Milling your own rice with the TwinBird Mill

This article discusses how homebrewers might use the TwinBird Mill to polish / mill their own rice for homebrew sake.

Rice milling is arguably the first step in sake brewing. For some, it may be the rice harvest or even the entire process of growing the rice. Then for others, most home brewers, milling is not much more than the story we hear about how the pros do it.  However, under the current situation this is hardly acceptable. As home brewers, we have, to my knowledge, only two options for milling levels; standard white rice (approximately 91-93%) and a rice milled for Ginjo grade sake (60%). For those who would like the chance of brewing other classes of sake other options are needed.

TwinBird MRD570
TwinBird MRD570

Recently, I got a rice mill from Japan; the TwinBird Mill. It’s not meant for sake but for home use with table rice. I was not sure that it would be able to mill the rice down as much as I wanted. That is that it would be able to mill rice to level far past those used for table rice. I wanted to be able to have rice ever where from table rice levels (about 90%) to daigingo levels (less than or equal to 50%). Maybe even lower levels. Could the TwinBird do the trick? The short answer is yes, but…

The TwinBird MR-D570 will mill up to about 4 go at a time. Hmmm, 4 go, well a go is 180ml volume. 1 go of Hitomebore white rice I measured weighs 161 grams. This weight will vary depending on the grain size and moisture. Anyway, 10 pounds is ~4540 grams or ~28 go, which is about what is needed for a “standard” small batch of sake. Let’s look at two cases for milling rates that we may want to do. This first is to mill to just under 70%. This is the milling rate that used to be required for both Junmai and Honjozo sake. While the legal milling rate restrictions for the Junmai classification have been removed most of the Junmai sake’s produced still follow this guideline. The second case is to mill the rice to just under 50%. This milling level will give us a Daiginjo or Junmai Daiginjo classification.

Case 1: Again using the Hitomebore as a concrete example, we want to know how much rice we need to purchase and mill. If we buy brown Hitomebore we simply divide the amount we need in the recipe by the milling rate we want:

14.3lbs = 10lbs / 70% or in go: 40.2go = 28.2go / 70%

However, it is more likely that we will want to buy Hitomebore that has already been milled to a white rice level. White rice is generally milled to somewhere between 91% and 93%. Making this adjustment gives:

13.3lbs = 14.3lbs * 93% or in go: 37.4go = 40.2go * 93%

Case 2: Still using Hitomebore and the same process as in case 1 we can strait away calculate the amount of white rice we need to purchase as:

18.6lbs white rice = ((10lbs milled rice) / (50% milling rate)) * (93% white rice milling)

This same calculation gives 52.4 go.

Why do I keep calculating the amount in go? Because the TwinBird mill’s batch size is measured in go and I want to be able to know how many batches I will need to mill. Since the largest batch size for the TwinBird is 4 go, case 1 will have 9.4 batches and case 2 will have 13.1 batches. For this discussion we will assume that each batch remains as a single quantity that does not mix with other batches. This clearly leaves open the possibility of more efficient methods but it will simplify our thinking and discussion.

There are a few more items we will need to consider while milling. First the goal is to mill the rice to the desired level while not breaking or cracking the rice. Broken grains expose the core / heart directly and may retain more of the outer layers than we would like. Cracked grains can also expose the heart or make it too easy for steam and koji mold to enter. If the steam gets into the heart it will produce overly mushy rice that is not good for growing high quality koji. Koji, to produce the best quality enzymes need to be able to get into the rice and grow but if it is too easy to get down into the rice or the rice is too mushy, the best quality enzymes will not be produced.

Commercial Brewers will slow the milling process down as much as they can with their equipment to more gently strip away the outer layers. This also limits the amount of heating and cooling that can contribute to rice cracking. As the rice heats up and its inner layers are exposed during the milling process it loses much of its moisture. This makes it even more brittle and vulnerable to cracking and breaking. While milling our own rice we need to be aware of these processes and avoid them and their negative consequences.

What characteristics does the TwinBird mill have in these respects? How much of the outer layer does it take off over how much time? How much does it heat the rice? Can I adjust these? Well, it depends, ~10F, not so much. Let’s look a little closer.

I did a limited experiment where I use 4 go of the Hitomebore rice in the TwinBird mill with the mill set to run for 4 go with brown rice. Setting it for brown rice makes it run its maximum time before automatically stopping. With this setting the mill runs for somewhere between 5 and 6 minutes. Running the mill repeatedly I captured the following data:

Mill Activations Seimai Buai Grams
-1 100 694
0 93 645
1 86 597
2 82 569
3 79 546
4 75 523
5 72 501
6 69 477

The -1 activation represents brown rice that has not been milled. It was calculated based on a 93% seimai buai for white rice. Seimai Buai is the percentage rice left over after milling. The 0th activation is thus the data for white rice, the real starting point. Each of the following activation data represent the rice after that number of times in the mill as it runs for its 5 to 6 minutes. One key point is that the more rice (by weight) the more effective and faster the milling process. In this discussion we will not make use of this fact but in the design of more efficient and less time consuming methods it would be important. This data is given in the chart below:

Effect of Rice Milling
Effect of Rice Milling

As the initial batch is milled its seimai buai drops by 7, 4 and 3, 3, 3, 3 finally reaching 69% after 6 activations. If run continuously, this batch could be mill to 69% in about 36 minutes. This means that we could mill all 9.4 batches needed in 6 hours non-stop. However, things are rarely so simple. Recall that each run is heating the rice by about 10F. The mill basket is also heating up significantly which implies the rice will heat more each run than the last.

Leaving the rice to cool in a bowl will work but is slow (slow may be better with less cracking but I did not look at this). It would be faster to spread it out over a cookie sheet or the like. The following chart is based on a single batch cooling in a bowl over time.

Rice Cooling Time
Rice Cooling Time

Interleaving batches may be a good way to go to ensure the rice does not take on too much heat. This would give each batch about 40 minutes to lose the approximate 10F it gained in the mill. While more complex than exclusively concentrating on a single batch until it is done, this would still be able to get you close the 6 hours time but without risking heat related damage. Now that we have a time period that the milling could, theoretically, be complete in, we need to acknowledge that the milling machine was NEVER meant to operate on a continuous basis like this. Nor have I explored the limits of the machine as I did the above 6 activations over a period of 2 hours.

After the rice has been milled to the desired level it needs to rest to move back to an ambient temperature and to absorb enough moisture so that when washed and soaked it will not crack from taking on moisture too rapidly. A day exposed to the moisture in the air should do the trick.

Finally, in other experiments I was able to get some rice down under 20% and even below 10%. What I cannot tell though is how much of these gains, so highly polished, are truly from the center of the rice. To my eye they look pretty good but I do not have a trained eye. At these levels I notice that the nuka (rice flour or shavings) is getting pretty coarse. I did not see this as much with larger grains.

I conclude from this that it is possible for those who really want to, to use this mill for their sake brewing. Practically speaking, it may take many sessions over several days to mill your own rice for a batch of sake. However, for those with less time this mill is not a practical solution.

P.S. I am considering carrying this mill in the shop for about $250. If I did, would you be interested?

Click to Enter a Comment

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

22 thoughts on “Milling your own rice with the TwinBird Mill”

    1. Hey Tony,

      Are you using this mill? What are your experiences with it? It says that it auto-stops at the selected milling levels, does this mean that you can’t get below the selection option or just that several passes with the mill are required to get high milling rates?


  1. Hi Will,

    Do you know what small Sake craft breweries use to mill their rice? Or are they mostly outsourcing this to larger milling operations?

    I’ve been looking into getting something shipped over to Australia but it’s very hard to figure out what is best. (I’ve been brewing with 90% semeibuai CalRose.)

    Alibaba has a few interesting looking devices:

    1. Hey Mitch,

      I don’t really know the answer. The two US brewers that I do know for sure about have their own Japanese type sake rice mill. One of these has a side business based on the mill,

      I’ve also looked at the mills on They are not the type used for sake rice milling so I don’t know if they will do a good job for highly polished rice.


  2. Hi Will,
    Have you had any further insights into home milling? I wonder what style of machine would be the next step up from the basket style home mills, which seem more designed for bran removal than grain size reduction. I am even interested to experiment with a home built machine but I’m finding it hard to map out a reasonable starting point. Vertical or conical mills with rubber abrasive wheels seem to be used at the large scale, but I just can’t glean much detail on the basics of their construction and operation.

    1. Not really.

      I wonder what kind of mill Valley Select is using. If it is the same style used for the Japanese Sake industry or something from the food industry. Minnesota Rice and Milling uses the Japanese type. Just wondering out loud, but I wonder if there have been any experiments comparing a food type mill and the sake type; rice vs. rice comparison at 90%, 70%, 60% or something like that.


  3. Hi

    Are you running the mill with any kind of step-down transformer? I understand that the voltage in Japan is 100 V and in the USA 110 to 120 V, so I wondered how this unit behaved at the higher voltage. I live in the UK where the power is supplied at 240 V, but I have a step down transformer with a 120 V output, and wonder whether that will work. I’d hate to import one of these units only to blow it up on the first outing!

    thank you


    1. Hey Ian,

      The twinbird, as listed on its back, is 100V, 150W at 50-60hrz. This is close enough to the US standard that there is no problems. In fact, while all the text on the unit is in Japanese, the plug is US standard (Not Japanese) and it was made in China. Just checking (US not Japan as before) I find a twinbird mill, Zojirushi mill and others. You may check the UK site for easier access.

      Let us know if you get one and your experience.


  4. Hi Will,
    I’m looking for a mill, not for sake production, rather for lab uses. We’re interested in the content and location(s) of heavy metals in rice and looking for a way to separate the husk and the various layers. From your note it appears that the TwinBird ( MR-D570 might well fit our needs.

    Do you know of anyone using the product for measuring what we’d like to?

    Any help would be very much appreciated.

    Best wishes,


    1. Hey Ken,

      I don’t know anyone using the mill in the way you’d like to. I do think that it should do what you need. You can do a bit of milling, then extract the milling for testing and cycle in this way to look at various layers of the rice. As you get to, say less than 50% of the original size, the rice seems to break up a bit more rather than just shaving a bit off the outer most layer.



  5. Hi guys,

    I happen to have two only slightly used mills. I bought them both at Yobodashi while in Kyoto and shipped them back. As I said, I only slightly used them maybe 2 hrs of use each. Should you want one or both I paid $243 each and $60 each to ship to the USA. The manuals are useless unless you can read JP. Tiger Mill Model ASE-R100-CU. If you want both, I’ll let them go for $500 plus shipping or separately for $300 plus Shipping.タイガー-TIGER-魔法瓶-RSE-A100-CU-アーバンベージュ/dp/B00BGMIH8I/ref=sr_1_55?s=kitchen&ie=UTF8&qid=1362428727&sr=1-55

    Dan Ford
    Brew Master, Toji, Advanced Sake Professional

    First East Coast Sake Brewery http://WWW.FACEBOOK.COM/BLUECURRENTBREWERY Make sure to like us to stay updated.

  6. It’s been awhile since anyone posted on this subject and I am curious if anyone has further comment in terms of experience with this machine or better alternatives. I am considering purchasing a TwinBird but apart from the two excellent articles on this site there’s very little comment out there on home milling in general and this device in particular. Thanks.

    1. Yes, if anyone as anything to add, it is of interest to the whole community. Small batch milling is one of the missing pieces for this hobby. Mills like the Twinbird are too small unless you are willing to put in a lot of time for milling.

  7. Very interesting article, Will. I appreciate your scientific approach. I’d probably be interested in owning a mill like this someday. I’m still a newbie at sake brewing, but the idea of gaining more control over the brewing process is appealing.

  8. I would likely not buy it. I would rather have a mill that works faster. If I made more smaller batches then I probably would buy it.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.