TSP #199 - Digilent Analog Discovery Pro 3000 Series (ADP3450) Review, Teardown & Experiments

[Music] hi welcome to the signal path in this episode we're going to take a detailed look at the analog discovery pro we're going to do a teardown a whole bunch of different experiments i'm going to push it to its limit and see what kind of capabilities this particular usb based instrument has now these usb based instruments rely heavily on the software that's going to run on the computer but this has some tricks up his sleeves in terms of the scripting capability and the sophisticated waveform software that governs this operation so i'm excited to put it into test let's take a look at it so let's take a look at the analog discovery pro the adp 3450 model and see what kind of hardware capability it gives you out of the box now in the front we have a four channel oscilloscope this is of course the four channel version and it's nice to see that we have 14-bit converters behind these bncs that means that you're going to get very nice dynamic range and some fine resolution in the measurements you're doing and 14 bit is really quite good for an oscilloscope now they're running at 125 mega sample per second and if i guess i would say that a pair goes into a single adc and we will see that when we do a teardown that means that there are some limitations when two channels are on potentially now the 3 db band of these channels is about 50 megahertz and of course they're one mega ohm 15 picofarad inputs there is no 50 ohm option and that can be an issue if you connect a long bnc to it so that you go from 50 ohm to high impedance and even at 50 megahertz you could potentially run into some problems but with it you know in one mega ohm probe it's going to be no issue at all this does support half a giga sample per second over sampling but that's of course for repetitive periodic signals that's essentially offset sampling and then you can build a away from that much higher equivalent sample rate having four channels is very nice this is i think a must essentially nowadays if you want to do anything serious you need at least four channels now on the right side we have a waveform generator this is also 14 bits at 125 mega sample per second and most likely coming from a single chip as well again having 14 bits for generating arbitrary waveforms is very very useful you will get again very good dynamic range from those signals and we will test that and these are 50 ohms of course coming out and the analog bandwidth of the channels is about 15 megahertz now on the left side is where we have our 16 digital ios now these are bi-directional so they can do pattern recognition and do digital decoding i squared cspi and all the classic typical things that's all done in software you can also program them to be output and that we use for arbitrary waveform generations of some kinds these are digital pattern generations that you can program directly into the software we will take a look at that too there is two ports here vio these are digital ios that can be configured for digital power supply as well they go from 1.2 volts to 3.3 volts 300 milliamps it's unfortunate that they don't go to 5 volts because all of these audios are five volts tolerant but you can create five volt power supply here and this is essentially for digital circuitry it's useful to have it it would be nice if it went a little bit higher all of these inputs are also protected for high voltage inputs plus or minus 50 volts internally so you don't damage something if you accidentally connect it in a weird way but overall yep some basic functionality is included and of course the software is what's really critical with something like this because that's how you can do mixed signal measurements combine different capabilities of different channels and create some complex measurement capabilities with the software on the side we have our power button of course which is nothing unusual at the very back we have two triggers this is very nice to see you can create some complex trigger conditions using the dual trigger input there is a whole bunch of usb ports available for external storage as well as wi-fi so you can add your own usb wi-fi to it and there is an ethernet port standard as well as a usb device now one of the strengths of this instrument is the scripting capability we'll take a look at that and if you run a script it will run natively on the instrument that's very different than having to run it on a host pc because the latency between getting the data in and out and synchronizing and so on is eliminated when it's running locally so that's also something that's included and the power supply of course comes with it as well so i think we should take this apart now and take a look inside of it and see how it's constructed and see if some of our guesses in terms of its architecture were correct and once that's done we'll do some experiments with it there's a minor complaint i have with what comes with the instrument now you can get this with or without probes but it always comes with this and this is the interface to all the digital ios as well as the built-in power supply that i just talked about but this is really too short and it's not a really easy thing to use i mean everything terminates into these headers and it can be useful in some situation but it would be better if you saw this in a nice longer ribbon that can get you a little bit further away from the instrument and have some label on these so you know exactly what you're dealing with i mean they're all color coded to some extent but i think this is such a simple thing to change and i think the user experience and the digital channels will be vastly improved if this cable were to be a little bit better so let's take a look and see how this instrument is put together now this is a fairly complex port because of the mixed signal nature it has it has some sensitive analog things like the oscilloscope input it has analog outputs from arbitrary wave from generators but it also has a lot of digital activity that needs to be taken care of going on through let's say the fpga the ethernet the usb controllers and so on now because they have a lot of space they've managed to push all of the noisy things to the perimeter of the board decoupling them from the sensitive analog inputs and then also of course all the dc-dc converters if you look in the corner here we have a lot of these cdc converters by the inductors and capacitors you can see here some on the right side as well some are adjustable nicely labeled these will go to the user accessible power supplies there are also some linear or voltage regulator based ones and those are allowing you to power things that are sensitive like the dax and the adcs and you don't want to those to run from the cdc converters so this nicely decouples everything so now we start from the analog inputs we have a four channel oscilloscope here now just like every other oscilloscope you have single and inputs coming in and the very first thing they see or some kind of signal conditioning there's going to be a bunch of relays here and these relays allow you to switch things like ac couple dc couple potentially impedance switching as well as various gain stages you may want to switch between let's say a low gain stage and high gain stage and these are done in a hybrid fashion sometimes sometimes some of the gain is done with solid state control and some of the gain is done by attenuation by doing relay switching in the front end but the signal coming in still needs to be amplified at some point especially if it's small and the very first thing it hits is this component here this is an analog devices this is what they call a fast fet op amp one gigahertz bandwidth ada 4817 now this is going to give us the gain and potentially adjusting it and so on but it's still a single ended signal and if you want to convert it to a differential you also have to prepare it to drive an adc for that they're using a linear technology ltc 6406 that's this component here it's a three gigahertz low noise rail-to-rail input differential amplifier this allows you to drive the input of the data converter which is right over here in a nice differential fashion and get all the benefits that comes from differential signaling and of course these adcs require differential input for their ultimate performance specs to be met and this is a dual adc you can see barely the trace is coming from the two differential inputs one goes here one goes here and this particular one is also an analog devices part this is an ad let me see if you can remember the part number ad9648 this is a 14 bit 125 mega sample per second total a data converter so if you turn on both channels you're going to get only half of that this is a 14 bit converter that's a lot of data still being generated by each of these data converters and we have a symmetry over here on the other side we have exactly the same thing we have a slightly different positioning of these relays which is interesting but nonetheless it's exactly the same behavior we have another identical component now all the data is generated by these data converters that's now 250 mega samples per second 14 bit in total has to be routed to something and of course eventually makes its way to this fpga potentially under the heatsink and there's some memory associated with that fpga as well so that covers this entire portion which is our oscilloscope input now coexisting with that we also have this dual channel arbitrary wave from generator on the left side and that's based on another analog devices component this is an ad9717 which is a bit 125 mega sample per second dual output that i believe is also half the rate per output so that's why you get two channels of waveform generation so then the data from that has to come from the fpga as well but because all of this data is handled through the fpga you can do some clever things you can synchronize the data capturing and the data generation and therefore you can do a frequency response and impedance measurements which are some of the capabilities that this instrument has using the software that comes with it which is quite nice and of course all the digital inputs that are also going to be buffered at some point and then directly controlled by this fpga we have a whole bunch of other chipsets over here these are going to be ethernet interfaces usb controllers nothing really super exciting these are just the digital ios and power coming in over here this has ethernet and device so multiple ways of controlling it now i wish this had a wi-fi module now wi-fi modules and instruments is always a tricky thing because wi-fi is obviously radiating so you're going to have a hard time coupling and isolating it but at the same time 2.4 gigahertz is above the frequency of pretty much everything on this board that you're looking at so with some filtering i think it's possible now note also as i mentioned earlier that there is nothing isolating these input analog channels from each other there is no cage around them so even though there is some signal coupling through the board and that you might be able to handle with some board design signal coupling from channel to channel just over the air is very hard to block without a cage so in this early production unit they didn't have that and i'll show you what they did to improve it and to make sure that these channels are well isolated from each other i also want to take this out to see if there's anything on the other side i don't think so because most things seem to be accounted for but let's take a look at anyway and then we can look at the modified version and as we suspected there's not much else going on in the back we have potentially some ethernet chipsets and so on and then some additional memory near the footprint of the fpga but yeah looks quite simple on the other side and here is their solution to isolate the channels from each other without having to redo the board again so without having a new board revision and not throw away a whole bunch of stuff they've already designed so normally if you redo the board you can move things around create some ground pads and then you can solder in a shield or clip in a shield between the channels now if you don't want to do that you can still make the shield like this find some clever places to bolt it down using some existing screws and of course missing not touching any of the components that are there not shorting them and then finding a way to ground it it looks like they're using these copper woven wires to go to the other side of the board where there are some exposed grounds and then basically clamp it down there and that way you get grounding on this now this is going to get you pretty close to where you would be if you read on the board but this saves tons of money and tons of redesigned resources and they can use all the boards they have already and meet the isolation between the channels so it's a nice clever design to overcome an initial oversight so here is our test setup i have acquired this board here which generates a variety of analog and digital waveforms it's a mixed signal board for testing various type of signal imperfections and so on so we can use this to evaluate the performance of our analog discovery pro here now this requires a fireball power supply that's coming from a rode ensures power supply because as we said this guy can only go up to 3.3 volts in his adjustable power supply section so there's a lot of things we can look here so let's start first with some am signal modulation and see if we can detect any problems with it and of course i'm always going to be using this board and different experiments so we can see how well the software behaves and how well it can catch various issues from this board and here we are with the waveforms software this is the central platform used by digilent to control all of their instruments and perform all the different kinds of measurements it's an impressively rich software there's so much to do in here sometimes the user interface is a little bit difficult to use but in terms of getting things done there is really almost no limit to how much stuff is hidden within the capabilities of this software there's also a very good help built into it allows you to see not just the architecture of the instrument but exactly how to use it and how various parts of the waveform software actually works this is a free software that you can use as long as you have one of their equipment this is how you would control it so i've already connected everything up and we're going to try to create an oscilloscope measurement here so you can create multiple measurements and it will show up in a tapped fashion you can combine them as we will see so let's try some oscilloscope measurements here we go here's our scope you can open it and as soon as you do that the instrument will be automatically configured with the four channels at the inputs i've already partially changed the settings and turned off channel 2 and channel 4 which we're not using let's go ahead and run this as soon as you run this you get the live update of course and right now the waveform is jumping all over the place because we don't have the correct trigger the trigger is set to channel one and there's a several different trigger types that you can define and conditions for those to be met right now it's on automatic triggering that's why the waveform is moving let's go ahead and set this to 2 volts so that we can stop this from jumping around there it is so we have two signals here the cylinder at the top is the synchronization signal that's synchronized with the am envelope of the signal at the bottom is the am modulation itself and you can see that it has some envelope and some high frequency am carrier embedded in it now you also if you look closely you can see some aliasing artifacts you can fold the software and push it into corners where it would make a mistake and produce weird artifacts which are the consequence of aliasing and the reason is because your sample rate is 5.9 megahertz over here and the reason is at the 500 microsecond per division we're capturing 32 768 samples so the instrument will always keep that buffer full so depending on what sample rate you put in you will have to have a certain horizontal divisions so we can zoom in for example you can see that these two numbers will move together in order to meet the condition of the number of samples if you go over here and manually enter it yourself let's say 125 mega samples per second you will be then fixed at a certain horizontal division so let's zoom back out and take a look and see what we can do with this let's do some basic measurements some measurements let's say on the signal at the top so we can go ahead and open the measurement tab there's a measurement tab we can add a whole bunch of measurements to it let's say we define some of our measurements on channel one let's measure the vertical peak to peak let's add that let's also add the rms value for example let's close this let's go to horizontal let's add the frequency so we can see the frequency of the envelope of the am signal and let's go ahead and close that there it is you can see very nicely and very fast update rate it is every time it captures a new 32 000 samples it will update those values as well we have a three and a half 3.6 volt peak-to-peak signal frequency 2.4 kilohertz that's our envelope signal okay looks nice we can go ahead and close that so if i ever open it again if i go under measurements again it will keep those measurements there you can of course resize this and put it in the corner if you want to okay let's get rid of it for now now if i zoom in a little bit i will notice something interesting if you look every once in a while it looks like there's something going on in the am signal itself if i zoom out a little bit more and put this back in the center you can see it happens here occasionally doesn't happen here in the middle it happens again so it looks like it skips every other am envelope we want to take a look and see if we can capture that a little bit better now you can zoom in here and look at it or you can use the zoom function which i like quite a bit in the software click on zoom there you go you get this zoom window and that's the window on the right side now we can make this one larger and i can move this way from anywhere i want let's center it right over around here and we can zoom in there we go very nice look at that so it is very clear that we have a glitch in the am signal itself and these are typically could be hard to find depending on the software that's why i'm going through it step by step so you can see how the software navigates another thing interesting thing to note is that that glitch causes another tiny glitch in the synchronization signal as well and the reason is because this glitch has a very high frequency content to it because of the sharp edges and that couples into the other traces on the pcb and that tiny variation is caught with the high resolution adcs that are built into this device so really quite nice you can capture that you can do a single capture for example you can stop it there's our waveform we can add some markers over here these markers will give us the frequency of the am signal quite easily or we can turn this marker off turn this one on and then we can see the exact voltages at different times throughout the signal i really like the how fast and smooth the cursors work on this i think they've done a fairly good job on that let's go and close this you can see on the right side the histogram of the signals this is how how often the signal is present at a certain amplitude it's very good we can actually show it here at the bottom there it is so because the edge for example the top and the bottom of this waveform flatten out before it comes back you can see that histogram shows a stronger presence at these ends than in the middle because the slope here is the highest so all these are quite useful similarly with the am signal the histogram has a higher probability in the center of the waveform so yeah just a quick overview of how you can move around and navigate on the oscilloscope i think they've done a pretty good job so here's a different waveform and this waveform looks like a simple square wave but it does have a problem every once in a while you can see that it shifts around and something seems to happen to it now this event that's happening is actually quite infrequent but because we have a reasonably fast update rate we're able to actually catch it if i zoom out further you'll be able to see that yep indeed something happens every once in a while and i want to catch this if i turn on the persistence the persistence will fill up very quickly showing that there is a trigger point that's shifting around jumping around and causing the waveform to overlap in a different place okay no problem let's close that let's change your trigger type instead of looking at the edge trigger let's look for a pulse that potentially could be a problem with this waveform let's turn on the pulse so we're looking at a positive pulse now nothing happened you can see it's doing exactly what it was doing before and the length of the pause is quite long one microsecond you can see the other side which is the length of the pulse being shown over here now one microsecond is a very long time and that's probably not where the pulse is let's make it shorter let's bring it down to let's say 48 nanoseconds and there you go as soon as you do that no problem we can catch this very infrequent event now if i zoom out quite a bit you don't see this anymore because it doesn't happen very often once you catch it the next time it happens is quite a long time afterwards yeah it's very easy to catch this let's go ahead and turn on the persistence on and there you go now with the persistence on you can wait for as long as you want and you will not get an overlapping region so it is quite useful to be able to use this kind of triggering as you know in oscilloscopes these kind of infrequent events are are hard to catch and having fast update rate of this instrument seems to help a lot so let's take advantage of some of the digital capabilities of this instrument i have here eight digital bits that are coming from our board and i also have a dac output both filtered and unfiltered and we want to look at the correlation between the signals and the cross triggering capability of the software so let's go ahead and do some make signal measurements so i have also opened the logic analyzer right next to the oscilloscope and because these are tabbed and windowed applications you can essentially order them and arrange them any way you want in the oscilloscope we have two signals on channel one we are looking at the raw output of a dac and on channel 3 we're looking at the filtered version of the same signal the raw output of the dac has the very flat regions between the different digital states going into the dac and that's what you would expect to see and of course once you filter it those things disappear because there are high frequency content if i go ahead and enable the histogram you can clearly see the different flat regions of the raw output dac being quite obvious in the histogram as well now we're triggering here on channel one so if i move the trigger point of course you can see that the waveform will shift with it that's what you would expect to see on the digital side we're not triggering on anything i have the digital selected but i'm not triggering any of the channels let's go ahead and try to change that let's trigger on channel d7 there it is so now everything is locked together you can see that all of these bits essentially have transitions that are correlated to each other but this doesn't tell us anything about the correlation between these two domains now we can also go ahead and investigate that because not only can you select the source of your trigger to be digital you can select it to be any of the other applications that you may have enabled so if i go ahead and select scope now these digital bits are actually triggered correlated to the trigger point on the analog channel so if i move this up and down you can see that they move together this is really convenient of course and this is what normally makes signal oscilloscopes do allow you to measure correlated points in time with respect to the two domains now this is perhaps not as interesting in this case because we want to find out if these bits at the top are actually our bits that correspond to the most significant bit of the output of the deck in order to find out we will switch this guy back to the digital but then we go back to the oscilloscope and switch that one to be the logic so now you can see we are on the rising edge of bit zero that's where our trigger is and right at that point we have the zero crossover point of the raw output of the dac so indeed this is the most significant bit every time it is high we're on the upper waveform and every time it is low we're on the lower side of the waveform and that's what we'd expect to see of course from a dac the reason the pink waveform on channel 3 is a little bit phase offset is because the filter has a phase response so once the filter the output of the dac it's going to have some time delay as it propagates through the filter that's why that is there so this can be quite handy you can change that to any of the combinations you have a whole bunch of triggering capabilities but the fact that you can trigger across any of the different applications that includes waveform generations and so on can be really really useful when you want to analyze something very complex now of course once you have access to the digital data the software can do various triggering on different protocols so here we have a situation where the dac is being engaged every once in a while and it's very hard to catch the individual pulses that are generated on the left side you can see these pauses just come and go and i want to catch the beginning of the process in a very repetitive and predictable way now i know that happens when i squared c interface writes to a particular address so i've defined i squared c on channel 0 and channel 1 of the logic analyzer and you can actually define i square c protocol very simply just like you would do on any logic analyzer and i've already defined here you can see various packets come and go now i can change the way the triggering is done and it's done on a protocol and since i have the i squared c already defined i can go over here and define how i want the triggering to happen so i have it set to a read condition where one byte of h41 which happens to be the address of the eeprom for the dac is being triggered so once i do that the system is going to look for that particular pattern in the i squared c data coming in and only trigger at that time i'm going to press ok and there we go we're triggering precisely at that point and now i can look at the wave formula in a much easier way and of course the waveform on the oscilloscope is also now settled because we're using the source as the logic to trigger this so we basically have now again a correlation but this is quite helpful because this pulse is is not common at all and in fact you can see on this pulse we only have two and if i go to the next one we only have one so we have a three two one sequence and we're able to identify whichever one of these we want to trigger on based on the i squared c data coming in so these are typical but it's nice to see that the waveform software from digital and implements this in a nice usable fashion we can also use the oscilloscope to measure the waveforms coming out of the arbitrary wave from generator just to show what kind of basic capabilities has so i'm running both channels here at the same time and you can run them in a synchronized fashion and since all the data is coming from the same memory essentially you can sync them perfectly and have total phase control between the two channels this is common for most arbitrary waveform generators that have more than one channel so therefore i can for instance apply an offset phase to channel two and if i do that you can see that the two channels of course drift with respect to each other that's quite convenient and this is normal we can also do some interesting visualization we can do for instance an xy measurement and measure channel 1 versus channel 3 which in this case would be a line since they're exactly on top of each other but if i apply phase you can see that i'm able to create this oval shape which is what you would expect to see if you plot two sinusoids with respect to each other and go and close this one and if you try to exceed the bandwidth it does give you a warning so for example at 10 megahertz it does give us a bit of a warning that one of the channels has getting close to its bandwidth limitation which is nice so you don't make erroneous measurement not knowing what it is you're measuring let's go ahead and turn one of these channels off since we don't need it and see what kind of modulation capabilities we have so we need we can change this from a simple mode to a whole bunch of different types and i like the functionality built into the way from generators so under the modulation we can do fm am modulation of different kinds we can combine am and fm on top of each other so here we have a one megahertz carrier frequency with a 100 kilohertz fm modulation directly on top of it so if i go ahead and turn the fm off of course all we'll end up is just a one megahertz carrier we can also look at the fft of this signal press that there we go here's our fft let's go ahead and make some adjustments so we will stop here at let's say five megahertz and we should see our tone right there now because the ffd is calculated under 32k samples coming in it is basically subject to how you adjust your timing so right now we're at 125 mega sample per second and if i go ahead and zoom out as soon as the rate changes from 125 megahertz to something else we will have a longer duration to capture and then we will get more spectral resolution in the fft so you'll see it in just a second there you go you can see that we get now more spectral resolution we get finer bins in our fft and i don't see much harmonics anywhere which is a good sign because this harmonics are not compounded so the non-linearities of the arbitrary waveform generator and the non-linearities of the adc are both being simultaneously shown here and we have a spurious free dynamic range of about 70 db which i'm quite happy with again you that's what you would expect from a 14-bit data converter we can go ahead and now apply our fm modulation and as soon as we do we will see the fm tones sitting on both sides exactly what you would expect now you can do some pretty neat things with the visualization as well so if i close this one and get rid of the fm modulation the 3d spectrogram can be quite interesting and let me go ahead and make this window a little bit smaller so that we don't run into this issue it would be nice if these two windows actually snapped to each other and moved with respect to each other but they don't actually quite snap so in this visualization you can see the frequency content of the signal as a function of time so we have time on one axis frequency and the other one and amplitude on the final axis so this is a three-dimensional plot right now we have a single frequency so we only have one frequency content across time and nothing shows up in the rest of this but as soon as i apply our fm modulation and zoom out so that we can see all the frequency content you will be able to see our fm tones there it is so and this view right here you're actually looking at the frequency content this is basically your fft across time so yeah you can do some pretty neat visualization as a result of this i quite like this and it runs reasonably well actually it runs even better if you do a single capture let's see how smooth this is very nice i really quite like it let's go ahead and run this again if you go back over here we have a whole bunch of other capabilities in terms of our waveform generation we can do things like sweep we can sweep from one frequency to another here we are sweeping from 100 kilohertz to one megahertz in a 50 microsecond time and if i go ahead and zoom in a little bit you can indeed see the sweep happening in three dimensions as well which is quite neat now because we have arbitrary waveform generator and you can do essentially any data and you can sync them between the oscilloscope and the arb you can now measure frequency response of course and you can do all kind of interesting characteristics of external devices using these two functions let's give it a try so here's the next device under test that i'd like to characterize using the analog discovery pro this is an audio amplifier it's based on a vacuum tube it's actually a combination of a solid-state front-end and a vacuum tube driver it's meant essentially for headphones it's not a very good amplifier but nonetheless we want to characterize not only his frequency and phase response but also it's non-linear behavior and see what happens when it's driving a load so the input of it simply comes from one of the waveform generators of the analog discovery and then one of the headphone outputs is terminated into a load that's going to emulate the headphone and then it's going to go into one of the oscilloscope channels so let's see what we can get so i really want to stress this audio amplifier so on the right side we have channel one of the wave from generator producing a one kilohertz tone at a two volt amplitude so this is a four volt peak to peak signal going into the line in it's not a typical situation we want to see what happens to the amplifier now on the left side i'm capturing the same waveform i'm putting into the amplifier that's on channel one and on channel three i'm looking at the output of the amplifier once it's already terminated into an appropriate load to emulate let's say a pair of headphones so you can see that the gain of this amplifier is essentially -1 there's a phase inversion but the amplitude is about the same now i can look at the linearity of this amplifier too let's try an fft first so i'm going to turn the channel 3 off for a second now we have to zoom out in order to produce better spectral resolution from the fft and i go all the way back and you can see at this scale we're looking at some small nonlinearities that are now poking out of the noise floor and this is because the input of the amplifier is also nonlinear so we can catch that because we have so much dynamic range from the r as well as the oscilloscope if i disconnect it from the amplifier these tones will actually disappear it's an interesting thing to note but let's go and turn this channel off and now turn channel 3 on look at how non-linear that channel actually is even though the waveform looked pretty normal by eye it's very difficult to see thds that are let's say better than 25 so this thd looks like about -25 dbc or so so i can go ahead and put a marker on it and we can read the exact values for example there you go so we're looking at here 998 hertz which is the frequency coming out of 2.6 db volts and you can see the next one we're looking at a value of about minus 30. so you can calculate thd from this very easily so this is one way to do it but there is also a spectrum analyzer app that's built into the instrument as well it's essentially doing the same thing it's taking ffts but it has additional averaging capabilities that emulate the behavior of a spectrum analyzer a little bit more closely so let's go ahead i don't want to look at all these other channels we can just simply look at channel three so i configured it already and you can see we're getting essentially the same kind of result now there is something a little bit unusual with the spectrum analyzer if i go under the channel options it's forcing me to put the range at 10 volts if i go to the range of one volt which is what i was using on the spectrum analyzer i get an overload condition and then this result is gibberish but there must be a bug of some kind because what you want is you want the smallest range possible and the reason for that is because it affects the noise floor of course so this is why i'm not so happy that i have to be at the 10 volt range in order to get a meaningful measurement because on the oscilloscope fft it was working fine so i wonder if that's a bug or something unusual is going on with this application but nonetheless you can see that we can very easily measure the thd of this amplifier and the frequency response of the amplifier is also quite interesting now as i said earlier because we have full control of the oscilloscope as well as away from generator and we can synchronize in any way we want we can measure frequency and phase response there's a tiny little diagram up here and a more enlarged version of it in the help that tells you how to connect it so i have the device under test sitting between the waveform generator as well as channel three of the oscilloscope i'm also plotting channel one at the same time which is just a through so we can see the baseline so we're gonna run this and see what it tells us so it tells us that very at the very low frequencies at about 50 hertz the gain of the amplifier is essentially a zero zero db and you can see the phase is minus 180. it's interesting that at some point the phase flips over i think this is just a wrapping around of the phase calculation and right when we get to around 100 kilohertz or so we can see we are almost down seven or eight db so that's normal you don't want to process 100 kilowatt signals that much anyway because you can't hear it so and this there's also some harmonic distortion behavior of this as some people find a particular type of harmonic distortion of audio amplifiers pleasing to the ear i'm not an expert on that but nonetheless we can measure the frequency response and it's very nice and you can see how easy it is to set up something like this and once you have all of these hardware capabilities a good software can take the most out of it so this is a basic measurement you can have multiple of these at the same time you can have several device standard tests turned on at the same time because you have multiple waveform generators and multiple oscilloscope channels you can make this quite sophisticated in how many things you can measure in parallel so let's now use the impedance analyzer so the instrument will use a known resistor and an unknown impedance in order to be able to calculate the frequency response of that particular impedance you're looking for and it will give you all of its characteristics so for that we're using the arbitrary wave from generator channel connected to this capacitor over here and we're using a one kilo ohm resistor in series with it at the same time we're monitoring the voltage on one side of the capacitor and on the other side of the capacitor and since we know what the value of this resistor is we can use some math to figure out exactly the equations that govern the behavior of our device across frequency let's go back to the software and see what it tells us so let's take a look at the impedance analyzer software and it's quite sophisticated allowing you to measure a wide range of type of devices and if you know what kind of device you're testing then you can set it to that so that you can get more relevant information or if you don't know what it is you can just get basic impedances as a function of phase and amplitude so we're looking at the capacitor so i've set it to capacitance here we're measuring from 100 hertz to one megahertz and the system measurement block diagram is shown over here again very simple you can do things like compensation for open and close and save all of those if you want much more precise measurements for now we can stick to this and we are exciting it with a one volt amplitude which is also important because if you have non-linear devices uh you could you would have to control this to make sure exactly what it is you're measuring that resistor is one kilo ohm here so it's going to run this now at 100 hertz the value of the capacitor we get is going to be pretty flat with respect to frequency because we are so far away from the self resonance of this particular capacitor there is no problem in measuring its value so if i hover over here you can see that it's reporting the cs to be 476 nanofarad which is exactly correct that's what you would want it that's what you expect and the phase is minus 89.8 degrees this is a capacitor of course one over j omega c and the phase is 90 exactly what you would expect now as the frequency gets higher and higher at some point the value of the capacitance shoots way way up that's the resonance of this capacitor and tells us that the self resonance of this capacitor the way it is implemented on the breadboard is around 585 kilohertz and in practice you want to stay about a factor of 10 away from that so that the value of the capacitance you're measuring is reliable and repeatable and that you can actually use it as a capacitor otherwise it doesn't even behave like a capacitor in the first place so a software like this combined with the sweeping of the arbitrary wave from generator gives you all of these measurements in a really straightforward way so i quite like this and this can be very helpful if you're designing a board you want to make sure that your devices are actually operating as expected and just for completion we can measure the 3 db band of this amplifier i have put a marker at low frequencies and we will bring this other marker until the delta is about -3 db there it is so we're looking at a 3 db frequency of about 50 kilohertz which is typical for this kind of audio amplifiers now i've only just barely scratched the surface with regards to the waveform software there's a lot you can do and other hidden capabilities and each of these functionalities has a lot more to offer there's also an iv tracer that's quite useful because again we have two waveform generators you can create a multi-dimensional plot and you can excite for example the base emitter of a transistor and the collector emitter of a transistor and get its iv characteristic can be really handy when you want to analyze a device that you don't fully understand or you want to check it's not linear behavior across dc so we'll save this for another time now as i also mentioned there is a scripting functionality too now this opens a whole new set of possibilities because these code runs locally on the machine and not restricted by the interaction between the instrument and host pc you can write your own code you can do all kind of analysis on the data that comes from various hardware capabilities of the box so you have full control over the waveform generators and oscilloscopes and so on i have some plans for this for a different video for some binning and some automated testing so we can leave that for that but just be aware that this also exists and there's a bunch of documentation available for the api so you can take advantage of it and there you have it i hope you enjoyed this detailed review of the analog discovery pro there are some minor limitations to this instrument but the software is definitely sophisticated and takes the most out of what this hardware can give us of course you should do your own homework make sure that this fits your application compared with some of the competition but i think it does make a compelling product especially with the scripting capability if you're interested in that kind of work and you want to explore it further and have something small on your bench you can carry around and do a lot of different measurements on i think this is a good option as always i'll see you in the comment section [Music]

Channel: The Signal Path

Views: 18,572

Rating: 4.9458241 out of 5

Keywords:

Id: ryxxYGmRCpQ

Length: 38min 17sec (2297 seconds)

Published: Sun Sep 26 2021