a child's drawing of me as line-art. i have a boring man head and stubble, with assymetrical features, and am wearing a ghostbusters t-shirt.

Turning it up to 10.1


my frederick post 1460 versalog 2 with the cursor set to 10.1 on the C scale (or is it 1.01.. hmm..)

Welcome to the next post in my series about decimal placement on slide rules!

In this post I'll discuss the role played by numeric formats.

Let's first introduce some terminology. Recall that when x is the base-b logarithm of a number k, it solves the equation k=bx. The slide rulers of old had names for the parts of that logarithm, x, to help them reason about decimal placement. The whole number part before the decimal place is called the characteristic, and the part after the decimal place is called the mantissa. Unfortunately, those were overloaded terms across math, science, and engineering back then, but I'll leave that side quest to you and carry on.

Right away we see that the logarithm x is the sum of the characteristic and the mantissa. Let's call those c and m respectively, noting x=c+m. If we substitute this into our original equation k=bx, we get k=b(c+m), and by exponent rules then k=bcbm.

Now, since the characteristic c is a whole number, the bc term is a nice round power of the base. In base 10 for example, 10c will always be things like thousands, millions, hundredths, billionths, and so on. So on slide rules in base 10, the characteristic c tells you exactly where to put the decimal place. This is key, and my future posts on decimal placement per operation will largely demonstrate how to find the result's characteristic under various operations.

On the other hand, the bm term also has some properties worth noting. Since 0m<1, it follows that 1bm<b. Again, in base 10, that means 10m is at least 1.0 and strictly less than 10. If you look at your C and D scales, you can see they increase from 1 to 2 to 3 and up to 8 to 9 to .. one?? What on earth!

Using the ideas we discussed above, let's reason through what's happening on the slide rule's scales right there at C's right margin where m is almost 1.0 and 10m is almost 10. Let's pretend 10x=9.95, which has a characteristic of 0 and a mantissa around 0.998. On my Versalog's C scale, 9.95 is the very last tick mark on the far right.

If we want to increase 10x from 9.95 to 10.1, we'll have to add a little bit to x. Notice this takes us past the end of the C scale. What should we add to x to increase 10x from 9.95 to 10.1? Let's use our exponent rules: 10.1=10(x+t) for some new t, so 10.1=10x10t. But we know 10x=9.95, so we substitute that and divide both sides by it: 10.19.95=10t. That leaves approximately 1.0151=10t, and taking log10 of both sides gives t=0.0065.

Wonderful, but what the heck does it mean? Well, let's add x and t together and see what we end up with: x+t=0.998+0.0065=1.0045, a logarithm with characteristic 1 and mantissa 0.0045!

Ok, so now that we're familiar with characteristic and mantissa values with base 10, and the exponent rule we used earlier, let's put 10.1 into normalized scientific notation. Scientific notation factors out a leading coefficient and the order of magnitude from a number to approximate numbers of any size. In this case, 10.1 is not very big but it does have a power of ten, so in scientific notation it is written as: 1.01101. In normalized form, the leading coefficient doesn't exceed 10.0.

That's amazing, because it corresponds neatly with the concept of a logarithm's characteristic and mantissa! Observe that log(10.1)=1.0045, with c=1 and m=0.0045. Now notice that 10.1 in normalized scientific notation is 1.01101=10m10c for those c and m values. Marvelous!

To bring this home, we wanted to get at the question of how to deal with numbers that are smaller than 1.0 or bigger than 10. We've learned that in normal form, the leading coefficient is always between 1.0 and 10, because the mantissa of the logarithm of the number is between 0.0 and 1.0. And we've learned that when the mantissa exceeds 1.0, the characteristic increments and the mantissa restarts at 0.0. The same is true on the slide rule: you never really go off scale, you just shift the decimals around appropriately. That is, using the normalized scientific notation not only solves the problem of keeping your numbers on scale, it also gives you the tools you need to track the decimal of your results. This is WHY slide rule tutorials out there tell you that scales repeat infinitely.

At the end of the last post, I encouraged you to make your own rulers and try adding numbers with those rulers. Do you still have those rulers out and ready? This time try thinking of the numbers on the rulers as mantissas that you're adding together. Just be sure to think of 1, 2, 3.. on the ruler as 0.1, 0.2, 0.3, etc.. Try adding some mantissas and use your calculator or WolframAlpha to find 10x!

Until next time!

Next post: slide rule multiplication!