First oscilloscope thoughts

[stj]

Re: First oscilloscope thoughts

waste of time, it will be dependant on the port interface and the cpu speed.
also if it's not buffered there will be a load from the pullup resistors.

if you need a cheap L.A. then you can get a chinese Saleia compatable usb one for about 6$ on ebay.
they work really well upto about 24MHz.

[magistertc]

Re: First oscilloscope thoughts

Look at the readme -- author has a point about unsuitability of USB, though ver 3 USB is somewhat faster. I used this paralllel port interface to sort out some timing issues in an MCU set up for matrix switching. With all possible services stopped, the only interference was from the PC clock signal, which I had to occasionally work around. I don't know what jitter would be like unless your signal is introducing it.

IMHO a cheap USB logic analyzer is probably wasted.

[stj]

Re: First oscilloscope thoughts

but it costs less than making a parallel-port one, and the software is good!
https://sigrok.org/

[eccerr0r]

Re: First oscilloscope thoughts

Well, USB should be fine but it will require more hardware to implement than parallel port, though the isochronous modes will help allowing a parallel port like PC-based sampling system (versus hardware based). The comparison he writes is making the assumption that it is PC based and needs to poll the USB subsystem for each sample, which is the wrong way to do things. Instead using hardware to set up iso mode will result in much higher sample rate. Better yet, trigger and sampling should be done/queued in the USB hardware, and the 1ms delay is no longer relevant.

I'm kind of concerned of jitter because we're not running a RTOS (i.e. Windows(/Linux)) to do the sampling. Having the complex USB hardware is one way to make this OS-based jitter go away. The 24MSps sounds like it's using the iso mode, or packet mode and buffering within the device and probably doesn't have OS-based jitter problems.

500KSPS parallel port is actually fairly crappy, and sounds like it's worse if using all three read channels on the parallel port, probably due to hardware emulation limitation. Unfortunately parallel ports still need to emulate ISA...

[stj]

Re: First oscilloscope thoughts

the usb ones use a dedicated usb microcontroller that uses dma to stream an 8bit port at full speed.
it's not polled.

[eccerr0r]

Re: First oscilloscope thoughts

I kind of figured as much, making the 1ms/1KSps reported by the USB vs parallel discussion simply bunk.

Are these USB devices continual sample and the PC does the "trigger," or does the onboard circuitry have the ability to trigger?

[Curious.George]

Re: First oscilloscope thoughts

I don't know what jitter would be like unless your signal is introducing it.

Jitter is the sample-clock to sample-clock variability in the measuring device/system. If there is a "clock" that is actually sampling the signals, then the jitter is essentially the stability of the oscillator behind that clock (and any variation introduced by the circuitry that qualifies it). If there is some software involved in the actual sampling of the signals, then the jitter is the variability between the "loop iterations".

The latter can be caused by different opcode sequences in each loop iteration (e.g., if you have any conditionals in the loop), different execution times for fixed opcode sequences (does "0*27" execute faster than "34234234*27"?), the affects of cache (first time through the loop may prime the cache -- at a slightly higher cost in time), NUMA, etc.

IMHO a cheap USB logic analyzer is probably wasted.

That would depend a lot on what you were expecting from it. If, for example, all inputs are known to be sampled simultaneously, then you can still extract transition ordering and/or causality even in the presence of jitter: signal A changed state before signal B (sampling frequency has to exceed the rate of signal changes).

Surely you could use even a dog-slow LA as an RS-232 protocol analyzer (e.g., to note the relative times that handshake signals change wrt the data streams).

When I design a processor or other complex state machine, I will drag out a logic analyzer to monitor several dozen points in the circuit. Then, step the machine with a push-button-clock (i.e., run it at DC). Because I can now see those dozen points, I can watch all of them changing as I press the button ("Yes, the program counter is being incremented; the instruction address is being placed on the address bus; the op-code is being fetched; and decoded; etc.")

Doing this with a 'scope -- or even a logic probe -- is just torture! And, prone to much error as you are constantly moving the probe around and may probe the wrong point this time...

[stj]

Re: First oscilloscope thoughts

I kind of figured as much, making the 1ms/1KSps reported by the USB vs parallel discussion simply bunk.

Are these USB devices continual sample and the PC does the "trigger," or does the onboard circuitry have the ability to trigger?

they are dumb serialisers, they just stream at the selected speed.
the pc does the rest.

[eccerr0r]

Re: First oscilloscope thoughts

Odd... Saleae has a website but the cheap $6 Saleaelogic LA's are not on their website... Are these hacks? Oh my, I think they are... someone reverse engineered their LA and put a fake name on it?

[stj]

Re: First oscilloscope thoughts

neither,
they are based on the original 8channel saleae that is no longer made.
https://sigrok.org/wiki/Saleae_Logic

they arent the first to clone it either.
http://www.usbee.com/zx.html
lol

[eccerr0r]

Re: First oscilloscope thoughts

I mean, what is this? This is a clone/fake that should be illegal?

https://www.ebay.com/itm/24MHz-8CH-U...0/272337085235

No relation of course... I'd say this should be illegal and should be reported? Actually nevermind this should be OK. I thought I saw one that actually printed Saleae on it... now that would be a trademark violation.

[stj]

Re: First oscilloscope thoughts

it's neither.
it just happens to be compatable.
you see, salaea had the cheek to use a reference design from the cypress datasheets for the hardware.
so they arent in a position to complain or sue other than use of their VID/PID in the eeprom.

the USBEE has it's on VID/PID so is untouchable.

[eccerr0r]

Re: First oscilloscope thoughts

Ah... crappy, I guess then saleae should have never gotten recognition for the hardware design, it's a reference design after all. Then the software is another issue if Saleae wrote their own software and people are copying it. At least it looks like usbee has their own.

I might have to get one alas wish there was more than 8 channels...

[eccerr0r]

Re: First oscilloscope thoughts

Speaking of sigrok, I wish it supported GPIB devices, or does it do so? Would be nice to use it with my GPIB 'scope...

[stj]

Re: First oscilloscope thoughts

if your usb is upto it, you can use the cheap la's in sets

[eccerr0r]

Re: First oscilloscope thoughts

That is definitely asking for problems, end up with skew between the cheap LA's unless there's a way to synchronize them...

[Dannyx]

Re: First oscilloscope thoughts

Ok guys, it's been nearly a year now since we've had this discussion, so here's a little follow-up:

Still haven't got a scope. Not because money would be an immediate obstacle (I could certainly dish out for a pro one at that), but because it wouldn't pay for itself, so I figured it's a better idea to just wait until I could get some actual use out of it, instead of just fantasizing how it would be to have a scope.
However, in the meantime, check out this thing I found in the junk pile at work: Leader LBO-508A...What you see is what you get: it's pretty beat up, the case is not on properly, dials and buttons have flown off...all sorts of nasties. Amazingly, the screen seems to be intact (at the front at least), since I'm not sure if the tube inside is still OK.

The ultimate question is: does it work ? Don't know: haven't plugged it in, since it's pretty messed up on the inside as well. Some traces broke off from what the chap who took it apart before me told me, but haven't had a look myself yet.

Did some searches and it turns out it's only 20Mhz which is rather meh, but hey: free scope...can't argue with that. Trouble is: should I bother to even crack the case open, or just toss it back in the corner where I found it ?

[stj]

Re: First oscilloscope thoughts

looks good for a free scope - get in it and see if it's all there.

[Dannyx]

Re: First oscilloscope thoughts

Can't do that just yet thanks to my "shop master 3000" who won't let me touch anything around here (wrote a whole rant about that in another thread), so I'll just let it sit there for now...sorry Santa, but I hope it's not long until the old geezer kicks so I can get some peace and quiet...that's one heck of a wish from someone who is usually peaceful, shy and tolerant AF...

[Dannyx]

Re: First oscilloscope thoughts

4 years later and yo' boy here is still struggling to get his first scope

I'm browsing a local site for second-hand ones, since there's no way I can afford a new one that's got decent specs without spending fortunes on it which wouldn't pay off. I found a Rigol DS1102e at a price I'm actually willing to pay....provided of course it IS any good, hence why I'm asking for advice. From the very little I know about scopes, a bandwidth of 100Mhz is on the low side....