Hi. This is a follow-up video to the Rygaard Es 1054 Z firmware issues that were found in a previous video. If you haven't seen it, click here or link down below and watch that first because otherwise it may not make too much sense. Anyway, Rygel have finally released some new firmware that they claim fixes not only the 5 microsecond jitter issue which we can see here, but also the AC trigger coupling jitter as well.

So I haven't installed a new firmware yet, so we'll try it now. They have actually released a beta trial firmware fix for it before they put it on the Eevblog forum, so huge thumbs up to ride off actually engaging directly on the EEV log forum and and actually releasing it there so people could trial it, but unfortunately that one didn't work and I didn't get around to trying it. Myself and a lot of people on the forum tried it and they said nah, didn't fix it and had other issues that was locking up the keyboard and doing all sorts of weird and wonderful stuff. so Reigai pulled that before I got a chance to try it, but they're finally in the new Year now! So happy! New Year to everyone by the way and Roy Go have released the new firmware, so let's give it a bill, but before we do that, a few users on the Eevblog forum actually went a lot of trouble to try and figure out exactly what was causing this 5 microsecond jitter issue and they pretty much came to the conclusion.

And I think they're right that it was something to do with the PLL The phase lock loop inside the 10:54 says that generates the main, our sample and FPGA clock. Now this is a lower frequency crystal and there's nothing wrong the crystal itself, but it's the PLL chip itself and the loop components around that and this is a programmable chip and they discovered that it looks like it's not locking properly and it's getting jitter and doing all sorts of weird things so that would explain it. If your main clock for your ADC and sampling system and you know everything else, your main FPGA is jitter and badly due to a poor PLL lock or something, you know some sort of issue to do with the PLL then well, that's going to upset the whole apple cart and we're pretty much and we're pretty sure I think most people are pretty sure that's actually what's at fault here and Rygel have They haven't admitted that, but they have said that they have updated the coefficients which get programmed into the PLL chip when the unit boots up and well, that's you know most likely to be the issue or us could be that plus a combination of some stuff in the in this sampling FPGA as well. We don't know unless Rygel actually come clean and admit exactly what the issue is and what they did to fix it.

but I thought we'd not only reproduce the problem here before we update the firmware, see if it works, but crack this thing open. have a quick sniff around with an AMC probe, see if we can get that signal. Let's give it a try. So I consider this our before shot for the five microsecond Jeddah.

Here it is: we've got a 20 megahertz square wave going in. We've got our 5 microsecond precisely offset and I've got infinite persistence on there and you can see pretty much it's doing just maybe a Devere Maybe you know one and a half little graticule positions there. So it's five nanoseconds per division. you know? looks like it's maybe 1.5 Almost two nanoseconds jitter on there.

Not a huge amount, but as I said in the previous video, my unit is not a particularly bad one. Some people have it really horrific aliso and if I push the horizontal to go back to no offset at all. bingo, you'll see it vanish. All right.

So I cracked this open and what we're going to use here to measure the PLL clock in this thing I'll show you a close-up in a minute. We're going to use this you've seen in a previous mailbag. This is the tech box EMC probe set and we've got the 3h field probes here, the magnetic field probes, and the one we're going to use is the E field probe. here.

the magnetic field probes will will work, especially the smaller loop one here, but in this particular case, we really want to get down isolate components right there. So the E field probe is better with its tiny tip. And really, we're not talking about large magnetic you know, loot currents and things like that. So the E-field probe is a better choice for this.

And we've got our wideband ampere so that goes from three Meg up to three gig, so that should be pretty good. We've got our rogue old Esa Eight One Five spectrum analyzer here and we can now sniff around. What I'm going to do though is I think I'm going to have to set up a second camera so that I can show you probing around here at the same time as the result we get on the screen here. Now here's the main culprit right here.

The Analog devices ATF 4360 - seven here and this is an integrated synthesizer and voltage controlled Oscillator is basically a PLL which multiplies the main crystal here. 25 megahertz. You can see that oscillator. Now, this oscillator is going to have bug or jitter.

That's not going to be the issue, It's going to be this phase lock loop and you can see it. So it multiplies that frequency based on these loop filter components around here and the internal coefficients that have been programmed into it by the firmware at startup. and then you can see it's a differential output here and that is the main clock which goes into your ADC over here. I'm not sure if it actually goes anywhere else, whether it's goes into the FPGA or such, but it looks like it might just be for the ADC here.

So let's give it a bell. So if we can sniff around here and get the signal out of this thing now, this is certainly not a tutorial. So I won't bother explaining how parallels work and how the loop filter components and coefficients and all that sort of jazz works. But these these components here they will be our loop filter components.

These are looks like we've just got some output AC coupling here and that's about it. So what we can do is use our probe to sniff around here and see if we can get that output clock signal. Alright, let's have a probe around with our ear filled probe here and you can see how I said. It's got a tiny little tip on there so it really allows us to get in here and look at our components now.

I've set up the DSA 81:5 spectrum Analyzer. Well, I haven't set it up. It's just like pair on defaults. It's just full span like this so we just want to have a look to see if we can find the signal.

Now if we get our probe in here like this and we go very close to our the output pin here to single ended output on our clock now at 25 megahertz clock, there we there we go You can see the clock and all its harmonics right down there. No problems whatsoever. So if we can now go over our PLL chip, we can't see anything on the PLL chip. You can see how discriminatory this probe is.

We've hardly moved it and it's vanished. Now let's move the probe over these loop components up here. Bingo. Look at that.

There is our PLL output frequency right there. Massive spike. A look I moved that away just a little bit from those components and or back over the chip and advantages vanishes. So this is how useful these damn probes can be.

You can get right in there. and of course we can go to the output traces here. not nearly as the differential output here, not nearly as high amplitude as what we get with those loop components. So really, anyway, we can eat quite a decent signal on there.

so now we cannot Center our spectrum analyzer here and just have a look for any outside components on that see how clean that is. But as I said before, we can use this tiny H field probe here. So let this is the smallest one out of the three. I've got so good smallest loop area so you'll get the most the best start discrimination there.

So let's put it over here and look, we can put over the eight ADC here for example and you can see like various components on the ADC there which we couldn't actually see with the E field probe. So this is going to pick up more, but it depends on the loop area and the current flowing. So here we go. What? Hey there we go as a spike that's actually I Think bigger that? yeah, oh yeah, look at that.

it's really going to town there. Check that out. We're really picking that up, but we're also picking up so we picking up much higher amplitude than we're getting before. but we're also picking up lots of other crap down here as well.

So for the purposes of measurement, I think the E-field probe is probably going to be the better choice here. we can really get localized in. but yeah, as far as amplitude goes, we are actually getting a higher amplitude there with that with that tiny H-field probe. But yeah, look, you can pick up a lot of crap at the same time if you're not careful.

Oh, look at that. and if we turn our marker on there look, we can see that is bang on one gig because hey, this is a one gig sample per second scope. is it not? That's exactly what you'd expect and it is bang on according to the Rye goal. Now I've got my probe over the luke components there.

I've got some auto scale happening. We've got one gig Center and it looks pretty clean like that with a span of 10 megahertz. But aha, our bandwidth. Here we go.

Our resolution bandwidth is Bi is the default are 300 kilohertz here, so we need to really knock that down before we can start seeing some components in there. So let's have a bill at that and see what we get. And here we go. If we get a four megahertz span on here with a resolution bandwidth, the 300 Hertz takes a while to build up there, but look at that that doesn't look to crash hot does it? And here it is.

that'll do now. I Don't really care about the particular details of what's actually going on here. All we want to do really for the purposes of this video is to see if this stays the same after our firmware change. With that, with the same setting, same probe position, everything else I Think we're going to see a fairly dramatic change.

Now here's the interesting bit: I've changed the span here to 1 megahertz total and look, we can see these Peaks here. Here's our fundamental here. Okay, and look at the deviation from here to here: The Delt: I've set up a delta of those two markers: 100 kilohertz. Why is that significant? Well, our issue is a 5 micro second multiple jitter problem.

That means at every 10 microseconds, it goes away. multiples of 10 microseconds, it goes away, in, comes back, etc. etc. Well, what is 10 microseconds? invert that? It's bang on a hundred kilohertz? Ah, please excuse the crudity if the model didn't have time to build it to scale or to paint it now.

I'll try and explain kind of what's going on here Now with our little Dave card right now. if we've got our trigger point here. okay, this is our I Know, it doesn't kind of make sense the way I've drawn it. but stick with me.

Anyway, if you've got the trigger point here, you've got your 1 gigahertz sample clock. Okay, now because that is the trigger point of the oscilloscope. you're not actually going to see any jitter there because the clock can be jitter in like crazy. You're not going to see it because at this point on the screen and on your waveform because it's the same sample point each time.

you're only going to see it when you shift the waveform in this direction and then view this one in the center 5 microseconds later. Okay, this is the problem. and let's say you've your 10 microseconds later. Well, there's our 100 kilohertz.

Okay, if this as main sample clock is jittering effectively being modulated by a hundred kilohertz, that's a period of 10 microseconds. So once again, you're not going to see it there because it's going to match up precisely with the jitter time. But if you go halfway, ye, half that value of 10 microseconds. 5 microseconds? Bingo.

You're going to see all your jitter in there. And then likewise, at 15 and so on. So I hope that's as clear as mud. Hmm I think I could have come up with a better explanation that anyway, right? So we've got that little bass line there now.

let's upgrade our firmly. For the record, this is what I am running now 4.01 dot SP 2 on 1.1 board revision. Ok, let's update the firmware here: I've installed the file on the stick that's always supposed to do the GTL There we go bang. Caution: Newer software version detected.

we're in Story 4.0 2.0 Four Point Oh Seven. Do you wish to continue? Please do Not remove. Come on. You can do it.

could take a while. Should I twiddle my thumb's I think it's worthy? Isn't it? Come on, Come on. Anyway, what we really want to see is a change in that PLL the PLL coefficients right? I'll have said they have changed them so they have admitted that so we should expect to possibly see a a changing that spectrum there. Congratulations Update was successful.

Now you can restart the oscilloscope and enjoy it I will enjoy it if a bloody well works. So yeah, I guess we just do you power and power backup. Remove the stick of course. And as usual this scope does take a little while to boot which is a downside of modern scopes, but all with their all new modern whiz-bang operating systems.

you know? Yeah, what do you expect? Anyway, we're in, we're in like Flynn so the utility system system information so no problems at all. 4.0 to SP for I Have no idea if this keeps the you know any upgrades and things you've got or any any software upgrades or anything presumably it would Alright here's our test: I'm putting the probe basically where I had it before and let's go in there and okay here's the test, putting the probe pretty much where I had it before. so let's do a single shot and see what we get being go significantly different one megahertz span there I Haven't changed anything else on here. so I What you the waveform you were saw before was this is what it has changed to.

there you go. So they certainly have updated the coefficients of this down in here. They've They've changed that because it's got a spy bus on there which interfaces to the main micro and then it has to program in those coefficients based on the particular loop filter components they got on there. and there's a lot of tricky theory which all goes behind this.

and well, yeah, I won't go into it, but hopefully my God have done their homework. now. let's repair it and see if there's any jitter. All right, there's our 20 megahertz signal at no delay.

Zero Point Zero Zero Zero Zero Zero. Pico Seconds Got to be me Anyway, that's just crazy. So we'll change your horizontal here up. Sorry.

we'll go this direction so we'll tweak that up. I'm holding my tongue at the right angle just in case. and five microseconds and tada. look at that it that looks pretty good.

You can't complain about that. Like I can't zoom in any faster than that because that's the lowest time basis. Five nanoseconds per division on this site. 1054 Zed You might get next year step there if you've got the hundred Megahertz model.

but I've only got the 50 Megahertz model. So there you go, that is that is gone. and I've got infinite persistence on there of course. and let's just all hit the horizontal button so we'll jump back so like we can turn off the infinite persistence there.

There we go. but it's certainly it's certainly changed. so I can certainly consider that fixed. Although I would have to go in and check for any other multiples in there or anything like that, you can pretty much be sure if you can measure it on the spectrum analyzer, then well, you should be able to actually calculate where any of these issues should be, but looks like they fixed it.

but that's fixed it for my scope. They had the same issue before when they're at least the beta version of this. Fix it, fix some people's machines and everyone said yeah, problem fix Thank you very much for I go And then other people came out and said no made it worse. All sorts of things because those loop components there, they're gonna have very significant tolerance.

like five ten percent tolerance on those things, let alone over-temperature in and everything else that can be involved. So you know, Really, Yeah, it's okay. it's fixed mine. So fantastic.

I'm a happy little camper. Let's see if it's fixed the AC triggered. You do put that back to the start there. that's going to our trigger jitter menu.

You've got to go down here to set in coupling. and this is trigger coupling as I've done a video on it is not input AC coupling. It's AC trigger coupling. A big big difference.

So let's do that. And Tada, we're a happy little happy little camper there. If I let's turn off our infinite persistence because I was I just varying the offset level control there. So let's go into display, change that back to minimum.

thank you very much. So I can change our level there. and of course it's going to lose it because I'm adjusting the just in the trigger level so it's obviously like going above the waveform and it just doesn't know what to do There We go. set it properly.

the X one scale factor on Channel 1 here. So yeah, you'd expect it to where I go off trigger at when it gets to about 2 and a half because we're talking 1 volt, 2 volts. You know, two and a half there. And that's what we start to see and bang at about.

Yeah, two point seven or something. So that's right down on the negative side. So our AC trigger coupling is now coming back down. Let's go.

sorry it's going negative and about the same 2.6 about 2.5 volts or there abouts. Yep, that's not too shabby at all on first glance here I see trigger coupling problem also fixed. Very nice. and if you're curious to know about the new PLL signal here, well our Delta there is about 68 point 3, 3, 4 kilohertz so we need to invert and then have that and that gives us an offset around about seven point three, one microseconds.

and here we go: seven point three one microseconds. Well I don't see any noticeable jitter in there at all. So yep, I think that's pretty well fixed. So although the problems fixed I did.

actually, if you probably watch this in HD you can actually see a little bit of fuzziness to the line. There is a slight bit of juda there compared to the regular one. but I really, you know you're really pixel peeping there. But anyway I still really have I Do really have a still have a problem with this, although hey, the jitter problem on my particular unit is fixed.

Okay, fine, but this is a pretty awful clock. I mean you know ordinate in this a little bit I've now got a 500 kilohertz span here and and I've got our dealt. There's a Delta on there, there's sixty seven point five kilohertz there, and then we've got another little peak which is shown itself here and really none of this should be there. and it's like only like 37 DB down from the carrier here.

I Mean that's it's pretty piss-poor so the output from that PLL is still not great. So I think my goal need to fix that I Mean they need to fix that solid whether or not that, whether or not they can even tweak the coefficients even further perhaps and get a better or whether or not they have to actually change their loop component values in hardware. And it wouldn't surprise me if Rygel just didn't mention this. And like all existing unit and this firmware, fixes all existing units, but secretly in the background, newer units might actually have dear front loop value components.

That wouldn't surprise me at all because this is pretty bloody awful. I Mean you know a proper one should be like it should look. You know it. None of this stuff should be here.

It should be like right down here like this. it should be very nice like that. so we shouldn't have these components in here, but it doesn't seem to. I mean I've checked the multiple frequency of that and it seems to be okay.

but yeah, that's that's a still a pretty bad clock. but hey, you know I mean my goal between a rock and a hard place here. they don't want to have to recall every unit that's already been sold to actually fix any hardware loop component values that would be. You know that would be absolutely horrendous and they have to weigh it up against the actual performance of the scope.

and well, if it's proven hasn't been proven yet. But if it's proven that every it fixes everyone's unit out there and there's no reports of it, you know, different spreads of units. We've already seen that issue. A lot of people have a huge spread of issues, so if we can get all those same people and even more to report whether or not this is an issue, we'll know if it's a solid fix and if it's a solid fix, Well, it's a solid fix.

even if this clock is still pretty pretty crusty. Well, you know now I think Rago do need to fix this though I don't like it. Still gives me the heebie-jeebies So there you have it. I'm pretty happy with the fix on mine.

It seems to work well, but yeah, the proof is going to be everyone else when we've got actually, you know, students, statistically significant numbers and spreads of units and component spreads and all that sort of stuff. So if it, if it works out to be a solid fix, great, you know I'm happy. But yeah, I there's still a chance that it may not fix, everyone's unit. So that's going to be Rygaards problem.

Oops, we've retreated there. It shouldn't have done that. Wonder why I did that? So yes, right. I'll still need to be are vigilant on this issue and as I said I still think they need to fix that clock? That's pretty piss-poor but hey, you know it's as arguments like is it good enough to do the job if the scope.

if there's no measurable difference on the scope world, does that clock matter? Is it good enough? Yeah, you know. Ultimately, yeah, it's a bit academic that okay, it's not the world's greatest clock, that's for sure. It's pretty piss-poor but if it does the job on this particular hardware and nobody can measure the difference on the scope itself, then yeah, what does it matter? So I hope it's a solid fix if it is fantastic, but yet. wouldn't be surprised if they sneak in a little hardware fix and they you know, like this thing shouldn't have happened like the design engineers should have damn well measured the output of the ATC clock there.

but hey, maybe they're like, maybe they did. And and and it was fine because a lot of units out there show absolutely no problems whatsoever in the PLL locks and everything's You know, everything's hunky-dory so but these didn't test a statistically significant number. So yeah, it slipped through. Um, oops, it was pretty embarrassing, but yeah, big thumbs up to Wry Goal for engaging the Eevblog community, ultimately fixing this relatively quickly.

so there are a bit premature on the initial beat up firmware release, but hey, you know they did pull it back which was great and now they're well on my unit. So everyone please, if you've got one of these, try it out. and well, yeah, the review is still out, but you know this is a pretty damn good scope. As I said, even with those issues, a lot of people still give this that scope a big thumbs up and on mine.

Um, I could maybe test it across our temperature or something like that. But anyway, keep reading the Eevblog forum down below for updates on this because that's where all the action is gonna happen. So sorry about the length of this thing. It was supposed to be quick as all my videos are, but you know I get a bit carried away sometimes.

Oops. So there you have it. if you liked it, please give it a big thumbs up on YouTube because that helps catch you next time you.