Logic Analyzer Basics

Question

I am buying a second hand logic analyzer (Agilent 1661A) and I see the terminology such as the following:

• 100MHz State Speed
• Maximum Conventional Timing Rate: 250 MHz full channel, 500 MHz half channel
• Maximum Timing with Glitch Rate: 125 MHz half channel

And there are many more. I expect to measure 50MHz signals between FPGA and other custom ICs.

I have the following questions:

• For 50MHz measurement, what is the minimum speed I should pick up? (If this were a scope, for 50MHz, I would probably need 500MHz scope to see the signal very clearly)
• I need to measure set up and hold times and some other timing parameters, how does the speed of the scope impacts these.

Answer 1

A logic analyzer generally has two modes of operation, depending on where its sampling clock is coming from. The clock can be coming from your circuit, in which case, the samples are synchronized to the operation of the circuit, or it can be generated internally by the logic analyzer itself, which means that the samples are not synchronized. I tend to refer to these as "synchronous" and "asynchronous" operation, respectively, but Agilent is using the terms "state" and "conventional timing", respectively.

State mode is useful for tasks such as capturing the bus activity of a microprocessor to trace how it's executing software. Conventional timing mode is used to measure actual delays between signal transitions.

The specification is telling you that in state/synchronous mode, the 1661A can accept a clock of up to 100 MHz, but in conventional timing/asynchronous mode, it can generate a clock of up to 250 MHz, and by interleaving the channels it can achieve a 500 MHz sample rate.

At the highest rate, the analyzer is taking a sample every 2 ns, so this is the limit of the resolution you can get for measuring things like setup/hold times. Also, keep in mind that any transition you see on the screen could have actually happened at any time ± 2 ns from the position shown, so if two transitions are shown within one sample interval of each other, you can't be certain which one actually occurred first. This makes it tricky to measure small intervals with a logic analyzer, and an oscilloscope is often a better tool for this sort of task.

Answer 2

I don't know the Agilent, but a logic analyzer is not a scope. It will show you what is happening at the bit and protocol level, not at the signal level like a scope. The trigger level for where a logic analyzer considers a signal a "1" may not even be the same as that of the target/destination chip, although good analyzers let you set this (and other similar) parameter.

If you want to measure slew rate, ringing, etc, and see signal transitions, get a scope. The analyzer will give you 1s and 0s -- but a lot more of them, with more smarts, than a scope.