Tek sale

Nice scope, but I really like long record lengths. When the event's infrequent, and long after whatever trigger event, a deep record lets you zoom in. A LeCroy turned me onto that, and I've loved that ever since.

-- Cheers, James Arthur

"John Larkin" wrote in message news: snipped-for-privacy@4ax.com...

After using an Instek scope, I'd buy the Tek without a second thought. Even if its $700 more.

Cheers

I have the 60 MHz dual-trace Rigol, under $400, and it's a fine bench scope. I was going to get the 200 MHz 4-channel version, but the Tek is almost the same price. Tek is now including stuff like FFT that they used to charge extra for... not that any scope's FFT is anything like a decent spectrum analyzer.

"John Larkin" wrote in message news: snipped-for-privacy@4ax.com...

Dave's teardown...

Cheers

I think that an RF spectrum analyzer has a hard time measuring things that FFT scopes get. Anything close to dc is usually tough for rf spectrum analyzers, whereas that is exactly where FFT o-scopes shine.

hnology.com=A0 jlarkin at highlandtechnology dot com

Huh? Why? I have an Instek GDS-2204 and it runs circles around more expensive Tek scope. It has 25000 point record length to boot where the Tek only offered 4000 back then, which made the Tek not very useful for ultrasound jobs. It spews out less EMI (some Teks are notorious for that), and I found it to exhibit lower noise as well. A downside is the low TFT-screen resolution but I typically use it via the laptop anyhow. The built-in screen is amazingly bright though.

I don't do pulse echo work. We have some scopes around with lots of memory, and I can grab one if needed, but I rarely do. For long-record stuff, a cheap USB logic analyzer is what we usually use.

Ten percent off?? Unacceptable; they need to go back to the financial calibration lab!

What cheap USB logic analyzer?

We use this one:

We have the Mictor connector adapter. We can design a Mictor onto a PC board, like on an important bus, and then plug the LA directly onto it, which avoids connecting a lot of loose leads.

We recently used this to verify timing margins of a 250 MHz ADC connected into an FPGA. We came out of the FPGA into the mictor and the LA, and snooped the data that the FPGA had latched from the ADC.

Thats not long record! It only has 2kbit per channel. Compression helps a little bit. And the loading is quite heavy (around 5pf). The protocol decoding is a nice feature but the short memory more or less kills it.

How can you verify such timing margins with an LA capable of 500Ms/s max?

I have a Tek TLA700 with a TLA7AA4 module. A neat feature is that it can route any input to 4 analog outputs which can connect to a scope. I seldomly use more than 8 channels though.

Kills it? It's helped us ship megabucks of stuff.

By running a sine wave into the ADC, buffering in the FPGA, and snooping the ADC data through the Mictor/LA. We tweaked the PLL that controls the offset between the ADC clock and the ADC data latch, looking for the limits where the sine wave started to have hickeys.

That's kind of a monster. A USB LA connects to a little networked laptop PC and saves data easily. The Intronics is tiny and costs $389.

A lot of instruments don't really need a big box and a screen and all that.

The software that goes with the TLA sure is a monster as well. Had to use it because a client sent me their LA results.

That's for sure, I love USB tools. Best of all, that makes them carry-on suitable. So far I am most impressed with this little guy:

I have the 4.4GHz version and the tracking gen. Weeks after I bought it they came out with the 12.4GHz version. Would be kind of nice for stuff that requires EMC up to 6GHz but one has to give up

Perhaps but if you really want to make a trace of two devices communicating over a longer period then the short memory becomes a nuisance.

Thats a bit trial & error :-)

I spend less than $1000 for the whole setup I have. Its the size of a big scope and has a handle so its still portable. And it has network connectivity as well.

I have used the Intronics Logicport device but I wasn't impressed. I also experienced some issues with conflicting USB devices. Another problem is that when the software loses connection to the Logicport only for a short period (USB glitch) the software freezes and you lose all your settings. Happened to me more then once. That sort of behaviour from 'equipment' is a short route to getting me agitated.

So far my experience with USB instruments hasn't been too good. In my dictionary 'pocket size portability' means giving up on features and stability.

Thanks for the info.

What bugged you about the Instek? Or was it that the Tek had something you really liked?

Cheers, James Arthur

We had one problem like that, low bit error rates on a fast serial link, but we fixed that by thinking. Most of the time, a logic analyzer just points out that something wasn't thought through.

It's easier to think than to hook up and use a logic analyzer.

The data window is 4 ns wide, so a 2 ns clock/data shift in either direction breaks it. I wanted to verify the timing margins on every bit. Of course that's trial and error.

How would you do it?

message

Many times it helps al lot if the problem is right in front of you. In many cases I have to deal with stuff other people put together and figure out whether the problem is in hardware or sofware.

Hook up my LA to the ADC bus capture a trace and just see how the data is aligned to the clock. The LA module I mentioned before has 16k of hi-res memory with 125ps resolution.

message

I can't connect a logic analyzer to the 26 traces in question (12 data, one 250 MHz clock, LVDS) and I'd trash the timing if I did. And what matters isn't what's on the PCB traces; what matters is the clock/data alignment *inside* the FPGA, at the first-level data latches. Even more important is to decide if the clock tweak PLL is set dead-center in its sweet spot. There's no way a logic analyzer can get at that. So the thing to do is have the FPGA pump out the latched data to the Mictor connector at a reduced rate and examine the sinewave data.

You can see the ADC (center) and the FPGA just below. I don't think there is any practical way to probe the signals between them, and it wouldn't really help if we could. The Mictor is to the right of the FPGA. We can compile test versions of the FPGA to run any signals we want to the Mictor.