Calling Bored@Work, and Others re: Arb Generators

[Electro Fan]

Hi Bored@Work (and Other EEVers),

I saw your post below (I'm looking at Arb Generators).  I agree with your reasoning that the PC software is important.  Of the Arb Generators that have good software, what are the best entry level (lowest cost) models in your opinion?  Thanks, EF

"Problem with the DG1022 is the PC software. You don't want to type in up to 4000 values, each requiring 5 or maybe 10 key presses, for an arb waveform on the front panel. So the PC software is important."

[tautech]

The Siglent SDG series have a few owners on this forum. Mostly good feedback IIRC.
The front panel UI procedure to set up is to select waveform type, then Frequency, Amplitude and Offset if needed. Each parameter can be set with either the MP knob or the numeric keypad. Thats just a few buttons plus a few turns of the knob.
Quick and easy for all I have needed it for. To date I have never needed its PC software.

[SArepairman]

a north korean arb generator involves many switches and a big stadium full of people.

[Bored@Work]

Of the cheap arb generators non I know has usable PC software for the arb function. Of course you can try your luck, and maybe you happen to belong to those for which such software not only installs without problems, but provides the features you need and happens to successfully communicate with the instrument. If that happens, also consider playing the lottery

If not, prepare to write your own software. The good news is you can customize your software to the task at hand. The bad news, depending on the support or absent of support you get with the arb generator, this can be a hellish task. Look for programming documentation, libraries, drivers, etc.

Quality manufacturers typically have usable software. E.g. Tek with ArbExpress. HP went through a number of product names, BenchLink Arb, IntuiLink, and their latest is BenchVue. I haven't seen it, but people complained it has some restrictions similar to trial software.

TTi has something, too. It looks promising, but I never worked with it. Keithley also has something. So does Hameg. I seriously can't remember if I ever touched those softwares.

An exception within the quality manufacturers might be LeCroy. Some what they sell is rebadged Atten/Siglent stuff. I have no idea if they fixed the Atten/Siglent PC software, but I doubt it.

Anyhow, even with the quality manufacturers it can happen that you need to write your own software. Your task might just be too exotic, or you happen to run into a bug in a manufacturer's software. And there is the big operating-system obsolescence issues. With any software, quality manufacturer or not, the next OS version, update, patch or whatever can kill the software. If you don't get an update from the manufacturer you are back to square one, write your own software. Or buy a new arb generator ...

[nowlan]

What is so special about software for arb?
ive not used any, but as from what I understand, you are generating/saving a waveform to a bin or csv.
Then uploading it to the function/signal generator via serial or lxi.

[Bored@Work]

What is so special about software for arb?

From a software development point of view nothing. But Chinese manufacturers in general have problems to write reasonable working software, and because developing software costs money, which they don't want to spend, you typically get a steaming pile of crap with cheap Chinese instruments.

[miguelvp]

What is so special about software for arb?

From a software development point of view nothing. But Chinese manufacturers in general have problems to write reasonable working software, and because developing software costs money, which they don't want to spend, you typically get a steaming pile of crap with cheap Chinese instruments.

I, for one, welcome our new overlords!

All this china this china that, I do travel quite a bit and guess what, China has a middle class and they have disposable income. They ate my puerto rican red snapper before I even had a chance to even smell it.

They are as capable as any western industry and now they have a decent income (comparing what they had before)

3 billion strong well, I'll say it again: I, for one, welcome our new overlords!

[TerraHertz]

On the topic of software for generating waveforms, have a look at these links.
I've had these sitting around for a while, meaning to play with the puredata stuff, but that's pretty unlikely to happen any time soon.


Jerobeam Fenderson - Nuclear Black Noise (oscilloscope / lissajous music)
http://www.jerobeamfenderson.net/nuclearnoise
http://puredata.info/

In this part only sine waves (discrete frequencies) were used. Using the harmonic series, i.e. integral multiples of the fundamental frequency, an unlimited variety of sound spectra can be created (additive synthesis). To produce these waveforms the open-source audio programming software Pure Data was used.

[PlainName]

typically get a steaming pile of crap with cheap Chinese instruments.

But, nevertheless, it can sometimes be a usable steaming pile of crap. I bought one of the cheapies off Ebay - a MHS-2300A:

eBay auction: #121239086225

The software that came with it allows you to control all the functions and enter arb data. In the example pics, I scribbled some rubbish in the software and uploaded to the wavegen:



Then played it back through a Tek scope:



Easy peasy, and there are various predefined functions to make the data, or you can sort something in Excel and blat the cells in.

Sure, the software could be better but it works fine. The major downer is that it use the Visa crap, so that has to be installed as well, but I guess you're going to get that with whatever software you use.

[Bored@Work]

Sure, the software could be better but it works fine.

We have apparently different definitions of fine.


I mean seriously, people wouldn't accept this level of quality as "fine" from any other profession than software programming. Rubbish user interface and fonts messed up. Very typical for Chinese software.

The font probably only happens to be present on the programmer's version of windows. And the software didn't get tested on anything but probably the programmer's own PC. And it raises the question why on earth did he feel the need to change the font at all? And what is wrong with using a normal menu bar? Or why the need to place buttons in the tabs? Pointless shadow effects around the center area? Items above that area not aligned? Etc. ...

[SArepairman]

Sure, the software could be better but it works fine.

lots of anger

you do realize this was probably some how based off of stolen soft ware (in 1997) used in something like food canning interface equipment that was modified to run a nike shoe factory. oh yea, and some mine sweeper code.

it probobly makes franken stein look original

can you PLEASE post what happens when you push the help button?? 

[nowlan]

Is it not possible for manufacturers to simply provide a VISA driver and some SCPI commands?

Then users can use whatever applications they prefer.

[PlainName]

happens when you push the help button??

[PlainName]

Is it not possible for manufacturers to simply provide a VISA driver and some SCPI commands?

I thought the use of VISA indicated that it has VISA drivers, otherwise a) why bother needing ViSA at all, and b) VISA has to have drivers to talk to the kit...

On the command front, can't recall if they are SCPI but I think probably not. My original intention was to roll my own arb data generator (I don't need instrument control, just upload of waveform and graphical editing), and it looked quite easy to do. Indeed, I am loath to pollute my system with VISA crap and elected not to look at their software. Nevertheless, curiosity got the better of me so I ran it up in a VM to take a look, and from that figured it was safe enough to migrate to my working PC.

As I said, it's not the best but it does the job fine. Of course, your requirements may be different to mine, and a misplaced comma or tpyo could be enough to cut off your nose to spite your face. Only you can know

[PlainName]

this was probably some how based off of stolen soft ware

I am amazed that you can tell that from a single screenshot, and yet I can't after having actually used it. The software is closely intimate with the hardware: where the hardware doesn't allow certain combinations, or has varying options depending on setting, the software does exactly that too. I know of no other arb gen like this (it is hardly a yellow Fluke clone, after all), so where would the software have been nicked from?

The problems with it, such as they are, seem to me to be exactly as B@W says: the Chinese programmer has fonts installed that I don't, and the look'n'feel style is not Windows as we in the West know it. But I have seen much worse from software that was a significant cost, and pretty much any web designer worth his salt will design pages to suit his monitor resolution. I have paid-for software that starts up straddling my two monitors because the programmer had a single monitor setup and thought it looked cool for his stuff to start in the middle of the display. The problem isn't restricted to nicked software, so is not an indication that this is either nicked, dodgy or flaky in real functionality.

Which brings me to the point that the original question was about entry-level and low-cost arb gens. I am not selling this one and have nothing to do with it other than putting my money where my mouth is. I didn't post initially because I thought someone with more of a clue than me would point out stuff like this, but all we've seen is actually expensive stuff (which may still be low-end, but it's not cheap low-end). So I just popped up to point out alternatice.

Don't forget that this is £65 for the hardware, and the software is free. That is bloody cheap, isn't it? And what, exactly, are you expecting for that kind of remuneration? Frankly, I am gobsmacked at the value, and whilst I would lurve some of the previous ones mentioned I can't justify them. This does what I need it to do and a price I was willing to pay, and whilst I would prefer the programmer to have a better grasp of internationalisation, I find it hard to slag him or his bosses off for supplying free software that functions well.

[PlainName]

I was asked if the scope could decode pulse waveforms made to represent ASCII bytes. The answer is no, that scope doesn't do that, but I have a Logic that does...

The question was via PM but I thought I should post the results here in case anyone else was curious. The phots are, in order:

1. The arb waveform screen. Doesn't look much because I elected to have each bit a single sample (there was another question as to how sharp an edge you could define), so the data is scrunched up on the left. With fewer than 1024 samples the data is enlarged to fit, but then each bit wouldn't be a single sample. Swings and roundabouts, and something that can screw your timing calculations.

2. The sig gen output control, just FYI. The strange frequency is because the entire arb waveform is treated as a pulse, and the pulse is output at the specified frequency. So each bit of this waveform is freq * 1024, hence a simple /1000 to bring it back to (well, close to) 9600.

3. The result seen by the Logic. Recall that the pulse (that is, everything from the start of data through to the start of the next data) is 9.6Hz

4. The result zoomed in to show the data decode.

[Electro Fan]

Nice work dunkemhigh,

It was a little hard to see the actual decoding but I finally recognized it:  EEVblog 

Any chance you can experiment with making the bits slightly longer with a larger sample (maybe 2-4 samples or more per bit)?

It would be interesting to find a way to treat each waveform as a pulse and be able to control the frequency of each pulse (each bit) but what you have done is a very good start with the generator you have.

I'm wondering if anyone out there with perhaps a Rigol or Hantek or other arb generator could do some similar experiments so we could see what additional level of pulse creation control might be available in the ~$300-$900 class of arb gens.

On the Rigol 2072 I have found that some ASCII character strings decode well but other strings (or parts of other strings) exhibit various errors that I haven't been able to figure out.  With a little more control over the pulse generation side of the setup I think it might be easier to understand the timing and other relationships involved and figure out what causes what.  I think this type of setup - ie, test equipment that provides pulse generation control plus test equipment (LA or scope) that provides waveform display and decoding -  is a valuable platform for observing, testing, and especially learning.

Clearly, it's doable to make such a platform - I'm kind of interested in determining what is the low (or reasonably low) cost way of doing it that provides good control and display visibility on the generation (input) side;  there are pretty good choices on the waveform display and decoding (output) side.  It doesn't have to be the very lowest cost; it's probably worth paying more if it provides better generator control and waveform display visibility on the input side.

Thanks again for sharing your results.  EF

[KJDS]

I've got some Rohde and Schwarz AMIQ units in the back of the storage unit. Details here

http://www.rohde-schwarz.com/en/product/amiq-productstartpage_63493-7564.html

Has anyone used these? I think they should be capable of being used as an arb generator, just haven't had time to find out.

[grumpydoc]

I think they should be capable of being used as an arb generator, just haven't had time to find out.

Just had a glance at the user manual - certainly looks that way as long as you don't mind that the user interface is a command line via RS-232 or GPIB.

Looks like 4 or 16M points at sample rates between 10Hz and 105MHz

[PlainName]

Any chance you can experiment with making the bits slightly longer with a larger sample (maybe 2-4 samples or more per bit)?

Maybe, but you could say what you're hoping to see or check - my time is limited at the moment. I originally used two samples per bit, so I could shove in half a stop bit, but I am manually creating the data file so don't really want to do more than typing than necessary

One thing I would quite like to do, but will probably take some time, is to generate and decode I2C and/or SPI.

[Electro Fan]

Any chance you can experiment with making the bits slightly longer with a larger sample (maybe 2-4 samples or more per bit)?

Maybe, but you could say what you're hoping to see or check - my time is limited at the moment. I originally used two samples per bit, so I could shove in half a stop bit, but I am manually creating the data file so don't really want to do more than typing than necessary

One thing I would quite like to do, but will probably take some time, is to generate and decode I2C and/or SPI.

No worries, let's forget the longer samples, but your idea of I2C and/or SPI was going to be next on my list too   If you get a chance to do those please let us know how it goes.  Thx

[alexwhittemore]

this was probably some how based off of stolen soft ware

I am amazed that you can tell that from a single screenshot, and yet I can't after having actually used it. The software is closely intimate with the hardware: where the hardware doesn't allow certain combinations, or has varying options depending on setting, the software does exactly that too. I know of no other arb gen like this (it is hardly a yellow Fluke clone, after all), so where would the software have been nicked from?

The problems with it, such as they are, seem to me to be exactly as B@W says: the Chinese programmer has fonts installed that I don't, and the look'n'feel style is not Windows as we in the West know it. But I have seen much worse from software that was a significant cost, and pretty much any web designer worth his salt will design pages to suit his monitor resolution. I have paid-for software that starts up straddling my two monitors because the programmer had a single monitor setup and thought it looked cool for his stuff to start in the middle of the display. The problem isn't restricted to nicked software, so is not an indication that this is either nicked, dodgy or flaky in real functionality.

Which brings me to the point that the original question was about entry-level and low-cost arb gens. I am not selling this one and have nothing to do with it other than putting my money where my mouth is. I didn't post initially because I thought someone with more of a clue than me would point out stuff like this, but all we've seen is actually expensive stuff (which may still be low-end, but it's not cheap low-end). So I just popped up to point out alternatice.

Don't forget that this is £65 for the hardware, and the software is free. That is bloody cheap, isn't it? And what, exactly, are you expecting for that kind of remuneration? Frankly, I am gobsmacked at the value, and whilst I would lurve some of the previous ones mentioned I can't justify them. This does what I need it to do and a price I was willing to pay, and whilst I would prefer the programmer to have a better grasp of internationalisation, I find it hard to slag him or his bosses off for supplying free software that functions well.

What I CAN tell from one screenshot (okay, the help window confirmed it, so really two) is that it was written in LabVIEW and exported as a standalone executable. Definitely the cheapest option, but actually probably better output quality than if they'd tried to roll their own with some other software toolkit. I mean, you get a graph view right out of the box. That alone is pretty great. Of course, this will never win any Apple Design Awards, but shit, it's also not a 100 line processing sketch. Come to think of it, making tool interfaces for PC-connected hobby gear this way (like the re:load pro, perhaps) wouldn't be a bad idea.

[PlainName]

it was written in LabVIEW and exported as a standalone executable

Ah! That would no doubt explain the reliance on the VISA stuff. Thanks for the insight