TSP #190 - Siglent SSG5060X-V 9kHz - 6GHz RF Vector Signal Generator Review, Teardown & Experiments

[Music] hi welcome to the sigma pad in this episode i have another product review for you guys we'll be taking a look at this siglin ssg 5000 series vector signal generator which is a six gigahertz synthesizer plus a built-in arbitrary waveform generator on inq channels allowing you to perform complex modulations and a whole other types of signal manipulation on the fly it's really exciting to see sigmund push so hard into the rf domain it's a difficult market to get into because you need to have a very strong reputation before people can trust the performance and the functionality you say you have in your instruments now i've spent a lot of time with this unit and i have a long review for you guys of course and you can see in the video description different parts of the video you can jump directly onto so there's a lot to talk about let's get started and here's the front panel of our siglin vector signal generator this one is up to six gigahertz and it has all the bells and whistles so there is a lot to talk about now keep in mind that this because it has an iq generator built in there are a lot of advanced functions as well therefore the design of the gui and the choices of the buttons and how easy it is to interact with the instrument without using any computer is quite important so we will take a look at that i think they've done a reasonably good job at that we have a fairly big touch screen our cd screen here there are no contextual buttons at the bottom anymore they're expected to use the touch screen i think it works reasonably well and there's buttons here to bring you back to the main menu which is kind of like the center of degree where you can jump to different functions various important things are also brought out into the buttons on the side which i appreciate and then we have a clickable rotation knob for selection and a keypad and everything i think they're very good these four corners also have raised edges which is a big welcome in these type of instruments not only does it protect the unit it can also be put face down on the floor without damaging any of the connectors or knobs or anything and we have our main rf output that's a type n it's quite common at these frequencies it is 50 volt tolerant which is good because if you're going into let's say a high power amplifier like a gallium nitride amplifier with a strong input bias you should be okay in that case and we also have the low frequency output here coming directly to a bnc and one usb port in the front we also have the rf output and the modulation output both with a line underneath which actually lights up which is also very good so you can see right away if your rf is enabled and if your modulation is enabled you don't have to search for it on the screen and the main power button of course is a soft switch not a hard switch yeah i think it looks really nice and we're going to take a look at the gui again that's what really matters but they've done a beautiful clean industrial design here and here's the back of the instrument and there is a rich set of analog connectivity to the back which i do want to talk about because it's quite good but first we have ethernet usb host and device there is no gpib option natively in the back of the instrument there is an ocx option which this one does have of course and power going in there's no hard power switch which i wish there was and they could have fitted somewhere around here so you can turn it off completely and basically disconnect it from the grid if you need to but look at all the bncs and finally we have an instrument with differential i and q output natively from the unit in this class which is fantastic so you can see these four connections over here give you a true differential output directly tied into the arbitrary waveform generator we do have single landed i and q inputs as well so you really can use this in a lot of flexible ways you also have a whole bunch of iq events and pattern triggers which you can set up as a arbitrary wave from generator you need to sometimes synchronize the modulation outside which all of these allow you to do that there's also pulses in and outs and of course various things like external modulation input for the analog basic analog modulation and the 10 megahertz references and trigger so there's a lot of connections and they're all really important i'm quite happy to see that this makes this instrument really flexible especially for the iq interface to basically getting a two channel arbitrary waveform generator with differential outputs built into this unit which is fantastic so now that we're kind of familiar with how it looks on the outside let's take it apart and do some analysis and understand how it works on the inside so let's take a look inside this instrument and see how they've broken the functionality of the instrument into different sub blocks so we have the power supply here on the left in its own cage to power directly from the back nicely insulated against em pollution of everything else the power going over there into a main pcb at the bottom and that divides it again into this board over here this board is very clearly the iq generation so the modulation functions are all on this board all the iq external inputs and outputs are at the back and we can see one two three rf cables that communicate from this board to the other one that kind of gives you already a hint of what the overall architecture of this is with respect to the main board which we will take a look at we have an application processor a board here which can separately be plugged in and now this is something we've seen signaling do in some of their other units and that allows them to create basically the same computer and use it over and over again it's a common technique and the board on the other side we'll take a look as well we have two fans over here which are partially blocked at the bottom by the other board but you know it's a reasonable thermal solution air coming over here flowing over the power supply as well and i think it's good heatsink orientation is all correct that's also pretty good let's flip it over to the other side and here's the other side and you can see one more cable coming from an external ocxo board that's over here so this is an option that you can add to the instrument and if you don't have it this cable is simply not there and the instrument will use the internal oscillator whatever that might be and over here we have a battery some buzzer a couple of memory a lot of dc-dc converters we also have the display connector as well as the touchscreen connector so interestingly even though the application processor is on a module they still have to route all of the other peripheral controls through the main pcb that has the rf functions this is okay because everything is well isolated but that application processor doesn't have directly display connectors and everything and that kind of makes sense because that's not a very big pcb allows you to put it into any environment in a much more flexible way and then you just route whatever signal you want to the main pcb so now the fun part we're going to take a look at these boards in detail and see how they work all right let's start from the main pc i've already taken a whole bunch of screws off from the other side and there we are we can take a close look at this now so now let's go ahead and take a look at the rf portion of the main pcb now there's a lot of components on here obviously but we can break down the functionality of the synthesizer into sub fundamental components so we know we need a vco and a pll and some kind of a synthesizer then we know we need a whole bunch of different filtering sections so that you can get a nice clean cw tone and potentially frequency multiplication and mixers in order to reach all the frequencies that are beyond the vco frequency we're also going to need amplifiers and attenuators and so on in order to control the output power of the circuit and then there's a modulation section that's separate so let's figure this out working backwards so the output here we have the main sma output now we saw that this main sma output is actually connected to the front panel rf output but there is another one that's not connected and there is really no difference between them there's a solid state switch between the two of them so you can switch this into any of these ports probably this is maybe a rear panel connector so that you can switch the rf from the front to the rear of the instrument or maybe some for some other testing purposes but either way in my unit this is not populated now working backwards we're going to need output amplitude control as well as attenuation so we have a switch over here which allows you to bypass this component this is a gallium arsenide analog devices power amplifier with up to 28 dbm p1 db which is quite high so i'm curious to see the maximum output power and linearity of this instrument in particular when it's doing modulation because we have this component here so you can bypass it by going this way or you can going right through it then after that going backwards we basically have a solid state step attenuator normally in very high performance synthesizers we see a mechanical attenuator where there are electromagnets switching mechanically individual attenuations in and out of the circuit on the signal path in this particular case we have switches all over the place that switch trf into various attenuation determined by this surface mount resistive dividers that are on the board or you can just bypass it completely so this entire section is nothing more than attenuation as well as output power amplifier then working backwards we have another section which is another bunch of amplifiers here and then we get a couple of switches that allow you to switch in some other paths into the main rf output these are probably some of the lower frequency paths that are generated over here eventually being fed to the output because everything at the end needs to reach to the output sma across the entire frequency range of the synthesizer if i work back a little bit more there is an analog control antenna right here then we see some more amplification and this looks like an array of pin diode attenuator as well so all the output power control is done partially in this section then we have a switch here here and all over these components so this piece is a little bit interesting because the signal coming over here by using this switch can be either switched to the right or to the left if it's to the left it goes to this connector if it's to the right it can be completely bypassed but if i switch to the left i have the option of going out into another board and come back and then switch back onto this path well this is obvious by now that this is the external modulation board that can be plugged into these two connectors now this doesn't cover all the frequencies it doesn't look like it which means some of the modulation at lower frequencies would have to be done some other way at least that's what i think at first glance then working backwards over here since this is our common port at this point we can see that we have a couple of other switches allowing you to switch different filters into the path we've seen that before too these filters are able to get rid of unwanted harmonics that may be generated through mixers and vco harmonic generation as well as multipliers so there are different filter sections over here and of course no surprise we see here a frequency doubler chip and this frequency doubler is probably what gets us all the way to six gigahertz because the main vco does not work all the way to six gigahertz so therefore we have a doubler and after the doubler you have the filtering to get rid of the fundamental leakage as well as some of the other harmonics that you don't want some more filtering and switching and so on and on and then we reach eventually after a bunch more switches into this portion if you look at some of the other paths as well some of the other low frequency frequency generation portions seem to be over here there's a mixer over here as well that's probably what this section is going on let's work all the way back to the edge of the vco so the output of the vco this is zcom this is two to four gigahertz the output is fed into this which then goes into the rest of the section there's also a coupler here this coupler steals some power from the vco and allows that to be amplified and fed into this chip which is a fractional end synthesizer from analog devices and that then can close the loop around the vco with using this 10 megahertz tcxo as its reference so this 10 megahertz therefore determines the fundamental actual frequency precision and the frequency accuracy of this entire synthesizer that's what's running this synthesizer ultimately at the end and this guy is going to lock to that so that's our main pll loop and if you inject an ocxo signal here it will override the tcxo and then this will become the actual frequency accuracy which is the ocxl in my case and that's about that let me see what else did i want to say so there's a couple of other components that are important here we have a tx tag here this is an analog devices 1.2 gigahertz per second 16-bit dac which is an interesting choice because this needs a a complete fpga to run it actually and all i see is a cpld so they're probably running this with some basic waveforms only but this can cover up to about 450 megahertz or so so you can generate a lot of the low frequencies directly from this normally we see dds's direct digital synthesis blocks to do this it's interesting i wonder if some of the modulation at lower frequency is done through this stack instead that makes some simplification on how you would use the external modulation board as we will take a look at in a second then we have another chip over here this is a ti part originally from a national semiconductor i believe and this is a dual pll multi-loop synthesizer clock generator this is a nice component to use here because it allows you to create the clocks for a lot of different things on this board so for example they can generate the clock for the stack it can also generate the iq clock which is labeled right over here for the synthesizer board and the all the other clocks required to synchronize different peripherals on this board by using this ultimately being fed from the synthesizer over here everything else on this board will be coherently locked to each other and the iq clock of course needs to be coherently locked because it's going to modulate the yellow carrier over here so that's why they're using this and in fact they're using one of its alloy outputs through a saw filter into this mixer and this makes there is a dc to 1200 megahertz if from mini circuits and that probably gives us some of the lower frequency capabilities too so it's quite interesting this portion especially with this stack over here and other than that there's really not much going on on this board of course the cpld it controls everything now there's also the low frequency output right over here that could also be coming from the dac and this connection is on the other side there's some relays which you can switch it in and out and this is for very very low frequencies lower than for example the dc blocks of the main rf allows you to have and that all of that then ties together some you know voltage regulator dc-dc conversion over here on the right side and all the different uh reference overrides that you can externally apply yeah it looks quite nice on the other side there's not much going on a few couple couple of other filtering sections that are fed over here are also present most of it is just basically dc and biasing and here's the main computer board that is plugged into here okay so let's take a look at the modulation board here so what do we expect to find on this board so obviously you're going to need a vector modulator mixer here otherwise you won't be able to create complex modulation so that needs to be there then there needs to also be two dacs because this is a complex iq modulation capable and therefore you're going to need two independent dax to be able to create any modulation you want and then you will of course need some digital circuitry to support all that so let's look for those so here are the two connectors that interface with the other board and this component and this component here are nothing more than digital attenuators so there's essentially power control but in the middle we have an iq vector modulated mixer from analog devices where the signal i and the signal q differentially come directly into this mixer so which means that the signal that goes into it can benefit from this mixing and then eventually you get complex signal on this other side by by playing this antenna you essentially get output power control now this is difficult to do because if you do this modulation in the wrong place in the chain all the non-linearities will add on top of it so it's a difficult thing to maximize the dynamic range of a synthesizer that has iq modulation built into it then after that we have a whole bunch of switches again with different filters in between this allows you to get rid of any potential images of the dax as well as any other harmonics that you don't want so you filter them depending on the frequency range you're operating all of them split and then eventually recombine back into the same connector on the other side so the signal comes in here as a single cw tone gets all the mixers comes out of here with a complex modulation on top of it it's quite nice and compact then what you need is you need some kind of buffering circuit to handle the iq analog signals and that's exactly what these three components are these are do adc drivers these are very high dynamic range very linear amplifiers that interface with this mixer on one side and get their signals from dax on the other side so why do you need three of them well that's because you also need to route iq signals to the outside because you can pass the iq modulation to the connectors in the rear you can also accept iq modulations from the outside and these three handle those tasks there are some analog muxes here that allow you to switch and route the iq analog signals into whichever of these directions you want so this one is the one that is interfaced with the mixer ultimately but using these analog muxes you can switch them in and out from the other ones so yeah nice and simple then on the left side we have two dacs these are identical 1.2 giga samples 16 bits from analog devices they're the same ones that i same one as the one that i found on the other board which is interesting and they're coupled with this xilinx fpga which provides all the data and there's some memory over here so that you can read the memory and then stream it into dax this has to happen in real time because these dacs are supposed to generate any arbitrary waveform that you want you can program anything you want into them so this is basically this entire section is essentially a two channel arbitrary waveform generator that's what the equivalent of it is and the rest of it is really nothing else there's some bcdc converter module over here some voltage regulators and all the interfaces required to get the iq signal in and out of this port those are the iq signal connectors over there and if i look on this side yeah there's not really much on this it's just mostly some analog support circuitry yeah nothing more so it's quite nice and that's a nice and elegant way of adding iq modulation to this and this board by itself because it has differential iq outputs in the back which is fantastic yeah you get yourself a a dual channel differential iq generator which is quite nice so now that we have a very good idea of how everything works together it's time to do some fun experiments let's take a very brief look at the gui which i think they've done a fairly good job allowing you to control the complex capabilities of the instrument from this grid of the home screen so this allows you to jump into any of the sub functions of these that you see such as rf and iq modulation analog modulation by simply tapping on them and you get a sub menu that allows you to control them i don't think this sub menu is really necessary because as soon as you go over here you get those functions at the bottom of the screen anyway so may as well just save yourself a click and directly take you there and the buttons on the side of the instrument directly take you to those menus as well so here in the rf i can go into frequency which is the same here or i can go into level and sweep and so on so really yeah it's quite nice and the menu is fast and accessible and the touchscreen works really well a couple of additional changes i think will improve it for example these things do not need to be so tiny you should make a big button that says on or off and as soon as you touch it you know the text changes and the color changes so you can see and very easily access those functions that can be changed of course there's some misalignments over here for instance at the very top when you enable the rf this just changes to blue it's really hard to see it should really become yellow just like the the selection or menu over here so degrees obviously overall quite good some minor changes definitely can be made to it but i really want to do some measurements with it now when you're dealing with instruments like keysight and rodent shores and tektronix they're very well established especially in the rf domain so you don't you know you don't have to double check to see if they made through specification but i'm often asked to measure some of the high performance capabilities of these and to make sure that they really do match the data sheet so let's measure the phase noise and a couple of other parameters before we do some custom experiments so let's quickly verify the phase noise performance of the signal synthesizer directly on the azure e5052b which allows us to do correlation measurements as well if you really need to and we can compare it with the datasheet of this instrument and see if it meets that especially across different los where the multiplication internally is kicked in so between one gigahertz for example to six gigahertz there's a different multiplication factor so let's give that a try and here's the measured phase noise of the carrier of one gigahertz and zero dbm apple power and i'm measuring the phase noise from 10 hertz offset all the way to 100 megahertz offset with 10x correlation as well as 16 times averaging so the correlation and averaging is still running the average is 8 out of 16 because we're starting from a 10 hertz offset the cross correlation takes a long time to complete now if you look at the shape of this and the values of the phase noise it's essentially one to one with what is in the data sheet which is fantastic to see i'm going to put up what's in the data sheet on the side so you can compare it and here for example at a one kilohertz offset we're at minus 106.5 dbc per hertz at 10 kilohertz offset minus 118 dbc per hertz there's a few other things to note for example we can see the noise floor of the phase noise at around 100 megahertz offset we're at about -135 dbc per hertz for this particular correlation that i have enabled and i don't think it's going to go any lower than this i believe this actually is the noise floor of the instrument and then over here if you look at the low frequency offsets 10 hertz 100 hertz one kilohertz there are absolutely no spikes so it's very clean at these low offsets and that's tough to do because in the lab there are so many dc-dc switching power supplies from all the leds there is a lot of em pollution present here and we see none of that coupled into the core of the synthesizer so excellent filtering and excellent shielding is definitely present in the instrument i've seen a lot of synthesizers that have a good phase noise but occasionally there's a big spike somewhere around here that's very hard to get rid of so i'm very happy to see this even this little tiny peaking in the phase noise is actually present in the datasheet too so you can read the rest of the numbers here with the markers the carrier frequency you can see it's reported that's almost almost one gigahertz keep in mind that the synthesizer and the azure and phase noise analyzer are not locked really to each other they're completely running separately and the power is minus half a dbm there's some loss in the cables so that's also correct i'm going to measure this also at six gigahertz just so that you can have a good picture and compare with the data sheet but i'm very happy to see that the measurement is very accurate and here's the measured phase noise at 5.5 gigahertz you can see at the very top and it looks exactly like again the data sheet and we do have a better noise floor at about minus 140 dbc per hertz but of course the phase noise in the middle of the band is a bit worse because we're running at a 2x multiplication internally and we again have a very good low frequency behavior as well now there's something for you to keep in mind here this phase noise is obviously not nearly as good as something like the keysight mxg but you have to compare this phase noise against the cost profile of the instrument mxg is quite expensive and this instrument with this iq modulation the equivalent keysight instrument is much more expensive but also at a much higher performance when it comes to phase noise what matters here is also that the data sheet is a good representation of what the instrument can do so you can evaluate to see whether it fits your application and if it's the specification you need for your work one of the features of the signal ssg is that it supports external power sensors there are a few manufacturers like keysight and rodent shorts and if you have a usb based power sensor you can directly connect it to the instrument here i have a u-2000a from keysight and when i connect it to the unit i basically turn this into a power meter at the same time it will basically show whatever power this unit is measuring directly on the screen for you so you can see how already that would be useful you can measure different powers let's say coming from an amplifier or so and this will increase the usability of this unit but you can take that a step further you can do a lot more with that because the instrument can directly and continuously access the power from this power sensor it can use that as an external leveling circuit meaning that it will measure the power and the input of the power sensor and automatically adjust the output power here in order to meet whatever desired power you want here continuously running in a background loop so i have set up an experiment to demonstrate and see how you can compensate for losses of an external system before it reaches the power sensor and that can be quite useful so here i have a power spirit this is a resistive power splitter and therefore it has some internal loss in 3 db to be exhaust which means that if i put 0 dbm here i'm going to get -6 here and -6 here because this is resistive is broadband but it's also lossy so if i turn on the external leveling it would mean that if i enter 0 dbm into the instrument i will actually get zero dbm here the instrument would automatically increase the power at this point until 0 dbm which is achieved here here on the left side i have the azure n1923a and that's directly connected to a power meter up here so we can simultaneously see if that leveling function is actually working by enabling it on the unit so let's take a look at how we do that so as soon as i connect the power sensor via usb this little section over here lights up shows that there is a power sensor indeed connected and you can see the power sensor model right over there and if i turn on the output i go back you can see the output here right now it's measuring nothing because the output is not enabled you can see it's turned off so let's go back over this menu now we can also turn on statistics you can zero these are all the functions you would need to have when you're dealing with the power sensor but this level control is quite interesting let's go under the settings of the level control here so i can turn the level control on and it will search to hit a particular target level so target level here is at zero which is the same thing as entered here but it also allows you to set what is the maximum output power the ssg would attempt to do this is essentially a protection so that in case there's something wrong in a setup the ssg wouldn't ramp out its output to maximum power and potentially damage something you're using so let's turn the level limit here and make it you know make it 10 dbm so make sure it doesn't go above that and let's change the catch range to 10 dbm also say we'll search within 10 db of whatever level that you are set here okay so before i turn this on let's go home over here and let's just enable the output at zero dbm if i turn it on what do i read i read minus six and a half which is the total loss of the output there's a little bit in the connectors and the cable and the power sensor has an additional a 6 db of loss so that gives you -6 at the input and if i look at the other power sensor it's returning exactly the same value it's also saying about minus six and a half but now let's go back here and let's go under here make sure this is set up correctly and let's turn that on okay now look what's going on now it's setting it to 6.6 because it's trying to reach a target of zero go back over here look at that it's reading exactly zero now if i go ahead and look at the other power sensor let's do that right now and let's see what we read look at that zero that's exactly what we expect so the instrument has automatically adjusted and this is dynamic by the way so no matter what i do it's always going to be that adjustment so for example i'm going to change the desired output instead of zero i'm going to change it to something else let's say instead of 0 i want you know 3 dbm as soon as i enter 3 dbm we get 3 dbm if i go back to the instrument you can see that the instrument has now adjusted it to 9.6 dbm so it's going to do that automatically for me this is pretty useful in a complex setup but you don't know what the losses are and you want that to be compensated for automatically but there is something else you can do which is even more useful than this consider now this slightly different setup what we have here is this black box this is actually an amplifier powered from a usb power source which i'm using the front of the instrument for and the output of the instrument is directly connected to the input of the amplifier and the output of the amplifier goes into our keysight u2008 power sensor which itself is then connected back on the instrument let's say you want the gain of this amplifier as a function of frequency we can measure it you know one point at a time across frequency and record the values or you could use the automated function built into the instrument which is essentially measuring the response of this this can have gain or it can have loss and it can give it to you as a function of frequency so it's another way of measuring the gain of whatever you want between the output of the unit to the input of the power sensor characterizing this over frequency let me show you how to do that so let's give this a try we're going to go under the power sensor and we're going to go under level and you can see that there is an option here called flatness this flatness is essentially the compensation the instrument has to do to account for whatever is between the sensor and the unit itself in this case the amplifier so let's go over here into the settings and there is a list of table which he needs to calculate this of course is empty because i haven't done anything yet but we can go over here and we can decide how to fill this table so right now it's in manual step you can also fill it with a flatness list that you define yourself or a sweep list that you can define directly under the instrument if that's what you want the unit to be doing but flatness in this case using the manual step makes sense because we want to just manually calculate the response there so i'm going from 50 meg to 6 gigahertz every 50 megahertz steps that's 120 points so 120 points is a lot if you want to do this manually of course and then we go all the way up here and we can go fill this flatness with sensor data now it's going step by step applying -10 dbm and measuring what is the power of the sensor and it's going to calculate how much it is to back off from that in order to achieve the minus 10 dbm at the sensor it's important to define this because for an amplifier if you have put too high of a value you're going to saturate it and get the wrong result so let's go ahead and see what we get there it is look at that so at 50 megahertz it has to back off minus 20 db which means that the gain of that amplifier is 20 db which is exactly correct that's indeed what it should be and the 3db bandwidth of this amplifier is three gigahertz let's verify to see if that's correct so right now this is 300 megahertz we can keep going all the way one gigahertz one and a half gigahertz keep going until we hit three gigahertz and here's three gigahertz and look at that minus 17.29 db that's about 3 db down from -20 that's exactly correct this is indeed what it should be we can continue on see what happens at the end of the list all the way at six figures there you go six gigahertz minus 14.8 so it's dropping by about 5.2 db or so at that point which is also correct so there you go you just measured the scalar response of the amplifier it could have also been a loss if you had loss indeed so it's very useful to have this at the same time because you can define the level you want this to be done at you can calculate even the compression of the amplifier you can find out what is the saturated output of the amplifier if you increase the level to a higher value and do a sweep then in this case you'll get a different result of course so pretty handy having this sensor connection having the function built into the gui makes a lot of these measurements easier okay so this is what i have to deal with in the middle of my review come on get out of here pooch so now let's take a look at the analog and pulse generation capability of the signal ssg i have the output of the instrument again connected to a power splitter this allows us to look at the raw output of the instrument on one channel of the oscilloscope and on the other side of the power speeder we're going into our amplifier which we just measured the gain of and the output of the amplifier then also goes into a different channel of the oscilloscope we're going to use the tektronix 6 series b which is a fantastic instrument for taking a look at rf versus time aspects of a particular signal so let's take a closer look and see what the instrument can do in terms of pulse generation so setting up a pulse on this is actually quite straightforward under the analog modulation tab you can go directly to pulse and here you have a couple of options you can create a single pulse a double pass or a train of pulses right now i have it on a double pulse which is essentially one pulse followed by another one where you can adjust the the width and the duration of those pulses independently so for example on a double pulse if you look here i have a pause period of 1.5 microseconds the first pause is as narrow as it can be which is 40 nanoseconds on this instrument the second pause is 300 nanoseconds so it's quite a bit longer and we can see what happens to the amplifier when it's exposed to very narrow pulses and somewhat wider pulses and the trigger is enabled as well this is very useful so the trigger of the pulse comes at the back of the instrument in case you need to do synchronization for example and you can even set a particular type of triggering on the the pulse train that's coming out right now it's on auto so it's repetitive it's just going to continue doing that if you want to have for example just a single pulse you can put it on a key and then press simply the trigger key and it will generate one pulse train for you now if you go under the actual train of the pulses it's even more interesting because here you can then define a train of pauses as you want so in this case right now you can see there's only one pause defined and you can even view what that would look like the graph of that and you can view for example how many pauses you have and what they would look like this is really handy if you have a complex set of pulses that you want to send out you may have a radar application or you may have some kind of a frequency encoding into the pulses so lots and lots of opportunities there to play around with it for our purposes we're just going to use the double pulse which is good enough for what i want to demonstrate but just keep in mind that indeed it can be very advanced and have you can load and save pulse trains as well okay let's go back to the double paws so let's keep these numbers in mind 1.5 microsecond and the duration of this and we can look at the amplifier i'm curious what it would happen when we saturate the amplifier with a very narrow pulse can it support that pulse and what happens to the output as a function of time and let's also not forget that of course am fm and pm modulation are all fully supported as well these fall all under the analog modulation capability of this instrument we haven't even gotten to the iq modulation which we will take a look at next all right so here we have the screen of the tektronix 6 series b and you can see that the first channel channel one at the top and channel 5 at the bottom channel 1 is the input of the amplifier and 5 the output of the amplifier right now there's no signal so we don't see anything i'm going to turn on just the cw tone first so here's the cw tone there you go so that's two sinusoids sinusoid going in and coming out nothing exciting so far we need to enable the actual pulse generation to see the effect we're looking for and here are the pulses take a look and it looks like it's doing exactly what we expected to do let's zoom in a little bit so here's our longer pulse it's about 300 nanosecond and here's a shorter pause which is about 40 microseconds now right away you notice that at the end of the process we have this big overshoot and especially we have quite a bit of ringing in the narrow pulse when we're trying to hit 40 nanoseconds i raised this issue with signaling and they've actually already solved it and so if you purchase one of these now it this problem would not be there you're going to get a nice rectangular shape pulses there so putting that aside we can now analyze the amplifier by itself if you look at the amplifier it doesn't seem to be having a very hard time following these pulses you can see it ramps up quite nicely follows the duration of the pause it even replicates some of the ringing and overshoot that's coming from the synthesizer that's good but i want to see more than that i want to enable the amplitude versus time function of the tektronix 6p so we can take a look at this in a lot more detail so i have it already set up here i'm going to turn it on on this one i'm going to turn it on this one as well and we're going to look a little bit further back there we go and i'm going to adjust the trigger put it somewhere in the middle of this pulse so look at that looks so nice we see exactly the pulse input and pulse output of the amplifier so if i zoom in further again you can see the duration of the pulse matches the duration of the pulse coming from the amplifier even a narrow pulse with this rise and fall time matching what's coming out of the amplifier also this all looks very good but what happens when i push the power to the amplifier to be higher and higher to a point of saturating the amplifier what happens to the pulse shape in that case let's give that a try i'm going to increase the amplitude steadily so let's see i'm going to go under level and let's increase it so as i increase the level i'm going to have to go back to the instrument and of course adjust this as well let's go all the way to 500 you can see how sensitive the tektronix 6 series b is even at this i look at this signal the signal is tiny and he has no problem detecting this amplitude let's keep going i'm going to really saturate this amplifier i'm going to put 10 dbm at its input and look at what happens the behavior of the amplifier completely changes so i'm going to drag this down a little bit as well so we can see the top of the pulse and drag this down too and let's look at it with feather out look at that so the amplifier no longer is able to ramp up its output at the beginning of the pulse so why would that be well it's most likely because the amplifier has a hard time capturing enough current from the power supply to follow this pulse immediately when it's saturated you see this ramp up is probably once the amplifier begins to accept the input at a high level and ramps up the current it needs into a steady state remember these are very very narrow this is a common problem for amplifiers if they don't have a very good power supply design the power supply of this amplifier is coming from the usb and then very narrow pulse it completely gives up you can see it doesn't have enough time to ramp up to the final output power it can actually achieve so during the narrow pause it doesn't hit its output saturated power and if you look at this waveform at the bottom you can see the ramp up in the amplitude as well i can change that into a different scale pull it out there you go you can see the difference it's also interesting to note how this is backwards so the when the pulse is nearing really really narrow the behavior of the amplifier is actually non-intuitive again this is normal for amplifiers that you saturate so this is one of the advantages of being able to create narrow pulses and pulses of different widths so at the same time what you can do is you can create a pulse train where the amplitude of the pause progressively gets larger and larger so you can find out what happens to your amplifier different pulse inputs very useful feature of the signal ssg for you to analyze and figure out if the power supply of the amplifier is good enough or not so when it comes to iq generation and arbitrary waveform generation especially with an rf unit like this what you're interested in is the raw performance of the signals coming out of the unit because when you pass your signal through any device under test the difference between what this guy puts out and what the device under test gives you is the performance of the duty you don't want any of the instruments limitations to show up so we're going to look at the output directly from the unit into another instrument and for that we'll use the excellent rodent shorts zn l6 which is a vector network analyzer spectrum analyzer hybrid up to six gigahertz and first we'll take a look at the multi-tone generation which is a very important tool for characterizing various rf components okay so let's take a look at the iq modulation sub-menu here so we have custom multi-tone arb and so on and these are various ways of creating an arbitrary waveform which is then applied to the rf signal and some of the built-in functions so let's start with multi-tone here so multi-tone is the idea that you can generate multiple tones around your carrier so right now the category is set to six gigahertz i can change that let's let's go into the ism band let's say a carrier of 2.4 gigahertz we can change the level here while we're at it to zero dbm okay so depending on how many tones we get in this situation it's going to try and put those tones around the 2.4 gigahertz carrier it's important for the gui to give you the options of what tones you want to turn on and off whether on the upper sideband on the lower sideband or you can offset them from the carrier so you can do im3 measurements and oip3 measurements for example on a device under test now the gui for generating multitone here is a little bit too basic because it doesn't have the flexibility that i think it should have so there is the sample rate you can set that all the way up to 240 mega samples per second i believe that's the highest i think if you try to enter anything higher it just goes back to that value and the frequency spacing now normally this frequency spacing intuitively refers to the spacing between the tones but in this situation it actually refers to the spacing between the first stone and the last stone so if i set this for example to 10 megahertz and i put this to five tones what this ends up doing is it's going to create five tones between minus 10 megahertz from the carrier two plus ten megahertz from the carrier i think it might even be plus or minus five megahertz from the carrier so it's not very intuitive i think they need to change that this frequency spacing should refer to the space between the tones another limitation here is that there is no way of selecting which tones you want to be on which tones you want to have off in this configuration the tones are always close to the carrier but normally what this should be doing is to create a table for you and then you can go through the table and individually turn the tones you want on and off on either side of the carrier it does have a single side button here which when you turn on puts all the tones on the upper sideband but you can choose for example to put them in the lower sideband if that's what you wanted i think a table would fix all of these problems you simply set your spacing you set how many tones you want you generate a table and then from the table you turn whichever you want on and off there's also no phase control for each of these tones and that's pretty important because when you create a lot of tones let's say 10 tones if they all start from zero degrees you're going to get a huge peak to average ratio which you don't want you want to randomize the phases so that they don't produce that effect and that is also missing again these are very very simple for them to add in code because the everything else is already there so hopefully they will add that make this multi-tone a bit more sophisticated so now that we have kind of a rough idea of what it looks like let's take a look at the spectrum analyzer and see how good things like image rejection and the low rejection and how clean these tones are for our purposes okay so first thing first let's turn on the carrier here's our carrier right now it's set to two and a half gigahertz in this case about zero dbm nice clean tone i'm going to turn on the modulation i have votes i have set the modulation to generate only two tones right now 10 mega spacing and there it is look at that that looks pretty good that's a very very good linearity and the lo rejection is also reasonable the yellow is coming out as minus 50 dbm and the output tones are around minus eight or nine so that's very good and the image rejection is excellent so let's push the tones a little bit spaced further apart let's put them at 25 make our spacing here again you can see now it pushes them further so it looks right now that the first and the last tone so this frequency here and the frequency here that that delta is now 25 megahertz it looks like look at that very nice very nice intermodulation terms so i'm going to now turn on the single side on so that it only pushes the tones only on one side of the spectrum there we go look at that that's an excellent image rejection so i'm really quite impressed with how well this performs at least at this frequency let's push the spacing a little bit further let's go ahead to let's say 50 megahertz there you go you can see now it pushed further so that's what i was talking about there's no real easy way of setting which tones you want there is no way for me for example to turn this tone off or put it on the other side because i cannot control that from a table that definitely is something they need to address so let's increase the number of tones and make this a little bit harder so here's five tones there we go i'm going to turn them on both sides i can begin to see some of the intermodulation terms and as i turn more and more tones on this becomes harder and harder if the phase is not randomized let's make that a little bit more here's 10 tones there we go 10 those again squeeze in the same rate still very good this is a tough tough problem by the way because of the fact that the total number of tones all of them intermodulate with each other creating all of these tones on the sides and then we can turn that also on one side only there we go you can see the intermodulation terms indeed get better when there is less tones and there is no tones on the other side image rejection remains very very good let's push the frequency spacing can i put that at 75 megahertz there you go you can see and again can i go to 100 megahertz there we go yeah i have to say it's working uh quite well if they make those modifications in terms of the multi-tone generation capability i think they have a winner here okay so for our final experiment there is a lot to talk about here i want to characterize the baseband generator iq circuit of the signal ssg while simultaneously looking at the rf signal it produces as well as the baseband signal it produces at the back of the instrument but as you also know this thing accepts iq externally too so we can even test that and that requires a lot of instrumentation to be connected at the same time we're also going to use the web interface that's built into the ssg to control it which is really really good signal is generally very good at that and so that helps us to be able to also record it a little bit easier so in order to capture all this data and analyze it we're going to use the amazing keysight mxr 608a which is an eight-channel oscilloscope with two gigahertz of down conversion on each channel independently when you use the vsa software so in order to do that we're going to connect the differential iq outputs of the ssg to port 1 to 4 and on channel 5 we have the rf input directly connected and this instrument will be then controlled by the vsa so that takes care of the internal iq generation both on the base pan as well as on the if now in order to apply our iq to the back of the instrument we're going to use the m819a from keysight and those have two single outputs going directly to the back of the unit too so we can do all of this experimentation all essentially using the same setup and of course control everything from the computer at the same time okay let's take a look all right so here is the gui interface which is accessible through the internal web server built into the unit and it works really well it works right right off the bat i didn't have to do anything special and it's a replica of the gui of the instrument one to one i think sigmund does a really good job at this rodent shorts is also very good at doing these kind of remote interfaces so right now the rf is turned off and the iq modulation is turned off so if you look at the spectrum coming out of the instrument it's basically nothing there this is a spectrum of channel five and you can see that we have a severe center at three gigahertz and a span of 781 megs okay looks good so let's go ahead and turn on our rf channel as soon as i turn on rf channel you can see we get a single tone here that's of course what is expected let me put this over here go back in here so then if i look at the iq modulation if i turn the iq modulation on right now again nothing happens because we haven't enabled any internal iq modulation let's go over here let's go under custom and if i turn the custom state on you can notice that there is some peaking in the noise floor coming from the instrument now this may or may not be a problem depending on what it is that you're trying to do but it's something to keep in mind this is probably coming from the noise generator by the basement circuit as well as the dac they might have filtered out some of it and pushed it to the outside frequencies okay looks good now what kind of data are we putting in here right now we're only at one mega symbol per second which is not that much let's go to 60 mega symbol per second which is halfway to the maximum this instrument can do okay so here's our 60 mega symbol per second okay and it's 4096 symbol length and there's different types of data you can feed to it this is on prbs15 you can define your own pattern as well okay let's go to the modulation type we have 64 core 64 qualm is easy let's push it all the way to 512 qm which requires a lot more linearity of course and much better noise of the instrument as well and phase noise as well of course so let's go back here we're going to update this as soon as i update this we should see that look at that here's our 512 qualm constellation at 60 mega symbol per second and the evm is around 41 db which is very very good already but this has no equalization on it so i'm going to enable the equalization as well to see what the best case scenario is if isi is removed if i do that you can see the equalization is really gentle right this is 0.15 db per division so it's very very little but already the snr is basically 45 db our evm is half a percent which is very very very good so this is what i would expect from a you know very good synthesizer a very good arbitrary wave from generator now we can go ahead and push it a step further let's go all the way to the absolute maximum data rate this thing can support which is 120 120 mega samples per second okay let me go over here oops i have to update that it does tell you that you can see update this okay so it's going to get a wider bandwidth and we're going to have to change this thing to tell the vsa software that we have changed at baud rate i'm going to reset the equalizer and let it do that thing again now it's not going to be as good as it was before but it's going to be pretty close so there you go the equalization is already fairly established we are now at an estimator of about 41 42 db maybe evm is still sub 1 percent there's no gain imbalance there's no quadrature error this is all very very good you can see the noise is still present here now the total output power from the synthesizer right now is minus 1.5 dbm there's some loss in the cable you know roughly close to 0 dbm but of course when i increase the power from the synthesizer it becomes more and more difficult to maintain such a good snr because of the non-linearity let's see how much we can push that i'm going to do that directly from the instrument front panel here so i don't have to change the screen okay let's go ahead and increase the power now the oscilloscope is overloaded we're going to do an auto scale wait for it to be done there you go now we're at 2 dbm the sr hasn't changed here's 10 dbm we're going to do this again i think we are now beginning to be fairly non-linear no no we're still good no problem 8 dbm that's pretty good let's go okay 10 dbm is actually the maximum okay i didn't realize that so when you're doing modulation you cannot go higher than this amount of power so it basically sets itself to 10 dbm as maximum so right now total output power you know 809 dbm if vm hasn't changed at all they probably limit that in order to maintain a very good evm throughout the entire sequence of powers that you're going through so that's pretty good not that you can only do the very high output power in cw it seems let me change the modulation to a lower modulation and see if that makes any difference let's go back home here let's go over here to the custom again let's change our modulation let's say to 64q and let's see if that allows us to go any higher in output power or not i will also change this to 64 qualm like so so we are definitely measuring the right thing okay let's try again can we make the power any higher 11 dbm no it limits you to 10. okay so that's just the way it is this design so keep that in mind that under modulation you cannot go over 10 dbm but look at how clean this is it's very good so now let's look at the iq coming from the back of the instrument how does that look in terms of performance so the iq outputs of the instrument from the rear panel are quite flexible because you can enable them or disable them independent of what the rf output state is so even if you have no rf output or if you're doing no rf modulation you can always get the iq outputs in the rear connectors let me go ahead and turn that on and there they are there's a 256 qm interestingly enough the snr is actually not as good as the modulation the modulation snr was a little bit better than this in some cases so anyway some there's some differences between the two channels but they are almost one to one which is great so if you look at the home screen here i don't even have the rf enabled or the iq modulation enabled this is just purely the internal arbitrary waveform generator generating the signals and then capturing it in the back so that covers almost everything except for when we go under custom we can go under iq control and we can change the iq source from internal to external as soon as we do that the internal arbitrary waveform generator completely turns off and now we can inject our own modulation and look at the rf signal so let's try that okay so let's switch our iq source from internal to external but before we do that let's keep it on the internal and enable it take a look at what happens to the cw tone that we have at the output as soon as i enable the internal iq modulation yes we get this noise peaking we saw before but the allo completely disappears and that's what you expect it to do because you don't want to have that low tone anymore you're applying an iq modulation and therefore it does allow rejection but as soon as i switch this from internal to external you can see that the noise gets quite a bit larger and the lo comes back so under this condition if i apply a zero if external iq i will not be able to get rid of this lo i think this is a bug in the firmware that they have to fix but nonetheless we can still apply our own external iq signal to this so i have a multi-tone signal i have prepared so let's take a look and see what happens here's our multi-tone signal and let's make sure that we level the oscilloscope so we're not overloading anything and yeah you can see that the image rejection is not as good and it's putting out a huge amount of power i think there is a minor bug in here that kind of the gain yeah you can see i just turned it on and off and the gain changed completely now it's doing 0 db there's a minor bug in there where you switch between internal and external it maxes out the gain of the signal and that can saturate everything but putting that aside you can see the lo is still present that indeed also needs to be fixed i think these are fairly easy to to eliminate of course but with this lo being here i cannot do a qualm modulation with 0if anymore because this is going to sit in the middle of my spectrum so indeed something that needs to be addressed now there is a lot more that this instrument can do for example if you load your own waveform into it you can apply live on the fly a gaussian distributed noise on top of your signal and that can be pretty useful in order to stress test the receiver against noise and if you go under the arbitrary waveform here you can see under the arp setup you can do real time awg and add it to your signal now the actual capability of this instrument to playback a lot of standard waveforms is very good you know wi-fi bluetooth and so on you can play all of those back through here you can even do various peak to average control directly on the signal as being played which is also pretty handy full iq control is also available when it is set to internal you can do iq adjustments directly on the signal so gain balance iq offset and angle adjustments between inq can be applied directly from the transmitter here this is again very useful if you're measuring a receiver and you want to know what happens in the presence of iq offsets this supports all of that you can even do iq output control which means you can apply various offsets directly to the ports in the back of the unit which is very useful too if you have an up converter and you're applying iq to it from the signal ssg you can correct for some of its imperfections these are all expected from a very good arbitrary waveform generator rf source essentially you have to have these features in a vector signal generator and they do offer all of that now i'm not going to test every single one of these it's going to take forever in the interest of time we can leave it behind but just keep in mind that indeed some of these advanced functions are also included and there you have it i hope you enjoyed this extensive review of the signal ssg vector signal modulator now from a performance to price ratio i think this thing is in a very nice sweet spot there are some minor performance limitations around the edges and some few firmware changes that need to be applied which i think they will be but it would be hard pressed to find something with the functionality of this one and the performance at this kind of price point you have to check the data sheet to make sure of course that it fits in your needs and in your lab the way you need to use it as always let me know what you think in the comment section i don't get anything in terms of the sales of this unit of course but it would help if you let cigna know that for example you saw my video and that might have pushed you to purchase this unit it's always good for me to have a strong relationship with the vendor so i can bring you the latest and greatest instrument here for teardown and review as always i'll see you next time

Channel: The Signal Path

Views: 13,457

Keywords:

Id: k_Zpr_CbSRc

Length: 55min 59sec (3359 seconds)

Published: Sun May 23 2021