Fun Stuff > CHATTER

Learning has occurred

<< < (44/112) > >>

osaka:
As somebody who spends half his life listening to music and the other half watching romcoms with his family, it strikes me that I hadn't noticed until today that the beat to Outkast's "Ms. Jackson" is a remix of the Bridal Chorus.

GarandMarine:
I was originally going to toss this in gunsmithing, but it's kinda mindblowing from a mechanical perspective. So I'm putting it here.


This is the safety switch of a K98 Mauser, which we are about to discuss.
From left to right the positions are "Safety off" "Safety on, bolt unlocked" and "Safety on, bolt locked"

There are 22 mechanical steps in the act of turning off your K98 Mauser's safety. You just flip a switch, a single movement on your part, but here's what's happening mechanically.

K98 Mauser Cycle of Operations
Weapon Condition 3 (Empty chamber, ammunition in magazine)
Step 1: Safety Off
1. Safety switch moves CCW to "Off" position
2. Cocking piece moves rearward
3. Firing pin spring compresses
4. Firing pin move rearward
5. Cocking piece stops on safety switch
6. Firing pin spring stops on cocking piece
7. Firing pin stops on cocking piece
8. Cocking piece moves forward
9. Firing pin spring decompresses
10. Firing pin moves forward
11. Cocking piece stops on safety switch
12. Firing pin stops on cocking piece
13. Firing pin spring stops on cocking piece
14. Safety switch stops on  bolt sleeve
15. Cocking piece moves rearward
16. Cocking piece stops on safety switch
17. Cocking piece moves forward
18. Firing pin moves forward
19. Firing pin spring compresses
20. Cocking piece stops on sear
21. Firing pin stops on cocking piece
22. Firing pin spring stops on cocking piece

Safety is off.

Grognard:
compared to setting the safety for a British STEN gun.

"take yer booger hook offn the bang switch!"

http://en.wikipedia.org/wiki/Sten

GarandMarine:
If I had one handy I could go through the mechanical steps for THAT simple action too.

Probably something along the lines of (assuming the individual was firing)
Trigger rotates CCW
Sear Rotates CCW
Sear spring decompresses
Bolt Carrier Assembly moves forward
Bolt Carrier Assembly stops on sear
Sear stops on receiver tube
Sear spring stops on sear
Trigger stops on sear

Then of course the mechanical actions in the human hand for the act of moving your finger...

osaka:
Have you ever wondered how computers truly handle decimals? If you have, you probably stumbled upon some info on floating point decimals and IEEE754, but I had to wrangle with it for some guys I TA. Turns out that floating point sum is a really slow and painful process, that can take up a massive amount of software cycles. That little task is something that makes me respect hardware engineers even more.

(click to show/hide)Where s is signum, S is significand, X exponent and N a number to represent:
N = 0bsXXXXXXXXSSSSSSSSSSSSSSSSSSSSSSS
The "significand" represents the decimal part of the actual significand. Since it's binary, all "scientific notation" or "normalized floating point notation" will have 1 as the integer part of the significand, with only the fractional part explicitly represented. If the biased exponent is 0x00, then the integer part is considered 0 as well, and exponent takes the value 0. If the biased exponent is 0xFF, then it's considered a special value, and will represent either +inf (if fraction is zero and signum is zero), Not a Number (with only signum zero), -inf (with only fraction zero) or an error code (if neither signum nor fraction are zero)

On the actual process of the sum, there are 4 main steps:

Step 1: Equalizing orders of magnitude. The Float single precision format represents a number that's like (1.)<23 bit binary> * 2^<8 bit binary>-127. So the sum will always have the exponent part of the greater number. The fraction of the smaller number is then moved right n bits, where n is the difference between the greater and lesser exponents.
Note that if the difference between exponents is greater or equal to 24, then the value of the sum is defaulted to max(a, b).

Step 2: Sum the fractions. Keep in mind that if one of the values is negative, you need to sum the 2's complement of that fraction with the other one.

Step 3: Normalize the fraction. To do this, you sequentially move the fraction left or right until the 24th bit (or position 23) is the first 1 in the fraction, and for each bit moved right you add 1 to the exponent, and for each bit moved left you substract 1 from the exponent.

(click to show/hide)Compare to 0x800000. Strictly greater when moving right, lesser or equal when moving left.
Step 4: Put everything back together and check for inconsistencies. Keep in mind that, even if it hasn't been specified explicitly here, you should handled un-biased exponents. Exponent bias is inserted to keep exponents as positive and consistently increasing, so the exponent should be considered an unsigned byte and its real value is the one in the unsigned byte minus 127 (in the case of single precision)

Navigation

[0] Message Index

[#] Next page

[*] Previous page