HP1650B logic analyzer vs Open Logic Sniffer

[quanttrom]

I have the opportunity to buy a HP1650B logic analyzer and I am wondering if it is worth it.

I currently own a Open Workbench Logic Sniffer. I have used it a couple of times but it feels more like a toy rather than a real world useful logic analyzer.

From it's page I notice the Open Logic can analyze signals up to 70MHz and a sample size of 24k samples when using 8 channels

The HP 1650B seems to have a memory for 1024 samples/channel and max frequency of 100MHz.

Do you personally think the 1650B is a logic analyzer that I should spend money on ?

PS. First post on this forum I hope it is in the right section

[alm]

I believe I've written on this last month or so. The short version is that the old boatanchors usually have more channels, higher sampling rate, often more sampling memory (not in this case) and more advanced probing solutions with less circuit loading (the pods, which often go missing, which makes them worthless unless you can acquire the pods). They were usually designed for connecting to the address bus of a micro processor, and would need support for each type of processor. Analysis of serial protocols is usually not their strong point. Connectivity is GPIB if you're lucky, which requires expensive ($150+) interfaces to connect to modern computers.

The FPGA-in-a-box type of logic analyzers is cheaper, smaller, allows you to analyze data on your computer and usually has support for modern serial protocols like I2C, 1-wire or CAN. In ten years, the old hardware will probably be more useful, since the software for the FPGA-based ones may be no longer supported and the computers may not support USB anymore. Stand-alone equipment is less susceptible to this. But for something like $50, I wouldn't worry too much.

If you have to ask this question, the answer is no in my opinion. Unless they fulfill some specific need or you work on the old hardware it was designed for, there are probably more useful things that you can spend money and bench space on.

[Balaur]

At work, I'm using a variety of older Tektronix LAs (basically PCs with all the acquisition logic).
There are two redeeming factors in my opinion for using dedicated instruments:
- better input stages: works with a variety of logic standards, thresholds, presumably good timing control, etc
- sophisticated triggering: algorithmic sequences from the simplest value-based ones to very complex scenarios. This will minimize the need for very deep sample buffers.

I don't know much about HP user interface, but Tek's is a standard Windows program, fairly comfortable to use.

That's being said, I'm not seeing myself using one at home. Too big and over-qualified for my simple needs. I'm planning to build a Open Logic Sniffer-like clone using some FPGA board that I have. While I foresee a lot of headaches and discomfort using SUMP, it is something that I will have to deal with.

Cheers.

[mikeselectricstuff]

I think it depends a lot on what you want to use it for.
Standalone gear has a lot of advantages, but things get painful when you want to pull data onto the PC for analysis, or just handle a lot of data. Even doing a screen print from older kit like this can be tricky. The older kit has very limited acquisition memory (1K I think for the HP), which can be very limiting, and require a lot of setting up of fiddly triggers.

One thing deep memory gives you (on a scope or LA) is you can often just grab a big chunk of data with little or no trigger and then zoom into the area of interest.

Given the choice of the 2 producs  you mention, I'd definitely go for the PC+FPGA over the HP - 24K is a lot more useful than the HP's 1K.
Note the HP does do some tricks to save memory by only storing when a channel changes, so you can get fine detail on long aquisitions, i.e. 10ns resolution on a 1ms pulse. I don't know if any of the  FPGA based ones do this.

[alm]

Note the HP does do some tricks to save memory by only storing when a channel changes, so you can get fine detail on long aquisitions, i.e. 10ns resolution on a 1ms pulse. I don't know if any of the  FPGA based ones do this.

Yes, most of them can. It's only simple run-length encoding, though, so if you probe a clock signal, you'll only get <512 cycles with 1k storage, even if nothing is being transmitted.

[quanttrom]

Thank you for the replies guys.

Based on what you have told me I think I will hold out for now. Like alm said, I don't have a specific need for this LA and it won't serve me any specific purpose.

Most of the protocols I will be analyzing will be the serial I2C, 1-wire, etc and deep sample memory is definitely useful.

That said, I am fresh out of school and I am just starting to acquire equipment that might be useful. If I absolutely need to, I can always run to school and do testing there.

[ArtemisGoldfish]

Definitely worth buying a USB-based logic analyzer compared to the 1650B. I have a 1652B (not sure how it's different from the 1650B, the 1652B has 80 channels, 100MHz timing/35MHz state, 2 channel oscilloscope) and it's alright, I suppose. I like that it can print through my LaserJet 5, and I basically got it for $30 (I got it for free and had to recap the power supply.) The logic analysis part is alright, though it'd be nice to have SPI/I2C/UART analysis capabilities on it. The oscilloscope capability is weird, too, not going to replace my 30MHz analog deal any time soon.

[baljemmett]

Definitely worth buying a USB-based logic analyzer compared to the 1650B. I have a 1652B (not sure how it's different from the 1650B, the 1652B has 80 channels, 100MHz timing/35MHz state, 2 channel oscilloscope) and it's alright, I suppose.

Sounds like the 1652B adds the oscilloscope; my 1650B sounds the same apart from that, and the 1650A is 80-channel 100MHz timing/25MHz state.

I like the 1650B for timing analysis -- works pretty well as long as you don't need too great a depth.  It only seems to want to store 128 post-trigger transitions; would be nice if that could be expanded but I haven't figured out if it's possible so far.  It's enough to check the timing of a short serial datastream but for real work a more modern system with effectively infinite depth and decoding will be significantly more use.

Its state analysis is geared towards parallel buses, and so far has struck me as a bit fiddly -- I can't, for instance, get it to capture 4-bit LCD communications and show ASCII bytes, although from the manual it sounds like it shouldn't be beyond its capabilities.  Have only had it since the summer (bought from eBay on a whim because it was cheaper than a Logic and I could do with more channels) and not needed to do too much on that side of things though.

In short -- yeah, get the USB first and leave the hulking great pieces of kit for when you need dozens of channels instead of modern decoding!