Digital Root circuit using tree of 4-bit CLA's with Cout fed

[lrl]

Hi,

In order to generate hexadecimal Digital Root of a number (i.e,
sum-of-hexdigits until it reduces to a single hex number - a nibble)

For e.g 0x123a = 0x1 + 0x2 +0x3 +0xa = 0x10 = 0x1 + 0x0 = 1 is the
hex digital root of (0x123a)

I am planning to use tree of 4-bit Carry-Look-Ahead adders with the
Cout fed into Cin of the same CLA.

Question: Can I directly fed back the Cout signal to Cin at each CLA
to get the Digital Root . Will it not form a combinational loop? Due
to feedback will it affect the next level CLA's o/p.

Thanks
lrl

 

[Glen Herrmannsfeldt]

"lrl" <lrl@eth.net> wrote in message
news:c8f17d2e.0307160957.2d21c698@posting.google.com...

Hi,

In order to generate hexadecimal Digital Root of a number (i.e,
sum-of-hexdigits until it reduces to a single hex number - a nibble)

For e.g 0x123a = 0x1 + 0x2 +0x3 +0xa = 0x10 = 0x1 + 0x0 = 1 is the
hex digital root of (0x123a)

I am planning to use tree of 4-bit Carry-Look-Ahead adders with the
Cout fed into Cin of the same CLA.

Question: Can I directly fed back the Cout signal to Cin at each CLA
to get the Digital Root . Will it not form a combinational loop? Due
to feedback will it affect the next level CLA's o/p.

I believe that can be done. Consider how a ones complement adder works.

Well, I would think that you would need G and P in, but other than that it
should be fine.

I thought about this some time ago wondering why anyone would build ones
complement machines, such as CDC did. If they couldn't do a fast add that
would be a big disadvantage. I then decided that CLA should work.

Now, it might be that there is a better way to do the problem you are asking
about, but that is a different question.

-- glen

 

[Glen Herrmannsfeldt]

"lrl" <lrl@eth.net> wrote in message
news:c8f17d2e.0307160957.2d21c698@posting.google.com...

Hi,

In order to generate hexadecimal Digital Root of a number (i.e,
sum-of-hexdigits until it reduces to a single hex number - a nibble)

For e.g 0x123a = 0x1 + 0x2 +0x3 +0xa = 0x10 = 0x1 + 0x0 = 1 is the
hex digital root of (0x123a)

I am planning to use tree of 4-bit Carry-Look-Ahead adders with the
Cout fed into Cin of the same CLA.

Question: Can I directly fed back the Cout signal to Cin at each CLA
to get the Digital Root . Will it not form a combinational loop? Due
to feedback will it affect the next level CLA's o/p.

I believe that can be done. Consider how a ones complement adder works.

Well, I would think that you would need G and P in, but other than that it
should be fine.

I thought about this some time ago wondering why anyone would build ones
complement machines, such as CDC did. If they couldn't do a fast add that
would be a big disadvantage. I then decided that CLA should work.

Now, it might be that there is a better way to do the problem you are asking
about, but that is a different question.

-- glen

 

[Glen Herrmannsfeldt]

"lrl" <lrl@eth.net> wrote in message
news:c8f17d2e.0307160957.2d21c698@posting.google.com...

Hi,

In order to generate hexadecimal Digital Root of a number (i.e,
sum-of-hexdigits until it reduces to a single hex number - a nibble)

For e.g 0x123a = 0x1 + 0x2 +0x3 +0xa = 0x10 = 0x1 + 0x0 = 1 is the
hex digital root of (0x123a)

I am planning to use tree of 4-bit Carry-Look-Ahead adders with the
Cout fed into Cin of the same CLA.

Question: Can I directly fed back the Cout signal to Cin at each CLA
to get the Digital Root . Will it not form a combinational loop? Due
to feedback will it affect the next level CLA's o/p.

I believe that can be done. Consider how a ones complement adder works.

Well, I would think that you would need G and P in, but other than that it
should be fine.

I thought about this some time ago wondering why anyone would build ones
complement machines, such as CDC did. If they couldn't do a fast add that
would be a big disadvantage. I then decided that CLA should work.

Now, it might be that there is a better way to do the problem you are asking
about, but that is a different question.

-- glen