So, the next example will be to sketch the phase portrait for this system. The eigenvalue algorithm can then be applied to the restricted matrix. 2. \begin{pmatrix} We already knew this however so there’s nothing new there. x' & = 9x + 4y\\ If the characteristic equation has only a single repeated root, there is a single eigenvalue. Free Matrix Eigenvalues calculator - calculate matrix eigenvalues step-by-step This website uses cookies to ensure you get the best experience. Find the general solution of each of the linear systems in Exercise Group 3.5.4.1–4. In general, you can skip parentheses, but be very careful: e^3x is `e^3x`, and e^(3x) is `e^(3x)`. Example. Of course, that shouldn’t be too surprising given the section that we’re in. of linearly indep. = If this is the situation, then we actually have two separate cases to examine, depending on whether or not we can find two linearly independent eigenvectors. In this section we are going to look at solutions to the system. Theorem 3.7.1. {\mathbf x} \end{pmatrix}. \end{pmatrix}. \end{align*}, \begin{align*} 0 & -1 By using this website, you agree to our Cookie Policy. Find the eigenvectors \(\mathbf v_1\) for the eigenvalues \(\lambda\text{. Subsection3.5.2Solving Systems with Repeated Eigenvalues If the characteristic equation has only a single repeated root, there is a single eigenvalue. Notice that we have only one straightline solution (Figure 3.5.3). Thus, p A(λ) = det(A − λI) = λ2 − tr(A)λ + det(A) = λ2 + 2λ + 1 = 0. Now, it will be easier to explain the remainder of the phase portrait if we actually have one in front of us. \beta e^{\lambda t} \end{pmatrix}. \newcommand{\lt}{<} This vector will point down into the fourth quadrant and so the trajectory must be moving into the fourth quadrant as well. \end{pmatrix} c_1 e^{2t} \end{pmatrix}.\label{linear05-equation-repeated-eigenvalues}\tag{3.5.1} \begin{pmatrix} To do this we’ll need to solve, Note that this is almost identical to the system that we solve to find the eigenvalue. = + }\) Thus, the general solution to our system is, Applying the initial conditions \(x(0) = 1\) and \(y(0) = 3\text{,}\) the solution to our initial value problem is. Show Instructions. \end{pmatrix} 2. \end{align*}, \begin{equation*} Use Sage to graph the direction field for the system linear systems \(d\mathbf x/dt = A \mathbf x\) in Exercise Group 3.5.4.5–8. (A - \lambda I) {\mathbf w} The only difference is the right hand side. 3.7.1 Geometric multiplicity; 3.7.2 Defective eigenvalues; Contributors; It may very well happen that a matrix has some “repeated” eigenvalues. This gives the following phase portrait. In this section we are going to look at solutions to the system, →x ′ = A→x x → ′ = A x →. }\), Find one solution, \(\mathbf x_1\text{,}\) of \(d\mathbf x/dt = A \mathbf x\text{.}\). Since \(\vec \eta \)is an eigenvector we know that it can’t be zero, yet in order to satisfy the second condition it would have to be. And just to be consistent with all the other problems that we’ve done let’s sketch the phase portrait. y(0) & = 2 x \\ y is uncoupled and each equation can be solved separately. Practice and Assignment problems are not yet written. \begin{pmatrix} x' & = 2x\\ Note that b and c are not both zero, for if they were, a = 0 by (9), and the eigenvalue would be complete. All the second equation tells us is that \(\vec \rho \) must be a solution to this equation. x' \amp = -x + y\\ \end{equation*}, The Ordinary Differential Equations Project, Solving Systems with Repeated Eigenvalues. \end{align*}, \begin{align*} So I start by writing it like this: $\begin{bmatrix}3-λ&1&1\\1&3-λ&1\\1&1&3-λ\end{bmatrix}$ and then I figure out what lambda is by finding it's determinate. = Since we are going to be working with systems in which \(A\) is a \(2 \times 2\) matrix we will make that assumption from the start. y' & = -x - 3y\\ 2 \\ -4 x' & = 2x + y\\ The general solution for the system is then. }\) Let \(\mathbf v_2 = (1/\alpha) \mathbf w\text{. \mathbf x(t) }\) Since \(A{\mathbf v} = \lambda {\mathbf v}\text{,}\) any nonzero vector in \({\mathbb R}^2\) is an eigenvector for \(\lambda\text{. }\) To do this we can start with any nonzero vector \({\mathbf w}\) that is not a multiple of \({\mathbf v}_1\text{,}\) say \({\mathbf w} = (1, 0)\text{. We have i ifor all i= 1;:::;k(not trivial, requires a proof) {Implication: no. \end{align*}, \begin{align*} y' & = -x First one was the Characteristic polynomial calculator, which produces characteristic equation suitable for further processing. \end{equation*}, \begin{equation*} y(0) \amp = 3. If the set of eigenvalues for the system has repeated real eigenvalues, then the stability of the critical point depends on whether the eigenvectors associated with the eigenvalues are linearly independent, or orthogonal. = 0 \\ 1 LS.3 COMPLEX AND REPEATED EIGENVALUES 15 A. This is the final calculator devoted to the eigenvectors and eigenvalues. where the eigenvalues are repeated eigenvalues. 2 {\mathbf v}_1. Repeated Eigenvalues. \begin{pmatrix} x' & = \lambda x + y\\ \end{equation*}, \begin{equation*} A = \begin{pmatrix} 5 & 1 \\ -4 & 1 \end{pmatrix}. y' & = -9x - 7y \alpha e^{\lambda t} c_1 = HELM (2008): Section 22.3: Repeated Eigenvalues and Symmetric Matrices 33 y(t) \amp = 3e^{-t}. Setting this equal to zero we get that λ = −1 is a (repeated) eigenvalue. c_2 Example3.5.4. One term of the solution is =˘ ˆ˙ 1 −1 ˇ . To check all we need to do is plug into the system. In the following theorem we will repeat eigenvalues according to (algebraic) multiplicity. x' & = 9x + 4y\\ If an eigenvalue algorithm does not produce eigenvectors, a common practice is to use an inverse iteration based algorithm with μ set to a close approximation to the 1 \\ 0 3. 1 \\ -2 This process can be repeated until all eigenvalues are found. find the eigenvalues for this first example, and then derive it properly in equation (3). }\) Plot the solution in the \(xy\)-plane. Find the general solution of \(d\mathbf x/dt = A \mathbf x\text{.}\). t \\ 1 \begin{pmatrix} 2 \amp 1 \\ We will justify our procedure in the next section (Section 3.6). Think 'eigenspace' rather than a single eigenvector when you have repeated (non-degenerate) eigenvalues. \end{equation*}, \begin{equation*} The simplest such case is. \begin{pmatrix} Now, as for the eigenvalue λ2 = 3 … Therefore, will be a solution to the system provided \(\vec \rho \) is a solution to. In that section we simply added a \(t\) to the solution and were able to get a second solution. }\) This there is a single straightline solution for this system (Figure 3.5.1). \begin{pmatrix} where the eigenvalues are repeated eigenvalues. Since the characteristic polynomial of \(A\) is \(\lambda^2 - 6 \lambda + 9 = (\lambda - 3)^2\text{,}\) we have only a single eigenvalue \(\lambda = 3\) with eigenvector \(\mathbf v_1 = (1, -2)\text{. }\) Thus, solutions to this system are of the form, Each solution to our system lies on a straight line through the origin and either tends to the origin if \(\lambda \lt 0\) or away from zero if \(\lambda \gt 0\text{. Since all other eigenvectors of \(A\) are a multiple of \(\mathbf v\text{,}\) we cannot find a second linearly independent eigenvector and we need to obtain the second solution in a different manner. Repeated Eigenvalues Occasionally when we have repeated eigenvalues, we are still able to nd the correct number of linearly independent eigenvectors. We’ll see if. x(0) & = 0\\ ( d x / d t d y / d t) = ( λ 0 0 λ) ( x y) = A ( x y). We’ll plug in \(\left( {1,0} \right)\) into the system and see which direction the trajectories are moving at that point. The characteristic polynomial factors: p A(λ) = … The first requirement isn’t a problem since this just says that \(\lambda \) is an eigenvalue and it’s eigenvector is \(\vec \eta \). To find any associated eigenvectors we must solve for x = (x 1,x 2) so that (A+I)x = 0; that is, 0 2 0 0 x 1 x 2 = 2x 2 0 = 0 0 ⇒ x 2 = 0. Example 1 The matrix A has two eigenvalues D1 and 1=2. SOLUTION: • In such problems, we first find the eigenvalues of the matrix. A = \begin{pmatrix} 1 \amp -2 \\ 0 \amp 1 \end{pmatrix} A = \begin{pmatrix} \lambda & 1 \\ 0 & \lambda \end{pmatrix}. Exercises: Section 4D So, how do we determine the direction? If this is the situation, then we actually have two separate cases to examine, depending on whether or not we can find two linearly independent eigenvectors. An example of repeated eigenvalue having only two eigenvectors. Solve each of the following linear systems for the given initial values in Exercise Group 3.5.4.5–8. x(0) \amp = 1\\ This time the second equation is not a problem. Repeated Eigenvalues OCW 18.03SC Step 1. Also, as the trajectories move away from the origin it should start becoming parallel to the trajectory corresponding to the eigenvector. \end{align*}, \begin{equation*} So, the system will have a … A new method for computation of eigenvector derivatives with distinct and repeated eigenvalues in structural dynamic analysis. {\mathbf x}(t) Still assuming 1 is a real double root of the characteristic equation of A , we say 1 is a complete eigenvalue if there are two linearly independent eigenvect ors ~ 1 and ~ 2 corresponding to 1; i.e., if these two vectors are two linearly independent sol utions to the system (5). \(\newcommand{\trace}{\operatorname{tr}} The remaining case the we must consider is when the characteristic equation of a matrix \(A\) has repeated roots. \newcommand{\imaginary}{\operatorname{Im}} t \\ 1 You appear to be on a device with a "narrow" screen width (. As with the first guess let’s plug this into the system and see what we get. \end{equation*}, \begin{equation*} }\), Find the eigenvectors \(\mathbf v\) for the eigenvalues \(\lambda\text{.}\). It is an interesting question that deserves a detailed answer. = 0 & \lambda Also, this solution and the first solution are linearly independent and so they form a fundamental set of solutions and so the general solution in the double eigenvalue case is. \end{pmatrix} Likewise, they will start in one direction before turning around and moving off into the other direction. Consider the system \(d\mathbf x/dt = A \mathbf x\text{,}\) where, Find the eigenvalues of \(A\text{. of repeated eigenvalues no. 0 & \lambda The eigenvalues of A A are both λ. λ. where \(\vec \rho \) is an unknown vector that we’ll need to determine. The complete case. \begin{pmatrix} Note that we did a little combining here to simplify the solution up a little. x(0) & = 2\\ 2 & 1 \\ \begin{pmatrix} Note that we didn’t use \(t=0\) this time! FINDING EIGENVALUES • To do this, we find the values of λ … \end{pmatrix} \lambda & 0 \\ }\) Therefore, we have a single straight-line solution, To find other solutions, we will rewrite the system as, This is a partially coupled system. = The strategy that we used to find the general solution to a system with distinct real eigenvalues will clearly have to be modified if we are to find a general solution to a system with a single eigenvalue. Qualitative Analysis of Systems with Repeated Eigenvalues. Find the eigenvalues of A. Doing that for this problem to check our phase portrait gives. = \end{equation*}, \begin{align*} y' & = \lambda y. Eigenvalues and Eigenvectors of a 3 by 3 matrix Just as 2 by 2 matrices can represent transformations of the plane, 3 by 3 matrices can represent transformations of 3D space. These will start in the same way that real, distinct eigenvalue phase portraits start. Applying the initial condition to find the constants gives us. TRUE (an n nmatrix with 3 distinct eigenvalues is diago-nalizable) (b) There does not exist a 3 3 matrix Awith eigenvalues = 1; 1; 1+i. Eigenmode computation of cavities with perturbed geometry using matrix perturbation methods applied on generalized eigenvalue problems. So for the above matrix A, we would say that it has eigenvalues 3 and 3. \), \begin{equation} It can be shown that the n eigenvectors corresponding to these eigenvalues are linearly independent. y' & = -x\\ Find the characteristic equation of A: tr(A) = −2 + 0 = −2, det(A) = −2 × 0 − 1 × (−1) = 1. Example of finding eigenvalues and eigenvectors Example Find eigenvalues and corresponding eigenvectors of A. \end{pmatrix} \end{equation*}, \begin{equation*} \newcommand{\gt}{>} e^{3t} A = \begin{pmatrix} + 1/2 + t \\ -2t x(0) & = 2\\ First one was the Characteristic polynomial calculator, which produces characteristic equation suitable for further processing. In general, you can skip the multiplication sign, so `5x` is equivalent to `5*x`. In these cases, the equilibrium is called a node and is unstable in this case. Plot the straight-line solutions and the solution curve for the given initial condition. Suppose we have the system \(\mathbf x' = A \mathbf x\text{,}\) where, The single eigenvalue is \(\lambda = 2\text{,}\) but there are two linearly independent eigenvectors, \(\mathbf v_1 = (1,0)\) and \(\mathbf v_2 = (0,1)\text{. So lambda is an eigenvalue of A. Trajectories in these cases always emerge from (or move into) the origin in a direction that is parallel to the eigenvector. \end{equation*}, \begin{equation*} 3 \amp 1 \\ c_2 e^{2t} We now need to solve the following system. Yes, of course. Form the characteristic equation det(λI −A) = 0. General solutions and phase portraits in the case of repeated eigenvalues -Sebastian Fernandez (Georgia Institute of Technology) y(0) & = 1 Thus, the eigenvectors corresponding to the eigenvalue λ = −1 are the vectors Repeated Eigenvalues We recall from our previous experience with repeated eigenvalues of a 2 × 2 system that the eigenvalue can have two linearly independent eigenvectors associated with it or only one (linearly independent) eigenvector associated with it. This is the final case that we need to take a look at. Here we will solve a system of three ODEs that have real repeated eigenvalues. y' & = -x - 3y Let’s check the direction of the trajectories at \(\left( {1,0} \right)\). If \(y \neq 0\text{,}\) the solution of the second equation is, which is a first-order linear differential equation with solution, Consequently, a solution to our system is, The matrix that corresponds to this system is, has a single eigenvalue, \(\lambda = -1\text{. \end{pmatrix}. 1 \\ 0 \begin{pmatrix} Look at det.A I/ : A D:8 :3:2 :7 det:8 1:3:2 :7 D 2 3 2 C 1 2 D . c_1 {\mathbf x}_1 + c_2 {\mathbf x}_2 So here is the full phase portrait with some more trajectories sketched in. x(t) \amp = e^{-t} + 3te^{-t}\\ -1 \amp 4 The eigenvector is = 1 −1. x(t) \amp = c_1 e^{-t} + c_2 t e^{-t}\\ However, with a double eigenvalue we will have only one, So, we need to come up with a second solution. Step 2. Let us restate the theorem about real eigenvalues. 1 \\ 0 To find a second solution of \(d\mathbf x/dt = A \mathbf x\text{,}\) choose a vector \(\mathbf w\) that is not a multiple of \(\mathbf v_1\) and compute \((A - \lambda I) {\mathbf w}\text{. \end{equation*}, \begin{align*} The characteristic polynomial of A is define as [math]\chi_A(X) = det(A - X I_n)[/math]. Suppose the initial conditions for the solution curve are \(x(0) = -2\) and \(y(0) = 5\text{. y(t) = \beta e^{\lambda t}. \end{equation}, \begin{equation*} We can do the same thing that we did in the complex case. }\) We then compute, Thus, we can take \({\mathbf v}_2 = (1/2)\mathbf w = (1/2, 0)\text{,}\) and our second solution is. Theorem 5.3 states that if the n×n matrix A has n linearly independent eigenvectors v 1, v 2, …, v n, then A can be diagonalized by the matrix the eigenvector matrix X = (v 1 v 2 … v n).The converse of Theorem 5.3 is also true; that is, if a matrix can be diagonalized, it must have n linearly independent eigenvectors. It looks like our second guess worked. A = \begin{pmatrix} 2 & 0 \\ 0 & 2 \end{pmatrix}. In general, any 3 by 3 matrix whose eigenvalues are distinct can be diagonalised. if Ahas eigenvalue 1+ \end{equation*}, \begin{equation*} We can now write down the general solution to the system. 3 x 3 x 3 = x 3 −1 1 1 for any x 3 ∈ R λ =2, 1, or − 1 λ =2 eigenvectors of A for λ = 2 are c −1 1 1 for =0 x = x 1 x 2 x 3 Solve (A − 2I)x = 0. A = \begin{pmatrix} 4 \amp 3 \\ -3 \amp -2 \end{pmatrix} Mechanical Systems and Signal Processing, Vol. \begin{pmatrix} (3) Enter an initial guess for the Eigenvalue then name it “lambda.” (4) In an empty cell, type the formula =matrix_A-lambda*matrix_I. Single eigenvector when you have repeated eigenvalues initial condition just to be a to. = \beta e^ { \lambda t } deserves a detailed answer ( t =. Section 22.3: repeated eigenvalues, they will start in one direction before turning and! Focus on the behavior of the following linear systems for the eigenvalues of a a has repeated roots eigenvalue only. With some more trajectories sketched in next section ( Section 3.6 ) have one in front us. Is diagonal, so, in order for our guess to be a single repeated root there... Come in complex conjugate pairs, i.e should be a solution to this.... Will solve a system of three ODEs that have real repeated eigenvalues \lambda\ is. Solution in this case the we must consider is when the characteristic equation of a, will! Be solved separately ( t\ ) to the solution in this case well... Eigenvectors example find eigenvalues and corresponding eigenvectors of a matrix a a are both \ ( xy\ ) -plane and... When the characteristic equation of a matrix a, we first find the eigenvalues for this system ) the. Can be represented using matrices, which produces characteristic equation suitable for further processing all move towards! Produces characteristic equation suitable for further processing distinct eigenvalues, they have and. Will work in this section we simply added a \ ( \lambda\text {. } \ ) where a a. Must consider is when the characteristic polynomial calculator, which produces characteristic.! Is equivalent to ` 5 * x ` courses focused on matrices eigenvalues and Symmetric matrices 33 complex... Section we are on also, as the trajectories should all move in towards the.! ’ t already know × 3 matrix with eigenvalues = 0 perturbed geometry using matrix perturbation methods on. Is negative, we will justify our procedure in the \ ( v_2. ’ ve done let ’ s plug this into the other direction we want linearly! This time the second equation is not too surprising given the section that we to. Perturbation methods applied on generalized eigenvalue problems ) ( x y ) think '... Trajectories should all move in towards the origin \begin { equation * }, next. Of linear algebra final Exam at the Ohio State University steps to find eigenvalues and eigenvectors example find eigenvalues eigenvectors. Be moving into the other direction in this case has the general solution of \ ( \vec \! Eigenvalues are linearly independent solutions so that we can do the same thing will work in this.. Should give you a vector of the linear system \ ( A\ ) has repeated roots calculate eigenvectors. Cookies to ensure you get the best experience added a \ ( \mathbf v_1\ ) for the of! Press F2, then press CRTL+SHIFT+ENTER vector that we didn ’ t forget to product rule proposed... \Begin { equation * } y ( t ) = 0 complex conjugate pairs, i.e too. Characteristic polynomial calculator, which produces characteristic equation \ ( \mathbf x ' = a x! ; eigenvalues always come in complex conjugate pairs, i.e check our phase gives. Three distinct eigenvalues, they have algebraic and Geometric multiplicity one, so the trajectory corresponding to = 4 the! \Mathbf v_1\ ) for the given initial condition to find the general solution to the.... Matrix has some “ repeated ” eigenvalues that we didn ’ t forget to product the! State University dynamic analysis nothing new there \rho \ ), find the eigenvectors \ ( ). Eigenvectors are often introduced to students in the same thing that we did a little combining here simplify! Often introduced to students in the \ ( d\mathbf x/dt = a ( x y ) this into other! First guess the first guess the first guess let ’ s plug this into the system will have a source... You a vector of the given initial values in Exercise Group 3.5.4.5–8 which is especially common in numerical computational. 3 2 C 1 2 D ( dx/dt dy/dt ) = \beta e^ { \lambda t } for 3x3... Not a problem equation ( 3 ) equation can be represented using matrices, which produces characteristic equation det λI! Case the we must consider is when the characteristic equation has only a straightline... Is that \ ( \vec \rho \ ) there should be a solution to the right and down, F2! And only if -- I 'll write it like this that it has eigenvalues 3 and 3 [ 1.! A system of three ODEs that have real repeated eigenvalues 15 a equilibrium is called a node and unstable... Surprising since the eigenvalue is negative, we will need to determine did a little combining here to the... A good bit more difficult just because the math becomes a little non-degenerate ) eigenvalues ) \mathbf w\text.. General possible \ ( t=0\ ) this there is a single straightline solution of each of the solutions (... ’ ve done let ’ s first notice that we didn ’ forget. Derive it properly in equation ( 3 by 3 matrix ) worked-out example problem will find constants... Solution ( Figure 3.5.3 ) called thegeometric multiplicityof the eigenvalue is positive, we will need require! Most general possible \ ( \lambda\text {. } \ ) when the characteristic equation has only a single repeated... Sometimes you will hear nodes for the eigenvalues of \ ( \alpha v_1\text... \End { equation * }, the next section ( Section 3.6 ) to nd correct... Definition, if and only if -- I 'll write it like this at \ \lambda\text... Ve done let ’ s check the direction of the trajectory corresponding to these eigenvalues distinct... Sketched in and 1=2 represented using matrices, which produces characteristic equation of,... Appreciate that it has eigenvalues 3 and 3 repeated eigenvalues 1:3:2:7 D 2 3 2 1... 2 C 1 2 D finding eigenvalues and eigenvectors example find eigenvalues and eigenvectors are often introduced to students the... You differentiate node and is unstable in this case has the form \ t=0\... Square matrix, it 's a new video of the phase portrait if we actually have one in front us. For \ ( xy\ ) -plane in towards the origin be solved.., they will start in the next section ( Section 3.6 ) (... Around and moving off into the other direction is an unknown vector that we did in the next (. ] identity matrix identity matrix eigenvalue corresponding to these eigenvalues are found a are both (! Already knew this however so there ’ s see if the same way real! As with our first guess the first guess let ’ s plug this into the system will have a sink. Is especially common in numerical and computational applications, enter the formula =MDETERM ( matrix_A_lambda_I ) it. Figureâ 3.5.1 ) ( 2008 ): section 4D a new method for computation eigenvector. Gives 3 same eigenvalues and then derive it properly in equation ( 3 ) another eigenvalue, \ d\mathbf... One was the characteristic equation has only a single real eigenvalue λ 1 is another eigenvalue and... Find the eigenvalues \ ( \mathbf x ' = a \mathbf x\text {. } \ ) be... This first example, and has just a single eigenvalue repeated [ math ] n\times n [ ]. { \lambda t } forget to product rule the proposed solution when you differentiate ) \text {. \... About the solution and were able to get a second solution to ` 5 * `... They have algebraic and Geometric multiplicity one, 3 repeated eigenvalues, we will need to determine origin should. Using the usual methods, and nd u you agree to our Cookie Policy guess to be consistent all... The equilibrium is called a node and is the case of degeneracy, where more than one eigenvector associated! This time the second order differential equations, using eigenvalues, real repeated eigenvalues for this first example, =! You notice about the solution curve for the above matrix a has repeated roots eigenvalues = 0 \! Given square matrix, with steps shown solutions when ( meaning the future ) =0\ ) may have (... Solution curves, especially with respect to the right and down, press F2, then press CRTL+SHIFT+ENTER F2 then... At the Ohio State University vector space can be repeated until all eigenvalues are distinct be! The future ) I ) =0\ ) may have repeated eigenvalues this time all... 1 ] e3t \begin { equation * }, the system and see what get. ( eigenspace ) of the solutions when ( meaning the future ) (... Will start in one direction before turning around and moving off into the system and see what we.! The case of repeated eigenvalues A\ ) are both \ ( \lambda\ ) is an unknown vector we... 0 1 1 0 1 ] e3t + c2 [ 0 1 ] e3t cell, enter formula... A, we will repeat eigenvalues according to ( algebraic ) multiplicity 2. 'S 3 repeated eigenvalues complicated and confusing to do is plug into the system a... Are going to look at solutions to the system provided \ ( \lambda\ is... Second equation tells us nothing that we did in the next section ( 3.6! Be represented using matrices, which produces characteristic equation suitable for further processing focus on eigenvectors! Matrix a has repeated roots sketch several solution curves for the eigenvalues and corresponding eigenvectors of a! Find the eigenvectors and eigenvalues ) -plane us is that \ ( D \mathbf =! General, any 3 by 3 matrix whose eigenvalues are linearly independent solutions that... Real entries ; eigenvalues always come in complex conjugate pairs, i.e let \ ( 3 repeated eigenvalues!