Hi. Now this is a bit of a follow on from my mailbag uh segment decided to make this as a separate video. We're actually using a pulse generator here from Man, thank you very much man who uh sent me this and I'll provide uh links to uh, his uh, schematic and uh design and things like that. but he sent that's powered from a 1.5 volt battery.

so see the mailbag uh segment for that one and uh, but I Just decided that we'd do a separate video here of using a Uh pulse generator and uh, see what the bandwidth of my Uh Agilant Msox 3054a Oscilloscope is. It claims it's 500 MHz analog bandwidth. That'll be um, nominally. Of course, the bandwidth of an oscilloscope is minus 3db down at that frequency, so the frequency is not actually.

you know it doesn't give you a flat response to 500. MHz should be 3db down. but using one of these pulse generators, we can actually calculate the real bandwidth. Yeah, I've got a little uh Dave CAD Note here to explain the rise time versus the bandwidth.

Now the rise. There's actually a direct relationship between the rise time you see on the oscilloscope, assuming a perfect input. um, you know, a pulse which has a perfect input and the actual analog bandwidth of the scope. And for traditional gin response Scopes they're your old like you know, your basic analog.

Uh CR Oscilloscopes These non digital types. It uses the classic formula: the rise time is equal to 0.35 on the bandwidth of the scope. It's as simple as that. but uh, these more modern Scopes Uh, digital Scopes don't necessarily use a have a Gan response on their analog input channels.

They'll have what's called a maximally flat response, but even that will depend on you know what kind of rollof they're actually using on the the filtering of the analog uh, front end. So uh, basically it's You know, it's a little bit higher General but roughly around about 0.4 on the bandwidth. So slightly different variations. but assuming that we've got one of these which generates an absolutely perfect pulse with zero rise time.

Like zero Fto seconds rise time? Well, zero seconds or zero ftocs. Same thing, but assuming it's absolutely perfect then uh, this for will apply. We can if we know the rise. If we we can measure the rise time on the oscilloscope, we can calculate the bandwidth and it's not necessarily that same 500 mahz quoted.

It's usually better than that. but of course these pulse generators never give you a perfect rise time. Um, this one at Best is going to be probably you know, 350 p a seconds, 300, 400, PCS or uh, something like that. It's not perfect so it is actually going to have contribute.

It's going to have an effect on and contribute to the rise time you see on the oscilloscope here. But as a rough rule of firm, if it's five times better than what you need, then it's not really going to affect it much. Now in the case of this 500 mahz uh bandwidth. Using the formula of 04, we can calculate the rise time In theory of this oscilloscope is going to be uh .4 divided by 500 MHz That's 800 Peak a second.

So this one, although we don't actually know, uh, we. you know we haven't actually measured the absolute rise time of this thing, so we don't know what it is, but we know it's going to be in the order of you know, 300 p a seconds or thereabouts, it's You know, it's only like three times just on, maybe three times as good as the oscilloscope. So it's it's it. May you know, it's going to contribute a little bit bit.

Ideally, you're going to want five times better, but this will certainly do the job for most bandwidth most oscilloscopes up to. say you know 500 MHz bandwidth. And of course, to measure the True Performance of this thing, we need a really high-end sampling scope. You know, the type that you mortgage your house for.

You know, $50,000 $1,000 scope. Something like that that has you know 10 GHz 20 GHz bandwidth. Something like that, we need something really good so that the bandwidth of the oscilloscope um, you know, is basically so high it doesn't matter where measuring the true performance of The Rise and four time of this thing. So I'm going to have to try and get access to a better scope for that to actually measure this circuit here.

But anyway, we can figure out from this what uh, actual bandwidth we're getting on the scope and I bet you it's better than 500 MHz And as you can see, this uh pulse generator is pretty good. The rise and four times are essentially identical 520 p a seconds, 530 p a second. And of course, we can simply swap these two terms, the rise time and the bandwidth here. Uh to? because if we know we're measuring our rise time, we can calculate the bandwidth.

So if we go 0.4 divided by our rise time of 520 PCS here, that gives us about 769 mahz. So there you go. The bandwidth of this scope is at least um, as good as that because you know we've got uh things like the uh, you know, we got the coax a meter of coax cable on the Uh end of this thing, which is you know, might contribute a little something and uh, and of course we've got the Uh contribution of the uh, unknown rise time of this thing. but we can say it's at least that good.

It's much better than uh, what you would think at Minus 3db down at 500 MHz. So this thing's pretty much kicking ass and just in case you're wondering what the other channels are, I've chosen Channel three here and we're basically getting identical 520 530 PCS And you can see that uh, our signal Integrity is not perfect here. We have a little bit of a little dip there just before it starts and of course some undershoot and ringing at the end of it here. Ideally, it should be better than that.

That's probably due to our coax cable. would have been nicer if we had a direct connection or if this connector here um, wasn't this um uh socket type was actually a plug type? We could plug that directly into the BNC on the front of our scope here. that would have been the ideal case to uh, you know cuz we're using a meter of Rg59 uh cable here so that's probably causing um, you know that that sort of ringing there in the waveform I Doubt it would be. uh, the layout cuz the layout he's taken.

He's really stitched the uh, the Uh, front and the back uh, ground Planes together. Really nice. so you know I'm sure it's really nice. short paths there directly from the Uh transistor there and it's working quite nice.

So I think it's mostly the coax doing that. But check out this interesting uh phenomenon. If I touch the can here, it disappears. Look at that, it just absolutely vanishes.

Let's turn the uh, let's turn the average in off there. and so we getting all of our signals there. we get in our that gets wider as as it goes out. it's quite a bit of Jitter in that signal.

but as you can see if I put my finger on that it's slightly it gets wider, gets a little bit bigger as you as I'm not quite touching that and you can see it start to expand there as the C capacitive couple in between my finger and that can really kicks in. but I can just completely kill that so that's the 50 htz obviously pickup um from my from my my body there just absolutely swamping that uh oscillator circuit and the Avalanche breakdown of this uh transist in here. so really you know that's that's absolutely killing that. Ideally you want this thing in a proper shielded box but hey, even with a bear board like this with no Shield at all using a big meter long uh you know crusty bit of coax cable you can at least get you know a decent measurement on the bandwidth the scope I like it.

Highly recommend you build one of these suckers up. They're very handy and as it so happens, I'll be getting this scope upgraded to the 1 GHz bandwidth model um in the not too distant future so I'll take this along with me down to Melbourne going down there to get the scope upgraded and uh we'll be able to check the performance before and after and here's before 630 PCS and uh of course we're only got 1 nond maximum uh time base there. that's the fastest time base we've got. the 1 GHz version should go a bit quicker than that, but it'll be interesting interesting to see.

um what we get on the 1 GHz version. So thanks to Minut for uh, this little board that's excellent brilliant saves me having to build my own I was going to build up that classic Jim Williams uh circuit as many people have and I highly recommend you do it. It's good fun and you can learn all about the Avalanche behavior of uh transistor. um Avalanche Breakdown: It's terrific and get a very handy extremely fast rise and fall time pulse generator.

And as you can see, it's you know one, two, 3, 4, 5, 6, almost seven volts. uh you know, 6 and 1/2 volt uh pulse there at you know, at least uh 530 PCS It's likely a lot better than that, so pretty neat. Now let's take a look at the application. Note from Linear Technology A47 Classic application note: It's massive written by uh the late great Jim Williams of course very famous app Note It's got a lot more in it than just this pulse generator circuit.

so I highly recommend that you you uh have some bedtime reading here. A47 Classic Anyway, um this is a circuit designed for measuring the uh probe oscilloscope. uh response. It's a classic circuit.

um few people have done something better using like Garden variety uh Parts but figure d one is the one we want which is the circuit that uh Man used here and it basically provides a 1 nond pulse with Rising four times of approximately 350 p a seconds. But they're going to depend upon the uh uh selection, the specific hand selection of the transistor, the Avalanche transistor that we're using to generate this pulse, and here's the circuit here and it basically uh, just a single cell converter to generate for the 90 volts required to uh, break down Avalanche breakdown the transistor which is a 2N Uh 23 69 and that's the exact same transistor that man is using here and it's got the metal T5 uh package of course. uh, it still comes in that but uh, you really have to uh hand select this and it says C text there for a reason So let's go up and take a look at that and here it is. Q1 may require selection to get Avalanche Behavior such Behavior While characteristic of the device specified is not guaranteed by the manufacturer sample of 50 Motor Roller Uh 2 N 2369 spread over a 12year date code span yielded at 82% result, so some of them aren't even going to work at all presumably.

but uh, this one does. it's been hand selected by minut I'm sure he's tested it to make sure it works before he sent it to me and uh, and they Jim will claims all the good devices switched in less than 650 p a seconds. So um, our one is clearly getting better than that because it's at least uh, 500 and something. But you know I think it's probably going to be around that, you know.

rough 350 Peak a second figure. uh, wouldn't surprise me cuz you have to combine that 350 Peak a seconds with the uh rise time of the oscilloscope and stuff like that. But I think we need access to a better scope to measure this thing. And here's the result that Jim Williams got.

and as you can see B the construction of his is basically you know he's got it in a metal box like this. Just A Rat's Nest construction like that. it looks messy, but that's you know, the absolute lowest. um, you know, imped inductance, sort of uh build you can get and of course uses a Uh plug directly on here which plugs directly in the osilloscope.

so there's no coax cable at all. It it works a treat and that's of course why he's getting no very little ringing. There's a little bit of ringing at the tail end there, but it's tiny as opposed to uh, the one which me we measured which is probably due to the long coax used on that thing cuz that's going to have something duting. Sure it's working like a transmission line, but it can only go so far.

So really, you need the utmost in Signal Integrity So if you're going to build one of these things, not only you need to Shield it as you saw the 50 HZ just uh, swamp the thing. And of course he's even got the battery box uh, shielded here. Do Not Drop me. Do Not Drop me.

And there you go. I Love it. So that's a classic build from Jim Williams And of course, this is a very elegant circuit. As you can see, it's effectively just a high voltage source.

It's effectively only. Uh, four parts. Pretty much there's three resistors in here and one, well, sorry, five parts, three resistors, a a capacitor and transistor and that forms an avalanche breakdown. Uh, pulse generator or an oscillator? Um, that is.

You know the breakdowns determined on the individual transistor and the component values. and uh, let's have a look at how Jim Williams explains it. The regulator's 90 volt output is applied to Q1 Via the 1meg 2 pea farad combination. Q1 is a 40 volt breakdown device for the 2N 2369 so it breaks down at 40 volts and then if you go above that, it non destructively.

Avalanches when C1 charges to a high enough voltage. so it you apply 90 volts here, it charges up by the one Meg and then it breaks down because of course the base here is tied to ground. so there's nothing driving. Uh, the base of this transistor.

So this transistor Switched Off it's effectively, you know is Switched Off Not doing anything, but once it reaches that maximum breakdown voltage roughly 40 Volts for this device bang, then it Avalanches The result is a quickly Rising very fast pulse across the 50 ohm output. That's why you have to terminate it in 50 Ohms on your oscilloscope as well. and then of course, uh, there we go C1 discharges Q1's collector voltage Falls and the breakdown ceases and bingo C1 charges back up and it free runs at about 200 khz. And there's the figure that shows the output pulse, which is pretty much exactly what we've got with a little bit more uh, ringing on the output.

But apart from that, that's a very elegant, very simple circuit. and that's a Jim Williams classic. And as always, if you want to discuss this video, jumping over to the Eev blog Forum where there'll be a special thread to discuss this particular one. and uh, this Jim Williams circuit.

And if you like the video, please give it a big thumbs up. catch you next time.