Hi. This is just a quick follow-up video on the acoustic delay line. We had a quick look at inside the Sony Uh Video8 camcorder 1985 vintage and this is a a pal delay line. A glass delay line you had a quick look at and a lot of people, uh, wanted me to expand on that and uh, just do some measurements and actually show the delay through this thing and uh, just generally have a play around with it.

Now since that, uh, video, um, some people have sent through uh, various links and info on these uh, glass delay lines and they really are fascinating stuff. So I'll link in an interesting uh paper, uh down below that basically explains how these things work, all sorts of math behind the Uh Shear waves in these things and the bandwidth and all sorts of stuff. So it's rather interesting. So if you want to get more involved in the math and the technical details of how these things actually work, I suggest you have a look at that link.

Now, as we said before, if you haven't seen the previous video, there's two uh, acoustic transducers up the top There, you can see that little bit of gold on the top there, one transmitter, one receiver, but they are actually uh B You know you can use one for either. So uh, this one can be the transmitter or the receiver or vice versa and we'll actually, uh, test that. We'll hook, hook it up to the scope and we'll measure a delay going through this thing. So as we said, if this one is transmitting here, for example, then it goes.

Then the uh sheer wave inside this thing uh, goes down through here. bounces off this wall here. bounces off this wall because these are oblique angles. They start off at oblique and they come in at these oblique angles so they can't actually Bounce Off the Wall So it goes all the way through there.

Like that: Bounce, bounce, bounce, bounce, and back out. Now as it turns out this, um, these things are typically either a quartz uh glass or a potassium Le silicate uh glass. The thickness of this glass will be uh, basically one wavelength so it uh, effectively works as a wave guide and this black stuff on the back of the glass here. I Did guess correctly last time.

It is actually an epoxy uh compound which just, uh, dampens the glass a bit in that particular location and helps guide the sheer waves through the glass. Now, as it turns out, these um, Potassium, lead silicate glass delay lines like this one I believe they have an approximate um, uh, veloc wave velocity inside the medium and it does travel inside. It doesn't travel on the surface so that even that, though that epoxy is on the surface, it's just used to dampen the wave inside the thing and keep it channeled within that particular wave guide. uh function.

So it can. You know because it's folded back and forth and has this convoluted path. you need those damp in there just to help concentrate those Uh waves Those Shear waves inside the glass. now that that has it, has a velocity of propagation of about 2.5 mm per microc or 2500 m per second.

Now I Actually got out my ruler and I measured all of the paths in here from Center to Center and I added them all up and uh, it was Uh 26 + 6 + 22 + 21 + 7 + 29 + 7 + 21 + 18 mm for a total path length of 157 mm or thereabouts, you know it's with a ruler. It's not going to be horribly accurate, but if you divide that by 2.5 mm per microc, what do you get? You get 62.8 microc and this is supposed to be a 64 microc delay line so it basically measures out. Now what makes this thing actually directional and allows the way wave to travel? Uh, all this complex path is uh, not only the thickness of the glass which is Uh Nally one Uh wavelength, but the aperture as well, which is the distance of the path, the width of the path like that, and that is typically much larger as you can see than the thickness of the glass, so that allows the wave to be directional. The physical properties of that allow the wave to be directional and hence make that entire distance without uh, dispersing too much.

And it's helped by these epoxy dampener materials which just keep it channeled within there and stop coupling between the paths. Now, if we have a look at the edge of this thing, you can actually see that it's very, very smooth and that is required in order to get the reflection off the oblique angle like that. We got 45 coming in like that, and the sheer wave inside needs to bounce cleanly off that edge, which edge with as little little attenuation as possible. So they do that by keeping that uh, a very, very smooth edge on that thing.

Whereas the surface is going to be different now, the surface there you can actually see is quite, uh, opaque. It's not, uh, completely clear and they've clearly uh treated that probably due to through some uh chemical treatment process or something to roughen it up to ensure that those Shear waves actually stay within, stay, uh, sheer ways And they stay within the body of the material like that so that they can actually reflect off there. Because this thing is effectively a Uh High Fidelity uh transmission line so to speak. Because we we're not just passing digital one and zero through this thing where the signal Integrity of this transmission line matters, we're passing analog high bandwidth analog information through this stuff color video signals.

so we need to. We need uh, for this to be as a good a transmission line as possible. and that includes Uh Reflections and bounces and stuff to do with regular transmission lines, load impedances, things like that. So if this top surface wasn't uh treated in some way, we might have some uh surface effects happening as well as the Sho wave going through the body of the material and then that would cause different delay and we, you know it really wouldn't be a good transmission line at all.

We' get lots of different delays and phase and all sorts of things uh happening there. So really, um, for this to be a very good transmission line, they need to treat that surface and have a very good reflection off these edges because you're going to get a loss uh at each. Edge And because this thing bounces what around a dozen times or something off these various um, oblique edges? Here it comes in here, bounces off 45 goes across here and up like that and that happens. you know, a dozen or so times.

Um, Gez, you know it's a wonder these things can get through with the Fidelity that they do, but these are been specifically designed and engineered to be. High Fidelity Transmission lines Fascinating and really, you shouldn't get uh, too many. uh. endtoend Reflections I.E Um, if this is the transmitter here, you shouldn't get it going all the way through and then bouncing back off the uh end the transducer end here because a it's terminated properly and B because it physically has the transducer uh, glued or soldered onto that edge there.

then um, you know it's it. You've affected the Uh properties of that edge. It's not going to reflect nearly as much uh magnitude as it does when you get bouncing off these nice smooth edges that you do within uh, other aspects of the device. and it actually would be fascinating to uh if if we actually had suitable Uh transducers to try and attach them, um, to the various Uh points.

And if we did that, and the various reflection points. if we did that, we'd actually be able to see all of the multiple delays we'd able to see. You know that would be like a couple of micros delay. You know, five micros.

You know, delay. If we put a sensor down in there, we'd measure that. that and we could actually see it propagating all the way through the device. Now as it turns out, the bandwidth of these things is actually can be quite, uh, large in like, in the order of like, uh, 80% or more of the carrier frequency.

So uh, let's draw a little graph here of a typical response characteristic of these things. and they are tuned to the Carer frequency, but they don't just operate at that. they will have a very wide bandwidth over that. So if we have a little Dave CAD drawing here of Uh frequency on here, and uh, this is basically uh, the gain of the thing.

which let's go V out on V in here and we'll have a response looking something like this, there'll be a little Ripple sort of in the pass band like that, but it will be something like that over the frequency and basically that'll be the center frequency there and we'll have quite a large bandwidth over that and this might be in the order of, say, minus 20 DB or something like that. So um, these things have to work with a specific load or they or they're better, uh, operated with a specific load and tuned to a specific frequency. So when we hook this thing up in measurement, we probably don't expect more out than what we put in, especially if we, uh, you know, if we load the thing down and tune it properly, but we'll just do some crude measurements today of you know, I'm not going to tune this thing exactly with LC circuit so we'll just have a bit of play around with, but we don't expect anything any more out than what we get. but this will I Would expect that to change uh, fairly greatly between different models and different types of glass delay lines.

and if we have a look at another Dave cat drawing of basically how they Implement these things, they're going to have a tuned uh LC filter on the input and the output here typically with a variable inductor and they have to be loaded correctly. Now for these glass delay lines, Apparently, that load is typically in the order of 270 to 390 Ohms on the output and also the source impedance as well. So apparently you know these things work better. They'll have a more linear Uh bandwidth if you specifically tune them.

but H Today, we'll just whack on a resist. We won't worry about the L L's and C's we'll just whack on some some resistor uh Source impedance and load impedance and well see what we get. Bob's your Uncle Now these ultrasonic Uh transducers on the oblique Edge Down here these have to be very, very thin. They have to be Uh at least Uh in the order of like a quarter of the half to maybe a half to a quarter of the typical wavelength, which as I said, will be the thickness of the glass there in order to ensure good performance.

So these things really need to be as thin as possible on that edge to maximize the performance. And the potassium lead silicate glass they use in this thing is also become known as Isop Porti glass with a T Um. And that basically Uh means that these are zero temp Co very low I In in practice, they're going to be in the order I believe of a couple of Uh PPM per degree C but for all practical purposes, these are referred to as zero temperature coefficient Glass Solded some pins in it just allows me to uh, pull it in and out and uh, rotate it easily on the breadboard. I've got a 270 ohm terminat resistor on the output here.

270 ohm series input resistor here I've got no L's and C's to actually uh, tune this thing. but let's see what we get, shall we? Input here is connected to my Ryal DS 4000 Uh function generator that that will allow us to easily generate the waveforms we need. I've got channel One on the scope connected across that as well. Uh, so we'll use that for the Uh trigger and that will give us our reference waveform where we get our Delay from and uh, the output here.

I've got Channel B of the scope hooked up so let's hook it up. feeding some burst pulses of around about the resonant you know, frequency of 4.43 CU As I said, these things do have quite a large Uh bandwidth on them, so uh, we'll be able to measure that on the scope. Let's go now on the function generator here. What I'm doing is I'm generating a sine wave of approximately 4.4 mahz there 5 Vols Peak to Peak and we're turning on Burst Mode as well so we can turn Burst Mode on.

There it is and it shows that we're generating a burst and then a dead period. So uh, basically what we want is a Um is the period of the burst, the complete period. That green part there to be 100 micros. Uh, you know, larger than larger than the delay period because our delay period of this thing we think is going to be about 64 micros.

so we want it to be larger than that I can set it much much larger. But let's just go for starters, 100 micros. So that gives us a burst of waveforms, then a dead time and I can set up the number of uh Cycles as well. So if you go right, you know let's I don't know, 100 Cycles or something like that.

Whatever. So we're generating a burst of 4.43 MHz and we can adjust that uh frequency of course and we will to, uh, check the bandwidth of this thing even though it's not tuned. And let's see what we get on the scope and bingo Here it is and you know, notice that it's jumping all over the place. that's because of the uh trigger.

So what we'll do is we'll single shot capture that. There we go and we can see our delay here. The yellow waveforms our input of course that's our burst of 4.4 megahertz uh sine waves there and we've got if you actually count them, you'll get 108 or whatever it is we set over to there and you can have like a single sine wave. doesn't have to be this long, but uh, whatever just for purposes of test and you'll notice that.

look at that if we have a look at that delay time there I think you'll find that that's 64 microsc. In fact, we can go to the delay here because our trigger point is that little triangle up there which is right at the start there the first pulse. But we can get more accurately in there. But we're getting around about that 63.8 microsecond delay.

So uh, let's get in there with the cursors and get a little bit more accurate on that, shall we? But you'll notice that uh, where 1 V per division input and 50 Ms per division output. So our output signal really is quite low at Uh 4.4 mahz or the you know, the Uh intended uh Center frequency I believe the intended Center frequency of this thing. It could be operating uh, higher and then they um, you know they modulate the thing. but I don't know at 4.43 the output voltage is quite low and remember, we don't have the Uh tuned LC filter on the input or output either.

We just got that resistive load. So what I've done is uh I've turned on cursor mode here and we want uh, because we're using different channels for the cursors, we want the X2 cursor or the second cursor to be a s a source on Channel 2 so we can zoom in there and set our cursor. Let's set it like right at the start when that waveform starts going. Okay, so there we go and then we want uh, cursor X1 there to be channel one and then we can adjust that so we can zoom in on that and then take that right at that point there and bingo our delay.

Let's have a look at our total delay there and there's our Al X time 6383 or 829 microsc. So very close to that predicted 64 micros. although curiously it is a tad under it. But an interesting thing to check will be does that, uh, time delay actually change with frequency at all? I don't know it may we need to check that.

So what we should do is, uh, turn our frequency up here L I'm going over to sign. Okay, I'm going to go to 10 megahertz. Here we go. I'm going to 10 MHz and you'll notice that the amplitude has gone way up there.

It's gone a long way up. so we'll check that. But look, we're still zoomed. We're zoomed in on that thing and we're still look.

We're still spot on pretty much so the delay doesn't change with frequency. I Mean we' have more than doubled that that frequency of this thing. And basically, um, that's pretty conclusive that there is no delay change with frequency. it's completely constant.

And as I said before, this is an isop potic uh glass. so it has Z zero temperature Coe as well. So the delay is not going to change with temperature either. Doesn't change with temperature it doesn't change with frequency.

These things are pretty stable and the only variable left really would be aging. Like you know, this one is 1985 five vintage so it is very, very old. But um, you know, because this is based on the physical uh, distance. You know the physical properties of the glass that hasn't changed over the 20 uh plus years.

So really, these things are going to be incredibly stable over frequency, temperature, and time. Now there's one thing you'll notice here is that we do have an additional Uh waveform. It looks like we've got some reflected bounce or something happening in here and it's not aligned with that at all. and it doesn't seem to be a multiple of that.

So um, you know it's not like it's a second reflection or anything like that. So I'm not sure what that particular burst there is doing. What I'll do is I'll run it here and you there will be some capacitive uh coupling as well. Now what I'll do is I'll physically remove the delay line from the the breadboard and we expect these to vanish.

but we'll probably see still some capacitive coupling in there from the breadboard at that frequency. Bingo There you go, We've still got that capacitive coupling. There's no delay line in the breadboard at all. it's just it's just that couple in there.

so I'm not sure what's happening with that one cuz it's not a multiple there. So maybe some sort of near field surface effect or I don't know. I'm probably talking out my ass I Have no idea there's something physical going on there. now.

What I'm going to do is have a look at the output amplitude versus bandwidth here now. I got it set to one megahertz and if we Zoom all the way in there, you can see that it's 1 megahertz, but uh, it won't. Oh, that counter just doesn't just doesn't kick in there. It really is, uh, quite annoying there.

But yeah, it's picking up the repetition rate there instead of the uh burst frequency, but that is 1 MHz So we're going to start there and you'll see that we're getting no output and I'm right down at 50 MTS per division. There, we're really down in the noise. we are getting nothing. So I'm going to turn this going to wind the wick up here and uh, 1.5 mahz 2 MHz 2.7 So you know around about you know, 2 and a half.

We sort of start to see a bit of something. we're at 3 mahz now and we're let's go to the Uh nominal operation free frequency this 4.4 MHz and you'll see we're still at 50 m per division, but we do have a 64 microc uh, time in there and I I Trust me, guarantee you if we go in there, that'll be 64 micros of course, because we've already discovered that it doesn't change with Uh frequency. and as you'll see there now, I'm at 5 six 6 MHz now and you'll see the amplitude going up 7 Mahz Really? Whoa, We're really going Chang that to 50 Ms per division going up 9 MHz 10 mahz now. So if we zoom in there, we go.

we're 10 megaherz there. and one thing you'll notice is that when we get to that four I'm at 4 mahz now 4 say 4 mahz you start to see that little uh ghost pulse there. After that that little reflection pulse. It's not a complete multiple of the uh distance of this delay line, but you see that in over that 4.4 MHz Mark Now I've gone up to 10 MHz now and let's go right up and try and find the peak value of this thing on 12.2 MHz There we go.

So about 13 MHz So if we zoom in, we're 13 MHz there and that 30 MHz for this particular physical configuration cuz we don't have the tuned LC's in there. 13 MHz seems to be the peak frequency in there and you'll notice uh, that's 15 MHz and we go down where it's 16 17, MHz 20 MHz Etc So 21 MHz There we go and you'll notice that there's uh, some coupling through there as well at high frequency. But yeah, basically that it does seem to have a very wide operational frequency range. So if I set it to that, there's that 13 MHz which seems to be about the maximum there.

We're at 50 MTS per division, so we're still not getting if we set it to the same volts per division. they're both one volt per division. You see how small that Uh output there is now how high that would be with your tuned Uh LC filter in there I Don't know you would have to actually uh, build the thing and check that out. but you'll notice that the Um if you looked at the Sony schematic Uh diagram as we saw in the previous uh, tear down, we would have.

We noted that there was a differential amplifier in there. you know, amplifying the output of this thing, so you can expect it to be reasonably low. But as I said, uh, near the start of the video, I'd expect significant differences in that depending on Uh models and types of delay lines. And if you want to know if this thing is bir directional or not, I.E You can swap the input and the output.

Let me disconnect it from the breadboard. Here, you can see the coupling high frequency coupling as we saw before and I'm turning the thing around so it's now backwards and bingo, it's still exactly the same. We're still got our 64 microc delay in there and everything's hunky dory. And the other thing to note, let's have a look at our output here.

Here's our output and you notice that it is positive going exactly the same as the input all the way over here is positive going as well. Now let's see what happens if we reverse the polarity on our output Terminals and see if it goes negative and bingo it does. I Just swapped the polarity on our output and as you'd expect, it goes negative while the input still goes positive. And if you're wondering if we can actually do anything by physically touching the glass, well, the answer is, uh, not much at all.

Let me tap it. I'm tapping the thing there and well, you know, we can't really see much in there at all. Let's we, We could actually trigger off the output here and get it a bit stable. There we go.

that's and I'm tapping. You can probably hear that. I'm tapping that with my finger and you can. You can't really see anything.

There might be something in there, but that seems to sporadically pop up even if I'm not touching, even if I'm not tapping the thing. So really it? uh, they don't seem to be very vibration sensitive at all. and if I squeeze it with my finger, I'll stay away from the Uh wires there. So I'm not capacitively touching anything.

So I'm squeezing The Edge there and you can see the amplitude certainly does drop. It dampens that a fair bit as you'd probably expect. Just like the epoxy um stuff dampens thing you know works as a dampener, so does that. And if I really, really squeeze it.

There we go. but we haven't killed it. We've just dampened the thing and we've still got our delay. time is exactly the same.

Our delays line doesn't change at all because, um, of course, the shear wave that swave is flowing through the body of the glass. It's not actually on the surface at all, but uh, we can certainly dampen it. It's quite interesting, and if you're wondering what happens with no load on this thing where we're still at uh, 12 14 mahz there? actually 14? MHz Now let's remove the load resistor here so we'll pull that and Bingo! We have actually gone up a bit, but it hasn't changed the delay time at all. Now it's uh, lower the frequency there.

Oh 4 MHz There we go. We're still 11 MHz 10 so we still seem to peek around that 13 mahz Mark there. and if we go all the way down to 4 mahz or so, again, there we we go. 4.4 MHz We're still way down in amplitude, so that that load, uh, you know, doesn't really make much of a difference in this particular build.

So let's increase our burst period here up to 1 millisecond and have a look at some of these Reflections in there. We should be able to do that at 1 millisecond. Let's give it a go. There we go.

we now have a 1 millisecond period between these things. 500 microc per division. We got the same number of bursts as I said before, the number of bursts does does not matter at all. it makes absolutely no difference.

And let's turn that frequency up to where it Peaks there at uh 12 mahz 12.5 MHz even at lower frequencies I Can't see any notable Reflections in there now, so looks like this is a pretty good glass delay line. but if we remove our 270 ohm Source termination resistor Bingo we get a reflection in here and if we take out the let's take out the Uh load resistant. No, we're just getting excess noise in there which is of course just 50 HZ crap all the way in there. but we're not getting any additional reflection really due to the lack of a low termination resistor.

And if I turn on high res mode on the scope here, we can actually, um, see a couple of little you know, one reflection there, then another one you know right there. and I've got the source resistor. uh Source termination resistor in place here and that's uh, 3 and2 mahz. Basically go down to 3 mahz and you can see the reflections changing there a little bit, but not much.

Let's go up to 4.43 which is supposed to be its uh, nominal operating point. but really, you know that hang on, it's 4.43 There it is, and uh, you know really, we don't see a huge amount of difference there at all. So if I go up 5 mahz, that first one disappears there and we only get that second reflection there. So you know I mean you can Muck around with this until the cows come home with various uh source and termination, uh, impedances and tune circuits and all sorts of uh stuff.

But really, we're not getting uh, significant Reflections there at all. So I'm up at 6 mahz at the moment and we're up at 10 mahz there. And you know, really, we are. we are talking very low amplitudes out still and even much lower than that reflection there.

So if we up the frequency 10 11, 12, you know, 13 megahertz where you know it starts to the reflections start to completely vanished there. So really, there effectively is no, uh, resonant point for this thing, it doesn't. You know? There really doesn't seem to be much happening there at all. There's just a couple of little pulses, but um, they are effectively all over the entire bandwidth of this thing.

So there's just some, uh, simple initial experiments on these glass delay lines just having a bit of a muck around. Um, it might be interesting to do some more thorough uh tests on these things to actually, uh, see how they perform with their real uh LC uh loads on them as well actually tuned to the specific frequency, But as we saw, they do seem to have a very large uh bandwidth and that's backed up by the Uh Theory as well. And these things are really, uh, quite, uh, linear and quite, um, remarkable uh devices in terms of the uh information that you can push, the bandwidth and information you can push through you know, just essentially what is just a piece of glass with a couple of transducers I mean there's you know more to it than that, just the physics alone on and just the material physics alone on these things, there really complicated subject and you could do entire PhD thesis on just one aspect of these things. so they really are fascinating.

As I said before, I've linked below an interesting um uh paper on these things. an interesting section from a paper in anyway on a bit more of the theory of how these things work, so highly suggest you take a look at that if you're interested in these things. So I might follow this up with uh, some more videos playing around with these glass delay lines cuz they are quite fun and fascinating devices. If you want to discuss it, jump on over to the Eev blog Forum cuz that's where everyone hangs out.

And don't forget, give it a big thumbs up if you like it. Catch you next time. Yeah!.