Viewing entries in

Ask the Alchemist #211


Ask the Alchemist #211

We've noticed that different beans seems to have different amounts of intrinsic oil. The Peruvian Maranon seems to have quite a bit of oil and produces a chocolate that flows very easily but it tricky to temper correctly. Is there a way to know in advance the amount of oil in a bean so we can adjust the amount of cocoa butter we add?


Ask the Alchemist #206


Ask the Alchemist #206

I notice you advocate dropping your beans into a hot roaster.  I assume this is because you want Maillard reactions and Strecker degradation products.  Is there a certain temperature that works best for these products?


Ask the Alchemist #205


Ask the Alchemist #205

Finishing phase. 

This phase extends from a bean temperature of 232 F until you decide your beans are fully roasted, generally 245-270 F, and lasts 3-6 minutes with the temperature continually rising.


Ask the Alchemist #204


Ask the Alchemist #204

Level: Apprentice

Reading Time: 8 minutes

Self:                       Alchemist.  You mentioned not really liking the name for the 2nd roasting phase.  What is it that you don’t like about the name Momentum phase?

Alchemist:           Well, it seems that is the term I use a lot for the drying phase.

Self:                       So you want to call the Drying phase the Momentum phase? 

Alchemist:            I was really thinking Drying/Momentum phase. 

Self:                       Then what would you call the Momentum phase?

Alchemist:           I’m thinking the foundation phase.  It is where you start to lay the ground work for the flavor building in the final stage? 

Self:                       I see your point, but I don’t like it

Alchemist:           I thought it was kind of appropriate.

Self:                       The others phases are all verbs.  They are action based.  I get your point but Foundation just isn’t right.  The foundation of a project comes first. This is the second part.  Drying is first.  You are developing flavors.

Alchemist:           (pouts)

Self:                       You know, you could call it the  Development phase.  And before you jump in and interrupt me, I knowyou already are using that name.  But I’ve got to tell you, it’s a little off also.

Alchemist:           but…..

Self:                       I know.  You are worried that people are going to start mixing them up.  But it really is better.  Look, you dry the beans and set the momentum for the roast.  Then you set the stage for flavors by developing the chemical pre-cursors you need in the final stage.  Then you finish the roast.

Alchemist:           Ok smart guy, what do I call the last phase then?

Self:                       Really?  I need to tell you?  It’s the Finishing phase.

Alchemist:           Yeah, well, ok.  That was obvious.

Alchemist:           Can I still use End of Roast?  EOR.

Self:                       Sure, it is common in coffee roasting and it has no issues that I see.     

Self:                       And before you ask, yes it is going to confuse a couple people here and there that you’ve changed the names, but your readers are smart people.  They are on top of this.  Give them some credit.

Alchemist:           You’re right.  And it is going to make it better in the long run.  I’ve even been stumbling over the terms recently.  That’s not a good sign.

Self:                       Just make sure you tell everyone what you are doing and go back and edit your previous articles.  I know you are not a revisionist history fan but this is for clarity.  And the record that you changed things will be right here.  So suck it up, do the right thing and make the changes.

Alchemist:           Ok.

Alchemist:           Is it ok if I tell them a bit more of my thinking?

Self:                       Sure, just don’t get crazy.  You still need to lay out the new Finishing phase, so don’t overwhelm them.  Just an overview.  Think of it as a little review before the final push.

Alchemist:           Deal.

So there it is.  My apologies but I’m changing the names of the phases.  The concepts are staying the same but the names were not working for me.  Over the last month I’ve discussed roasting profiles with a few people and it became abundantly clear I was stumbling over the names.

I ran these ideas by a couple people and I was met with resounding agreement.

So here are the official new names for the Roasting Profile Stages:

Drying   - Ambient – 212 F – It is where you built and set your momentum for the following stage.

Development – 212 – 232 F.  In this stage you will develop the foundation of the roast forming flavor and aroma pre-curors that you will finish in the next stage

Finishing – 232 F – EOR.  Here the flavors finish developing and you ensure the beans are fully roasted.  It is also the time to remove any remaining excess acids

EOR – The End of Roast temperature that may or may not contribute a final character to the taste profile.

It looks like this.

I think about roasting in the same way I think about cooking.  There are lots of ways to do it and depending what you do and what you want you can produce radically different flavors.   And no one way is really right or wrong but you simply can’t reliably get some flavors with certain methods. 

Let’s talk soups.

You can bring all your ingredients to a boil in a pot of water and serve it immediately.  It will be edible and possibly good.  It works pretty well if the ingredients are fresh and you want a simple uncomplicated soup.  Rarely is it stunning though.

You can also sauté some of the ingredients first, the carrots, onions, celery, and garlic, letting them start to release flavors and nuance.  When the rest of the ingredients are added the previous flavors combine to give a depth of flavor not originally present.

There are some styles of Japanese soup making that have you create a stock chock full of these precursor flavors.  A common one is called dashi.  On its own it is pretty bland and nearly flavorless.  I didn’t understand it for years but loved the results.  Once combined in the soup, deep, rich, satisfying umami flavors can develop almost like magic.

I suggest thinking of both these techniques as Development phases.  They add a dimension to your roast/soup. A roux does this also.

Regardless of what soup you make and how you started it, you then have to decide how long to simmer it or finish it.  Some soups need very little simmering, some a lot more.  And some will just go neutral and unexciting if simmered too long.  It is up to you how the finish the soup and it is usually informed by what you want it to taste like.

Finally there is garnish.  It will affect the final flavor in most cases, otherwise why would you do it.  But it can be done with a light or heavy and and sometime you don’t want any garnish at all.  Again it is your call and how you decide at what temperature to end the roast.

So that is where my head is when I think and talk about Profile drum roasting.  I’m not a fast boiled soup fan nor do I favor soups simmered to bland inoffensiveness. 

I like them dynamic and satisfying.

Next week we finish up with Finishing! 

And if all goes right details about the Roasting Seminar I will be offering on June 24th and availability of the new 30 lb Ozturk cocoa roaster.



Ask the Alchemist #203


Ask the Alchemist #203

I am roasting to reduce acidity and boost chocolate flavor: I use a 3k drum coffee roaster. Would you recommend roasting a bit longer or hotter maybe 25 minutes building up to 290 degrees?  For the Ugandan I've been enjoying a 21 min roast at 275 tops. 


Ask the Alchemist #201


Ask the Alchemist #201

Is testing the moisture level of the beans before roasting needed? I thought it may help cut back on the dry phase stage to insure you don't over roast. It's just I've seen coffee roasters do it.


Ask the Alchemist #199


Ask the Alchemist #199

Could I ask your opinion about how the Gourmia compare to the Behmor roaster? I'm thinking of getting either from you. I'm hoping to roast both cacao and coffee beans.


Ask the Alchemist #194


Ask the Alchemist #194

Coming from a coffee roasting background, it's difficult to get away from the roasting precision requirements and theory minutia.  I keep telling myself, its not as important with cacao beans, but I still get a little lost trying to plan out what I'm going to do.


Ask the Alchemist #191


Ask the Alchemist #191

Is 'Behmor' impossible to adjust temperature in any profiles?
I read your articles of Behmor and all I could see was the 'Power(%)' of heat and time.
Not exact temperature.
So it looks as if I can roast beans with fixed profiles only.
Because I like to roast with flexible (manual?) temperature and time running.


Ask the Alchemist #183

1 Comment

Ask the Alchemist #183

Level: Apprentice

Reading time: 8 minutes

I have been making a batch of chocolate, but I am unable to lower the acidity taste (vinegar like) that i sense on my chocolate. i have been trying to find a solution, and nothing so far. Maybe you can help me lower that acidity taste I have. The mix is 65% cocoa and 35% sugar so far and no powdered milk.

1 Comment

Ask the Alchemist #182


Ask the Alchemist #182

Ask The Alchemist #182, Question: I got a Behmor (exciting new toy!!!) just did a roast of 500g beans on 400/P2 and they have come out black? Is that normal for drum roasting as I have never had this in the oven and worried that I have burnt them??



Ask the Alchemist #177

Level: Alchemist

Reading time: 13 min

Hi, John - I just watched your roasting video. I am one of those people (for the last 1.5 years since I moved to Seattle) who doesn't currently have an oven. I hope to move in the next few months, but until then I've been taking my beans to my son's house to roast them in his oven. After watching this video, I'd like your opinion.

The oven is a Wisco - 1300 Watts.

Somehow I made the assumption that using this for my beans wouldn't be a good idea because the air circulation is so strong inside it; much stronger than a traditional convection oven. When using it for typical baking I need to drop the temp about 25* from what recipes recommend. Now I'm not so sure. This thing is small - about 15in wide, so these are the Pyrex dishes I have that will fit. Of course, I don't want to take a chance on wrecking a batch of beans. Given what I've described, do you see any problem using this for cacao?


Let us get this out of the way early.

I am going to lie to you today.

That is a lie.

What I mean by that is that in a binary world, something is either the truth or a lie, and if I am not telling you the whole truth, I therefore must be lying to you.

This is basically I joke I tell to get the attention of a group of 8th graders when I talk about the science of chocolate making.  Go watch my youtube video ( ) where I tell the whole story.  Suffice it to say, I am NOT going to actually lie to you (on purpose) but I am going to leave a lot out since what I am going to talk about is full of maths and thermodynamics that are going to make many people’s eyes glaze over.  My goal is to get you to understand what is basically going on and why I am making the statements and assertions I make.


And it’s very possible many of you CAN handle the truth.

With that, hold tight, and let’s jump in.


First the answer.  Yes, you can use that oven.  For 1 to 2 lbs of beans.

Now I will tell you why I say that.  But I need you to be somewhat conversant in some basic terms and how I think.

From Google – “The British thermal unit (BTU or Btu) is a traditional unit of work equal to about 1055 joules. It is the amount of work needed to raise the temperature of one pound of water by one degree Fahrenheit.”

I am starting there because I know from empirical experience that my Royal #5 coffee roaster produces about 42000 btu/hour with propane.  How do I know that?  I measured the orifice of the propane outlet and looked up the BTU/hour connected to standard liquid propane.

Multiplying that out, that translates to 44310000 joules/hour.

I can roast 35 lbs in my Royal in about 20 minutes or one third of an hour.  That means I need  14770000 joules of energy for 35 lbs regardless of time it takes me to get it there.  Dividing by 35, that means 422000 joules per pound are required.

Part of this kind of number crunching is verifying I am in the ballpark.  I’m going to do that with checking my pre-heat requirements against how long I know it takes me to pre-heat my roaster.

Iron has a specific heat capacity of 0.45 j/g C.

I know my roaster weights about 300 lbs, but estimate the portion I heat up to 350 F is only about one third of it or 100 lbs.

100 lbs = 45 kg iron =45000 grams

350 F is a change in temperature of about 150 C.

0.45 x 45000 x 150 = 3037500 joules to pre-heat.  If I divide by the heat input (44310000 j/hr) and convert to minutes (multiply by 60)

3037500/44310000 * 60 = 4.11

4.1 minutes to pre-heat.  Yep, that is not a lie.  Some days it’s a touch longer, but it’s in the ballpark, so I am counting that as confirmation that my heat input and roasting requirements are about right.

Next, we can use those numbers to see what a 1300 watt convection oven should be able to do.  But we need to get everything in the same unit.

Again, from Google:

“One watt of power converted to joule per second equals to 1.00 J/s. How many joules per second of power are in 1 watt? The answer is: The change of 1 W ( watt ) unit of power measure equals = to 1.00 J/s ( joule per second ) as the equivalent measure for the same power type.”

That just means that 1300 watts = 1300 j/s

We know we need   422000 j/lb.  A little quick math canceling out joules and the minimum time needed in seconds falls out.

422000/1300 = 324.7 seconds.

Dividing by 60 seconds/min tells up how many minutes we need to roast a pound with perfect efficiency.

5.4 minutes per pound.

On the surface, that sounds great.  Except you really don’t want to roast that fast.  If you try, you are going to over roast the outside and under roast the interior.  So you need a slower roast.  No faster than 12 minutes for a roast is what my experience tells me.  But you don’t want to turn the heat down and not use the oven to its capability so that means you should put in more than 1 lb of beans.  How much?  Just divide 12 by the 5.4 minutes for a reasonable estimate.

12/5.4 = 2.22 lbs = 1 kg

Which amazingly enough is right what I said at the top of the article.

“First the answer.  Yes, you can use that oven.  For 1 to 2 lbs of beans.  “

But you will note I said 1-2 lbs.  Why 1 lb?

Well, that delves into modes of heat transfer.  That means when you add heat to a system, there are three ways for it to get from the source to the item you want to heat up.

Convection, conduction and radiation.

This is a convection oven.  That basically means air heating up by the heat source, circulating to the air to what you are roasting and that heat transfers to the cool items (cocoa beans).   It’s pretty efficient since hot air can surround the beans except where it is touch the pan or other beans.  In a thin layer (what I advocate, and why I advocate it) 70-80% of the surface of the bean can be exposed to the hot air.

Conduction happens with contact.  From the surface the beans are on to the beans.  It’s actually pretty inefficient for anything that isn’t perfect flat.  In the case of a cocoa bean, there is probably only 10-20% of the bean touching the surface of the pan.  That is why it doesn’t do well.

Radiation is basically direct line of sight.  When the sun comes out and shines on your skin and you feel warm, that is radiation.  You are not touching the sun and it isn’t because it caused air to move and warm you.  So if there are elements shining on the beans from above, then they are being heated that way too.  If they are heating from the bottom, then very little radiative heating is happening.  Instead the elements are heating the pan and the pan is conducting the heat (per conduction – above) and heating the beans.  In this case, it’s probably the least efficient way to heat.

Ok, deep calming breath.  I know that was a lot.  We are almost done.  And I can now use those terms to talk in a more efficient manner about why you might only be able to roast 1 pound.

Even though this is a convection oven it doesn’t mean it is efficient convection or convection as good as it can be.  Also, conduction is going to be very low due to the very low percentage of bean surface area actually touching the pan.  And where radiative heat might be effective in some cases, the convection in a way circumvents it by distributing the heat around the chamber.  In this case though, that is not a terrible thing and is actually pretty good or the top of the beans might get scorched.

After all is said and done there is probably only 50-60% efficiency going on here.  The rest of the energy is being lost.  Either through the non-insulated walls and glass door or through direct energy loss when you open up the door to stir the beans every 5 minutes or so.  Remember, you still have to stir. If you don’t, the top of the beans will be over roasted and the underside that is in contact with the pan will be under roasted.

And how do I know so much is being lost?  Again experience and puzzle solving.

Years ago I had a drum roaster that I built.  Over time I modified it for efficiency.   It started off at 2000 watts, with a slow 6 rpm motor and no insulation.

If we do the same calculations, recalling we need 422000 joules per pound  and have 2000 watts we find we, in theory, could roast  a pound of beans in 3.5 min.

422000 / 2000 * 60 = 3.5 min/lb.

The drum contained 5.5 lbs of beans.  So if this roaster was working as well as my Royal, I can apply this math:

3.5 (min/lb) x 5.5 (lb) = 19.25 min

And predict it should have taken me a touch over 19 minutes to roast 5.5 lbs of bean.  But the reality of the situation was that I could only roast about 3.5 lbs of beans in that time.  Doing that efficiency check :

3.5 lb / 5.5 lb = 63.6%

This showed me my system was inefficient.   A little over 35%.

It was not until my 6 rpm motor died and I replaced it with a 45 rpm motor did I discover how important REAL convection is.  As soon as I did it, my roast time dropped to 14 minutes. It was like a 35% boost in power.  Efficiency really.   Consequently I was able to add more beans to my roaster.  When I put in 5.5 lbs, my roast time went back to 19-20 minutes – exactly where a good efficient system should be.

See, it was not enough that the beans were tumbling.  Too many were touching and protecting each other from the heat.  Just like in a table top convection oven with beans on a try.  It was not until I got them lofting did real, full convection kick in.  And it is worth noting that after I insulated my roaster, the roast time dropped to about 17 minutes showing again how much heat gets wasted out the walls and over time.  Just like when you open the door to stir.

So, sure, it is a convection oven, but the beans are only partly benefiting from the moving air.  Sure, if it was not convection, then the roast times would go way out to 30 minutes like a classic oven.    But you really can’t have too much convection in a non-tumbling roasting situation.

That is where I get the 50% efficiency count from.  30% or so from non-ideal convection (no loft) and 20% loss from continuously having to open the oven to stir and there you go.

A bunch of lies.  Well, half truths of omission.  And it was STILL a ton of reading.

Yes, you can use that oven.  For 1 to 2 lbs of beans.

And  now , hopefully, you know why and how to work it out yourself if you need.

Go have some chocolate.  You deserve it.



Ask the Alchemist #163

Level: Alchemist

Read time: 9 minutes

We roast 300g (10 oz or so) in our Genecafe.

We are trying a new batch at 350 F for 15 minutes, and one for 20 minutes. The beans start to pop really quickly after we start (around 3 minutes) though, is this unusual? Also, after roasting, some beans have their skin cracked open, and others don't. Should they all be cracked open, or maybe it does not matter at all?

We get a chocolate cake/brownie smell quickly as well, which we've heard is a good sign during roasting. 

Regarding the 30 minutes at 300F , we decided on that from what we read online and what we saw in various videos... 

You are bringing back memories of me using the Genecafe.  I too seem to recall popping really, really early on.  3 minutes in is the sign that the heat is quite a bit high.  And that does not correlate with 300 F unless the beans are that hot.

Let’s get a couple simple things out of the way.  You don’t have to have pops or cracks in a roast, but can.  You absolutely don’t and won’t ever have them all pop like popcorn.  At most I don’t think you will ever get more than 2% popping. And maybe only 0.1%.  That said, I personally have found I like chocolate made from cocoa that has popped.  But that is comparing against the same beans that did not pop and are vetted beans that I know are good.

So it looks like we are talking roasting again. That is good. In my opinion, it can be the make or break step after fermentation.  And it seems to be the hardest to grasp in it’s totality.

Let’s see if we can make some connections.  Have a look at this graph.  It is a version of the oven profiles I put up a few weeks ago.  We will talk about it in a moment.


Here is the thing with what you are finding for roasting profiles and what you are using.  300 for 30 minutes sounds like oven roasting.  Not drum roasting with forced air, which is basically what the Genecafe is. You have control over heat and profiles and should be trying to use them if you can.  That means not using one temperature for a set amount of time.  I like to call that baking, not roasting.

At first I thought you were under roasting with a 30 minute 300 F roast.  I am going to back up and say that it may indeed be that you are/were over roasting.

Why?  The key here is understanding how the heat is applied and absorbed by the beans.  In an oven (even convection) the heat transfer is very slow.  Only the top of bean on the top layer of beans is really getting heat.  It means most are not in contact with heat so they can’t heat up quickly.  It’s why it can take 40 minutes at 400 F to get the beans to 250 F.  In a drum roaster, and even more so in one with blowing hot air, the heat transfer is MUCH faster.  And that is what you are seeing in your roaster.

I recall having trouble stretching roasts out in the Genecafe (one reason I never pursued it or recommend it).  Actually, if your bean temperature was indeed 300 F then you over roasted.  250-260 if as hot as you want for  a bean temperature.

Keeping in mind I've not used this roaster in nearly 13 years, here is where I would start.

I would first start at looking at that graph.  I’m going to decide I want about a 15-18 minute roast and find the curve/line that corresponds to it.  The 400 F OVEN temperature looks good.  We picked a line.  Now forget the 400 part.  It is for an oven.  We don’t care.  It is immaterial.  What is important is the bean temperature and the time it takes.

Shoot for this profile. I'm mostly just taking numbers off the chart.  Meaning adjust your temperature knob as you go, keeping an eye on your bean vs set point.

65 C/150 F for 5 minutes 80 C/175 F  4 minutes 96 C/205 F 3 min 110C/230 F 2 min 120C/250 F 2 min

In general, those are the set point and I'm trying to estimate where the beans will be at those time points.  So, in this case, if everything goes right, those are also the bean temperatures.  Or to give it its technical name, that is my “roasting profile”.  To say it again, a roasting profile is the plot of bean temperature vs time.  NOT  ambient temperature vs time.  Yes, sometimes you will be given an ‘oven profile’ because it is easier, but as you can see it isn’t that useful whereas a bean profile is.  It is transferable to ANY method of roasting if you can get bean surface temperatures.

And that is where you went wrong.  I mistook an oven temperature profile for a bean profile.  The later is transferable to any roaster, the prior almost never so.

So what is going on up there in my suggested profile is that I hope to hit those bean temperature markers in that allotted time.  I’m not just putting the ambient temperature at 300 F because this roaster transfers heat to quickly and you can indeed burn or scorch the beans.  I’m effectively assuming heat transfer is going to be very fast and so I keep the difference between my target (say 150 F in the first step) and the initial temperature (ambient in this case) pretty small (only about 80 F).  I stretch it out to 5 minutes so there is plenty of time for the heat to soak into the beans before I start my next step or ramp.

A couple final things.  This roaster has a bean temperature thermocouple also.  You get actual feedback.  If at 5 minutes the temperature is at 120 F, then I want to increase my temperature set point to something a little higher.  Press on the gas pedal more as it were to go faster.  If they hit 150 in 2 minutes, the I need to back off the temperature a bit so they are not just hanging out at 150 so long.

I realize this may be way more that many people want or can use, but I also rather hope that if you read it enough, look at the situation different ways, that eventually there is be a click or ah-ha moment where you get it.  With that, see if these make sense, and if not, give them some thought.

  • The more your beans move, the quicker the beans will heat up.
  • The faster beans heat up, the smaller the difference ambient temperature has to be to the bean surface temperature. This can be as little as 50 F in a good drum roaster or as much as 300 F in the oven
  • The faster your beans move, the cooler you can keep your ambient temperature.
  • You can only heat up beans so fast before the outside scorches.
  • If your cocoa beans are not moving much you have to have a higher ambient temperature to keep the heat transfer moving at a reasonable pace.  Make it as big as you can without scorching the beans.
  • Cocoa beans can scorch if the difference in temperatures is more than about 300 F.  A 200 F difference is safer.

Give that a try and report back.

Oh, and one last thing.  I don’t sell or really suggest the Genecafe roaster. It’s not because it can’t do a good job.  It’s because I find it expensive to the amount of beans you can roast.  It’s kind of off putting to my mind and my goal is all about approachability and affordability wherever possible.  If you have one, go ahead and use it, but otherwise don’t go out and get one just for cocoa.




Ask the Alchemist #161

Level: Apprentice

Read time: 7 minutes

I have been making a batch of chocolate over and over.  It is sharp and astringent.  I think I keep over roasting it but nothing works.  I have tried roasting from 250 to 300 degrees for 15-30 minutes.  What am I doing wrong?


Somewhere along the line the trend has been to roast cocoa lighter and lighter for fears of not over roasting it.  That is an admirable goal.  You don’t want to over roast your cocoa.  Unfortunately this is the classic result that I am seeing every single week.  Effectively under roasted beans or possibly even raw beans that are sharp, astringent and lacking in chocolate flavor.

Before we going any further, I want to define who I am talking to.  If you like your chocolate and how you are roasting, then this article isn’t aimed at you.

If there is that little voice in the back of your head wondering why you chocolate isn’t quite right and you are fearful of over roasting because all the experts (self proclaimed?)  out there who have never roasted a cocoa bean in their life warn against over roasting, then maybe you should read on.

I’ve tried to over roast.  It is REALLY hard.  I’m talking  you have to try to over roast and you still might fail.  I’m not joking here.  I am kind of dumbfounded where the idea came from that it is easy to over roast cocoa let alone that it burns easily.

Let’s talk about roasting.  As in really what is happening.  To do that I need you to put away what you think you know about roasting cocoa and instead engage what you know about cooking and baking in general.  Because the rules are the same.  There is nothing special or magic about roasting cocoa beans.

Ok.  Agreed?  Great, let’s begin.

What has to happen when you roast?  You need to take a cocoa bean that is around 70 F and take the whole thing, all the way to the center, to somewhere in the range of 230 to 260 F.  From experience, I have learned that I personally like it best when that can happen in 15-30 minutes.  And anecdotal evidence suggests so do most people.  So that is what we are going to talk about.

I said this is like any other cooking or baking.  So instead of a cocoa bean, let’s talk about roasting a hunk of meat.  Or loaf of bread.  Either works fine.  I’m picking those because they are usually 2-3 lbs, start at room temperature and come to some higher final temperature.

What is the classic way to roast these?  350 F oven for an hour is not uncommon for either one.

You put your loaf/roast into a 350 F oven.  Here is the first important thing.  The surface does not immediately become hot.  After 10 minutes, the surface is probably only 100 F and the interior is still room temperature.  And you are not surprised, right?  As time progresses, the heat sinks in.  Two inches in becomes 90 degrees, 1 inch in is 100 degrees and the surface is 110.

Notice the gradient?  That is how the whole roast is going to proceed.  At 30 minutes in the surface might finally be starting to get hot to the touch.  150-160 F.  But If you put in a thermometer to the center you will find the very center is still under 100 F.  Raw.

Not until nearly an hour later is the center getting to 150-160 F for a small roast and maybe 200 F for the bread. The meat has more water so it heats more slowly if you were wondering.

Roasting cocoa is the same.

But cocoa beans are not a loaf of bread I hear you saying.  Well, they kind of are.  They are a solid like mass in the pile they are in.  You can speed the roast along by stirring.  Distributing the heat.   This is exactly why I suggest stirring every 5 minutes.

And let’s look at a few other common things you bake.

Muffins or cup cakes.  350-375 F for 15-20 minutes.

Biscuits 400-425 for  12-15 minutes

What would have happened to either of those things if you had put them in at 250-300 F?  Totally under baked,  right?  Yep.

It’s all the same thing.  Cocoa behaves exactly the same.  Why wouldn’t it?  They are not magical.  It takes quite a bit of heat and time to heat them all the way to the center.

You are worried about burning them at 350-400 F I hear you say.  You notice that gradient I talked about?  That is why they don’t burn.  The oven can be 450 F even at the start.  The only way for the bean (or you bread) to burn is if the whole entire thing is that hot.  But it isn’t.  The heat is continuously sinking in, in effect keeping the surface from burning.

I know many of you are shaking your head in disbelief.  This isn’t just my theory.  I spend a few weeks recently testing just this.  Look a little of the data.

Temperature plot

Those are actual surface temperatures.  From everything we have seen the interior has to be cooler.  They don’t have a choice.  Absolutely none of those beans were even close to burned, let alone even over roasted.

What that means, even at 300 degrees for 30 minutes, your beans probably never made over 150 F on the surface.  Which means 100-115 F inside.  Raw by any definition for the majority of the bean.

That is why they were so sharp and astringent.  They were massively under roasted.  They were still raw.

It’s really that simple.  It takes a lot of heat to roast cocoa beans in any reasonable amount of time.  And to keep it to a reasonable time, you have to have a hot environment.  Heat flows and how fast it flows is proportional to how big the difference is between your beans and the oven.  At 300 F, although it seems hot, it  isn’t that different from your target of 250 F.  The consequence is it takes a long time to get there.

Just think about baking bread or a roast of beast and go from there.   Be fearless.

So get the beans in the hot oven.  Stir them often to help the heat distribute.  Take temperature readings (I use this IR thermometer) and stop worrying you are going to burn the beans.

Be fearless.  It’s only chocolate.



Ask the Alchemist #149

I heard you are getting some porcelano in. How do I roast it? I know it should be very light because it is puree Criollo, but I’m lost after that. I don’t want to ruin it.

First the bomb. Yes, I have some Porcelano in that will be available next week. Next, if you roast it very light your chances of ruining it go WAY up.

Let’s talk about assumptions, pre-conceived notions, patterns, extrapolations and the fallacies of trying to use patterns we think we have found to make predictions. It’s human nature. I get that. But it’s a nasty trap to fall into. It is a variation of ‘correlation is not causation’. What that means is just because you see a pattern, “My knee hurts. The last time my knee hurt, it rained,. It must be about to rain” doesn’t mean that the two facts are related.

The variation I tend to see is slightly different than the knee pain one. It is usually associated with around general cocoa types, their availability, worth and how to handle them. It usually goes like this.

There are three types of cocoa (not really true, and part of the problem, but moving forward). Forastero, Trinatario and Criollo.

Forsatero is the most common accounting for around 85-90% of cocoa grown.

Trinatario is next about 10-15%

Criollo is the most rare in the lower single digits.

Next, there is a general trend we see in price. Forastero is the least expensive, Trinatario and Criollo are more expensive.

And similarly, painting with a broad brush, there is a general trend of cocoa quality is Forastero at the bottom, Trinatario next and Criollo at the top.

Finally, when you roast cocoa, there is a very generalized trend that emerges.

Criollo 235 - 270 F Trinatario 250 - 285F Forastero 260 - 310 F

And what we now have are three groups of data that appear to reinforce one another and that is a very powerful thing in the human mind. It makes us want to draw conclusions and predictions where they don’t exist. The ‘conclusion’?

“Forastero is the cheapest bean, of the lowest quality and roasts the hottest”. Therefore, (trumpets sound), “Criollo is the most expensive, best quality and must be roasted very cool”.

I cannot tell you how often I hear this. It is so ingrained. And so very wrong.

Forastero is the cheapest NOT because of its poor quality but because it is produced in the most quantity. Basic economics. Supply is high, so price is low (again a not quite true premise, but helpful in this case). Notice I never said Criollo is more expensive than Trinatario? I’m will to wager though that is what you thought I said. Your brain forced the pattern. I only said it was the rarest. The same goes for a lot of Forastero being ‘bad’. If only 25% is bad, then just by the nature of there being so much of it, the ‘bad Forastero’ out numbers all the other non-Forastero.

And there it is. The issue. Rarity and how it does not relate to quality (or roasting). Porcelano is the rarest of the rare. Therefore the logic goes, since it is the rarest it MUST be the best and MUST be roasted to coolest…..and it is totally wrong. It doesn’t work that way.

There is a range of roasting for Criollo of 235-270 F. It is because each bean is different. It is NOT because there is a pattern in the pattern. It does not mean that the more pure the Criollo is the cooler it much be roasted. It’s just an empirical observation from roasting Criollo over the years. If you look again at the temperature ranges I gave up there you will see they all overlap in the 260-270 range. Roast any bean to 260 F and you have a better than not chance that the roast will be fine. Any more fine tuning than that and you are asking for trouble.

Back to the original question and the assumption:

I know it should be very light because it is pure Criollo.

By now you should see how wrong that assumption/conclusion is. The only thing you can say is that it is rare. And that is just because it cross breeds easily, is a low producer and isn’t as hardy as many cacao trees. That’s all. Just because it is on one end of one parameter (availability, i.e. it’s rare), does not mean all its parameters are shoved over to one end of the graph.

I’m going to repeat this.

Porcelano is rare. End. Stop. That doesn’t mean it is the best, must be roasted the lightest, has the most antioxidants, will give you the best endorphin rush or anything else. It just means it is the rarest.

Ok, so the rant is over.

So how do you roast the rarest of the rare of cocoa beans? You certainly don’t want to roast it super light because it is super rare? Right? Right!

Might I suggest you treat it like any other bean. Hrm, I bet shooting for 260 F would be a GREAT place to start. Maybe take it a little more gentle because it is Criollo, but note, I said a LITTLE. Not “a lot” because it is Porcelano.

It is already starting out a little light on the chocolate flavor as it isn’t Forastero, so you want to encourage those flavor developments by giving it a good solid roast. Develop those flavors. But nice and easy. At the end of the day, enjoy the chocolate for what it is. Don’t fall into the trap of thinking that because it is rare, and because it is sought after that it is because it is going to be the best chocolate ever. Those are all unrelated items. Maybe it will be, maybe it won’t be. You won’t know until you try it and decide for yourself. Because, after all, that is all that counts. Do YOU like it.

And that brings up the final point. It is very possible you WILL want to roast it light. That you like the bright snappy flavor, the low chocolate level and that too is perfectly ok. Just don’t think you have to.



Ask the Alchemist #145

I tried different roast profiles ranging from 250-325F 4-28 minutes. I am trying not waste many beans by putting only a cup full of beans in at a time. I realize that the roast wouldn't be representative of a full oven, but my goal was to try to find a good dwell temperature and time. Once I found my dwell time and temp I would increase total roast time to account for the increased ramp time of a loaded oven. Do you think this approach is acceptable or should I always experiment with what will be my typical roasting weight?

Roasting. Some of you will love this. Some of you will hate it. Some will eat it up and suddenly everything will make sense. Some will fall into despair that you will never get it and you are doomed to wander the wasteland of uncontrolled roasting. I really hope the former outweigh the later.

I understand your reasoning about roasting just a cup. Unfortunately, the reality is that to effectively use that technique you will probably end up using more beans in the long run as you scale up. The reason is this. You learn best from an experiment by changing one variable at a time. Otherwise the permutations get too great and you can't tell really what effect your changes had. You are changing two (or possible three things) as you progress up in weight. Weight, time and temperature/profile. It's possible to glean information from this but it actually takes more runs and more cocoa unless you get lucky. I know you said your plan was to increase roast time. Experience tells me that won’t work. Fundamentally because by changing the time, you are changing the roast profile which means you are in the apples and oranges situation. They are not comparable. More on profiles in a little bit.

The main problem is that by roasting only a cup of beans, the data you get isn't transferable to a larger load expect to say you will know you will need some combination of more time and heat. But you already know that, right? In essence you used a cup of beans and have no more useful information that you did before. If your 1 cup of beans took 8 minutes at 325 F you have nowhere to go with that information if you try and roast 2 cups. It's not linear. It won't take 16 minutes @ 325 F, nor will it take 8 minutes at 650 F. It will be somewhere in the middle....which you already knew most likely.

On the other hand, if you roast 1 lb for 16 minutes @ 300 F, and determine they are under roasted, you have somewhere to go. You should either increase the roast temperature (my first choice) OR time. If 1 lb is still under roasted at 16 @ 350 F, then for your next roast you increase the temperature to 350 F for 16 minutes. If that is still under roasted, you can start to increase time as you don’t want the oven much hotter, generally speaking. In short, yes, it will take a few runs, but you will get useful data from the methodical approach.

So to directly answer your question, I think you should test roast the weight what you will be roasting in your standard batches. Here is what I hope is a good analogy.

You will never master baking a 2 lb loaf of artisan bread by baking 2 oz bread rolls. Rolls need a cooler temperature or you will burn them. And they need a shorter time. 350-400 for 10-12 minutes will do the trick. A 2 lb loaf could well take 60 minutes at 350 F. To compensate for that large mass, you can put it in at 500 F for 15 minutes. It won't burn since there is so much there that has to heat up and soak in. But you can't keep it there or is will burn. After 15 minutes, you turn it down to 425-450 for another 15-20 minutes and it continues to cook without burning the outside. Sounding a little familiar?

In short, you have to learn your over for a given batch size of beans. You have to actually load the oven enough so it's doing some work. And you have to find the sweet spot, batch size wise, for your oven. All ovens have a setting for 350 F. But they all don't apply the heat at the same rate. Some might take 5 minutes to get an empty oven up to temperature, and another might take 15 minutes because it is lower wattage or has less BTU/hr (electric vs gas). That right there is the crux why I don't and simply can't give people exact temperature 'profiles' to roast in an oven. We have zero knowledge that your oven matches my oven and without that, a temperature is useless.

I'm going to keep repeating this different ways. Let's say you need to go 1 mile in your car. 1 mile is well roasted beans. How am I to tell you (without an odometer) how long to push on the gas pedal to the floor when you only have a speedometer? I should make the note that we can't talk about partly pushing it down. We have to push it to the floor as that is how ovens work. They apply full power until they hit temperature, then turn off until the oven cools some amount below temperature. In short, I think you can see I can't give you that information.

My car might make it to 60 mph in 5 seconds. If that is the case I need somewhere around 55 seconds to go 1 mile. But if your takes 25 seconds, you clearly are going to take well over a minute total to get there. And without that exact data, and pulling out some algebra equations, I can't tell you how long it will take.

But let us assume we have that data for both cars. What can we predict if we hook up a 1 ton full trailer to the cars (i.e. roasting more beans). I think you can see that the answer is nothing at all. Maybe my car is light and can accelerate well under a light load but has no extra capacity so takes 60 seconds to get to speed under load. Yours on the other hand is huge and that is why it took so long to get to speed the first time. Yours, because it has plenty of extra capacity hardly notices the extra load and takes only an additional 5 seconds and you are up to speed in 30 seconds. Clearly yours will make the mile in just about the same time, whereas mine will take 50% longer.

At the end, what I am trying to get across is that AMBIENT oven temperature (i.e. the dial setting) is a terrible gauge to describe roasting profiles....but it is all we have!! By preference I would like to say this to describe a roasting profile.

For 6 lbs of beans, apply 2000 watts of power for 12 minutes. Then apply 1200 watts of power for 3 minutes. Finally, apply 1000 watts of power until the beans either smell done or the beans are at 265 F. Why? Because it is scalable! If you have 2 lbs of beans, then you need to apply 1000 watts for 12 minutes, 600 watts for 3 minutes and 500 watts until you hit 265 F. Half the beans, half the power. Because THAT is a roasting profile. Energy input in a given time. Not temperature settings.

And just so I don't skip the detail, that is for a well insulated system, neglecting heat loss. The reality (based off a given roaster's insulation - see, more variables again) is that I may need only 900 watts for the 2 lbs of beans in the first leg of the profile, or I might need 1100 watts. And how do you know which it is? NOW temperature comes into play. Because we are getting to how I roast day to day. For the above profile, for me, this is what is actually going on.

Apply enough energy so my surface bean temperature goes from room temperature to 210 F in 12 minutes. For MY ROASTER this is 2000 watts.

Turn down the energy input so the bean temperature goes to 245 F in 3 minutes. For my roaster this is 1200 watts.

Turn down the energy again so I reach final target temperature of 265 F (which I have predetermined by SMELL and TASTE of previous roasts) in 3 minutes.

That now is an honest to goodness roasting profile that anyone can take and apply to their own roasting situation and with any amount of beans. Assuming they have a way to control the power (energy input) of the roaster AND know the bean temperatures at different stages along the way. Without all three of those items, my profile is useless to you. Which is why I don’t offer them up as a matter of rote. It is also why I give the very generic ‘profiles’ I do give.

From my Roasting page:

In general, if you try oven roasting, you will start hot (350-400) for a short amount of time and slowly lower it to you target temperature (300-320 F). The more you are roasting, the higher your initial temperature can and has to be.

Remember, you want to roast the cocoa beans, not bake them. This is how this looks: Whole cocoa beans 375-400 5 minutes 350 5 minutes 325 5 minutes 300 until done. Look for the aroma of baking brownies and/or pops. Both are good indicators you are there.

I’ve tried the best I could to give something usable based on what I know and what the limitations are to oven roasting.

So, there. What does that get you? Hopefully a peak into my thought processes about roasting, profiles, what they are and what they are not. Hopefully there is a nugget or two in there on what you need to develop your own profiles, or at least an understanding the limitations of your tools (your oven). And why I’m not being obstinate about sharing my profiles. To rift off of A Few Good Men, “You want the truth? You can’t handle the truth!”

I’m not withholding the truth because you can’t handle it. Or maybe I am. I’m withholding it because in most of the cases out there, it’s useless information because you don’t have the background and equipment to use it.

This is my first concerted effort to get you the background so you can handle the truth.



Ask the Alchemist #134

Would spraying water via a water bottle work or would a small pipe running the length of the roasting drum with small holes in it to spray water be better? Of course, there's the question of how much water and when to add water during the roast. I know this can be used for flavor development and better micro kills at lower temperature; as I have read.

This seems to be a very common misconception. That being that adding water during roasting will give a better micro kill at lower temperature. This seems to be a classic variation of the telephone game, as so many things are. Someone hears something and repeats it slightly changed until it really does not resemble the original facts.

In this case, from the best I can piece together, it comes from a combination of two things. One is that you can have ‘moist’ heat and ‘dry’ heat kill curves and that the ‘moist’ are lower than the dry. The second is the fact that some coffee roasters are equipped with water spray systems. The issue is that to my knowledge these are completely unrelated.

In the first case, it has been taken way out of context. Here is context in full.

Of all the methods available for sterilization, moist heat (steam under pressure ) in the form of saturated steam under pressure is the most widely used and the most dependable. Steam sterilization is nontoxic, inexpensive, rapidly microbicidal, sporicidal, and rapidly heats and penetrates materials”

When you read that I think you will see very quickly that ‘moist’ converted to the lay term of moist or wet when in reality it means steam under pressure when read in context. In short, steam carries a lot of energy. But you have to keep it around, and really, the only practical way to do that is to have a pressurized vessel. None of our roasters are sealed. If you spray steam in, it quickly goes away and has no effect on the microbes. And notice I said steam. If you spray water, you are potentially doing less than nothing as the water, in order to turn into steam, must absorb quite a bit of energy. That energy comes from your beans and your roaster. The result is that if you spray water in, you cool off your roast. In no way does this help microbe kills.

But it does explain why water sprays are in roasters. It is a very effective way to get a very hot roast (one potentially out of control) in cooled down and in control again. Adding steam adds energy to a roast and helps kill things (if you can keep it contained for minutes). Water takes away energy.

As for flavor development, that is news to me. I would postulate that water’s only purpose is to keep the roast profile where you want it. It isn’t like it is going to be absorbed by the beans. At anything over 212 F it is going to flash evaporate thereby cooling the roast, or at least decreasing the rate at which the profile is increasing.

Funny how the telephone game works, and what we think we know, huh?

Send in your Ask the Alchemist questions to