TDS-1002b Any good? Comments?

It's a simple matter of that if you have it available you'll find a lot of uses for it you didn't know you had before. If you don't have it available you make do and don't realise what you are missing.

Dave :)

On 3/1/2007 8:36 PM, The digits of john jardine's hands composed the following:

I sometimes need to see a signal that is hard to trigger on so I trigger on something else, which may be later. For example I do a lot of CCD design. I often trigger on the reset pulse which happens at a different time than the image data I am looking for. This helps me know that I am looking at the same pixel each time because the image is referenced to the reset not the picture. I would hate to think I am looking at a black reference pixel when I am actually looking at a live data pixel.

Hawker

That really surprises me. In my meager experience I have wanted to know what happened _before_ the trigger many times. I, for one, can't wait to find out. ;-)

As a master's project I designed a CCD camera from scratch. To be able to look at one CCD pixel I borrowed a HP scope from the RF institute next door because I knew someone there. By today's standard that scope would be considered pre-historic. Then I built myself a "crutch", a little precision delay circuit so I could trigger on the previous line end and scoot the scope trigger via a 20-turn potmeter (the scope did not have delayed trigger).

Sure enough, the millisecond I turned that scope on for the first time .. POOF. Dark smoke came out. One of the PS transistors had decided to become a rocket and this was the darling scope of the RF guys :-(

But luckily we had such a transistor and it hadn't taken much downstream of it into the grave.

I would bite the bulet and buy a new one. I bought 2 second hand ones from a dealer and they stopped working very soon after I got them. I would go for a modern new scope that fits what you are trying to do.

Seems to me you're missing something. Moons ago scope manufacturers spent a *lot* of money on delay lines and dual-timebase modes to do what is trivial with digital storage and $.29 worth of code today.

Do you never trigger logic analyzers on anything but the start?

that

the

sure do

donations

people

Yes indeed, for logic it's essential!. Maybe that's the difference I'm puzzling over, in that my scope gets used about 95% for analogue. For logic and micro stuff I've a HP1630D. Must admit though I'd rather run a mile, before being arsed to pulling it's bulk off the shelf, clip on the mess of pods and probes, battle with the lunatic menus and listen to the horrendous fan noise. john

Analog scopes can trigger off other than the source too. Delayed (run A after B, A delayed by B, etc.) are very powerful triggering modes for analog scopes. Digital scopes make these sorta obsolete.

Well... ;-) That wasn't my point.

You missed a pretty significant point here. The Intronix unit (

I shopped around quite a bit before I chose the Intronix product. There are several USB analyzers in this price range with more memory, but all sacrifice something in exchange for it. They either have much slower sample rates, less sophisticated triggering or a lot fewer channels. Some of them have pretty pathetic input characteristics as well. I saw one which claimed to sample at 500MHz, but only had a

Advanced trigger features are particularly important for digital analysis were the data stream tends not to be repetitive and can literally go on forever. In fact, I'd challenge anyone in a "deep buffer" versus "sophisticated trigger" contest. An appropriate trigger setup can capture data which no amount of buffer depth could find. For example, an intermittent problem which takes hours/days to occur. How much memory would it take to capture the entire period and then search back through it for the problem? When set properly, sophisticated trigger features allow an analyzer to monitor signals for days, allowing it to capture just that eureka moment ;-)

Jim

Millions?

I've looked at a bunch of stuff, and I think I'm going to get one of these fairly soon. One thing I do think they could improve upon IIRC is the speed over the USB link.

And yet another reason that I really liked this one.

The data analysis/decoding is another big plus for the LogicPort. I could really make use of that.

I see them on e-bay for $10 off the normal price, but I'm holding out for a better deal. ;-)

Oops, I didn't see that it used transitional sampling. Transitional sampling is good, but it's no substitute for a good deep memory. Transitional memory is only as good as the number of samples captured by the highest frequency (most transitional changes) signal being captured. i.e no point having one channel only changing slowing if the other one changes fast and uses up all the memory. Of course, you can also have transitional memory on a per channel basis too, but then you don't get to see the other channel data that have already had their memory used up.

That's what buffer probes are for, you can make your own low capacitance probes for any logic analyser if you need too.

Of course, good trigger features are essential for any logic analyser. Stuff like serial data decoding and sequencing is invaluable for instance.

No one uses a logic analyser like that, you *always* use smart triggering, that's not what big memory is for. Big memory allows you to capture *detail* at the trigger point and also *detail* much further back or forward in the signal. Transitional sampling can't do this.

Yep, that's the idea with *any* logic analyser. Goos logic analysers need comprehensive trigger features *and* big memory.

Dave.

Both techniques have pros and cons but without transtitional sampling, even 1 or 2 megasamples of "good deep memory" gets filled in a hurry. I have an Agilent 54622D sitting on my bench along side the Intronix unit. The Agilent unit has 2 meg of memory, but doesn't have transitional sampling. When I need to do pure logic analysis, I find myself reaching for the Intronix unit first. Mostly because it has more channels and more speed (you can never have enough of either). I've also become partial to PC-based instruments because I can view the signals on a 21" monitor instead of a tiny 6" screen. It's also very nice to use a mouse and full sized keyboard for setup. Naming signals on the Agilent unit is very tedious - one character at a time, scrolling through the whole alphabet to enter each character. I use the Agilent only when I need mixed signal analysis (or just a digital scope).

You must be joking. You're going to buy a logic analyzer and then have to turn around and build/buy an input buffer to get it to work adequately? Any buffer will add skew and potentially noise or other signal fidelity issues which can greatly degrade its high-speed capability. This characteristic is the manufacturer's responsibility. I wasn't shopping for a kit.

No, not *any* logic analyzer, because not all of them have that capability. As I said before, give me sophisticated triggering, and I'll take on any buffer depth ;-)

Jim

It can, but you missed my point about the limitations of transitional sampling, it is no substitute for a large deep memory. The best logic analyser would have a combination of transitional sampling and a big memory.

I can hook an external large screen LCD monitor up to my Aligent mixed signal scope, very nice. Yes, the single digit name setup is a pain, but for my use the portability of a standlone logic analyser is usually more important. Each to there own, but a PC based logic analyser aren't a bad compromise.

Skew is not an issue if all the signals are buffered equally including the clocks. Most good logic analysers come with external buffered probes, otherwise you are going to have some mighty short test leads ...

All logic analysers have at least the basics, word pattern matching and don't care states across all channels for instance. Good enough for most general use.

Depends entirely on your requirements. I personally find that for my use the deep memory and basic triggering is more important than transitional sampling, e.g. to see detail separated by large idol periods. No amount of fancy triggering and transitional sampling will let you do that. A tranistional sampling analyser with a small memory would be useless for many of my recent applications.

Once again, depends on the application. But transitional sampling is

Dave.

Just realised that isn't an issue with the Intronix unit due to it's small size, just move the whole unit close to your circuit.

BTW, I was saying you'd only make your own probes if you really needed too. Usually you buy them of course.

Dave.

I never said it was a 100% substitute. I said that it extends the capability significantly. I also said that given the choice of deeper memory or more sophisticated trigger features I would (and in fact did) choose the trigger features. The Intronix unit had the best of any I could find, along with transitional sampling and other nice features.

And it would also have a rich set of trigger features. But you will not find that "ultimate" logic analyzer in the sub $1000 range. In that price range you will have to compromise. I thought that was what this discussion was about - picking the right mix of features within a given budget. If you have an unlimited budget, then this whole thread is meaningless to you.

The $5000 Agilent 54622D won't do that.

A stand alone unit is much less portable than a typical PC-based unit (the 3 ounce Intronix analyzer for example). When I'm doing digital work - which is almost always, I of course have a PC nearby. What else would I be using to develop the software/logic I'm debugging? Moving a PC-based analyzer from PC to PC seems to me like much less work than lugging around a 20 pound "portable" stand alone unit. It also fits nicely in my laptop bag for field work.

If only that were possible. Anything you can use to buffer a signal has delay. Delay between multiple buffer channels will not be equal. With a lot of effort you can make them very close to equal, but again why bother? Just choose a properly designed analyzer to begin with, and it's no issue.

By the way, you keep mentioning an off-the-shelf "buffer probe" which you would buy rather than having to build one. Could you give me the model / brand? I'd like to have a look at that.

"the basics" /= "smart triggering"

If you believe that basic word pattern recognition constitutes sophisticated triggering, then I understand why you rely so much on deep memory. I will agree that deeper memory is sometimes a reasonable substitute for deeper thought.

You have described the ideal scenario for transitional sampling. Specifically, "detail separated by idol periods". That situation gives the greatest compression possible. Have you ever actually seen transitional sampling in use? It sounds like you haven't. Take a look at the Intronix website (under screenshots / sample compression). They have an example showing about 10 megasamples of data captured in the "2k" buffer. As a more extreme example, I have personally sampled several seconds worth of serial traffic separated by idol time, all in a single acquisition with 5ns resolution. That's hundereds of millions of samples. Will it always do that? No of course not. Can your deep memory Agilent unit do that? Not a chance unless it has transitional sampling.

Jim

Sure, it's all a matter of what features set you need for what price. I'm not bagging the Intronix unit, it looks like quite a capable unit.

The 54620 series has been replaced by the newer 6000 series, at a similar price level. I have one of each. I thought the 54600 series was excellent, but the 6000 is so much better. I still use both though. They could use some more advanced trigger features on the logic analyser thoush, but considering it's just got a logic analyser tacked on, it's very useful.

Once again, horses for courses, depends on what you are working on. For me, I don't particually like PC based instruments, I can't always bring things to the bench near a PC. I also don't trust the PC to stay up for a week if I'm waiting for an elusive trigger signal.

Of course, I'm just pointing out that that it *is* possible to make your own probes if you need super-duper low capacitance or whatever for a specific application. Just like it's not uncommon to roll your own scope probes, but digital ones are not as easy of course.

There really are no "generic" ones, either the manufacturer offers a range of suitable probes or they don't offer them. The Agilent 54600 and 6000 series MSO's come with very nice buffered probes.

I did not mean imply it was.

I don't think that, I was just pointing out that all logic analyser come with basic triggering and often that is suitable for most purposes for most people. If you need more complex triggering then you need it, simple as that.

I know that is the ideal scenario for transitional sampling, but in the case of the Intronix unit we are talking about it only has 2K points of memory. My needs often call for capturing say thousands of cycles, waiting a period, capturing a few more thousand, and repeat XX times. In this case the 2K points of memory on the Intronix unit would make make it useless for this task. That does not mean it's bad unit, it's just unsuited to some tasks I do. For some tasks I just have to drag out the $XX,XXX logic analyser, but it seems the more you pay the harder they are to drive!

I've done more than that, I've designed one, thanks.

Can the Intronix unit capture more than 2K transitions? No. Horses for courses.

Deep memory is good. Transitional sampling is good. Complex triggering is good.

Dave.

Oh so now you've changed your mind. Previously I pointed out that many of the PC-based units had excessively high input capacitance. You mentioned twice that a "buffer probe" would solve that problem, and that rather than build one "Usually you buy them of course". Now you're saying that only the manufacturer can supply them. None of the PC-based manufacturers actually do offer them - so you are on your own. I'll restate my original point: Be sure to take a close look at input capacitance when choosing a PC-based logic analyzer. Some of them are excessively high, and this will cause you grief.

By the way, the probes which come with the Agilent 54600 and 6000 series *do not* contain any buffers. They are completely passive, with no acitive circuitry at all in the probe itself. I would have expected an expert like yourself to know that.

Previously you said that "No one uses a logic analyser like that, you

Please tell us more. With your vast knowledge I',m sure you've designed a great analyzer. How much does it cost? Where can we buy it?

Never claimed it could. Most of the analyzers I've used over the past

Yes, all of those things are good, but you won't find them all in a sub $1000 PC-based logic analyzer. Spending $XX,XXX on a logic analyzer (and then bragging about it) is not the right solution for everyone.

Jim

If none of the PC based scope manufacturers offer them then you obviously don't have a choice. In which case be careful in what you buy up front or be prepared to make your own probes if needed.

I am not doubting your original post, yes it is an important point.

I could have sworn I opened the 54600 series probe box once and it had some active circuitry, but that was long time ago, so ok, my mistake, sorry. I may have been confused with another LA, I've used a lot of them.

You have misinterpreted what I said again. What was I was referring to was your inference that I simply hit the trigger button and was relying on the deep buffer of the scope, but of course no one uses a logic analyser like that (that I know of), they always use some form of minimal triggering to find the event they want and then get any required pre and post data. Sorry, the word "smart" probably wasn't too appropriate there, I should have said "standard" or something that would not confuse you.

It was a follow up design to my 40MHz 32 channel PC based logic analyser published in Electronics Australia in Oct/Nov 96, but it never made it to publication, sorry to disappoint you.

Well that's just fantastic for you, and that's all I need *most* of the time too. But recently (and at other times) such a logic logic analyser would be essentially useful for me.

What is your problem Jim?, I was just pointing out that I've had to use (many times) a logic analyser much more powerful and with much more sample memory than the 2K (transitions or otherwise), it does happen you know, that's why they make bigger and better logic analysers. If a logic analyser does not let me see what I want to see because it's got a small memory, then it's essentially "useless" for my task at that time. But if that's all you got, then, well, you make do with it some how.

I never said you could.

I never said it was! and I was not bragging, just pointing out that I've had to resort to such high end logic analysers on quite a few occasions.

Chill out a bit, I'm not out to get you, or to bag the Intronix unit, really.

Dave.

Ah, judging by your newly displayed email address you either work for Intronix or you *are* Intronix, in which case that explains why you have such a bee in your bonnet! ;-)

Dave.