WEBVTT 00:00.000 --> 00:11.000 Okay, I have a lot of covers, so I'm going to get straight into it. 00:11.000 --> 00:14.000 So we have a very simple agenda. 00:14.000 --> 00:16.000 There's some tedious stuff on Vita49. 00:16.000 --> 00:20.000 I'm going to take you through and then we're going to do cool stuff with the Wingerloat telescope. 00:20.000 --> 00:24.000 But it's a package deal, so you have to get both for me. 00:25.000 --> 00:28.000 So first, a few words about the Wingerloat telescope. 00:28.000 --> 00:31.000 I think a lot of you are familiar with it already. 00:31.000 --> 00:34.000 I've already presented here earlier on the Wingerloat telescope. 00:34.000 --> 00:41.000 Built in 1955, inaugurated 1956, mainly built for doing hydrogen line research, 00:41.000 --> 00:45.000 has been active for research until and of the 90s. 00:45.000 --> 00:54.000 Then decommissioned and early 2000s was taken over by a volunteer organization called Cameras, 00:54.000 --> 00:58.000 of which Paul and I are both volunteers, representatives here. 00:58.000 --> 01:04.000 So right now, let me see, does it work? 01:04.000 --> 01:06.000 Yeah, there we go. 01:06.000 --> 01:10.000 It's fully operational and it's probably better than ever. 01:10.000 --> 01:13.000 And we even have white rabbit link to a hydrogen measure. 01:13.000 --> 01:16.000 So not only a 25 meter dish, but also a good clock. 01:16.000 --> 01:19.000 And there's a lot of fun you can have with that. 01:19.000 --> 01:24.000 But before we get to that, the thing we ran into on a practical side. 01:24.000 --> 01:26.000 So this is all very practical. 01:26.000 --> 01:29.000 I've heard a few talks here with not very nice frameworks and program and language. 01:29.000 --> 01:32.000 This is much more just day-to-day stuff. 01:32.000 --> 01:35.000 Is that we are there in the telescope. 01:35.000 --> 01:39.000 We do something interesting following a spacecraft and deep space or doing something on the 01:39.000 --> 01:42.000 astronomical side. 01:42.000 --> 01:44.000 And there actually multiple things you want to do. 01:44.000 --> 01:46.000 You want to save the data to disk. 01:46.000 --> 01:48.000 You want to have a waterfall just to see it. 01:48.000 --> 01:52.000 If there's RFI, maybe you want to do some online analysis already. 01:52.000 --> 01:53.000 But it's really hard. 01:53.000 --> 01:57.000 If you only have one SDR, most tools just capture the SDR and that's it. 01:57.000 --> 02:02.000 And if you want to save an hour of data to this, then you're sit there, wait for an hour. 02:02.000 --> 02:04.000 And you have no idea what is happening. 02:04.000 --> 02:07.000 So we want to do multiple stuff at the same time with the same SDR. 02:07.000 --> 02:14.000 So basically, I brought to you one the main application of what we're trying to do here is 02:14.000 --> 02:18.000 to replace RFI letters with something digital. 02:18.000 --> 02:23.000 So instead of doing it on the analog side, I have multiple SDRs on the same dish. 02:23.000 --> 02:26.000 Let's see if we can do this on the digital side. 02:26.000 --> 02:29.000 The other thing is ultimately the data storage and metadata. 02:29.000 --> 02:33.000 So it happened me a few times that I did some measurements went home and then I happened 02:34.000 --> 02:37.000 you was visiting and it says I kill one, I kill two, I kill three, I kill four, 02:37.000 --> 02:38.000 and I know it anymore. 02:38.000 --> 02:40.000 What exactly I've captured there. 02:40.000 --> 02:44.000 So it would be really useful if we can store stuff and immediately have the data. 02:44.000 --> 02:46.000 Where was the telescope pointing at? 02:46.000 --> 02:47.000 Did we have the right pointing model? 02:47.000 --> 02:49.000 Which frequency was it on et cetera, et cetera? 02:49.000 --> 02:52.000 And not have to worry that yourself. 02:52.000 --> 02:54.000 The other thing is keep the SDRs running. 02:54.000 --> 02:57.000 So if you're going to do a very sensitive measurement, don't start your B210 02:57.000 --> 03:01.000 five seconds before the measurement because it's going to heat up. 03:01.000 --> 03:03.000 And the gain is going to change a lot. 03:03.000 --> 03:08.000 So keep it running for all the time that you're actually in the telescope or wherever you are deployed. 03:08.000 --> 03:11.000 This is not specifically to the granular telescope. 03:11.000 --> 03:16.000 Also, if you keep the SDR running, the face between different measurements is going to be continuous. 03:16.000 --> 03:18.000 Of course, it needs to be scriptical, scripted. 03:18.000 --> 03:22.000 So we want to have something CLI-based, accurate timing. 03:22.000 --> 03:27.000 And here I'm going to repeat something that Marcus said yesterday in the Knubolf is, 03:28.000 --> 03:33.000 why are there not so many or so few SDRs with proper timing and timed commands? 03:33.000 --> 03:35.000 Because that is what we need for all these things. 03:35.000 --> 03:39.000 And everybody builds another 936 something clone and this feature is lacking. 03:39.000 --> 03:41.000 They're all the same and they're all missing it. 03:41.000 --> 03:44.000 And the other thing really important buffering. 03:44.000 --> 03:50.000 So if you're writing to this, everything goes fine and certainly the disk sinks and your SDR starts dropping samples. 03:50.000 --> 03:51.000 That of course should note them. 03:51.000 --> 03:52.000 So this was the outline. 03:52.000 --> 03:55.000 And what we settle for is use VRT. 03:55.000 --> 03:58.000 So we feed the radio transport for streaming. 03:58.000 --> 03:59.000 You use sigma for storage. 03:59.000 --> 04:01.000 Now sigma for storage is already a given. 04:01.000 --> 04:03.000 I think everybody is on that page already. 04:03.000 --> 04:06.000 But VRT for streaming is somewhat less familiar. 04:06.000 --> 04:11.000 So if we look at VRT, so this is Vita 49 radio transport. 04:11.000 --> 04:13.000 And I took this from from their site. 04:13.000 --> 04:19.000 And they critically describe it as a transport layer of protocol designed to promote interoperability. 04:19.000 --> 04:22.000 I don't know what that means. 04:22.000 --> 04:26.000 Either it's unteroperable or it's not, but how do you promote it? 04:26.000 --> 04:29.000 And to be honest, they actually have somewhat of a point. 04:29.000 --> 04:32.000 It's not really a protocol like implement this and you're compatible. 04:32.000 --> 04:34.000 It's more like a framework. 04:34.000 --> 04:40.000 It's a set of definitions of frames and fields and you can use them for all kinds of purposes. 04:40.000 --> 04:43.000 But you might have different fields implemented. 04:43.000 --> 04:45.000 So you're not necessarily compatible. 04:45.000 --> 04:50.000 But it's a really great framework, I think, for doing exactly the things we need to do here. 04:50.000 --> 04:57.000 There's several versions, which are additions, so to 49.2 has all kinds of extra features on top of 49.0. 04:57.000 --> 05:01.000 We will actually be focusing on 49.0 here all the time. 05:01.000 --> 05:03.000 So what is this? 05:03.000 --> 05:08.000 If you're familiar, for example, with a ground station for geo-satellite, you have all kinds of receivers, 05:08.000 --> 05:14.000 then you have a physical IF distribution to all kinds of panels with something most of the time in L-bent, 05:14.000 --> 05:20.000 where you plug in your processors and they can duplicate and they can move things etc. 05:20.000 --> 05:24.000 It's a really complicated system, but there's an analog interface to it. 05:24.000 --> 05:34.000 So the modem side, let's say the digital side, analog interface interfaces to what comes from the dishes. 05:34.000 --> 05:40.000 Now, of course, that's what we want to replace with a digital network. 05:40.000 --> 05:46.000 So if we look at the received chain for any receiver, we have a lot of information that we want to capture. 05:46.000 --> 05:48.000 So it starts with the antenna. 05:48.000 --> 05:51.000 Where was the antenna pointing? 05:51.000 --> 05:54.000 Then we have an analog signal, which goes to a tuner. 05:54.000 --> 05:57.000 What was it tuned for? What was the gain? Was it locked? 05:57.000 --> 05:59.000 What is the bandwidth? 05:59.000 --> 06:03.000 On the ADC side, what sample rate was used? Was it overloaded? 06:03.000 --> 06:09.000 Et cetera, et cetera. So we have all this information and here in the chain, you get your IF data. 06:09.000 --> 06:12.000 And that's simply your IQ data. 06:12.000 --> 06:18.000 Most cases, this is the data that gets stored, and maybe there's a timestamp and a frequency, and that's basically it. 06:18.000 --> 06:22.000 This is all the stuff you want to capture, and also it can change. 06:22.000 --> 06:25.000 It's not necessarily static, it can change. 06:25.000 --> 06:34.000 So VRT actually merges those two things, and says you get one stream, and this stream has both your data and your IQ data, and it has your metadata. 06:34.000 --> 06:37.000 And then there's that in a fairly nice way. 06:37.000 --> 06:41.000 So just to have a look at the frame, this is the tedious part. 06:41.000 --> 06:45.000 This is the data frame, so here goes your IQ that goes in here. 06:45.000 --> 06:52.000 And then you have a header with some housekeeping, a little bit overkill and identify us with, okay, that's how it's defined. 06:52.000 --> 06:56.000 And really important, there's a very fine timestamp in everything. 06:56.000 --> 07:01.000 And there's some trailer where you can have some fields on what's everything locked and stuff like that. 07:01.000 --> 07:04.000 So most of your data actually goes in here. 07:04.000 --> 07:08.000 Now it's a big mistake if you take this frame and only use the IQ data. 07:08.000 --> 07:15.000 So you ignore all the metadata and you just use Feta 49 as a framing for your data. 07:15.000 --> 07:19.000 Because then you're throwing away all the good stuff, and then you think, well, what an over it. 07:19.000 --> 07:21.000 It's not a nice frame, let's not use it. 07:21.000 --> 07:26.000 So it's really the combination of this frame together with this metadata frame. 07:26.000 --> 07:29.000 So here you have a similar one, also with timestamp. 07:29.000 --> 07:32.000 But here you have all kinds of optional fields. 07:32.000 --> 07:35.000 So you can put a sample right in here again. 07:35.000 --> 07:38.000 The IF frequency is the bandwidth. 07:38.000 --> 07:41.000 Everything you want to have here, and it's also extensible. 07:41.000 --> 07:44.000 So this frame, you can send, for example, 10 times per second. 07:44.000 --> 07:48.000 So whenever you get a stream, you wait a few milliseconds. 07:48.000 --> 07:53.000 You get one of these frames, you know exactly what you're looking at, and you can interpret the IQ data that's coming. 07:53.000 --> 07:55.000 You can even extend this. 07:55.000 --> 07:57.000 So that's the whole idea. 07:57.000 --> 08:01.000 You combine this metadata and this IQ data. 08:01.000 --> 08:06.000 Now that sounds all perfect, and probably it sounds too good to be true, and there is lead a catch. 08:06.000 --> 08:08.000 And that's this one. 08:08.000 --> 08:09.000 It's not free. 08:09.000 --> 08:11.000 How can you imagine this? 08:11.000 --> 08:13.000 You need to pay for getting the standard. 08:13.000 --> 08:17.000 There's no limitation on what you can do with it, but you just need to pay to get to the document. 08:17.000 --> 08:21.000 And it's, it's watermarked when you get it really annoying. 08:21.000 --> 08:25.000 I searched, I hoped I would give you a URL here, said, 08:25.000 --> 08:27.000 Oh, but you can go to this URL and still get it. 08:27.000 --> 08:28.000 No, you can't. 08:28.000 --> 08:32.000 And I have a first in that I got from somebody who I'm not going to embarrass by giving it to you. 08:32.000 --> 08:34.000 So this is silly. 08:34.000 --> 08:36.000 I cannot do something about it. 08:36.000 --> 08:37.000 This is really stupid. 08:37.000 --> 08:41.000 But the good news is, we actually don't need this standard. 08:41.000 --> 08:43.000 Because it's already implemented. 08:43.000 --> 08:48.000 There's a nice library called Lip3RT from somebody called Emil Berg. 08:48.000 --> 08:50.000 Does anybody know him? 08:50.000 --> 08:52.000 Okay, I don't know him. 08:52.000 --> 08:54.000 It's really nice to have. 08:54.000 --> 08:57.000 It's really well written. 08:57.000 --> 08:59.000 I really easy to use. 08:59.000 --> 09:01.000 Well, document, et cetera. 09:01.000 --> 09:03.000 Even if you don't do anything with the stuff I've done, 09:03.000 --> 09:07.000 if you ever do something with Peter 49, this is the library you will need to. 09:07.000 --> 09:08.000 You need to have. 09:08.000 --> 09:09.000 So that's perfect. 09:09.000 --> 09:11.000 So we don't need to read this whole standard. 09:11.000 --> 09:15.000 We just use this library and it's API to do all the stuff we want. 09:15.000 --> 09:18.000 So we don't need to care about the standard itself. 09:18.000 --> 09:21.000 There's one caveat here in this library. 09:21.000 --> 09:25.000 And the standard says, all the fields are big and the end. 09:25.000 --> 09:28.000 This implementation says, I'm ignoring that. 09:28.000 --> 09:31.000 So I just do whatever the house does. 09:31.000 --> 09:38.000 That's not really nice because then it's not compatible anymore with between different platforms. 09:38.000 --> 09:42.000 In the other hand, I only use xc86 or arm and it's a little Indian. 09:42.000 --> 09:44.000 So it is compatible. 09:44.000 --> 09:51.000 So what I did in the beginning was I made a fork of this with a branch that is big and the end. 09:51.000 --> 09:54.000 So all the metadata is in big and the end. 09:54.000 --> 10:00.000 The IQ data I leave in little Indian because I see no reason for all these platforms that we use to do 10:00.000 --> 10:03.000 conversions all the time when it goes on to do wire. 10:04.000 --> 10:11.000 In the meantime, I found out that Epic as the R manufacturer also forked this library and that same thing. 10:11.000 --> 10:18.000 And we're actually in the description at the moment that we probably should collapse all three back to a mailbox original and make that one big and the end. 10:18.000 --> 10:21.000 Because we're talking here about a few lines of code. 10:21.000 --> 10:23.000 This is really nothing. 10:23.000 --> 10:25.000 Anyhow, so this library exists. 10:25.000 --> 10:32.000 So what we do here, we combine all of this and we're going to run this FIRT frames over zero MQ. 10:32.000 --> 10:37.000 Now, almost any other implementation I've seen that there's anything between the 49 including DV. 10:37.000 --> 10:38.000 I'll come to that. 10:38.000 --> 10:39.000 Use it. 10:39.000 --> 10:44.000 Use it over UDP, which seems obvious, but really why? 10:44.000 --> 10:47.000 I mean, at the data rates we're working TCP works fine. 10:47.000 --> 10:50.000 You have no packet loss and also it solves your buffer in problem. 10:50.000 --> 10:55.000 Because CMQ, you can use as a huge buffer, which is also client specific, which works really well. 10:55.000 --> 10:57.000 So we run this over zero MQ. 10:57.000 --> 11:01.000 The other thing is with zero MQ, you can connect from server to client from client to server. 11:01.000 --> 11:04.000 So you don't have this issue anymore that you have for UDP. 11:04.000 --> 11:05.000 It's so much easier. 11:05.000 --> 11:10.000 I've only found one case yet so far where UDP would have been better. 11:10.000 --> 11:16.000 So what we do, we have a set of tools that is something to VRT, basically that's your SDR. 11:16.000 --> 11:20.000 And in the other end, you have something that is VRT to something. 11:20.000 --> 11:22.000 For example, to write us to sigma. 11:22.000 --> 11:28.000 And then you have a pipe, which contains your IQ data, which is the majority of your frames. 11:28.000 --> 11:31.000 And so once in a while you interleave it with a context frame. 11:31.000 --> 11:35.000 That says, well, you're looking at this kind of band where it is the frequency, etc. 11:35.000 --> 11:38.000 And we put our own extended context in here. 11:38.000 --> 11:42.000 So this is the whole idea, and you can have multiple of these clients. 11:42.000 --> 11:45.000 The way this is implemented, it's all very, very, very simple. 11:45.000 --> 11:50.000 There's just an header library, nothing compiled, which you include. 11:50.000 --> 11:53.000 That actually interfaces with lib3rt. 11:53.000 --> 11:57.000 But the application itself is responsible for getting it on the CMQ stream. 11:57.000 --> 12:01.000 So you just get a CMQ packet, you give it to this library. 12:01.000 --> 12:03.000 It does all the parsing, and you get your data back. 12:03.000 --> 12:06.000 It's really simple, and then you can start building a toolset on top of that. 12:06.000 --> 12:08.000 And that is what we have done. 12:08.000 --> 12:11.000 So a few specifics for what we have implemented. 12:11.000 --> 12:15.000 Our data payload is always 16 bit complex. 12:15.000 --> 12:20.000 A little Indian simply because that's the easiest one on the platforms that we use. 12:20.000 --> 12:25.000 Every data packet is 10,000 samples, just fixed so that you always know, 12:25.000 --> 12:27.000 if you get a packet, this is what I'm going to get. 12:27.000 --> 12:30.000 So you don't need to know what you need to buffer for. 12:30.000 --> 12:34.000 And we have extended context classes for pointing, which is the telescope, 12:34.000 --> 12:38.000 to where is the telescope pointing, what were the settings for the pointing, 12:38.000 --> 12:40.000 and what were we tracking? 12:40.000 --> 12:43.000 For example, with Voyager, what we did earlier today, 12:43.000 --> 12:46.000 there would be two frames in there, one is where the telescope pointing to, 12:46.000 --> 12:48.000 and the other one is where it's Voyager. 12:48.000 --> 12:51.000 Ideally, those are the same, but if they're not, 12:51.000 --> 12:54.000 you really want to know that in your metadata. 12:55.000 --> 12:57.000 Good, then we come to the SDR. 12:57.000 --> 13:02.000 So we have a few platforms that we can, that we support right now. 13:02.000 --> 13:06.000 And again, this is mostly based, includes GPIO support. 13:06.000 --> 13:10.000 And you can actually use the GPIO on the SDR with your times commands. 13:10.000 --> 13:13.000 So the time you have in your IQ stream, 13:13.000 --> 13:17.000 and the time you have on the commands for your IO, are the same. 13:17.000 --> 13:19.000 So you can do really fancy things with switching, 13:19.000 --> 13:22.000 which I will show later how we use them. 13:22.000 --> 13:24.000 So with full, I mean, you have external reference, 13:24.000 --> 13:26.000 you have proper timing, everything. 13:26.000 --> 13:29.000 And the only tool that actually do that is, 13:29.000 --> 13:33.000 well, and I, and the RF space, cloud SDR from Peter, 13:33.000 --> 13:34.000 and it's versions. 13:34.000 --> 13:37.000 There's partial support, but that's more just for, 13:37.000 --> 13:40.000 for showing stuff, or developing, is the air spy, 13:40.000 --> 13:42.000 which actually works quite well. 13:42.000 --> 13:45.000 And the SDR, just for creating some IQ. 13:45.000 --> 13:47.000 Stuff, I have, but I've not released it, 13:47.000 --> 13:50.000 because I'm not completely happy about it, is the Pluto, 13:50.000 --> 13:52.000 which actually is works. 13:52.000 --> 13:56.000 The lime SDR, because I simply never ever can get it right. 13:56.000 --> 13:58.000 It always is off frequency. 13:58.000 --> 14:00.000 It's not, it just never works. 14:00.000 --> 14:02.000 This one, I really like, it's a really cheap clone 14:02.000 --> 14:04.000 that this direct sampling. 14:04.000 --> 14:06.000 You can run it at 200 megahertz. 14:06.000 --> 14:08.000 Perfectly fine. 14:08.000 --> 14:09.000 It's just not the SDR. 14:09.000 --> 14:10.000 It's an ADC, but who cares? 14:10.000 --> 14:12.000 And the epic sidekick, which I think is, 14:12.000 --> 14:15.000 maybe the only one that is on the level of this one. 14:15.000 --> 14:18.000 So that's what's currently on the SDR. 14:19.000 --> 14:21.000 So then, we have these tools. 14:21.000 --> 14:25.000 You have several of the SDRs. 14:25.000 --> 14:28.000 And your metadata is injected here. 14:28.000 --> 14:31.000 So there's a tool that actually sends from the control, 14:31.000 --> 14:34.000 the tracking software and in-dwing a low, 14:34.000 --> 14:37.000 into the stream, as well as the script that is the tracking. 14:37.000 --> 14:39.000 And then you have all these tools here. 14:39.000 --> 14:40.000 So you have, for example, Fiatty. 14:40.000 --> 14:41.000 It's a new radio. 14:41.000 --> 14:43.000 So if you want to play with a flow graph, 14:43.000 --> 14:45.000 at the same moment that you're doing a measurement, 14:45.000 --> 14:47.000 you just start a stool and you open a radio, 14:47.000 --> 14:49.000 you get your data into a radio. 14:49.000 --> 14:52.000 I have a sample flow chart to open it, 14:52.000 --> 14:54.000 because I didn't want to build the block. 14:54.000 --> 14:56.000 I wanted to have something that would work out of the box 14:56.000 --> 14:57.000 with every deployment. 14:57.000 --> 15:01.000 Fiatty to signal that is clear. 15:01.000 --> 15:04.000 Fiatty spectrum, I'll show more about that, 15:04.000 --> 15:06.000 but that was also the one I used for. 15:06.000 --> 15:08.000 It's just a quick show to spectrum. 15:08.000 --> 15:10.000 So for example, for Voyager. 15:10.000 --> 15:14.000 And for some reason, I accidentally put a twice on the sheet. 15:14.000 --> 15:18.000 But also, you can pipe this into or open this in other tools. 15:18.000 --> 15:19.000 This is SDR plus plus. 15:19.000 --> 15:21.000 So I made a plug in for that. 15:21.000 --> 15:22.000 And you just start it. 15:22.000 --> 15:24.000 And that automatically connects to the stream. 15:24.000 --> 15:26.000 Set a frequency, and you have a nice waterfall. 15:26.000 --> 15:28.000 We don't need to develop all everything ourselves. 15:28.000 --> 15:31.000 Whatever is there, we just interface to it. 15:31.000 --> 15:33.000 Here's an example from Dwinga Low. 15:33.000 --> 15:35.000 This is our Pills Heart Pipeline. 15:35.000 --> 15:38.000 So you click on the GUI, on the desktop, 15:38.000 --> 15:39.000 and it starts this. 15:39.000 --> 15:42.000 It starts the B210 software. 15:42.000 --> 15:45.000 It runs Fiatty Pills Heart, which is in demo. 15:45.000 --> 15:48.000 We have for live show and the Pills Heart. 15:48.000 --> 15:52.000 But in the back, it also starts Fiatty to data, 15:52.000 --> 15:54.000 which is a professional Pills Heart software. 15:54.000 --> 15:56.000 And it does call all kinds of analysis. 15:56.000 --> 15:59.000 So everybody from our volunteers will give us a tour 15:59.000 --> 16:01.000 automatically saves all kinds of data, 16:01.000 --> 16:03.000 which we can aggregate after a year, 16:03.000 --> 16:05.000 and do very nice analysis on. 16:05.000 --> 16:09.000 So here is already an application where you run two or three things 16:09.000 --> 16:10.000 at the same time. 16:10.000 --> 16:15.000 So you can also save the filter bank, which is a radio strongly format. 16:15.000 --> 16:17.000 So Fiatty Spectrum is the tool, 16:17.000 --> 16:20.000 where you can do easy stuff to make spectra. 16:20.000 --> 16:25.000 So, because not everybody, for example, in cameras, 16:25.000 --> 16:27.000 wants to do radio stuff. 16:27.000 --> 16:29.000 Sometimes you just want to have a spectrum and do something 16:29.000 --> 16:30.000 astronomically. 16:30.000 --> 16:33.000 You don't want to go through all the hassle of figuring out 16:33.000 --> 16:35.000 how to do it on the radio side. 16:35.000 --> 16:37.000 So again, you start the SDR tool, 16:37.000 --> 16:39.000 and you start the spectrum and you say, 16:39.000 --> 16:41.000 okay, I want to have, in this case, five bins, 16:41.000 --> 16:43.000 normally you would say 1000 or so. 16:43.000 --> 16:45.000 Give me 1000 bins, one second, 16:45.000 --> 16:48.000 start on the, on the integer second, 16:48.000 --> 16:50.000 so I have nice easy timestamps, 16:50.000 --> 16:51.000 and you just get a CSV file, 16:51.000 --> 16:53.000 and you can open it in Excel. 16:53.000 --> 16:54.000 I mean, for a lot of people, 16:54.000 --> 16:57.000 that is what, what they need. 16:57.000 --> 16:59.000 Or you open it in Python. 16:59.000 --> 17:02.000 So this is, we export actually an ECSV format, 17:02.000 --> 17:05.000 which is CSV with a little bit of metadata. 17:05.000 --> 17:08.000 So here we do a scan over something called 17:08.000 --> 17:11.000 from Wooden Cloud, which is discovered by, 17:11.000 --> 17:13.000 actually somebody in the Dwinga file. 17:13.000 --> 17:14.000 You open it. 17:14.000 --> 17:16.000 You actually get the frequency and everything, 17:16.000 --> 17:18.000 and just with a few commands, 17:18.000 --> 17:19.000 you can make a nice plot. 17:19.000 --> 17:21.000 So if you just want to focus on doing your radio, 17:21.000 --> 17:22.000 or strong enough, 17:22.000 --> 17:24.000 this is a room run at very long intervals, 17:24.000 --> 17:27.000 like several minutes per per line, 17:27.000 --> 17:29.000 and you have a few days of data and you plot it 17:29.000 --> 17:32.000 and you can see how the spectrum is used. 17:32.000 --> 17:36.000 So really simple stuff to get insight 17:36.000 --> 17:38.000 in your spectrum. 17:38.000 --> 17:40.000 So here's a list of some of the tools. 17:40.000 --> 17:42.000 We have built this way. 17:42.000 --> 17:44.000 We have VAT to filter bank, 17:44.000 --> 17:46.000 which saves in filter bank, 17:46.000 --> 17:47.000 which is in radio. 17:47.000 --> 17:48.000 So let me format. 17:48.000 --> 17:50.000 This is a fast radio burst, 17:50.000 --> 17:53.000 captures with, with our tools basically. 17:53.000 --> 17:55.000 So 200 megahertz of data, 17:55.000 --> 17:57.000 and so this runs at 200 megahertz, 17:57.000 --> 18:01.000 and this is like a single burst coming out of the universe, 18:01.000 --> 18:02.000 somewhere. 18:02.000 --> 18:04.000 VAT-Poser, that's our demo, 18:04.000 --> 18:06.000 VAT-P to SIG-MF, 18:06.000 --> 18:07.000 I cover it. 18:07.000 --> 18:10.000 VAT to SDRF, that's from case Balsa, 18:10.000 --> 18:12.000 where you can do satellite orbit fitting. 18:12.000 --> 18:14.000 So you can run that in parallel, 18:14.000 --> 18:16.000 really convenient, 18:16.000 --> 18:18.000 a new radio, we cover it, 18:18.000 --> 18:20.000 and then there's VAT tuner. 18:20.000 --> 18:23.000 This one actually extracts a sub band. 18:23.000 --> 18:25.000 So for example, you just start your SDR, 18:25.000 --> 18:26.000 roughly at the right frequency, 18:26.000 --> 18:28.000 and you say that to 10 megahertz, 18:28.000 --> 18:29.000 and you say, okay, 18:29.000 --> 18:31.000 now I want to have this 200 kilohertz out of that band. 18:32.000 --> 18:34.000 It's a polyphase filter, 18:34.000 --> 18:37.000 but it can also follow objects. 18:37.000 --> 18:39.000 So for example, this is what I use for Voyager, 18:39.000 --> 18:41.000 which I will have an example of. 18:41.000 --> 18:43.000 So if your stream already says, 18:43.000 --> 18:44.000 I'm following Voyager, 18:44.000 --> 18:45.000 this is the frequency, 18:45.000 --> 18:47.000 you can actually have this tuner, 18:47.000 --> 18:48.000 follow the object, 18:48.000 --> 18:50.000 and therefore the Doppler it. 18:50.000 --> 18:51.000 Then there's a channelizer, 18:51.000 --> 18:53.000 which is a polyphase filter bank, 18:53.000 --> 18:55.000 that can extract all the sub bands. 18:55.000 --> 18:59.000 But it can also do a non-minimal decimation. 18:59.000 --> 19:02.000 So for example, if you have 20 megahertz, 19:02.000 --> 19:05.000 you can split it in 10 times four megahertz, 19:05.000 --> 19:06.000 or something like that. 19:06.000 --> 19:08.000 So you get overlapping channels. 19:08.000 --> 19:09.000 The nice thing is, 19:09.000 --> 19:11.000 this also runs on a GPU. 19:11.000 --> 19:13.000 So we can easily run this up to 100 megahertz 19:13.000 --> 19:15.000 or something as split it in like 100, 19:15.000 --> 19:17.000 one megahertz channels. 19:17.000 --> 19:19.000 This is my favorite, Fiati to Void, 19:19.000 --> 19:20.000 doesn't do anything. 19:20.000 --> 19:22.000 It just connects and gets the stream. 19:22.000 --> 19:24.000 But it's also the basis for writing new tools, 19:24.000 --> 19:26.000 because somewhere in that tool, 19:26.000 --> 19:28.000 it says do something with your IQ here. 19:28.000 --> 19:31.000 So you don't need to start writing from scratch. 19:31.000 --> 19:33.000 Next one is a correlator. 19:33.000 --> 19:37.000 So basically it calculates the cross spectrum of two streams. 19:37.000 --> 19:39.000 There's a four-worder that's for a local fan out. 19:39.000 --> 19:41.000 So if you are at a remote site, 19:41.000 --> 19:43.000 you pull the IQ data in once, 19:43.000 --> 19:45.000 and then you locally distribute it again. 19:45.000 --> 19:47.000 There's a control VRT, 19:47.000 --> 19:49.000 which is the other direction. 19:49.000 --> 19:51.000 So you can actually send commands to the SDR. 19:51.000 --> 19:52.000 You can say, 19:52.000 --> 19:56.000 change your gain, change your frequency to change the settings, 19:56.000 --> 19:58.000 and there's play VRT, 19:58.000 --> 20:00.000 which is for transmitting signals. 20:00.000 --> 20:02.000 All of these I have examples of, 20:02.000 --> 20:04.000 so that's more interesting to look at. 20:04.000 --> 20:05.000 And as I said, 20:05.000 --> 20:07.000 we don't need to reinvent the wheel. 20:07.000 --> 20:08.000 We use set them, 20:08.000 --> 20:09.000 we use SDR++. 20:09.000 --> 20:13.000 I just wrote a plugin that connects to the stream. 20:13.000 --> 20:15.000 You can run both of them at the same time, 20:15.000 --> 20:17.000 no issue at all. 20:17.000 --> 20:21.000 So really convenient for doing a lot of stuff at the same time. 20:21.000 --> 20:24.000 So a few questions that I ask, 20:25.000 --> 20:27.000 why not use new radio for all of these things? 20:27.000 --> 20:28.000 Could you do that? 20:28.000 --> 20:29.000 I guess you can. 20:29.000 --> 20:30.000 And in some case, 20:30.000 --> 20:31.000 that might be the better solution. 20:31.000 --> 20:32.000 In some case, it's not. 20:32.000 --> 20:35.000 This is not a new radio versus VRT tools. 20:35.000 --> 20:36.000 This is really, 20:36.000 --> 20:39.000 how do you combine and what is more useful. 20:39.000 --> 20:41.000 Also, 20:41.000 --> 20:45.000 I think new radio is a great client for actually this stuff. 20:45.000 --> 20:47.000 Then there's something called D.V. 20:47.000 --> 20:48.000 You might have heard of, 20:48.000 --> 20:50.000 that's a normal implementation of Vida 49. 20:50.000 --> 20:52.000 We looked at that, 20:52.000 --> 20:54.000 but there were immediately all kinds of choices 20:54.000 --> 20:56.000 that didn't work for us. 20:56.000 --> 20:57.000 That's why we didn't use it. 20:57.000 --> 20:59.000 Also, it runs over UDP. 20:59.000 --> 21:00.000 So for us, 21:00.000 --> 21:01.000 it was not the right choice, 21:01.000 --> 21:03.000 but I'm still thinking about making a tool 21:03.000 --> 21:04.000 that converts between the two, 21:04.000 --> 21:06.000 because they're so close. 21:06.000 --> 21:09.000 It runs at easily at two times 100 megahertz, 21:09.000 --> 21:10.000 but we do it with X310. 21:10.000 --> 21:12.000 We also tried 400 megahertz, 21:12.000 --> 21:15.000 that's a little bit on the edge. 21:15.000 --> 21:16.000 And unfortunately, 21:16.000 --> 21:18.000 it's not really well documented, 21:18.000 --> 21:20.000 which is my fault. 21:20.000 --> 21:21.000 Good. 21:21.000 --> 21:22.000 Now, we got the cool stuff. 21:22.000 --> 21:23.000 What can we do with this? 21:23.000 --> 21:26.000 So this is running at several places already. 21:26.000 --> 21:29.000 So we've seen Bokum this morning. 21:29.000 --> 21:31.000 Dring a low, of course. 21:31.000 --> 21:33.000 Stock art is using it. 21:33.000 --> 21:34.000 It's not a 25 meter addition, 21:34.000 --> 21:35.000 Germany. 21:35.000 --> 21:36.000 I've installed it at the ATA. 21:36.000 --> 21:38.000 On the new radio server. 21:38.000 --> 21:41.000 And this is a very simple dish in my backyard, 21:41.000 --> 21:43.000 which I'll share later as well. 21:43.000 --> 21:45.000 So, 21:45.000 --> 21:47.000 but we've seen this morning was 21:48.000 --> 21:51.000 the Voyager reception. 21:51.000 --> 21:52.000 So in 2006, 21:52.000 --> 21:53.000 they had done this in Bokum, 21:53.000 --> 21:55.000 with a lot of equipment. 21:55.000 --> 21:56.000 And of course, I mean, this was a year ago, 21:56.000 --> 21:58.000 it was not so simple. 21:58.000 --> 22:00.000 But in between, there has been nothing. 22:00.000 --> 22:02.000 So in between last year and 2006, 22:02.000 --> 22:04.000 they have not seen Voyager. 22:04.000 --> 22:05.000 And with these tools, 22:05.000 --> 22:08.000 it's actually relatively easy. 22:08.000 --> 22:11.000 And then we did it this morning again here, live. 22:11.000 --> 22:14.000 So this is a good example on what we actually run. 22:14.000 --> 22:15.000 So here, 22:15.000 --> 22:18.000 we have to be to 10, which I use in all my examples. 22:18.000 --> 22:20.000 We have Pursa P to Vrt. 22:20.000 --> 22:22.000 We have the Dringler Telescope control 22:22.000 --> 22:24.000 that inserts the tracking data, 22:24.000 --> 22:27.000 or actually the data that this is pointing. 22:27.000 --> 22:28.000 We have a tracking script, 22:28.000 --> 22:30.000 which takes high-scurrenals as the input. 22:30.000 --> 22:33.000 It tells the control where the track. 22:33.000 --> 22:35.000 But it also inserts the same data 22:35.000 --> 22:37.000 into the IQ stream. 22:37.000 --> 22:40.000 Then we get the zero MQ stream, 22:40.000 --> 22:41.000 which is in this case. 22:41.000 --> 22:42.000 One megahertz, 22:42.000 --> 22:44.000 it's overkill, 22:44.000 --> 22:46.000 why would you not set it to that. 22:46.000 --> 22:48.000 And then we have a tuner, which is tracking. 22:48.000 --> 22:52.000 So this one is following the Voyager signal. 22:52.000 --> 22:54.000 So it deducts the Voyager in a continuous way. 22:54.000 --> 22:56.000 So every sample gets its own correction. 22:56.000 --> 23:01.000 It's not doing like one hertz steps every second or something like that. 23:01.000 --> 23:02.000 It's really simple by simple. 23:02.000 --> 23:03.000 It's continuous. 23:03.000 --> 23:05.000 And then you get a stream, 23:05.000 --> 23:07.000 which is only 10 kilohertz, 23:07.000 --> 23:09.000 and that's why you run into Vrt's spectrum, 23:09.000 --> 23:11.000 and you see the peak. 23:11.000 --> 23:12.000 The other one, we run one, 23:12.000 --> 23:14.000 because we also want to save data, 23:14.000 --> 23:16.000 but you don't want to dedopler the data. 23:16.000 --> 23:18.000 You want to have the original data, of course. 23:18.000 --> 23:19.000 So we run another tuner, 23:19.000 --> 23:21.000 which we set non-thracking. 23:21.000 --> 23:24.000 So that one just saved 50 kilohertz of the original data, 23:24.000 --> 23:27.000 because it doesn't make any sense to save one megahertz 23:27.000 --> 23:30.000 for a CW of one hertz. 23:30.000 --> 23:32.000 And that we save to sigma f. 23:32.000 --> 23:34.000 So this is really simple setup, 23:34.000 --> 23:36.000 but this is how we track Voyager. 23:36.000 --> 23:38.000 And you can run that on your laptop, of course. 23:38.000 --> 23:40.000 It doesn't take any CPU. 23:40.000 --> 23:41.000 Good. Now you might think, 23:41.000 --> 23:43.000 I don't have a 25 meter dish. 23:43.000 --> 23:44.000 Why is this important to me? 23:44.000 --> 23:47.000 You can easily do this on a 1.5 meter dish. 23:47.000 --> 23:49.000 Of course, not with Voyager, 23:49.000 --> 23:52.000 but this is a set of really simple dish. 23:52.000 --> 23:54.000 It's not optimized in any way. 23:54.000 --> 23:56.000 And I easily do, 23:56.000 --> 23:57.000 do you know, 23:57.000 --> 24:01.000 at Jupiter, which is 620 million kilometers away. 24:01.000 --> 24:03.000 And this is Beppe Colombo. 24:03.000 --> 24:05.000 This is actually the mercury fly by, 24:05.000 --> 24:06.000 just on a 1.5 meter dish. 24:06.000 --> 24:09.000 So you really don't need fancy stuff to do. 24:09.000 --> 24:10.000 If you also don't need to beat to 10, 24:10.000 --> 24:12.000 you can do this with an air supply. 24:12.000 --> 24:14.000 It just works. 24:14.000 --> 24:15.000 Good. 24:15.000 --> 24:17.000 Now that was about 25 meter dish, 24:17.000 --> 24:21.000 but now suppose you have two 25 meter dishes. 24:21.000 --> 24:24.000 And now the timing gets into the picture, 24:24.000 --> 24:28.000 because if we're the only dish that timing is somewhat less relevant, 24:28.000 --> 24:30.000 with two dishes, it starts metering. 24:30.000 --> 24:32.000 So we have exactly the same setup. 24:32.000 --> 24:33.000 And this is important, 24:33.000 --> 24:35.000 because even to beat to 10, 24:35.000 --> 24:36.000 it's not accurate in everything. 24:36.000 --> 24:38.000 If you set it to a certain frequency, 24:38.000 --> 24:40.000 it has a few million Hertz offset. 24:40.000 --> 24:41.000 There might be over that. 24:41.000 --> 24:43.000 There are offsets in timing, etc. 24:43.000 --> 24:45.000 But there are exactly the same. 24:45.000 --> 24:46.000 So they are reproducible. 24:46.000 --> 24:49.000 As long as you do the same setup on both sides, 24:49.000 --> 24:50.000 you get the same offset. 24:50.000 --> 24:51.000 So they cancel out. 24:51.000 --> 24:55.000 It's really important to set the master clock rate of these things, 24:55.000 --> 24:57.000 the same, because a lot of it depends on that. 24:57.000 --> 24:59.000 Then a new feature here is, 24:59.000 --> 25:03.000 we can also transmit with the B210. 25:03.000 --> 25:04.000 So we have a tool, 25:04.000 --> 25:06.000 which we can transmit files. 25:06.000 --> 25:09.000 So on one side, which is Ringelow, which runs metafile, 25:09.000 --> 25:11.000 we bounce it off the moon, basically. 25:11.000 --> 25:14.000 And then on both sides, we receive it, 25:14.000 --> 25:16.000 and we save it into a sigma file. 25:16.000 --> 25:18.000 So we have all the metadata, 25:18.000 --> 25:22.000 and we have a nicely synchronized set of data. 25:22.000 --> 25:24.000 Now we can do that on the moon. 25:24.000 --> 25:28.000 Later on, we can also see we can try this on Venus. 25:28.000 --> 25:31.000 So the physical setup looks like this. 25:31.000 --> 25:34.000 We have the B210 on the output. 25:34.000 --> 25:38.000 We have a small amplifier, which has an output of one watt, 25:38.000 --> 25:42.000 because we need to drive the amplifier here to 120 watts, 25:42.000 --> 25:43.000 and it's one watt input. 25:43.000 --> 25:46.000 And this is the sequencer for the Ringelow telescope. 25:46.000 --> 25:48.000 So this is all the switching of the relays, 25:48.000 --> 25:51.000 and the LNA is an everything to make sure you don't destroy anything. 25:51.000 --> 25:53.000 When you start transmitting. 25:53.000 --> 25:55.000 And we control that from the B210. 25:55.000 --> 25:58.000 So based on the Vita 49 stream, 25:58.000 --> 26:03.000 we send commands that actually do the relays switching. 26:03.000 --> 26:06.000 So it tells you just before you're going to transmit, 26:06.000 --> 26:07.000 switch the relays. 26:07.000 --> 26:10.000 And it's always in sync with your transmission, 26:10.000 --> 26:12.000 because it's using the same timestamps. 26:12.000 --> 26:16.000 And then over here, we have the relay and the feed. 26:16.000 --> 26:20.000 In the end, we only get 70 watt or so out of the feed. 26:20.000 --> 26:22.000 Keep that in mind when you're going to see the pictures. 26:22.000 --> 26:25.000 So here you see basically what happens? 26:25.000 --> 26:28.000 At one second, we start transmitting. 26:28.000 --> 26:30.000 And you see just before one second, 26:30.000 --> 26:32.000 you see the switch being made. 26:32.000 --> 26:34.000 Now the protection relay goes on. 26:34.000 --> 26:37.000 We are switched the receiver to a 50-on resistor, 26:37.000 --> 26:39.000 which is hotter than the sky we were looking at. 26:39.000 --> 26:43.000 So our noise actually goes up when we terminate the LNA. 26:43.000 --> 26:47.000 Here we transmit, of course, that gives a complete clipping of the signal. 26:47.000 --> 26:48.000 So you see all kinds of stuff. 26:48.000 --> 26:49.000 Ignore that. 26:49.000 --> 26:50.000 This is just a CW. 26:50.000 --> 26:53.000 Here we wait a little bit before we switch back. 26:53.000 --> 26:55.000 And here we go into the receiver mode. 26:55.000 --> 26:59.000 And here you see the CW being reflected from the Moon. 26:59.000 --> 27:01.000 You already see this as like offset and frequency. 27:01.000 --> 27:03.000 So there's a bit of Doppler from the Moon. 27:03.000 --> 27:08.000 But this basically is the whole flow for transmitting and receive. 27:08.000 --> 27:11.000 Okay, I'm going to not go to cover the details here. 27:11.000 --> 27:14.000 The Moon is not a point source. 27:14.000 --> 27:17.000 So you have different speeds to different parts of the Moon, 27:17.000 --> 27:19.000 which gives different slightly different Dopplers. 27:19.000 --> 27:22.000 And that's what we're going to use for our analysis. 27:22.000 --> 27:25.000 So we sent a CW, so a pure tone. 27:25.000 --> 27:27.000 But we get back, has some spreading in it. 27:27.000 --> 27:30.000 So we sent basically a single tone. 27:30.000 --> 27:34.000 And we get something in this case, which is spread around nine hertz. 27:34.000 --> 27:38.000 That's basically of this pseudo rotation of the Moon. 27:38.000 --> 27:40.000 We can actually predict it. 27:40.000 --> 27:43.000 And I might even have copied this code from you. 27:43.000 --> 27:50.000 So you can predict how this Doppler is going to be spread over the Moon. 27:50.000 --> 27:53.000 And you will see this, this widening. 27:54.000 --> 27:58.000 With that, you can do something called Delay Doppler mapping. 27:58.000 --> 28:00.000 Because basically we have two axes now. 28:00.000 --> 28:03.000 We have the Delay, because this is the front of the Moon. 28:03.000 --> 28:05.000 So our first echo will come from the front. 28:05.000 --> 28:08.000 Our later echoes will come from deeper parts of the Moon. 28:08.000 --> 28:09.000 And we have the Doppler. 28:09.000 --> 28:12.000 There's only one problem, there's an ambiguity. 28:12.000 --> 28:14.000 So where the Doppler and the Delay cross each other, 28:14.000 --> 28:17.000 there are two points that map to the same Delay Doppler. 28:17.000 --> 28:19.000 And we're going to see that later on. 28:19.000 --> 28:22.000 Basically, the two hemispheres will fall over each other. 28:22.000 --> 28:24.000 We have not invented this. 28:24.000 --> 28:26.000 This was done in the sixties. 28:26.000 --> 28:32.000 I think this one was with RECBO or Hillstone. 28:32.000 --> 28:36.000 So there they used a pulse, because they couldn't do the fancy stuff. 28:36.000 --> 28:37.000 We're going to do. 28:37.000 --> 28:40.000 But you already see the Delay Doppler map coming back. 28:40.000 --> 28:43.000 So we do this as well with these scripts. 28:43.000 --> 28:45.000 So we just had this play command. 28:45.000 --> 28:48.000 We play a sigma file, which is this one. 28:49.000 --> 28:51.000 Completely overloads the front end. 28:51.000 --> 28:53.000 And we receive an echo here. 28:53.000 --> 28:58.000 And this echo we're going to compress the single point. 28:58.000 --> 29:02.000 And now we can actually measure the depth of the Moon. 29:02.000 --> 29:05.000 So here we, this is the front of the Moon. 29:05.000 --> 29:06.000 So you see the peak. 29:06.000 --> 29:09.000 And then you see it slightly, slightly, 29:09.000 --> 29:11.000 dying out deeper in the Moon. 29:11.000 --> 29:15.000 And this is where this is basically the depth of the Moon. 29:15.000 --> 29:18.000 So what we use, and again, I'm not going into detail here, 29:18.000 --> 29:20.000 because the simply isn't the time for that. 29:20.000 --> 29:22.000 We use it of two sequences. 29:22.000 --> 29:25.000 If you want to learn more of those, go to Danny's block. 29:25.000 --> 29:29.000 He wrote some nice details on an analysis on how this works. 29:29.000 --> 29:32.000 If I have to summarize it in one line, 29:32.000 --> 29:34.000 it's basically an alias chirp. 29:34.000 --> 29:39.000 And if you look at the waveform, it looks like a chirp, 29:39.000 --> 29:42.000 but also you see there are actually more chirps at the same time. 29:42.000 --> 29:43.000 Kind of. 29:44.000 --> 29:47.000 But they have very nice outer correlation properties that we're going to use. 29:47.000 --> 29:51.000 And you can make them extremely long, indefinitely long. 29:51.000 --> 29:55.000 Doesn't ambiguity, which I will skip now as well. 29:55.000 --> 29:57.000 So we immediately go to the result. 29:57.000 --> 30:00.000 So we bounce this z of two of the Moon. 30:00.000 --> 30:02.000 And this is what we get back. 30:02.000 --> 30:05.000 So this is 60 seconds. 30:05.000 --> 30:08.000 So obviously this is between doing a low and stalker. 30:08.000 --> 30:11.000 Because the echo from the Moon comes back in like two, two and a half seconds. 30:12.000 --> 30:16.000 So we transmit for 60 seconds, stalker receives for 60 seconds. 30:16.000 --> 30:20.000 And because of all of the nice synchronization between the SDRs, 30:20.000 --> 30:22.000 together with the good clocks on both sides, 30:22.000 --> 30:24.000 we get a perfect match with what we expect. 30:24.000 --> 30:27.000 So our frequency is at zero, where we expect it, 30:27.000 --> 30:29.000 and then you see the delay going into the Moon. 30:29.000 --> 30:32.000 We can see the full depth of the Moon actually show along up. 30:32.000 --> 30:35.000 Every point here in this graph, every small point, 30:35.000 --> 30:39.000 is a correlation of the transmitted signal with the full received signal 30:39.000 --> 30:41.000 for a certain delay and offset. 30:41.000 --> 30:44.000 And then we look on how strong it is. 30:44.000 --> 30:47.000 I'm not the first one to do this in Dwingelo, even. 30:47.000 --> 30:50.000 This is from Peter Cherich, a long time ago, 30:50.000 --> 30:54.000 who did the same thing. 30:54.000 --> 30:57.000 But he actually managed to do it through the audio signal 30:57.000 --> 30:59.000 of the radio amateur equipment. 30:59.000 --> 31:02.000 So we only had a few kilohertz to do this with unlocked equipment. 31:02.000 --> 31:05.000 And he did, I think, a lot of these things on top of each other. 31:05.000 --> 31:08.000 And it's amazing that he got to hook out a shape out of it. 31:08.000 --> 31:10.000 This is a real achievement. 31:10.000 --> 31:13.000 And in his article, he writes a list of all the things 31:13.000 --> 31:14.000 that could be done better. 31:14.000 --> 31:17.000 And that's exactly the things we have done now. 31:17.000 --> 31:19.000 So thanks for that. 31:19.000 --> 31:21.000 So we have different frequencies. 31:21.000 --> 31:23.000 We normally do this on 23 centimeters, 31:23.000 --> 31:25.000 because we have the most power. 31:25.000 --> 31:27.000 And we have a good Doppler resolution. 31:27.000 --> 31:31.000 We can do it at 70 centimeters, but the wavelength is longer. 31:31.000 --> 31:34.000 So you're like a resolution in your Doppler. 31:34.000 --> 31:37.000 And 30 centimeters, we don't have enough power to properly do this. 31:38.000 --> 31:41.000 You also see some asymmetry here, which is interesting, 31:41.000 --> 31:46.000 because we're not eliminating illuminating the full moon anymore 31:46.000 --> 31:47.000 at this point. 31:47.000 --> 31:50.000 And yeah, we do all this work with 23 centimeters. 31:50.000 --> 31:52.000 Now, there's another interesting one. 31:52.000 --> 31:55.000 We get two echoes back, because stock art has two proliferation. 31:55.000 --> 31:59.000 So they have linear polarization, horizontal vertical. 31:59.000 --> 32:02.000 And from that, we can recreate our LHCP and RHCP. 32:02.000 --> 32:04.000 We transmit RHCP. 32:04.000 --> 32:08.000 So what we receive back from the moon, but we expect is LHCP. 32:08.000 --> 32:10.000 That's the normal echo. 32:10.000 --> 32:12.000 But the moon is not a perfect sphere. 32:12.000 --> 32:16.000 So we also receive, like I said, the wrong echo. 32:16.000 --> 32:19.000 And it's interesting, as that one has more detail. 32:19.000 --> 32:22.000 Because those are basically the features you see on the main. 32:22.000 --> 32:24.000 So this one is stronger. 32:24.000 --> 32:26.000 That one is more interesting. 32:26.000 --> 32:29.000 Let me see how I'm doing on time. 32:29.000 --> 32:32.000 If we look at the power of these, 32:33.000 --> 32:35.000 obviously, the LHCP is much stronger. 32:35.000 --> 32:40.000 The RHCP has less power, but has more features. 32:42.000 --> 32:43.000 This is nice. 32:43.000 --> 32:46.000 All these kinds of things you can compare from articles from the 60s. 32:46.000 --> 32:48.000 They were doing so many of these experiments then. 32:48.000 --> 32:51.000 And it's nice to see that you get the same plots, but I will not go and eat. 32:51.000 --> 32:53.000 Okay, we skip to the results. 32:53.000 --> 32:57.000 So this is basically the screen is a little over-illiminated, 32:57.000 --> 32:59.000 but this is what we achieved. 33:00.000 --> 33:04.000 This is 10 times 30 seconds stacked on top of each other. 33:04.000 --> 33:07.000 And here the timing becomes really critical, 33:07.000 --> 33:13.000 because you're talking minutes that you need to keep everything coherent. 33:13.000 --> 33:15.000 This is the LHCP. 33:15.000 --> 33:19.000 If I switch to the RHCP, you will see it's roughly the same. 33:19.000 --> 33:21.000 Here you see a bit more detail. 33:21.000 --> 33:24.000 I'm not a switch back and forth. 33:24.000 --> 33:28.000 You see one has more power, and the other one has more detail. 33:28.000 --> 33:32.000 But both are actually interesting. 33:32.000 --> 33:34.000 There's one thing though. 33:34.000 --> 33:38.000 This might look like a picture of the moon, but it isn't. 33:38.000 --> 33:44.000 I basically cheat a little bit by giving it the right aspect ratio 33:44.000 --> 33:49.000 to make you think it's the moon, but you have the two hemispheres overlapping here. 33:49.000 --> 33:53.000 So this is the front of the moon, and then you have the two hemispheres. 33:54.000 --> 33:59.000 If we basically recreate the map, so this is based on the map from the moon, 33:59.000 --> 34:04.000 we can see that all these craters are on the right place in our reception. 34:04.000 --> 34:07.000 But if you would look at the moon, you would not see this. 34:07.000 --> 34:11.000 This is actually two halves of the moon, and then fold it on top of each other. 34:11.000 --> 34:18.000 To make it more insightful, this is basically what is happening. 34:18.000 --> 34:22.000 So the next step is we need to have multiple measurements with different orientation, 34:22.000 --> 34:25.000 and then we can start removing those two. 34:25.000 --> 34:27.000 But that's still on the two loveliest. 34:27.000 --> 34:30.000 I have one more interesting one. This is my favorite. 34:30.000 --> 34:32.000 This is work. Yeah. 34:32.000 --> 34:34.000 I can look at this for hours. 34:34.000 --> 34:39.000 Basically, I was showing pictures which were like 30 or 60 seconds overlay. 34:39.000 --> 34:43.000 This is real time. So this is playing at real time on where on the moon, 34:43.000 --> 34:45.000 the echo is coming from. 34:45.000 --> 34:51.000 And you see that you get destructive and constructive interference. 34:51.000 --> 34:56.000 So you see these canals that's basically where the waves are destructive. 34:56.000 --> 35:01.000 And the yellow places are the ones where you get the reflection from. 35:01.000 --> 35:04.000 And you see because the geometry between us and the moon is moving, 35:04.000 --> 35:07.000 you actually see it changing all the time. 35:07.000 --> 35:12.000 I find this amazing to see this in real time. 35:12.000 --> 35:14.000 Twenty-three centimeters. 35:14.000 --> 35:16.000 Yeah. 35:16.000 --> 35:18.000 Okay. 35:19.000 --> 35:22.000 Next one. So I've already shown. So this was the moon. 35:22.000 --> 35:24.000 That's easy because you get a lot of power back. 35:24.000 --> 35:26.000 The next one is Venus. 35:26.000 --> 35:28.000 Now we try to do the same thing. 35:28.000 --> 35:30.000 So Tom and Jan made this one. 35:30.000 --> 35:34.000 We had a picture of it, but he automated the animation. 35:34.000 --> 35:37.000 What you see here in the middle is Earth. 35:37.000 --> 35:40.000 The yellow one is the Sun. The white one is Venus. 35:40.000 --> 35:46.000 And you see that Venus is in between us and the Sun every one year and seven months. 35:46.000 --> 35:48.000 And that's when it's very close to us. 35:48.000 --> 35:52.000 And we need that because the radar equation has the distance to the fourth power, 35:52.000 --> 35:54.000 which means you always lose. 35:54.000 --> 35:56.000 There's just no way you can compete with that. 35:56.000 --> 35:59.000 So we really need it over here. 35:59.000 --> 36:03.000 At the same time, this is also the moment it's between us and the Sun. 36:03.000 --> 36:07.000 So this is pointing at Venus and you see how close to the Sun it is. 36:07.000 --> 36:11.000 It works, but if your sight loops are just in the Sun, 36:11.000 --> 36:14.000 you will actually notice that. 36:14.000 --> 36:18.000 So we wanted to try the same thing as we did with the Moon. 36:18.000 --> 36:20.000 So basically, this is the same setup. 36:20.000 --> 36:25.000 The only thing we changed is that instead of the amplifier we had here, 36:25.000 --> 36:29.000 we put a 1 kilowatt amplifier in the front end. 36:29.000 --> 36:35.000 So right at the feet in the center, in the focus of the, of the Dwingelow dish. 36:35.000 --> 36:38.000 That works because then you get really 1 kilowatt of power. 36:38.000 --> 36:41.000 It has some issues that I will show later. 36:41.000 --> 36:43.000 So here we're installing it. 36:43.000 --> 36:45.000 We have a lift, so you can go up. 36:45.000 --> 36:48.000 You can do work, you can take the whole focus box out of it, 36:48.000 --> 36:51.000 and do the work on the ground, and then put it back in. 36:51.000 --> 36:57.000 Here you actually see the feet going back into the, into the box. 36:57.000 --> 37:00.000 Over here you see the 1 kilowatt amplifier. 37:00.000 --> 37:04.000 And this is going to be put into that hole. 37:04.000 --> 37:08.000 And you can already imagine that there will be a lot of heat coming from this. 37:08.000 --> 37:10.000 That there's not going anywhere. 37:10.000 --> 37:12.000 That's what's indeed the issue. 37:12.000 --> 37:14.000 I'm going to skip. 37:14.000 --> 37:20.000 I have still how many minutes do I have? 37:20.000 --> 37:22.000 Okay. 37:22.000 --> 37:25.000 Then I will just. 37:25.000 --> 37:32.000 So we're a little bit looking with Venus in the sense that the huge Doppler we get from the moon is much smaller on Venus. 37:32.000 --> 37:34.000 Also because of the retro gate, 37:34.000 --> 37:36.000 we have a rotation of Venus. 37:36.000 --> 37:38.000 So while it's coming towards us, 37:38.000 --> 37:45.000 it's rotating in the other direction, which cancels the Doppler spread a bit. 37:45.000 --> 37:51.000 So most of the energy we get back from Venus is actually in a very small bandwidth around one herd. 37:51.000 --> 37:55.000 So even half of the power are so we get back in one herd, 37:55.000 --> 37:58.000 which means that we have a chance of doing this. 37:58.000 --> 38:01.000 So we had a whole list of things we're going to send to Venus, 38:01.000 --> 38:07.000 and we start with 4 CWUs, and then we had a whole set of interesting set of two sequences, etc. 38:07.000 --> 38:11.000 And after the 4 CWUs, the set of field. 38:11.000 --> 38:14.000 And we only had one window of two hours to do this, 38:14.000 --> 38:18.000 because we had a permission to send with one kilomatt kilowatt in a bandwidth, 38:18.000 --> 38:20.000 which is very close to the Galileo, etc. 38:20.000 --> 38:26.000 So we, yeah, it filled and we couldn't redo it. 38:27.000 --> 38:29.000 Skipping to the results. 38:29.000 --> 38:31.000 So we did see an echo, luckily. 38:31.000 --> 38:34.000 So we have four transmissions, four receptions, 38:34.000 --> 38:37.000 because we were both receiving ourselves, five minutes later, 38:37.000 --> 38:39.000 and Stockat was receiving. 38:39.000 --> 38:41.000 Stockat reception was much better. 38:41.000 --> 38:46.000 Maybe also because we were heating up a little bit at that moment. 38:46.000 --> 38:50.000 But still, we see the CW coming back, 38:50.000 --> 38:52.000 and if we summarise all of them, 38:53.000 --> 38:57.000 we actually get a very nice as an R on the return. 38:57.000 --> 38:59.000 And again, I'm going to refer to Danny's blog, 38:59.000 --> 39:01.000 because he read it the whole analysis, 39:01.000 --> 39:03.000 with a better match filter. 39:03.000 --> 39:08.000 So if he got even higher as an R's out of it, then we did. 39:08.000 --> 39:12.000 And I will skip this one as well. 39:12.000 --> 39:16.000 Oh, yeah, this was the result. 39:16.000 --> 39:19.000 So the transistors, these sold at themselves, 39:19.000 --> 39:23.000 they simply, some components, 39:23.000 --> 39:25.000 even some of the capacitors, 39:25.000 --> 39:28.000 even to simply fell off the whole thing. 39:28.000 --> 39:30.000 But it's all repaired now. 39:30.000 --> 39:32.000 Okay, last topic, very quickly, 39:32.000 --> 39:35.000 simply because it's also fun and related to timing. 39:35.000 --> 39:40.000 Very long baseline in the Fermatry has been done since 1976, 39:40.000 --> 39:42.000 in the vanilla telescope. 39:42.000 --> 39:43.000 That's the nice thing. 39:43.000 --> 39:45.000 Everything you do there has been done before. 39:45.000 --> 39:47.000 And there's always history to it. 39:47.000 --> 39:50.000 For the white rabbit people, this is an old rebedium clock, 39:50.000 --> 39:51.000 with a battery. 39:51.000 --> 39:54.000 So instead of transferring the time over a white rabbit, 39:54.000 --> 39:57.000 you basically carry your clock with your battery. 39:57.000 --> 39:59.000 That's the alternative. 39:59.000 --> 40:04.000 Paul has done this and presented here in 2018, 40:04.000 --> 40:06.000 doing interferometry. 40:06.000 --> 40:10.000 So getting fringes between the vanilla telescope together with Joe Robank, 40:10.000 --> 40:13.000 and Resterbork, but it can be a radio of flow graph 40:13.000 --> 40:15.000 that created videos if I'm not correct. 40:15.000 --> 40:18.000 And then using the GIVE correlator. 40:18.000 --> 40:22.000 This is professional stuff, of course, with the other telescopes. 40:22.000 --> 40:24.000 It's ice-hawry. 40:24.000 --> 40:26.000 For you, it was hobby. 40:26.000 --> 40:30.000 I also got a note from the person actually in the photo 40:30.000 --> 40:32.000 on this experiment. 40:32.000 --> 40:35.000 The first fringes in the vanilla was on 3C73. 40:35.000 --> 40:37.000 That's the first one they have ever tried this on. 40:37.000 --> 40:40.000 So now we have, again, these tools. 40:40.000 --> 40:43.000 So I'm doing everything with the tools that I have described before. 40:43.000 --> 40:45.000 And we do this with stockers. 40:45.000 --> 40:47.000 Now it's basically the same as the EME setup. 40:47.000 --> 40:51.000 The only difference is we're not transmitting and correlating 40:51.000 --> 40:53.000 the reception with the transmission. 40:53.000 --> 40:56.000 Now we're simply correlating a quasar. 40:56.000 --> 40:59.000 So we're basically a noise source in this guy. 40:59.000 --> 41:00.000 But we need to tell the scope. 41:00.000 --> 41:02.000 So again, we get two files. 41:02.000 --> 41:07.000 Everything is synchronized simply by design. 41:07.000 --> 41:10.000 But we can also do this life. 41:10.000 --> 41:12.000 Because this is streaming data. 41:12.000 --> 41:15.000 We can stream data from one telescope to the other, 41:15.000 --> 41:17.000 run this VAT correlate. 41:17.000 --> 41:20.000 And at the moment we point, you live. 41:20.000 --> 41:24.000 See, actually, the correlation between the two streams. 41:24.000 --> 41:28.000 There's a script that actually calculates the whole geometry 41:28.000 --> 41:30.000 and updates for that. 41:30.000 --> 41:35.000 So with that, we can actually detect very far sources. 41:35.000 --> 41:39.000 So this is a quasar, basically the nucleus of a black hole. 41:39.000 --> 41:41.000 Let's say a black hole very far away. 41:41.000 --> 41:45.000 This is a quasar with a red shift of five-foot-four, 41:45.000 --> 41:49.000 which means we're looking back in time here, 12 billion years. 41:49.000 --> 41:54.000 So this radio wave has been traveling for more than 90% of the lifetime of the universe 41:54.000 --> 41:57.000 before we capture it with the 25 meters of this. 41:57.000 --> 41:58.000 And we do it with two dishes. 41:58.000 --> 42:01.000 And we look at the signals we put them on top of each other 42:01.000 --> 42:02.000 and they actually correlating. 42:02.000 --> 42:03.000 So they are the same. 42:03.000 --> 42:05.000 This is extremely far. 42:05.000 --> 42:10.000 It's mind-boggling how far you get with some such dishes. 42:10.000 --> 42:13.000 Another one, this was a discovery for me, 42:13.000 --> 42:16.000 is that, of course, we can do it with two 25 meter dishes. 42:16.000 --> 42:18.000 That's not hard. 42:18.000 --> 42:22.000 So I put my own one-and-a-half meter dish in the equation again. 42:22.000 --> 42:25.000 I can easily do this between a low and my one-and-a-half meter dish 42:25.000 --> 42:28.000 or between stockers and my one-and-a-half meter dish. 42:28.000 --> 42:33.000 Because as an hour of the correlation depends on the geometric mean of the dishes. 42:33.000 --> 42:37.000 So if you have one really big dish, the old one can be really small. 42:37.000 --> 42:42.000 So you still have basically a big dish if you calculate the mean of those. 42:42.000 --> 42:47.000 So even with a small dish, you can do interferometry or actually time transfer 42:47.000 --> 42:48.000 of all the things. 42:48.000 --> 42:53.000 I think this opens a lot of possibilities with relatively simple hardware. 42:53.000 --> 42:55.000 I must admit that I have a good clock. 42:55.000 --> 42:58.000 That's the other part of the equation. 42:58.000 --> 43:01.000 Another one over very long distance, so we can also do this 43:01.000 --> 43:04.000 between being a low and the alan telescope array. 43:04.000 --> 43:05.000 Well, but the single dish. 43:05.000 --> 43:07.000 That's a huge baseline. 43:07.000 --> 43:09.000 That's 78,000 kilometers. 43:09.000 --> 43:11.000 And it correlates fine. 43:11.000 --> 43:14.000 The phase is not completely stable, but this is five minutes. 43:14.000 --> 43:17.000 So there's still not bad. 43:17.000 --> 43:19.000 Okay, we're almost there. 43:19.000 --> 43:22.000 Last one, and this is actually the bridge to the next talk. 43:22.000 --> 43:25.000 We're doing this all on the full data. 43:25.000 --> 43:28.000 So we have 12 bits of data and we correlate that. 43:28.000 --> 43:31.000 But we can also do this on one bit data. 43:31.000 --> 43:34.000 So we just off the 12 bits IQ data. 43:34.000 --> 43:36.000 We only keep the sign basically. 43:36.000 --> 43:39.000 Especially if I do this between the two telescopes, 43:39.000 --> 43:41.000 so the Dwingalo and Marwanda and half meter this, 43:41.000 --> 43:45.000 I still have something like 76% of the as an R. 43:45.000 --> 43:46.000 Which is quite good. 43:46.000 --> 43:49.000 Interestingly, that's a different number than Jean-Michel is going to show 43:49.000 --> 43:51.000 in this presentation. 43:51.000 --> 43:56.000 And we'll leave it at that. 43:56.000 --> 44:00.000 So this is from a project to do with your small home. 44:00.000 --> 44:05.000 And we can definitely extend this. 44:05.000 --> 44:06.000 Okay, there was very quick. 44:06.000 --> 44:07.000 There was a lot to cover. 44:07.000 --> 44:09.000 And there was a lot I wanted to show you. 44:09.000 --> 44:11.000 Big thanks to everybody who helped Tomoyon, 44:11.000 --> 44:14.000 and case who contributed to the VRT code. 44:14.000 --> 44:18.000 Paul, for all the time and frequency stuff that is key for all of this. 44:18.000 --> 44:19.000 Otherwise, we couldn't do it. 44:19.000 --> 44:22.000 Wolfgang from Stockard's Peter from Abovecom. 44:22.000 --> 44:25.000 And of course, cameras is a big team of like 60 people 44:25.000 --> 44:27.000 who all keep the telescope running. 44:27.000 --> 44:30.000 There's no way you can do that on your own. 44:30.000 --> 44:33.000 And the other thing is, if anybody is interested in cameras, 44:33.000 --> 44:36.000 talk to me, talk to Paul, we're here, 44:36.000 --> 44:39.000 and it's a volunteer organization, so you can join us. 44:39.000 --> 44:41.000 Okay, that was it for now. 44:41.000 --> 44:51.000 Everything I've shown is online, data is online. 44:51.000 --> 44:54.000 Everything is available. 44:54.000 --> 44:59.000 Yes. 44:59.000 --> 45:07.000 We have looked at that, but you wouldn't need... 45:07.000 --> 45:10.000 It's a feedback loop, it's too slow. 45:10.000 --> 45:14.000 We didn't, we've tried it, we've not managed to do anything. 45:14.000 --> 45:15.000 But it's a good question. 45:15.000 --> 45:16.000 Yes. 45:28.000 --> 45:29.000 Yes. 45:29.000 --> 45:33.000 Actually, it built a such a way that you can merge any control data into that. 45:33.000 --> 45:35.000 But you have to define either your own format, 45:35.000 --> 45:38.000 or use one of our own formats, but it's completely generic. 45:38.000 --> 45:43.000 So I write this clip basically for every site that injects the control data into the, 45:43.000 --> 45:46.000 which is separated from the tools.