TM4C123GH6PM: Any Advice on Purchasing an Oscilloscope?

A scope is not really connected to a TM4C microcontroller - but useful anyway.

However, I would suggest to look at logic analyzers, too. Albeit bound to "logic levels", they have much more channels available (at least 8), and are quite affordable, starting at around 100 bucks. I have a RIGOL scope, 2 channel, 100MHz bandwidth. This company used to produce for Agilent before, and the scope did not let me down yet. Only the price had been more than double of that Picoscope ... And for a more comfortable configuration, you are usually charged a hefty premium - 4 channels, high bandwidth and extensibility will cut large holes in your budget. But perhaps you can locate a used device.

Or, perhaps this could be something: https://www.lpcware.com/content/project/Mixed-Signal-Logic-Analyzer-Oscilloscope-Lab-Tool-Solution

Bob,

Same thought as f.m. here: for serial communication, a logic analyzer will serve you better for less money.

Scopes will be essential when you get into analog development...

As for the scope, my personal workhorse is an Agilent (now Keysight) DSO-X-2002A with the 8-channel logic analyzer add-on. And I do use a Picoscope once in a while, but that requires the PC, and the control features are quite limited.

Nowadays, there's much more PC-side software involved in a logic analyzer than hardware. Actually, why don't we coordinate a joint effort and have the folks here on the forum develop a logic analyzer boosterpack for a TM4C123 launchpad - in case there ain't one in the market?

You might help on the PC software side - one of my MANY weak spots - and I'll probably be able to help with the hw layout, some specs, and what not...

(PS: order your off-the-shelf piece before you even consider this, as it might take a long time to become real - if ever!)

Bruno

Hello Bob,

If analog signal monitoring is not really the case, you may want to see if a Logic Analyzer can fit the budget.

Hi guys - Thanks for the feedback.  I didn't know what a logic analyzer was, but after doing some Googling and educating myself this looks to be the exact tool I need, and will very likely come in very handy with future projects.  I did a fair amount of Googling and looking around at different ones.  It appears people are very happy with these Saleae logic analyzers.  Very good Amazon reviews here and a nice YouTube review here.  I think I'm going to go ahead and get the Logic 4 for $109.  Just wondering if anyone thinks this I should consider something else before purchasing?  

Thanks in advance!

A couple of thoughts to add to your consideration.

Logic analyzers are powerful tools but they do hide information on the waveform and can be actively misleading as a result. You really need to have an oscilloscope as well. If I had to choose just one, I'd start with an oscilloscope.

Second, and you will not like this, although picoscope appears to be gathering a good reputation at $250 I'd wonder if they were good enough to be more than a toy. Remember these are relatively high frequency devices.

Robert

I did just buy a Salae. I haven't had the opportunity to use it yet. There do appear to be some limitations because of the way they divide responsibilities between the PC and the logic pod. I suspect you won't notice the lack. On the definite upside it was nicely constructed in a solid case.

Robert

Bob Jennings said:

Just wondering if anyone thinks this I should consider something else before purchasing?  

Besides the obvious (number of channels and, perhaps, price) I would check for the following points, and include what required:

• sampling frequency, should be higher than the frequencies of the signals you want to observe (Nyquist)
• configurable thresholds, to use with different supply voltages (5V, 3.3V, 1.8V)
• separate thresholds for channels (mixed supply voltages or trigger signals)
• 'luxury' functionality, like protocol decoders

BTW, there are mixed devices (scope + logic analyzer) on the market, but probably far beyond your stated budget ...

Thanks, f.m.  When I first posted I thought this might be a tool that I would use only once or very rarely.  The more I contemplate this, the more I think I might be using it more frequently with current and future projects, therefore I'm willing to spend more to get the right tool for me.  

With my current lack of experience, what I'm trying to best determine:

1. Finding a tool that will help me with my current mini-project of monitoring signals sent to/from an I/O expander.
2. Purchase a tool that I will not quickly regret it's limitation in the near future.
3. Not spend more than I need to. 

So, I'm definitely open to the idea of purchasing an all in one scope and logic analyzer but would love some recommendations since I'm such a newbie here.  You previously mentioned this device - is this something you would recommend?  Or is there possibly something else I should look at?

Hello Bob

I will not comment about any specific LA; but I have seen both Salae and Intronix in operation and prefer the latter for wide signal bus capture when it comes to pricing.

Thank you, Amit. I will look into Intronix products and compare.

Like Amit, I hesitate to give any "take this, not that" recommendation. It depends very much on your requirements - even the unstated.

The device I mentioned comes with open-source software and a LPC-Link 2 (CMSIS-DAP compatible debug adapter), to an affordable price. I mentioned this one as example, there are similar boards as well. I would always check foras and communities for support in general, and critical reports in particular. You know, advertisement tends to hide things that don't work. The same goes for commercial devices like the Picoscope or Salae.

But this devices, like other PC extensions, have another problem - the power supply by PC, and possible ground loops. Once you experienced the first spurious device resets or kills, you know what I mean. Standalone scopes use to be galvanically isolated.

Perhaps you can locate a used scope, not sure about that market in your country. Prices of  really good scopes are off the chart - usually between 10.000 and 50.000 € or $.

I have a Rigol DS1152E for about 750 € (at that time), which serves me quite well. Having only 2 channels complicates debugging sometimes, but on can get around this.

Okay, thanks for the additional info, f.m. All of your suggestions make sense and I will follow accordingly.

Bruno Saraiva said:

why don't we coordinate a joint effort and have the folks here on the forum develop a logic analyzer boosterpack for a TM4C123 launchpad

Zero echo on this suggestion?

Bruno Saraiva said:

Zero echo on this suggestion?

To increment your echo counter, here some basic ballpark figures and ideas.

To be of real value, the boosterpack should support sampling rates of at least half the max. MCU core clock, say, > 40MHz (better 50MHz). That would require a data throughput (bandwidth) of at least 40 MBytes/sec., either in real-time directly to the PC (hmmm, a bit off the charts I guess), or to a dedicated memory extension. The latter should preferably be in the MByte range, to allow for a useful storage depth. That would lead to a hardware solution with autonomous sampling and storage, where the TM4C only does the configuration and data delivery (to the PC). Putting the TM4C into the sample chain would IMHO decrease the max. sample rate considerably.

I'm no hardware guy, perhaps someone else can continue this line.

BTW, looking up PicoScope, Salae & Co, I saw a PCB image with a Spartan FPGA. Seems this guys use that first (hardware) approach ...

f.m.
In fact, there is no way that the TM4C123 can do anything about tracking 40MHz signals, specially when going up multichannel.
But my proposition for a boosterpack is more of a low-end multi-purpose signal analyzer... it is surely possible to record multiple channels at UART speeds (seldom faster than 921600), and it is probably possible to record SSI at 8MHz at least for one channel.
I wouldn't bother real time to the PC (the transmission to the PC, no matter how magic the protocol and communication channel, will take longer than that). Still, serial communications are normally bursts with lots of idle time in between, so some memory in between could take care of that and the PC screen will see results "in human real-time" - enough for debugging.
The idea is out there for a challenge. A TM4C123 Launchpad costs "nothing" and "everyone on this forum" has access to one... The booster would of course be compatible to TM4C129x, so we're up to 120MHz (and the new faster MCU further below the horizon...) - Anyway, something "useful at a low cost". I'll leave the idea open, and if some more support, can create a shared project landing document.
Time to go back to the calculations and designs that actually buy my children's milk...
Bruno

Bruno Saraiva said:

In fact, there is no way that the TM4C123 can do anything about tracking 40MHz signals, specially when going up multichannel.

I am quite aware of this fact. However, I was tackling this issue from a customers view, stating the most probable expectations of a user. My expectation would be to at least be able to sample SPI transmissions with about 25MHz clock rate (or more). Before starting designing/coding, one just needs to think what to achieve, and what the limits are.

Bruno Saraiva said:

Time to go back to the calculations and designs that actually buy my children's milk...

You are right, me to (tomorrow, then). Albeit my kids are not satisfied with milk anymore ...

MSOs are very useful but tend to be expensive. The advantages are

• One instrument to carry around
• Can capture synced Analog and digital based on either an analog or digital trigger source. 'Scopes generally have a trigger out (and in) so it's possible to use that to sync a logic analyzer to the sync out from a scope (just steal an input) but syncing the other direction which I find more useful may not be possible. Definitely if using USB devices check to see if this limitation is present. You may decide to live with it but you should be aware in advance if it is possible.

Disadvantages include cost and one instrument to break. Also really capable devices are seldom combined together, that's unlikely to be an issue for you soon.

f.m has already mentioned this but watch your bandwidth requirements. For logic signals you want logic analyzer to be able to sample more than twice as fast as your fastest component (Note that for PWM that means the narrowest pulse width not the period so for an eight bit 20kHz PWM you need > 10MHz ). If measuring those digital as analogs (to see ringing or actual voltage levels for instance) you need something more like 5 to 10x the edge rise and fall rates. Even low frequency digital contains a lot of high frequency components.

As f.m. suggested take a look at some of the used refurbished test equipment places. They can have good deals, especially on older equipment.

And for the logic analyzer (or portion) look at the probes. Besides the typical clips and single pin sockets many logic analyzers allow the use of cable with standard connectors (usual 2xn) on them. If you design your board with those connectors in mind (you only need to populate for test) you have an easy, relatively sturdy connection for testing and measuring. Route a few extra pins to one and you get a simple profiler/trouble-shooter.

Robert