Menu Close

Four-spheres-with-radii-a-b-c-and-d-touch-each-other-Find-the-radii-of-their-circumscribed-sphere-R-and-their-inscribed-sphere-r-in-terms-of-a-b-c-and-d-




Question Number 46481 by MrW3 last updated on 27/Oct/18
Four spheres with radii a,b,c and d  touch each other. Find the radii of  their circumscribed sphere (R) and  their inscribed sphere (r) in terms of   a,b,c and d.
Foursphereswithradiia,b,canddtoucheachother.Findtheradiioftheircircumscribedsphere(R)andtheirinscribedsphere(r)intermsofa,b,candd.
Commented by MJS last updated on 27/Oct/18
plenty of cases. 3 spheres similar to 3 circles:  3^(rd)  one could be very small compared to the  2 others, then the 4^(th)  is still possible but no  circumscribed sphere could exist, even not  any sphere or plain touching the 4 spheres  from the outside.  inscribed sphere doesn′t exist in some cases  too. imagine three spheres: their centers  are in a plain; the 4^(th)  one could exactly fit  into the gap between them.
plentyofcases.3spheressimilarto3circles:3rdonecouldbeverysmallcomparedtothe2others,thenthe4thisstillpossiblebutnocircumscribedspherecouldexist,evennotanysphereorplaintouchingthe4spheresfromtheoutside.inscribedspheredoesntexistinsomecasestoo.imaginethreespheres:theircentersareinaplain;the4thonecouldexactlyfitintothegapbetweenthem.
Commented by MrW3 last updated on 27/Oct/18
thank you sir!  generally you are right.  but one can treat at first the case that such  a sphere exists. from the result one can  then see what is the condition that such  a sphere exists. let′s come back to the  question with three circles. from the  results  R=(1/(2(√((1/(ab))+(1/(bc))+(1/(ca))))−((1/a)+(1/b)+(1/c))))  r=(1/(2(√((1/(ab))+(1/(bc))+(1/(ca))))+(1/a)+(1/b)+(1/c)))  we can see that the inscribed circle  always exits since we get a valuefor  r with any values of a,b and c. but  the circumcircle doesn′t always exist.  if 2(√((1/(ab))+(1/(bc))+(1/(ca))))=(1/a)+(1/b)+(1/c) there is  no circumcircle at all, since R=∞. and  if 2(√((1/(ab))+(1/(bc))+(1/(ca))))<(1/a)+(1/b)+(1/c) there is  a “circumcircle”, but it tangents the  three circles from outside, since in this  case the value of R is negative.    so my opinion is that we should start  with the general case and then see  which exceptions there are.
thankyousir!generallyyouareright.butonecantreatatfirstthecasethatsuchasphereexists.fromtheresultonecanthenseewhatistheconditionthatsuchasphereexists.letscomebacktothequestionwiththreecircles.fromtheresultsR=121ab+1bc+1ca(1a+1b+1c)r=121ab+1bc+1ca+1a+1b+1cwecanseethattheinscribedcirclealwaysexitssincewegetavalueforrwithanyvaluesofa,bandc.butthecircumcircledoesntalwaysexist.if21ab+1bc+1ca=1a+1b+1cthereisnocircumcircleatall,sinceR=.andif21ab+1bc+1ca<1a+1b+1cthereisacircumcircle,butittangentsthethreecirclesfromoutside,sinceinthiscasethevalueofRisnegative.somyopinionisthatweshouldstartwiththegeneralcaseandthenseewhichexceptionsthereare.
Commented by MJS last updated on 27/Oct/18
you′re right  we can start with a≥b≥c≥d  put C_a = ((0),(0),(0) )  C_b = ((0),(y_b ),(0) )  C_c = ((x_c ),(y_c ),(0) )  (I′m a coordinate−method−lover)  the results of the 3−circle−problem fit here  C_d = ((x_d ),(y_d ),(z_d ) )  and we must find out the borders for c  and d  depending on a and b so that a circumsphere  exists.
yourerightwecanstartwithabcdputCa=(000)Cb=(0yb0)Cc=(xcyc0)(Imacoordinatemethodlover)theresultsofthe3circleproblemfithereCd=(xdydzd)andwemustfindoutthebordersforcandddependingonaandbsothatacircumsphereexists.

Leave a Reply

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