correct.
Though I knew what to look for it wasn't as easy to keep the beers
apart since the lingering bitterness of A seems to stick with one for
longer enough to affect the taste of the next beer.

While this was a double blind triangletasting at a club meeting it was fairly unorganized. I didn't not get to start before many of the participants already had other beers. The setting was also not as quiet and free of distractions as one would expect for a taste testing.

The difficulty to differentiate the
beers in blind tasting may explain why some brewers, who have tried
this experiment before, found no difference and thus claim that it
doesn't matter if the brown Kräusen gunk is removed or not. The type
of beer may also play a big role. I can imagine that a hop dominated
highly bitter IPA may not show the difference or may even provide a
case where the beer, which didn't have the Kraeusen removed, is
preferred. Having done this experiment and tasted the difference I'm
convinced that the Kraeusen needs to be skimmed or blown off for any
German style beer. The type of harsh and lingering bitterness, which
I experienced in A, is considered a flaw even in the more bitter
German styles like Northern German Pils and Altbier. The bitterness
should be clean and linger only little. When it fades in the after
taste is should never reappear later. The only German beer where I
had this happen to me was Oettinger Pils which is one of the cheapest
beers you can by there.

The results are in line with similar experiment reported in Zymurgy. The article can be
read at the AHA forum.

When I sat down for lunch today I had 2
bottles of this beer and thought I poured the good one. After taking
the first gulp I noticed that I got the wrong one. To me the taste
was so bad that I poured it down the drain and poured the other beer
which I was able to enjoy. I'll likely only finish beer B and pour
out beer A.

Conclusion

Removal of the bitter Kraeusen gunk
makes a difference in the quality of the beer even though it may not
be detected by all brewers. The outcome of this experiment is enough
to suggest that interested brewers try this on their own to see if it
can improve the quality of their beers.

updates:

(1) to make up for my own failure to pass 3 triangle tests with this beer I set up a different taste test tonight. I took 12 identical glasses. 6 were filled with batch A and 6 filled with batch B. I then asked my wife to set up a random line of all 12 glasses. Taking my time and cleansing my palate with bread and water I went though each glass and took one to two sips to determine which beer it was. In the end I was able to separate them precisely based on both their hop taste and lingering bitterness. It shows that if I take my time I'm able to tell them apart reliably. 

(2) 1 month after the initial taste
testing I brought samples to a club meeting and was surprised to see
that the difference, which was very clear to me earlier, has aged
away to some extend. Knowing what to look for I was still able to
taste a difference but at this point I would not be surprised if
others can't tell them apart.

This is a blog entry I have been thinking about a while. How precise is the ppg (points per pound and gallon) based efficiency calculation really. The reader should see this as something that is interesting to know and more of an exercise in using Plato and sg rather than something that any brewer needs to worry about

When calcylating efficiency (American) home brewers usually use:

(1) Eff = 100 * (gravity points of wort * wort volume) / (grain weight * grain extract potential)

Wort volume is given in gallon, grain weight in pound and extract potential in ppg. But that's not how efficiency is actually defined. It is defined as the ratio between the extract weight in the kettle vs. the extract potential of the grain:

(2) Eff = 100 * extract weight in kettle / grain extract potential

The grain extract potential is simple. It is its weight multiplied with the extract content determined in the laboratory mash. For most base malts it is about 77% (80% dry basis extract and 4% moisture content). Going forward I will call the grains extract potential "e". The weight of the extract in the kettle is a bit more complicated. For that we have to look at the Plato scale. Many brewers know degree Plato as another way of expressing wort strength. To be exact: the wort strength in Plato is the ratio between the weight of the extract dissolved in the wort and the the total wort weight:

(3) P = 100 * extract weight in kettle / wort weight in kettle

Extract weight in kettle is what we need for (2) but I still need the wort weight weight in the kettle. For that I simply remember that sg (specific gravity) is nothing else than the density of the wort in kg/l. It follows that the wort weight in kg is the product of wort volume in l and its specific gravity:

(4) extract weight in kettle = sg * wort volume in kettle

Now I can calculate the actual efficiency by using (2), (3) and (4). First some clean-up and shorter notatons for the variables:

• Eff_ppg = Efficiency calculated using gravity points and ppg for extract potential
• Eff_% = Efficiency calculated using Plato and extract % for grain extract potential
• P = wort strength in Plato
• sg = wort strength in specific gravity (1.xxxx)
• GP = wort gravity points ( = (sg – 1)*1000)
• V_l = wort volume in liter
• V_gal = wort volume in galon
• m_kg = grain weight in kg
• m_lb = grain weight in lb
• e_% = extract potential of the grain in %
• e_ppg = extract potential of the grain in ppg

With that the two efficiencies are:

(5) E_ppg = 100 * GP * V_gal / (m_lb * e_ppg)

(6) E_% = 100 * sg * P * V_l / (m_kg * e_%)

Grain lab analysis results don't show the extract as ppg but as percent of dry weight. To get the ppg equivalent I need to find a formula that calculates e_ppg from e_%. Since it is assumed that both efficicncy calculations (5) and (6) are equal I can set them equal:

(7) E_ppg = E_%

Now the busy work. They both use weight and volumes but in different units. That will be fixed by assuming these conversions:

(8) V_l = V_gal * 3.78
(9) m_kg = m_lb * 0.45

For simplicity I'll be using the simplified Plato to sg conversion. I'll later discuss how much that makes a difference.

(10) P = GP / 4 = (sg – 1) * 250

After putting all this into (5) and (6) I end up with this huge equation:

(11) 100 * (sg – 1) * 1000 * V_gal / (m_lb * e_ppg) = 100 * sg * (sg – 1) * 250 * V_gal * 3.78 / (m_lb * 0.45 * e_%)

Luckily this can be cleaned up considerably. V_gal and m_lb exist on both sides and fall out. So does (sg-1). All the constants can be consolidated into one. What's left is this:

(12) 0.476 / e_ppg = sg / e_%

solving this for e_ppg gives:

(13) e_ppg = 0.476 * e_% / s_g

This equation means that the extract potential in ppg depends on the grains extract potential in %, which is to be expected, and the specific gravity of the wort for which efficicncy is calculated. This was not expected. Here are a few examples. If sugar, which has an extract potential of 100%, is used to make a 1.040 sg wort it has an extract potetial of ~ 46 ppg. If it was used to make a 1.080 sg wort it has an extract potential of only ~ 44 ppg

The same is true for a base malt with 80% dry basis extract and 4% moisture. The actual extract content is 76.8%. If used for 1.040 wort its ppg extract potential is ~36.0 ppg. When used for 1.080 wort the extract potential is ~34.6 ppg.

As a final exercise lets look at a chart that plots the two efficies over the gravity of the wort. The wort volume is held constant while the grain bill is scaled such that the "%" based efficiency remains constant. In addition to that, the sg to Plato conversion is done using the officicial ASBC conversion formula which is a polynominal fit of their sg to Plato tables [deLange]:

(14) P = -616.868 + 1111.14 * sg – 630.272 * sg^2 + 135.997 * sg^3

While there are many similar formulas out there, this is the official one given by the ASBC (American Society of Brewing Chemists) and it should be seen as the standard.

This is the chart I came up with

It plots 3 curves. "Eff_% using the exact sg to Plato conversion" uses (14) to convert between sg and Plato. It is constant at 70% because this formula is used to calculate the necessary grain weight for the given volume and specific gravity. "Eff_% using the simple sg to Plato conversion" uses (10) to calculate the sugar content (Plato) from the specific gravity. "Eff_ppg" calculates the efficiency using gravity points and an extract potential of 35.7 ppg.

Despite the existing discrepanacy and incorrectness of the ppg based efficiency calculation, which I discussed earlier in this text, it tracks very well with the actual efficiency of 70% over a wide range of specific gravities. The reson for this is simple: while I showed that technically the extract potential in ppg also depends on the specific gravity, I also simplified the sg to Plato conversion by using (10) instead of (14). Both errors compensate each other to some extend. This also becomes clear when looking at the efficiency which is calculated using the simple sg to Plato conversion. It already shows an error of ~4 percent point at a specific gravity of 1.100.

Conclusions

Does it really matter in brewing whether you use the ppg based forumla or the Plato based one? Not really. If you always use the same formula for efficiency calculation and subsequent recipe design it doesn't matter at all. It may matter when discussing and comparing efficiency with other brewers. In this case the ppg based approach is within 1% of the actual efficiency for all realistic gravities. That error, however, is too small to be a conern in home brewing. Using the % based efficiency calculation with a crude sg to Plato conversion, on the other hand, can overestimate efficiency significantly. Thus care needs to be taken when converting Plato or Brix readings into specific gravity readings. That is in particular true for high gravity worts.

One last word about ppg or "points per pound and gallon". It should be called "point gallons per pound" or pgp since it is an expression of how many "point gallons" (gravity points multiplied with gallons) one can get from one pound of grain, sugar, etc. Its actual unit is gal/lb. 

[deLange] A.J. deLange: Specific Gravity Measurement Methods and Applications in Brewing.

This was the first time that I compared
dissolved chalk against undissolved chalk in a 5-gal “production”
batch of beer. Up to this point I have only done small scale
experiments. Those experiments suggested that chalk dissolved with
CO2 would be twice as potent in raising the mash pH as undissolved
chalk is. As a result I new that I should cut the amount of chalk
needed in half when it will be dissolved with CO2.

To brew the Schwarzbier I used the
following grist. This is my standard recipe for a Schwarzbier:

• 53% Pilsner malt

• 40% Munich Type II malt

• 4% CaraMunich III malt

• 3% Carafa I special

The water was prepared from reverse
osmosis water by adding the following salts. Version A uses
undissolved (i.e. suspended chalk) while version B used dissolved
chalk

The resulting profile was calculated as
follows. Note that I do have an old analysis of the reverse osmosis
water which I included in the calculated mineral profile:

*) There is some ambiguity as to how
much calcium is actually contributed by undissololved chalk since it
contributes only half its alkalinity potential, it may also
contribute only half its calcium. These results assume that the chalk
contributed all its calcium. The result is a lower residual
alkalinity compared to the water with only half the chalk but
dissolved.

The salts were then weighed. For beer
A, they were mixed into the strike and sparge water. Since the chalk
was not dissolved the water remained cloudy. Water treatment for the
strike water was done in the mash kettle.

For beer B the salts were added to 2
liter soda bottles and reverse osmosis water was added. Then the
bottles were carbonated with a carbonator cap. Once sufficiently
carbonated the water cleared overnight which was a sign that the
chalk got dissolved. This water was then added to the remaining
reverse osmosis water for mashing and sparging. The mash water was
prepared the night before to allow residual CO2 to escape. No chalk
precipitated during that time, There was also no precipitation of
chalk during the heating of the strike water or the sparge water.

The resulting pH values during the
brewing process are shown in the following table. All pH values were
measured with a sample cooled or heated to 25 C

For both beers the pH dropped during
mashing which I contribute to the continued release of acidic
compounds from the dark specialty malts. One oddity is that for batch
A, which used undissolved chalk, the kettle full pH is lower than the
cast out pH. Generally the pH falls during boiling. This is something
worth paying attention to in future batches although it may also have
been a measurement error. The initial mash pH of batch B is greater,
which supports the fact that the residual alkalinity of its water
should have been higher. This is the case if all the calcium added by
the chalk is considered for undissolved chalk as it was done in the
aforementioned water analysis.

I have not yet done a final tasting
with these two beers. But preliminary tasting of both batches during
their fermentation and conditioning did not show any significant
differences

Conclusion:

To achieve roughly the same mash pH,
only half the chalk is needed when it is dissolved with CO2.

After last year's Maibock, this is the
2^nd experiment where I compared a beer brewed with
decoction mashing and a beer brewed with infusion mashing.

This time I wanted to see if there is a
more pronounced flavor difference if the majority of the grist was
composed of highly kilned base malts. This is one type of grist for
which decoction mashing is still fairly common in Germany. test test test . So I chose
a basic Dunkel recipe and the brewing process is outlined after the
mash diagram for the 2 beers (click the diagram for a larger version).

It should be noted that the Dark Munich
malt caught me by surprise and the mash for Dunkel II resulted in a
rather unfermentable wort (attenuation limit 71%) which was
compensated for during the mash of Dunkel III (see longer maltose
rest). As a result the wort for Dunkel III was more fermentable. But
both beers finished with a similar attenuation (67% and 69%). The
poor fermentability was attibuted to the enzymatic weakness of the
Best Malz Dark Munich which took a long time to convert (see the 40
min 70C rest of the decoction) and showed similar attenuation
problems in subsequent beers.

3 ½ months after brewing Dunkel
II and 3 months after brewing Dunkel III I tasted the beers
side-by-side. It should be noted that at the time of this tasting I
was not aware that I brewed one with decoction and the other one
without. I had brewed quite a number of other beers in between and
actually forgot how I mashed these beers and thought that they were
both brewed with decoction until I checked my notes.

As you can see I did notice differences
berween the beers but it is difficult to tie them to the decoction
alone. I contribute the better clarity, lower head retention and
thinner mouthfeel of the more intensely mashed Dunkel III to the
stronger protoelytic activity in the mash. Its increased sweetness
stems from the larger amount of residual fermentable sugars (see
attenuation delta) compared to Dunkel II. I even considered Dunkel II
(the non-decocted, more precisely only 3 min thin decoction boil) to
be the more malty of the two beers.

Conclusion: This experiment was
not as conclusive as the Maibock experiment and I would even call it
inconclusive. There were too many differences between the analytic
parameters (in particular the attenuation numbers) of the two beers
to tie their slight taste differences to the more intensive mashing
(including a 35 min decoction boil) of the Dunkel III. A future
experiment needs to increase the decoction boil time to 60 min and
attempt to keep the original extract, attenuation limit and
attenuation and fermentation the same.

(Comments have been disabled due to
spam being posted. Please send comments and questions to kai at
braukaiser dot com)