Author Topic: Sourdough starter quantity predictive model  (Read 21885 times)

0 Members and 1 Guest are viewing this topic.

Offline TXCraig1

  • Registered User
  • Posts: 11729
  • Location: Houston, TX
Sourdough starter quantity predictive model
« on: January 01, 2013, 01:53:26 PM »
>>> 1/3/13 NOTE: The chart below has been revised to include the effects of the Lactic Acid Bacteria in the SD culture <<<
I have also slightly adjusted the color zones. The changes to the data in the original green zone were negligable. I will post a recap of the changes down the thread.

>>> 1/6/13 I added a section here: http://www.pizzamaking.com/forum/index.php/topic,22649.msg230690.html#msg230690 that shows you how to use the tables to calculate the starter quantities and fermentation times with multiple fermentation temperatures.

>>> 1/6/13 Here is a link to a spreadsheet that automates the calculation of starter% given multiple fermentation stages: https://docs.google.com/spreadsheet/ccc?key=0AuvMQbzk5INUdGZScWx6U2lYSEtZVkJuVGJiR19NaXc#gid=0 more details here: http://www.pizzamaking.com/forum/index.php/topic,22649.msg230734.html#msg230734

This model should help identify a starting point for starter quantity and/or fermentation time in a new sourdough-leavened dough. It’s a generalized model, so as you would expect, changes in culture, hydration, salinity, dough mass, etc. will likely affect the fermentation time; therefore experimentation will be necessary to fine tune a specific dough formula. Notwithstanding, the model shows good predictive ability with typical doughs.

The model was built around real-world 75F data. This is a typical temperature coming out of the mixer, and I had a wide range of data at this point. I then used the temperature vs. growth model for C. milleri (a typical yeast found in sourdough cultures) from Gänzle et al. (1998) http://aem.asm.org/content/64/7/2616.full.pdf  to expand my model across the viable temperature spectrum.

The chart below predicts total fermentation time in hours at specific starter quantity and temperature combinations. I included all the predicted data because I thought it was interesting. I don’t however believe it is all equally reliable. For example, the model assumes 100% of the fermentation takes place at the specified temperature, and the farther away from 75F you get, the longer it takes the dough to get to that temperature thus skewing the results (not to mention the farther the model must extrapolate from the original regression). Also, as fermentation times become very extended and once you go much past 40% starter, I think there is significant risk of the proteolytic enzymes catastrophically denaturing the gluten matrix in the dough. I color-coded the chart to indicate what I believe to be the relative reliability of the predictions. The color codes are just my gut feel – your results may vary.

The graph following the prediction chart shows how the model compares to real-world data at 65F, 70F, and 75F. Only the 75F data (round black markers) was used to build the model. The black diamond and triangle markers (real-world 70F and 65F data respectively) were overlaid after the model was built as a predictive test.  

I’m curious to hear how the model output compares to other people’s experiences.

« Last Edit: January 06, 2013, 10:46:35 PM by TXCraig1 »
I love pigs. They convert vegetables into bacon.


Online scott123

  • Registered User
  • Posts: 6658
Re: Sourdough starter quantity predictive model
« Reply #1 on: January 01, 2013, 02:39:58 PM »
This is interesting, Craig.

I'm curious, how are you defining 'fermented?'  Is it volume?

Also, is starter this static or are variations in starter what you're calling 'changes in culture?'  Doesn't the plate count vary quite a bit depending on where the starter is in the feeding cycle?

Offline TXCraig1

  • Registered User
  • Posts: 11729
  • Location: Houston, TX
Re: Sourdough starter quantity predictive model
« Reply #2 on: January 01, 2013, 03:19:54 PM »
This is interesting, Craig.

I'm curious, how are you defining 'fermented?'  Is it volume?

Also, is starter this static or are variations in starter what you're calling 'changes in culture?'  Doesn't the plate count vary quite a bit depending on where the starter is in the feeding cycle?

The data was collected from my notes and from posts on the forum. "Fermented" is subjectively defined as "the dough is ready to bake." It is assumed that all starters are at their peak of activity and that assumption should carry over when employing the results of this model. Coming from multiple, uncontrolled sources, I had to normalize the data, and that was somewhat subjective as well. To make the homogenizing process as unbiased as possible, both the regression and test data were adjusted prior to building the model.

Again, I stress this model is designed to provide a very educated guess at where to start given typical assumptions across a wide range of uncertainties (various cultures, hydration, salinity, dough mass, etc). Some tweaking over several batches will likely be necessary to achieve optimum results.

I love pigs. They convert vegetables into bacon.

Offline taylorpetrehn

  • Registered User
  • Posts: 16
  • Location: Lawrence, KS
Re: Sourdough starter quantity predictive model
« Reply #3 on: January 01, 2013, 03:54:06 PM »
Great information Craig!

That's a very useful diagram. Excited to put it to the test

Online JD

  • Supporting Member
  • *
  • Posts: 968
  • Location: NE Mississippi, but NY born & raised
Re: Sourdough starter quantity predictive model
« Reply #4 on: January 01, 2013, 06:26:57 PM »
This is amazing.... excellent work! I just made a good NY street for new years eve, and your model was 100% accurate based on a 65* fermentation.

Wish I had this info a year ago. Very nice contribution to this forum. Thank you.
Josh

Offline Michael130207

  • Supporting Member
  • *
  • Posts: 138
  • Location: Maryland
Re: Sourdough starter quantity predictive model
« Reply #5 on: January 01, 2013, 09:05:01 PM »
Really cool resource Craig! Thanks for posting it.

FWIW In the past I tried to substitute ADY for starter in your original posted 48h workflow and found 0.02% ADY seemed to work alright. I did a 24 hour countertop ferment at about 65F-68F recently. After reading this post  I was curious if your data would fit with my results from that batch. I calculated a conversion constant from the ratio of ADY:Starter from my earlier attempt and arrived at 0.015 (ADY/Starter). I then checked your spread sheet and it seemed to fit well. I used 0.13% ADY and let it rise on the countertop at about 65F-68F and it was ready to go in about 24H.

I look forward to using your spreadsheet to try and predict other yeast temp combinations.
Michael

Online Serpentelli

  • Supporting Member
  • *
  • Posts: 1132
  • Age: 48
  • Location: Wilmington, NC
Re: Sourdough starter quantity predictive model
« Reply #6 on: January 02, 2013, 09:45:44 AM »
Craig,

What are your experiences with using higher % starter in terms of effect on taste/crumb/leoparding, etc?
Most of the time I would prefer to use my starter, but my schedule doesn't permit me to make the dough 48 hours in advance, and I end up throwing in ADY. :(

I will try a 10% starter dough this weekend and give it 25 hours at 65 degrees.

BTW --- Holy crap, man. That post is one hell of giant leap for pizzakind. :)

John

Offline TXCraig1

  • Registered User
  • Posts: 11729
  • Location: Houston, TX
Re: Sourdough starter quantity predictive model
« Reply #7 on: January 02, 2013, 10:45:22 AM »
Craig,

What are your experiences with using higher % starter in terms of effect on taste/crumb/leoparding, etc?
Most of the time I would prefer to use my starter, but my schedule doesn't permit me to make the dough 48 hours in advance, and I end up throwing in ADY. :(

I will try a 10% starter dough this weekend and give it 25 hours at 65 degrees.

BTW --- Holy crap, man. That post is one hell of giant leap for pizzakind. :)

John

Thanks John. You can make a great pie with a 24 hour dough for sure. Leoparding will probably be a little less. I've been doing 24 hour with my Detroit style with excellent flavor and texture results. I’ve been using 2% in the low 70F’s.

You will notice you can draw a line on the chart from 10%/65F to 1%/75F, and all the combinations will be 24 hour. I think any of them will work. Which you choose will depend on several factors including:

 - Your starter – if you are very confident in it, I’d go with a lower %.

 - Your ability to maintain the temperature – If you think your fermentation temperature may drift up, go with a lower % and warmer temp that you can hold or vice versa.

In any case, particularly with your first attempt, keep an eye on it for the last 8 hours or so and make corrections as needed (e.g. putting in a warm oven to speed things up or in the fridge to slow things down). Personally, I like to work with warm dough as opposed to cold dough; if I ferment at 65F, I have to be mindful of needing to get it up into the 70’s over the last couple hours.
I love pigs. They convert vegetables into bacon.

Online scott123

  • Registered User
  • Posts: 6658
Re: Sourdough starter quantity predictive model
« Reply #8 on: January 02, 2013, 11:35:00 AM »
Craig, does this model have a formula?

Offline dellavecchia

  • Lifetime Member
  • *
  • Posts: 2628
Re: Sourdough starter quantity predictive model
« Reply #9 on: January 02, 2013, 11:52:50 AM »
What a great resource Craig. My workflow for 68-70 degrees looks just about right.

What are your thoughts on slow vs. fast fermentation rates as it applies to temperature? Is one better than the other in terms of flavor, texture, etc.? You can ferment something at 60 degrees or 90 degrees with the same percentage of starter, you just wait longer for one. Does more time equal more flavor production (by-products), or is it all a wash since the yeast are consuming slower?

John


Offline TXCraig1

  • Registered User
  • Posts: 11729
  • Location: Houston, TX
Re: Sourdough starter quantity predictive model
« Reply #10 on: January 02, 2013, 11:57:21 AM »
Craig, does this model have a formula?

It [Rev1] has two three formulas.

The regression of the 23.9C (75F) data is:
Hours = -5.368 * LN(starter%) - 0.6014

The Gänzle et al. C. milleri growth model I used to expand the regression across the temperature spectrum is:
relative growth = 0.0124 * ((36-temp)^2.981) * exp(-0.3355*(36-temp))

The LAB growth model is:
relative growth = 0.1267 * ((41-temp)^1.5404) * exp(-0.1931*(41-temp))

Temp is in Celsius. LAB activity is weighted at 50%.

The model uses the curve calculated at 23.9C and then adjusts it based on the growth rate ratio at a given temperature vs 23.9C.

CL
« Last Edit: January 03, 2013, 01:51:57 PM by TXCraig1 »
I love pigs. They convert vegetables into bacon.

Online scott123

  • Registered User
  • Posts: 6658
Re: Sourdough starter quantity predictive model
« Reply #11 on: January 02, 2013, 12:28:33 PM »

Offline TXCraig1

  • Registered User
  • Posts: 11729
  • Location: Houston, TX
Re: Sourdough starter quantity predictive model
« Reply #12 on: January 02, 2013, 12:33:49 PM »
What a great resource Craig. My workflow for 68-70 degrees looks just about right.

What are your thoughts on slow vs. fast fermentation rates as it applies to temperature? Is one better than the other in terms of flavor, texture, etc.? You can ferment something at 60 degrees or 90 degrees with the same percentage of starter, you just wait longer for one. Does more time equal more flavor production (by-products), or is it all a wash since the yeast are consuming slower?

John

John, I have no experience doing that with pizza, but I've had great results fermenting at 90-95F with bread. Theoretically, it should give an advantage to the bacteria as their peak activity is in the low 90's where the yeast should really be slowing. My experience is that you get a little more sour flavor, but what I really notice is more of a creamy, buttery flavor. The texture is as good or better compared to a cooler fermentation.

You can't go straight off this chart if you go hot (>85F). It won't take as long as it says. I think it has a lot to do with the yeast really getting kicked up into high gear as they go through their prime temperature zone as the dough warms. It's not as much of a factor when you go with a cooler fermentation as you are always moving away from their optimum temperature as the dough cools.

CL
I love pigs. They convert vegetables into bacon.

enter8

  • Guest
Re: Sourdough starter quantity predictive model
« Reply #13 on: January 02, 2013, 04:34:28 PM »
John, I have no experience doing that with pizza, but I've had great results fermenting at 90-95F with bread. Theoretically, it should give an advantage to the bacteria as their peak activity is in the low 90's where the yeast should really be slowing. My experience is that you get a little more sour flavor, but what I really notice is more of a creamy, buttery flavor. The texture is as good or better compared to a cooler fermentation.

While >90F is towards the extreme of the active temperature range for Candida Milleri, it's optimal for certain LAB which release CO2 through the heterofermentative fermentation pathway.

enter8

  • Guest
Re: Sourdough starter quantity predictive model
« Reply #14 on: January 02, 2013, 04:45:38 PM »
@TXCraig1

Your model's predictions definitely seem to correspond with my previous experiences over a range of roomish temps (68-75F). I haven't done much cold fermentation so I can't really comment on that part of the chart.  Really impressed with this!

Offline Matthew

  • Registered User
  • Posts: 2228
Re: Sourdough starter quantity predictive model
« Reply #15 on: January 03, 2013, 08:53:29 AM »
Happy New Year & Great job as always.  

« Last Edit: January 05, 2013, 08:09:16 AM by Matthew »

Offline DannyG

  • Registered User
  • Posts: 129
Re: Sourdough starter quantity predictive model
« Reply #16 on: January 03, 2013, 10:06:43 AM »
Great chart but am I reading it correctly? To do a refrigerated fermentation at 40 degrees I would have to use 40% starter and it would take 169 hours or 7.05 days?

I've been doing my NY style with 5% starter, refrigerated for 3 days followed by room temperature for about 4 hours. The pies turn out fine but maybe I don't really know what complete fermentation is. According to the chart 5% starter at 40 degrees takes 654 hours or 27 days. Maybe I need to make some adjustments!

Offline dellavecchia

  • Lifetime Member
  • *
  • Posts: 2628
Re: Sourdough starter quantity predictive model
« Reply #17 on: January 03, 2013, 10:30:01 AM »
Great chart but am I reading it correctly? To do a refrigerated fermentation at 40 degrees I would have to use 40% starter and it would take 169 hours or 7.05 days?

I've been doing my NY style with 5% starter, refrigerated for 3 days followed by room temperature for about 4 hours. The pies turn out fine but maybe I don't really know what complete fermentation is. According to the chart 5% starter at 40 degrees takes 654 hours or 27 days. Maybe I need to make some adjustments!

No - The color coded areas in red are predictive only and Craig says they are not supposed to be reliable. Hydration, salt and time spent at room temperature has a direct effect on the maturation of dough. But it is interesting to see how non-optimal the refrigerator can be, and how it is better used as a retard instead of a fermentation technique.

John

Offline TXCraig1

  • Registered User
  • Posts: 11729
  • Location: Houston, TX
Re: Sourdough starter quantity predictive model
« Reply #18 on: January 03, 2013, 10:32:03 AM »
Great chart but am I reading it correctly? To do a refrigerated fermentation at 40 degrees I would have to use 40% starter and it would take 169 hours or 7.05 days?

I've been doing my NY style with 5% starter, refrigerated for 3 days followed by room temperature for about 4 hours. The pies turn out fine but maybe I don't really know what complete fermentation is. According to the chart 5% starter at 40 degrees takes 654 hours or 27 days. Maybe I need to make some adjustments!

I think you're basically reading it right. Every data point on the chart represents the theoretical hours needed for a particular starter%/temperature combination. 40%/40F/169h is one possible combination. I color coded the chart for exactly the reason you noted. I don't think the points in red are viable and the ones in yellow are questionable (based on my experience - not science per se). The examples you gave are all solidly in the red.

Remember, this chart assumes 100% of the fermentation takes place at the given temperature. It does not take into account the time it takes the dough to get to that temperature nor does it contemplate fermentation at multiple temperatures. I'm not an advocate of cold fermentation- particularly with SD. I only included those temperatures in the chart for discussion purposes such as this.

All that being said, I'm still surprised to hear that 3 days in the fridge + 4 hours at RT is enough. Is your starter sourdough or is it made from commercial yeast?
« Last Edit: January 03, 2013, 10:33:45 AM by TXCraig1 »
I love pigs. They convert vegetables into bacon.

Offline dellavecchia

  • Lifetime Member
  • *
  • Posts: 2628
Re: Sourdough starter quantity predictive model
« Reply #19 on: January 03, 2013, 10:39:44 AM »
All that being said, I'm still surprised to hear that 3 days in the fridge + 4 hours at RT is enough. Is your starter sourdough or is it made from commercial yeast?

Yes, I wondering the same. But do you think that bacterial fermentation is a factored into your chart? Or is it a non-issue?

John