Advertisement

Blog

Amateurish Equipment

A while back, Michael Dunn, Editor-in-chief, of EDN published an article, 8 Signal generators under $100 compared, which compared inexpensive signal generators.

In his last paragraph he solicited reader feedback with the question being: ‘Is the equipment too amateurish?’

Now I don't want to talk about those generators specifically, but more generally and then, in particular, my unique equipment problem and why it relates to the subject of amateurish equipment.

In the tradition of Application Engineering greats such as the late Jim Williams, I have a lab at home that I have equipped to do credible testing which results in a document that contains test equipment info, test schematics, and test results.

In the last couple of years I acquired an Analog Discovery II, by Digilent, now a National Instruments company. I have become quite impressed with this software-driven device which offers two generators, two scopes, two power supplies, and more (mostly digital). What it really gives me is an excellent network analyzer and spectrum analyzer.

So I need to list this device in my equipment list, which I haven't done so far. The reason I haven't listed it is if you go to National Instruments website to look at https://www.ni.com/en-us/shop/select/analog-discovery-2″target=”new”>this excellent device, you find it described as: “Gives students access to a 100 MS/s oscilloscope, a logic analyzer, and six other instruments in a pocket-sized lab device.”

I already hear my co-workers criticizing my work as being done with a student-grade instrument. All because National Instruments chose to describe it in that way. Is it really student grade, or something amateurish? Consider the specs. Detailed spec sheets are available on the NI website but let me bring out some high points:

  • Resolution: 14-bit
  • Accuracy (scale≤0.5V/div, VinCM = 0V): ±10mV±0.5%
  • Sample rate (real time): 100 Ms/s
  • Analog bandwidth with Discovery BNC adapter: 30+ MHz @ 3dB, 10 MHz @ 0.5dB, 5 MHz @ 0.1dB
  • Input protected to: ±50V
  • Sampling modes: average, decimate, min/max
  • Real-time views: FFTs, XY plots, histograms, and other

Even the much more expensive Virtual Bench has an 8-bit resolution (albeit it offers much higher speed).

My biggest personal limitation of Analog Discovery is that for my purposes it is a 10 MHz bandwidth device, and the spectrum analyzer noise floor is about -75dB. But that is sufficient for testing most of my products. Also, as it is shipped, it is a module with a connector and “flying wires”. I put mine in a proper housing including BNC jacks to the outside world.

I am one of those people that has zero patience for unintuitive, hard-to-use software. Analog Discovery software is superb in its ease of use, one of few software applications that have impressed me. Within 10 minutes of getting it in my lab I had it doing a Bode Plot of an amplifier.

I just wish that NI could give Analog Discovery the professional image that it really deserves. I personally consider it a very professional device.

So is Analog Discovery amateurish? If it is, then I'm going to be having to spend some big bucks this year.

4 comments on “Amateurish Equipment

  1. Work to Ride, RIde to Work
    January 24, 2018

    I like to think of it being prudent with limited resources.  I outfit my home lab with cast-offs and equipment that was discarded because it wasn't cost-effective for a business to repair.  The other thing is that if you don't NEED 2 GHz bandwidth, why pay for it?

    There is no sense in going mosquito hunting with a machine gun. (nod to Monty Python 🙂

    My company throws away equipment that we would have killed to have twenty years ago.  Ebay and Craigslist and other forums can be mined for some jewels.  Some people still want way too much for older stuff though.  You have to sort through the dreamers out there in 2nd-Hand-Land.

     

  2. MWagner_MA
    January 24, 2018

    0.5% accuracy for a measuring device is amateurish because many systems we wish to design want 1% accuracy.  Using standard measurement science, you generally want 4x the accuracy to have reasonable uncertainty in your system.  That would mean a device with an accuracy of at least 0.25%.  A cheap Fluke 177 has a basic DC accuracy of 0.1%, but it depends on how fast your signal is.  Now if you are doing a proof of concept for some quick data, I see no problem with using what you have.  You will just want to use better equipment during validation.

  3. Pesky Varmint
    January 24, 2018

    I should have mentioned that I do not use it for official characterization or validation of our parts. Us apps guys mostly duplicate customer circuits or requirements, or do a variety of testing basically looking for trouble. Precision usually isn't required (if it is I do have things like my HP voltmeter to cover me). The network and spectrum data I get is sufficient for most of the apps related testing I'm doing.

    Jerry Steele

  4. thelyonnais
    January 28, 2018

    I agree with anon9303122 : if you don't NEED 2 GHz bandwidth, why pay for it?

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.