 Yeah, well thank you for inviting me. It's a pleasure to be here. It's a pleasure to talk about my favorite topic. They always say that the best salesman is one who believes in his product and I hope I hope I hope my enthusiasm will come across. Okay, so here we have our first slide we see a little sketch of me on orbit with the large area telescope purchased on top of it and a sketch of gamma rays. And so here's one I'm going to talk about first of course, a quick review of pulsars. They're essentials for those of you who don't think about these every day. And then I'm going to tell sort of the history of gamma ray pulsars starting in the early 90s, personal history, and then just telling the story of how things went for us with Fermilat. So I'll be talking about me. Okay, and then, and then the main point is, is that detecting gamma ray pulsations is really, really easy. Once you've put all the tools in place. And the tools are simply that your clock times have to be good. You have to have an accurate model of the spin down evolution of your neutron star. And you need to make sure that all your code is good. And at that point, in addition to all the various cuts you're used to to bring out your gammas relative to some sort of broad background. You have this timing signature which gives you an extra boost of sensitivity, and a very ambiguous signature so it's very gratifying when you see it, because it's just crystal clear you know that you're seeing what you think you're seeing and not something Then I'll be giving friendly advice to CTA and I'm going to be drawing the contrast between what was our situation before launch and what's your situation before you come up to full full power. Alright, so first statement, I have to make it is a pulsar is a rotating neutron star. We'll talk about more of this in a moment, but to cut these beams, and when the beam sweeps your line of sight you see a flash like for lighthouse, like for the light on top of the police car. That's, that's most of what you need to know about all stars. So neutron stars are very cool. It's the insist that normal matter. It's any, any more gravitational pressure, and it will collapse into a black hole, and they only exist in a fairly narrow range of masses they go from a little under 1.4 solar masses to getting towards three solar masses. It's an open question how, how heavy are the heaviest stars. They spin incredibly fast. They have a huge moment of inertia and so their rotation is incredibly stable. And they've inherited a magnetic field from the massive star from which they were born. The magnetic field is now concentrating to a very small volume and so is ridiculously strong. Small volume says 13 kilometer radius so it's smaller than a large city. And we all know that if you have something magnetic that spinning, we call it a dynamo makes electricity and electricity accelerates particles, and then these particles in the presence of a magnetic field or other matter will radiate. So there you've got it. They're the endpoint of solar evolution, stellar evolution, excuse me. So in this drawing I have here maybe you can see my mouse moving clouds of best collapse to make stars. And then whether it's a star of a few solar masses or less, or several solar masses or more, everything is different. The top branch, it's a very slow process, you have stars like our son, after billions of years they'll give you a planetary nebula and the wind up as a white dwarf, white dwarf which is atoms compressed by gravity to roughly the size of the earth. The lower branch is a little sexier it's a lot more fun. More gravitational compression hotter denser access to more thermal nuclear reactions. So burns more burns faster burns hotter and the lifetimes are not billions of years but are millions of years. And then when it's over, it runs out of gas, and explodes in a violent supernova. The core of the star, depending on this mass will leave you a neutron star or black hole. Everybody here knows the crab scene with naked eye, even in daytime in July of 1554. Probably everybody here sort of ecological likes recycling and so you'll like the fact that all this stardust gets blown out into the interstellar medium, then contributes to the next generation of stars, and all these heavy elements that you've made in the lifetime of heavy stars accumulates and contributes to the world as we know it so. Very cool stuff. And then how many poll stars are there in the Milky Way how many poll stars are there in the universe how many, how many neutron stars are there well just a rule of thumb is that if you have one supernova per century. And your galaxy's been around for 10 of the 10 years, but that gives you a border 10 to the eight neutron stars. And the stars have been predicted in the 1930s, and then mostly forgotten. Coincidence or is this volume fellow in the 60s who did some theoretical calculations of what I said, rotating, rotating magnet making a dynamo making beams of stuff. And at the same time they were radio astronomers who were looking for transient transient radio signals they were looking for interstellar scintillation in order to discover quasars. And this is really important because when you have noisy electronics, what you do is you integrate in time. And what that means is you get a big resistance times capacitance RC, so that you can wash things out. But if you want to see quicker signals then you need to beat down the noise in your system. And then you can shrink your RC time constants, and they did this. And in addition to the simulation and discovering lots of quasars. I brought just one bell picture on the left. She saw something she called scruff weird little signals on the on this trip chart recorders that she couldn't couldn't really identify and so she pursued it and talk to her boss and they're doing a zoom and when they did the zoom, you got the kind of thing like you see there which was individual pulses. Very exciting. And here we are. We today know about 333400 poll stars. And it's very convenient to represent them on what we call the PP background. P is the x axis, it's a spin period P dot is the rate of change of the spin period. And it's on a log log scale, and you clearly see two populations poll stars, what we call normal poll stars are younger middle aged, and then off to the left here recycled and millisecond poll stars. This green dot up here at top left is the crab, born about 1000 years ago. And you can see that it's right on this blue dashed line that we call the characteristic age, and this will make sense to you. If you see something that's spinning. And you measure its rate of spin down where you can extrapolate backwards and figure out when it started more and it's an approximation but it works okay. So poll stars are born up here at top left. They have their moment of inertia times their spin rate times the rate at which they're losing rotational energy, we call the spin down power. And down here E dot is equal to four pi squared times I times the been frequency times the first derivative spin frequency or p dot over p cube. And we see that it ranges from 10 to the 38 down to 10 to the 30 ergs per per second. So they spin, they radiate, they lose power. Since they lose power they spin less quickly. And since they spin less quickly, they radiate less. So it's a sort of vicious circle, the more they slow down, the less they radiate, but they're still radiating so they slow down more and so forth and so on. And so in the course of a poll stars lifetime, it will evolve across the field, and eventually get to this black line, which is an approximation called the radio death line. At this point, it's spinning so slowly, we're at periods of a few to several seconds here. And with a magnetic field that perhaps is weakened so that the dynamo is not working well enough anymore to make radio emission. We call that the radio death line I said that. And, and so these 100 million neutron stars in the galaxy, well a lot of them are here beyond the death line. But they don't pulse, they don't radiate we don't see them. They're dead. We call that the graveyard cemetery. Now I drew some lines here with different. I got one scurvy and I got one's going straight they have different slopes. And that's because we don't know exactly what the breaking mechanism is to slow down the start. If it's purely dipole radiation, then there's something called the breaking index I'll talk about that in a minute, which is three, and you get a diagonal more or less like the one I did there. But what if the magnetic field decreases little curve, what if the magnetic field increases little curve the other way. The other thing we see is that in order the, I didn't say it, but all the colored dots are pulsars we've seen in gamma rays. The green ones are ones that we found the way that CTA will find pulsars. That is to say, somebody already knows about the pulsar gives you a rotation of emerus, and you fold your data, and you see the pulsations. The blue ones is something that's been very successful for me lap. We find a static DC gamma ray source on the sky and wonder, maybe it's a pulsar, and then we do a pulsation search. We will all be extremely pleasantly surprised if CTA can do this someday, and I won't talk about it for today. And then over here the red ones, middle second pulsars. So what you notice, and then the black ones, the black ones are ones where we infer me we have a rotation of emerus. We've folded the data to look for pulsations, and we haven't seen any. And I'll go into this in more detail. And then the gray dots, the gray dots are known pulsars where we don't have a rotation of emerus, and so we never have folded it. My personal ambition is to fold all pulsars someday so I'm trying to track down rotation of femurities for the gray ones about halfway there, about halfway there. So let's look at the left hand side of the slide. And there's a little sketch that says what I said a minute ago. The spin frequency rate decreases as the years go by. And the rate of decrease decreases. As the rate decreases. Let me say that again. Okay, so I said down here the slow down p dot slows down, we call that P double dot. Now, if you have a perfect magnetic dipole and no timing noise, you get a parabola. Even if it's not a magnetic dipole if it's some other mechanism slowing down your neutron star, you'll get a parabola with a different curvature. In real life, we measure the parabola for a dozen or so pulsars, maybe 20. It's difficult to measure the second derivative. Because of something called spin down jitter or timing noise. And that is I said that this this neutron star with this huge inertia and this huge amount of rotational energy. It spins down mostly very regularly but things are going on, especially in the young pulsars you've got a pulsar wind carrying away energy, and it's turbulent, and there's feedback with the star that applies torques in a sporadic chaotic way. You can have micro star quakes that the crust of the neutron star settling in and cooling off and, and that too can lead to turbulent chaotic micro changes in the moment of inertia. And all this means that instead of following the black line and having nice move regular evolution. You can have some some chaotic movement. There are pulsars for which that timing noise is very, very small. And once you've measured how the slowdown evolves, you can make very accurate predictions for years to come about just when pulses will arrive. This is what we call the good timers, the fantastic natural clocks millisecond pulsars that people use to search for gravity waves. And I'll only mention in passing, ask me questions about it later if you like. Right. And off we go. So this is just about reincarnation life after death. I made you all sad talking about neutron star graveyard off your life. Some of these stars will find a binary companion the primary companion it gets old it gets fat and it's will give its annual momentum to the dead neutron star and spin it up. And that's how you get the millisecond pulsars before for me. There were some predictions of gamma rays from millisecond pulsars there was even a detection of one one of them with data from EGRIT on the Compton Gamma Observatory. But most people thought it would be anecdotal marginal not much. And one of the big surprises with Fermi is that it's turned out to be half. Half of our primary pulsars roughly are our millisecond pulsars and everything I'm going to tell you today about getting the timing right accurate timing is it's all more so for millisecond pulsars everything has to be that much more careful and precise. So pulsars are mainly a radio phenomenon. 90% of the pulsars that we know about are only known in radio, but the power in radio is peanuts. It's like wisps of smoke. And that's what you see on this block here you have the power as a function of the electromagnetic power and radio way over here is a factor of a million million down from the power radiated in gamma rays, which is 1000 times more than the power radiated in X rays. So the gamma ray pulsars, the available power from the spin down. So gamma ray is carried away by the electron wind, but a substantial fraction of it which can be a tenth of a percent to 50, 60, 70% of the power goes off in gamma rays. And so gamma rays are good diagnostic tools for understanding pulsar emission. Another thing is very hard spectrum increasing and at some point just runs out of the mechanisms that generate the gamma rays just can't go any farther, you're not going to get any higher energy electrons magnetic field is only as strong as it is. And so you get gamma rays up to a few GV and then the flux goes way down and CTA your your vocation is to study the details of this cut off and also look for additional components. And this is almost my last year. That's about my last. Probably most of you when you think of a pulsar beam, you think of the radio beam which is here in pink. It's sort of a conical thing. And, and, and I made that worse because I told you to think about a lighthouse and one house you cut this conical beam going out. Well, Gamma Ray beams aren't like that. Gamma Ray beams look more like the dorsal fin on this fish here. They're very, very narrow in longitude. They're very sharp longitude. They're very, very extended in latitude. It goes from the North Magnetic Pole here on the fishes forehead, all the way to the South Magnetic Pole down towards the fishes bottom. And, and then it's curved, it's curved because all of these magnetic field lines are rotating at speeds near the speed of light and, and so the lags effects of relativistic aberrations are consequent, and therefore you get this this bent stuff. So this is this is just something to help you think about it as fish, but down here lower left from Ellis Harding you have a model where you see that indeed. The most intense emission is around the equator. There is a mission all the way up to high latitudes Finland, Canada, Fuego, but it's much, much less. And then two beams, going along with a with a half rotation, a little less than a half rotation between the two beams. And then this banana here. Well, the region in the magnetosphere, where the electrons are getting accelerated to high energy, and then forced to follow these incredibly intense magnetic field lines to then radiate by curvature that whole region is more or less banana shape. And depending on the speed of rotation, the back and the inclination of the magnetic field relative to the rotation access. The banana will be more or less curved more or less fat more or less compact. But the reason I insist on this is that many of you today are in rooms where there's lighting from light bulbs which are spherical and there's a lot of light coming on tubes and the geometry of the radiated beam is different. And the geometry of the beams from neutron stars, they come from this banana shaped region, it all is tangential to the curves of the banana. And that's what gives you these incredibly narrow pulses and longitude, which can however be quite extended in latitude. But before the challenge for CTA is to detail as well as possible, what's happening, the highest energies rated by a pole star, and in particular to look for other components components that we don't see very well with Fermi, such as inverse and the consequences are important because, well, just as an example, you all know that people have been looking for signatures and park matter, neutrality and annihilation in the diffuse radiation all these years and they thought maybe they saw it and well, it's not at all clear because there is a population of gamma rays and a population of gamma ray pulsars. And so the pulsars even if you don't care that much about pulsars, you do need to understand their contribution to overall radiation fields for understanding galactic ecology. Here we go. Last, not least of the overview. These are the four pulsars seen from the ground that Francesco mentioned. Thank you very much, Rache. Rache and I work together a lot on Fermi and Hess pulsars and he's my main resource for knowing what's going on in your field. And so here's the PP diagram. He dot he. No TV cameras, most of the pulsars yet. And then up here you have crab, Vela, cominga, and another eager pulsar. All seen rather well in a range of energies. And, and how many should CT expect to see well that's been studied by Rache and his colleague in the doorway. And in some slides they sent me it looks like it's of order a dozen, but it doesn't as I understand it spread between North and South CTA South CC in North. And then also spread sort of evenly and right ascension which means you'll see some in the winter, some in the summer. So at any given time, any given telescope will not have more than a small handful of pulsars that are observable. That's it for the overview. And now let's see and I've used to have my time. There's my time. Um, so I used to be a sharing cup person. I worked for 10 years on atmospheric sharing of telescopes first in La Palma, and then in the Pyrenees. And that was the time at which I got familiar with eager at their catalog and and pulsars in particular. And when we were working on cat and Celeste in the Pyrenees. Celeste's goal was to get as low an energy as any sharing coffee telescope could at a time we were the lowest, we were quite proud of that. And we had a very messy detection of the crab, which we wrote up in a thesis but never published in a journal. I kind of think we really saw it. Okay, who cares never mind. We published it as a number limit. And I think you all know this plot here left this is a compilation after after Veritas finally did see it in a convincing way at high energy. So here's our upper limit from Celeste. And what's really, really important when you're looking for a faint signal. And, and you're right at the borderline of threshold is you've got to be 100% sure that your analysis works. And so you need to validate everything as well as possible. And you're looking for your actual money plot. And, and so what we did here and what I encourage you to do is to use your instrument to detect the crab optical pulsar. And we did this and we saw the optical phase I granted us a lot of fun and let us understand a lot of stuff. I'll go back to that a few minutes. All right. So, um, after 10 years of Celeste and Sharon cough I decided I was tired of full moons and rainy weather and I wanted to go into orbit. And I had the opportunity to invest and I went up to Stanford for a year and we were testing the perimeter with atmospheric nuance. And in the meantime, I was studying, you know, what do I want to do with science do I want to do with class. And I learned that poll stars. We're going to be just a bonanza with Jeremy. And that there was a person who was supposed to be in charge, and that person had gone off to do other things other other career opportunities and the topic was an orphan. And I went to the IU and Prague and I organized a splinter session about pulsar timing. They all came. Some people were sort of angry. What took you so long, what took you so long, but we got organized and we wrote up a paper and we formed something called the whole start timing consortium. And we resolved that based on what we'd seen from eager it. We were going to try to gamma phase fold, all of the high dot pulsars, which was 240 at the time. Today it's 440. I went into at that time GMO there Lucas chemo was my student, and he's now a radio astrometer astronomer in say he started the methodical task of testing, testing, testing, all of the code. And he wrote something called the Fermi template to plug in at that time. I'll talk more about that later. And what I want you to take away from this slide is just that we've got Columbia, whoops, mouse. We've got Columbia we've got our CBO we've got Green Bank we've got parks we've got, we've got a few key people from the key media telescopes on the planet, who are signing up to help us get this right. So the formal agreement was that they would use the high E dot ones, but in fact, goodwill and informally they ended up also giving us another 800 with all the dots ones that they just happened to have. And that enabled us to test not only our idea that only high E dot ones admit but also look for the other ones. And then here in the abstract we talked about verifying the software and and testing the accuracy of our whole chain. So, beautiful radio telescopes. And a memo of understanding. So I visited the beginning people a little bit of play, and they wanted things written black and white and now who, who does what who shares what if I share what people share with me. But that didn't last very long. And once we got to work and results are coming in. People trust each other and, and the MOU expired after a year or two. And today we just keep working as if, you know, here we are 14 years later. We just keep going. And in addition to the pulsar timing consortium there's something called the pulsar search consortium, which is to a large extent the same people same instruments, but a different goal. The goal is to search for unknown radio pulsars in pulsar like an identified gamma resources. You can ask me about that later. The other thing I learned when I was at Stanford is that, you know, I went around I said so you know how do we, you know, I've worked on Celeste I've worked on cat. We had a VM, we had data acquisition we had GPS. I knew about all that stuff and I would just you know I'd say well hey you know. What does that do its timing. And it was really scary, because everyone would say gee that's really important question. I'm too busy. I don't know, go ask so and so. And so and so would say the same thing. And I quickly realized that the whole issue of making sure that the times are right was an orphan. And this was scary. And at the same time. It's really important. I learned that every single space mission before glass had had a timing goof. And never the same one twice. So I started to get scared. So I thought, and as I said before you know I spent my working hours watching me on sailing through the lap. And I said well gee you know we got to do all we got to do all we got to do is take a pair of simulators, put them next to the lap. In the, in the on the ground before we launch, and you look at an atmospheric muon go through the lap, and it'll generate a timestamp through the flight software and the really complicated electronics and everything. So these are simple acquisition system I took the VME GPS module from from Celeste, which previously had been on cat, and you know it's just easiest pie. And so what put all that in a suitcase and checked my bags and went to Arizona and, and there we are we're in the factory of the future journal dynamics. And we put our little simulators right next to the thing and you can see here I'm showing you the little rectangles there that's a, that's an antenna for GPS receiver. And it's plugged into a little box like this, and there's two of them on board. Everything is redundant in redundancy of the redundant spacecraft. And that's why you also have two on each side in case one breaks you use the other, and then data on the other side you have them on both sides, so that you can see the whole sky. And this is the test and my surprise. There was a major problem. Look here at lower right. This is the difference between my simple GPS and the spacecraft GPS. It has an amplitude of one millisecond, a period of 290 seconds, and it just oscillates, and this is a disaster. With the young pulsars, you screw up your times by millisecond you barely see it. Right, you'd still see sharp crab pulses, you know, well, crab maybe not crab is 33 milliseconds but we've got lots of pulsars with 200 300 500 600 millisecond periods. Who cares, what's a millisecond between friends. And we didn't really believe in the millisecond pulsars at the time. Well it turns out that it's half of our sample it's a large amount of our science, and if the bug had still been there, we would have only seen a few consistent with expectations, and we would have seen broad sinusoidal peaks, like people see for x-ray millisecond pulsars. So, there's a good chance that everyone has said, Okay, that's just what Mother Nature gave us. We fixed it, look a few months and the bug, I understood right away but there's mess of your accuracy. Okay, so I need to speed up a little bit. The point of this slide is that the references are here. If you want to read a documentation on our ground time tests. Here they are. And here it read says that we trust our timestamps to about 300 mic nanoseconds. So my friendly advice for CTA is make a procedure for your northern telescope. See the crab optical pulsar with the least perturbation to your standard data acquisition. I mean you'll have to change something probably as much as possible have standard data files, standard pipeline standard processing and and have your optical pulsar come out with the right phase at the end. Now in the south. Well, the crab is unique. The next brightest pulsars 1000 times 1000 times fainter. So it's a real challenge. But if there's an ambitious smart grad student postdoc listening today. Do the calculation figure out if if if a huge CTA mirror and the light full small pixels and exquisite fast electronics can let you do it. And I guarantee that if you succeed. Pulsar people will notice some people will notice. And otherwise I read it on the slide. Dream something up. Okay, so we launched 14 years ago. So next month, everything worked right out of the box. We've got the cover of science. And here's an advertisement for a paper that we just published in science in April, where it turns out that our timing is so good that we can actually contribute to the searches for gravitational waves, and we're just to go think about that. I told you that the pulse art to Simon timing consortium has lived over 10 years beyond its normal expiration date, and this paper from 2019. We face full of 1000 gamma ray pulse arts that's these black dots, and the color dots for once we see a new ones we discovered the main thing in this paper is that we demonstrated that the five sigma threshold we've been using early in the mission is too conservative. Four Sigma is fine with four Sigma you are comfortably at the edge of your noise floor. And therefore you can lower your threshold before Sigma, if you are very careful about the number of trials. You can execute when you look for pulsations. And in this paper, we apply six trials, three for waiting parameter to throw the choice of all of our data, or just when the ephemeris is formally valid. And in this way, you know what's going on. And that's why we're running because trials are can what can make your life miserable. You know when you're starting out with your analysis you're not quite sure what's the best. This cat that that cat you know we know that reconstruction of Schoenckhoff imagers and stereo managers is complicated and there's different recipes and you can be tempted to try recipe a recipe D k k. So if you start to see a three single signal in one of your samples and not in the others, your critics will correctly say, Well, did you correct this for the 17 trials you made. So you need to understand the trials ahead of time. And so a lot of people come to me and say Dave you fuss so much about your ephemerides. Why don't you just download them from ATF. It's because they don't work. And the time is running slight so it's all here on the slide. My slides will be online for you. So we're marching on the rate of new pulsars seen in Fermilab is quickly approaching 300. We've detected 233. We have a number 40 or so that we're pretty darn sure will become memory poll stars in the next year to ask me about it later if you care. Yeah, the curve is starting to roll over a little bit. We're still marching forward. All right. Let's talk about CTA. Let's talk about that first. So, as I said, Fermilab scans the whole sky, eight times a day during the seminar, we're getting more data on everything. And so we potentially can see hundreds and hundreds of pulsars. Even once you don't expect. And so we wanted, we needed hundreds of radio ephemerides, but it's worse than that. But that is small. Sort of a meter and a half by a meter and a half and then, and then there's acceptance. So for our famous pulsars. We're not even getting one photon per month. And to see a signal above background, you integrate for years and years and years. And so we need our rotation ephemerides are folding to be valid. And the radio community they said they said it couldn't be done, but we do it anyway. Well, actually they said it had never been done. And it would be hard. The contrast is with CTA where as we said, you're going to have small numbers of pulsars. But better than that, better than that. You're, you, you, you look at a given pulsar during a season. Springtime. Here's this bullspark. Autumn. Here's that bullspark. You'll look at it for three months. You'll look at it for five months. You're never going to need a 10 year ephemeris. Especially if you followed my advice and you're extremely careful and rigorous about the absolute alignment of your CTA pulses with respect to whatever instrument is providing your reference. Because then you can look at it in spring of one year, spring of the next year. Oops, rain and repairs the year after three years later, more pulses. And if you're sure that every epic, you have the same alignment, then you can some year after year and watch your signal grow. I hope and believe that Fermilat will fly forever. It's working great. And the funding committees are convinced that we should keep it working, but it could get hit by an asteroid or something. It could die. And so your plan B is to have timing from radio astronomers. And if you, if you'd like, we, we, we in lat who are doing this. A lot of us are radio astronomers and we're in good relations with them so we can, we can help hook you up if you want to talk to them. Wow. It's now a quarter to 45, they asked me to say everything 45 minutes. 45 minutes and then 15 minutes for questions. You have still as yes five to 10 minutes. Yeah, but we want time for questions. Okay, well, I'll get to the point. What's the rotation of embers. Well, we already talked about this parabola here. But it turns out this F2 term, the thing that makes it bend, the quadratic term, it's really small. And if you're looking at short epics, it's a straight line to our first approximation. The, I wanted to find phase. So you have a model for the frequency as a function of time, you integrate your model and you get the number of rotations as a function of time. Phase phase is the remainder, the integer remainder at a given time T. Don't confuse T zero, which is the reference epic for the spin down with T zero MJD which is one of three parameters in the ephemeral parameter files that help you get your absolute phase right. This is a convention and you understand the convention and respect it very centering. I'll be quick. You measure the time at your instrument that your instrument is on the surface of spinning earth, which is going around the sun. And so you need to translate all those times to the solar system very center. And once we use, I use tempo to talk about later, they all do this. And I'm not going to explain you about Jupiter and planetary family days because I'm running out of time. If do is what gives you the breaking rate next. It's really interesting topic. We can talk about it some other day. But I do want to show you what a road patient ephemeral looks like in the real working world. And so I picked one of the easiest ones we can get. This is a radio quiet pulsar found in a blind search by Colin Clark. And therefore it's camera bright. And since it's camera bright. gamma timing is easy. And it's a middle age pulsar. And so it's turbulence youthful turbulence and trauma and all that is all gone away. And you see this ephemeris F zero F one. This is what you see here this beautiful straight line in this is phase as a function of the years going by is straight as an arrow with nothing but position spin and spin down rate. And it starts when our first data this is August 4, 2008 when our data became available. And this is 11.3 years later. So, when there's no noise. Ephemerides are very simple. I highly recommend anyone interested in pulsars to get in the habit of making this vertical plot. Okay, so here I'm preaching about alignment again. Here I'm insisting on the fact that there's a validity for an ephemeris we'll talk about that in a minute. And now let's move on to the more complicated cases. Well, I bet a lot of you every time I say timing noise you're thinking oh he must be talking about glitches. No glitches and timing noise are related. And maybe there's a continuum connecting the two, but basically glitches are huge massive events, where everything changes for a while, whereas glitches is just jitter all the time. And glitches are just not that hard to deal with you but a few parameters into your parameter file, and we deal with it, where you block out that time. So glitches aren't a problem. Binary is not a problem either. For most of your systems, the binary orbit is extremely stable. It's described with five lines in a parameter file, and then it works for years. This is what the use for gravitational waves. And for me discovered large numbers of spiders. These are very tight systems where the companion stars pooping all over the neutron star and wind and it's just a mess. And so then we widened them to hell. Okay, so what I want to insist on here is this is a polynomial and everybody listening today has fit noisy data to a polynomial. And we all know that as you increase the number of terms in your polynomial you can get a better and better fit. But the more terms you have the more a more extrapolation is possible and be at some point even interpolation starts to be problematic. So we do use very high order polynomials. The physical meaning of F zero F one F two that I talked about disappears. It's now just the coefficients of a Taylor expansion. The code lets you go to F 13 F 23 just go as far as you want. When you start getting these very complicated noisy systems over long durations rather than using a stupid polynomial. There are more clever things like harmonics of sinusoidals and a spline approach and the third pulsar catalog with our noisy pulsars, we have some horrific affirmaries. But you people in CPA don't need to worry about that, at least not at first. Simple if I'm reading these are nice because you can probably chase us a few lines of code. It's very attractive. And here I was just going to talk about how. So, um, how do you actually do it. Well, some words on software tools, but they do, as I said, I used to put to the way you make it a femoris is you get a hold of radio to a radio times of arrival, or extra times of arrival, friends of yours can send them to your or you make them from the lab data, which is easy to do there's good software. And then you put them into this tool graphic interface and you click and you choose and it works very nicely. It's if you're going to do it, I recommend that you get help from someone who's done it. And then over here. What precision do we need on the femorities that's a very important question, and we decided before launch that probably 20 mil periods, a 50th of a revolution would work well. And on this plot, you can see, plus 20 minus 20 thousands of a rotation. And this is typical what we find is that if you get to 20 mil periods, it works. And there's no real point in trying to get better. And it's substantially worse than that. You should probably figure out what's going on. Um, once you have an ephemeris to actually fold with the lab data, it's incredibly easy. All 14 years of lab data, big 32 gigabytes on my laptop. I use something from the science tools called GT select to pick the ones within two degrees of the pulsar position. And then I run the tempo to Fermi plug in written by my grad student a long time ago. It calls a Fermi data file, and the F2 file which gives you when where exactly is the satellite in its orbit, and a parameter file, and it calculates the pulse phases and adds into the F2 one file, and then you can make beautiful plots. And here's an example of a non detection. You see the pulse significance going up weighted age test, he gets to 15, which apparently is a little under three sigma. And from the other slide I showed you three Sigma is a typical large fluctuation, and is no reason to think that it's gamma pulsations here furthermore, a sort of a sinusoidal profile, the first harmonic. Yeah. To make to ways from that data. You extract peaks from short stretches of data and find the time using code in a package called geo time of arrival. You run item plate to get your profile shape for this pulsar gamma ray profile shape, and then you run something called you poly fold. And it just runs, it's very nice, beware, you have to very centered the time first, and I do that using a science tool. It works great. It's written up in the app J. And in Fermilat, we have the FSC Fermi Science Support Center, and there's a whole page full of wonderful software tools. This is where you need to do a nicely written up. You need to use weights for that data, and misty written up. I feed well, there's the link. And what you get in this ad weights is it's a 10 line Python script that you call from your code and it's just, it's just easy. Okay, this is my last slide. I'm an old fart. So I like to put do you youngsters, and I use the science tools I don't like them but it's just what I use for me pie for young people for me pie for beginners and pint pint is the future. And so then I asked Paul Ray and Matthew Curr I said, can I tell these people at CTA that that this tool does what I said does, and pint says yeah yeah it's definitely good for building your ephemera is from the two ways. It's definitely good for folding for me data. And it's definitely good for folding TV data. You've got, they're all in there. He says, it doesn't make to ways yet, but if there's enough. He uses something from nicer, which is also available, but he says, if the CTA people want it, we can add it to pint. And Matthew chipped in his two bits about how he uses five. Okay, last slide. conclusions. Y'all are going to do beautiful stuff I'm looking forward to it. You don't have the radical needs for timing radio timing that we did. You can just use what we have done to get what you need with some radio input here and there. You can just overlay a radio profile from time to time you wanted to get today you want to do a sanity check. But so you will need some radio input but you won't need a massive campaign with an MOU with 50 people, 20 people and 10 observatories. And then I'm saying this to young people postdocs grad students once again. Use your imagination to think about what can go wrong. Use your imagination to invent tests to check if it works, because anything that hasn't been tested well, probably doesn't work. Thank you very much.