Nov 012009

I’ve begun a new experiment examining how to best scale up a yeast starter from a glycerol stock, plate, or slant. Typically when starting from one of these storage methods, a 5-10 ml culture is inoculated. From there one can either pitch that culture into the final starter volume (typically 1-2 L). Alternatively, one can scale up the culture until the desired final volume is achieve, let’s say 5 ml to 100 ml to 400 ml to 1.6 L. I haven’t seen any hard comparison data showing one method is better than the other. I’ve made good beer by both methods, but I haven’t directly compared methods to determine whether there are flavor differences between them. Therefore, I have begun an experiment to do just that.

I am going to brew a blonde ale, and split the batch. To one half, I will add yeast from a one-step starter. To the other half I will add the same number of yeast from a step-starter. Then I’ll compare the flavor differences between the two.


I’m going to make a 1.040 SG blonde ale.  For each 2.5 gal halves, I will need to make 800 ml starters, assuming 1e8 cells per ml of starter, and pitching 7.8e10 cells per 2.5 gal.

On 26Oct2009 I inoculated 5 ml of 1.040 SG wort with a small amount of Wyeast 1056 glycerol stock.  Kept at room temperature for several days, shaking several times a day.

On 29Oct2009, I made 1.6 L of 1.040 SG wort.  After boiling and cooling, I split the wort into four parts.  For the one-step starter, I filled a 2 L Erlenmeyer flask with 875 ml of wort.  For the multi-step starter, I filled a 12 oz bottle with 50 ml, a 750 ml-bottle with 200 ml, and a 2 L flask with 625 ml of wort.  Stored at 4C.

On 30Oct2009 I inoculated one 875 ml starter with 2.5 ml of the culture, and I inoculated one 50 ml starter with 2.5 ml of the culture.  I added a stir bar to each, and stirred on stir plates.

On 01Nov2009, both cultures are growing.  I stepped up the 50 ml starter to 250 ml by adding the 50 ml culture to the 750-ml bottle containing the 200 ml wort.

On 02Nov2009, stepped up the culture to 875 ml by adding the 250 ml culture to 620 ml wort in a 2 L Erlenmeyer flask.  Put on stir plate and let it go.

Cultures had been going well for a few days now, and were probably winding down.

Yeast Count

On 05Nov2009 I counted the number of cells in both starters using a Beckman Vi-Cell XR 2.03 automated cell counter.

  • One-step starter – 8.9e7 viable cells per ml, 96.4% viable
  • Multi-step starter – 7.4e7 viable cells per ml, 94.2% viable

I continued the cultures stirring overnight to determine whether the cultures were finished dividing.

On 06Nov2009, I counted cells again.

  • One-step starter – 8.6e7 viable cells per ml, 95.3% viable
  • Multi-step starter – 7.3e7 viable cells per ml, 94.9% viable

The one-step starter had about 18% more cells than the multi-step starter.  I doubt that this is significant, and I imagine if I repeated this several times I will see no significant difference in the cell numbers between the two starter methods.

Cultures were stored at 4C overnight to flocculate the cells.

Brew Day

I brewed an American blonde ale for this experiment, so that any differences in flavor between the starters can be easily detected.  Here’s the recipe:

Briess 2-row: 8 lbs
Weyermann Munich Malt: 0.75 lbs
0.5 oz Cluster (7% AA) – 60 min
0.5 oz Czech Saaz (2.8%) – 30 min
0.5 oz Czech Saaz (2.8%) – 10 min

Starting Gravity: 1.042
Efficiency: 73%
Calculated SRM: 2.2
Calculated IBU: 19.2

Mashed at 152F for 1 hour with 1.25 qt water per lb grain.  Collect 3.5 gal wort, batch sparged with 3 gal water at 168F.  Boiled for a total of 1 h.  Added whirfloc at 45 min.

Chilled wort, transferred to fermentor, and aerated by shaking for 15 min.
Split wort into 2.75-gal aliquots, and pitched equal numbers of either the single-step starter or multi-step starter.

On 21Nov2009 the final gravity of both was 1.010, yielding 76% attenuation.  Therefore, in this experiment there was no difference in attenuation between the two starter methods.

I bottled with 2.5 oz dextrose, to give ~2.7 volumes of CO2 once carbonated.


My wife and I did a blind tasting of the two batches of blonde ale.  The beers were poured into three glasses per beer that were marked in such a way that they could be identified after the tasting, but not during.  Then we each chose three glass randomly, each receiving two of one and one of the other.  We then tasted to determine whether there were any perceivable differences, and to determine whether we could match the two glasses that were the same beer.  The results were thus.  Meredith correctly matched the same beer.  She thought the single-step batch was more watery, and less flavorful.  She found the multi-step batch to be more fruity.  I incorrectly matched the beers.  So, one can conclude that I could not distinguish the two batches.  An overview of the beer: it has a malty aroma, with a subtle hint of hops.  It has a medium-low body.  There is a slight fruity flavor with some maltiness.  No hop flavor.  It’s a nice lawnmower beer, in my opinion.


Overall this was a nice experiment to see if there was a significant difference in the final beer using yeast grown from 2 different starter methods.  I personally could taste no difference between the two, and will likely only use the one-step method in the future.  However, this is the result of only one experiment, tasted by only two people.  To really validate the experiment, it really does need to be repeated and the results tasted by a much larger panel.

 Leave a Reply

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>