Multi Function Test Device?

I've been using a Digilent AD2 PC test device for a few years. I'm thinking it might be time to purchase some actual test equipment. Other than my dmm, I mostly use only a scope, wave generator, dummy load, and voltage divider. The AD2 provides the scope and wave generator and I built a 4/8 ohm dummy load and a 2/1 voltage divider. When I test something I've usually got a ton of connecting cables and scads of adapters for for joining different types of connectors. My work space usually look like a Rube Goldberg creation; and I spend a lot of time chasing down illogical readings caused by loose connections, etc. Anyway, a device that combined some or all of these functions could be appealing. If it was available used, that would add to the appeal.

Any thoughts would be appreciated.
 
So what is wrong with the Digilent Analog Discovery 2? The connections are flimsy? I am actually thinking about getting one of these multi-function devices, but if the execution is a lot worse than the idea, I should probably rethink that.
Could you expand a bit on your experience with the AD2, please?
 
I have an Analog Discovery 2 that has worked great for me. If you buy one, be sure to spend the extra money to get the BNC adaptor board for connecting the probes. No flimsiness, at least so far.

BNC Adaptor

That way your connections are the same as they'd be on any other oscilloscope designed for use below 1 GHz.

A little background, more than you wanna know, I'm sure.

Around the time I bought the AD2, I'd decided to add test capability here rather than dragging whatever I was working on at home to the lab at work for testing at lunch or after hours. Lots of practical reasons for that. But, one of my criteria for buying test gear was that I wanted it all to be small. I don't run a business from home or anything like that. It's all for my own use. There are times when any one piece of test gear won't get used for months, but then it will get a few hours or even weeks of use. The rest of the time, I want it to sit on a shelf in the closet in a case, not collecting dust.

The AD2 does all that for me. I haven't found a real need for higher bandwidth than the AD2 provides for what I do. For real RF testing, I have (small) RF test gear that gets used the same way as the AD2. For serious audio measurement, I have products from QuantAsylum, E1DA, and a few other random pieces. Plus, there's the ubiquitous computer that's always on the workbench, which also doubles as the CAD system for simulations, pcb design, mechanical design, and so on. This is the entire lab when not in use:

Lab.jpg


If I was using a scope all day to repair gear or for a business, I'd get a scope for that. Same for the other stuff. But, for me, this works fine. I do have a separate DMM because that's just easier to use and, like with most people, that gets the most use.

One other point - I received an email from Digilent just this week letting me know that they are coming out with an Analog Discovery 3 later this month. There's more information on their web site, but no pricing. It looks different, uses a USB-C connection, and I guess has more buffer memory which provides some additional functionality with their latest software. I'm not sure about any other details. Nor am I sure that I'll upgrade.
 
  • Like
Reactions: 1 user
Thanks a lot for the details, CG! I got the same email about the AD3, and I'm thinking of getting the discounted AD2 (with BNC adapter and impedance test board) instead of the AD3. 125 vs 100 Msamples/s does not seem like that big of a deal, and as you implied, for serious audio measurements (incl. FFT) one would use REW with other hardware, so the increased memory in the AD3 might also not be that important.
 
I see there's a price attached to the AD3 now - $379. Not a huge jump, but not zero, either.

There's a suggestion that with the new hardware, Waveforms will now also let you use FIR filters. OK...

An audio measurement software package you may also want to check out is: Multitone

It's somewhat different from REW, but both are quite useful and great applications. I use them both with a Topping D10s as a source, an Autoranger, and an E1DA Cosmos ADC. Plus a proper USB Isolator. Remarkably good. Not quite as good as an Audio Precision system in some ways, but also not as much money as a new automobile, either. Here is a plot of that system made with Multitone with no device under test (analog loopback).

D10s > Autoranger > Cosmos ADC.png


By the by, I made that plot to show the value of a proper USB Isolator. Here's the before picture:

No ioslator.png


Subtle differences...

I didn't really want to hijack this thread, so that's enough from me!
 
Gee, I thought the thread was about closets. We never seem to have many posts about them here.

I built my own isolator using an ADUM4166. That's what is being used in the plot shown in my last post. (It also works very well on improving the sound quality from this desktop computer system as well as on the main listening system in the living room. Despite what the bits are bits, ahh, enthusiasts insist. I also thought they believed in measurements, but, alas...) There's at least four commercial offerings using the part. Of course, mine is better. :D

I've read that IVX is updating the design of the Cosmos ADC a bit since he's running out of the first generation. It will have USB isolation built into it. (Cosmos ADC v2) I think he's planning to use the TI USB isolation chip (TI USB ISOUSB211), based on a posting following the one in the link. I'm sure it will work great, because his products always do.

The QuantAsylum products have isolation built in, but in a different fashion. The latest models use the same A-D chip as E1DA uses. (QA403/404) E1DA doesn't offer a Cosmos DAC, at least not yet. Similar performance and there's pluses and minuses to each product set from the two manufacturers. And, both are small. I guess you end up with fewer boxes and cables with the QA solution, but are more limited in the software choices.

Again, maybe more than you wanted to know.
 
Last edited:
No, no, this is all very very interesting. So that is a high speed device. Do you have a schematic and PCB layout for that isolator you designed that you would care to post? Or even better, do you maybe still have a spare PCB sitting around (assuming/selfishly hoping the minimum order quantity was >2)? ;-) I'm not sure a Cosmos ADC v2 is in my immediate future.

I did order the Digilent AD2.
 
Well, I don't think you'd need a USB isolator unless you are going to be making distortion measurements below -90 dB or so. For that you'd also need either a QuantAsylum product, which already has isolation, or a Cosmos ADC. (There's some other commercial products that can do this, but they cost way more.) The Cosmos V2 is likely to have isolation already built into it.

Not to discourage you, but are you sure you really want to build a board with tiny SMT parts on it? (I didn't!) The crystal and the transient protection devices don't lend themselves to hand soldering very well. Have a look at the ADI eval board for the ADUM4166: ADUM4166 Eval Board

The minimum order for an assembled PCB with parts from JLC happened to be 2! Ask me how I know... In the end, although it isn't in a snazzy looking box, the eval board is economically more sensible.

But, good news! There's a thread on this very website that describes an isolator using the ADUM4166. Lots of off-the-rail conversations, of course, but generally I think he may have what you're looking for: Somewhat Off The Rail ADUM4166 Isolator Discussion Full disclosure: That design is somewhat different from my own. I have no idea how well it works.
 
All good points, CG. I no longer have access to a reflow oven, but at some point I actually bought some solder paste for use at home, to try with a hot air gun. That hasn't happened yet. I'm ok with working under 10x magnification (or even a microscope), but I know that fine pitch devices can be problematic, and 2x1 mm parts can go flying from just looking at them...

Regarding the need, I actually thought of using a USB isolator for my tablet-monitor connection (it needs two cables, one to drive the 5k video, and one for the USB hub inside the monitor), which introduces a really bad buzz well above -90 dB. But maybe that is a USB ground loop of a different kind?

I had a look at the eval board earlier when I was looking for high speed isolation, but thought it might be possible to find something cheaper than $66. But maybe that's not true. Topping is actually marketing the HS01 and HS02, both USB isolators.

For now I ordered a slow (full speed) isolator, and I'll see what it does to the buzz and the measurement noise floor.
 
Last edited:
It's pretty hard to compete price-wise with what the Asian based or sourced companies can do. Volumes really matter.

For example, due to some logistical reasons, I installed the ADUM parts and the connectors myself. (Those aren't that tiny.) The ADUM4166's cost about $17 each in the quantity I bought - way less than 10. That's without shipping, too, which isn't trivial these days. You can get great pcbs from JLC PCB for only a couple dollars. But, the shipping adds about $20 to that. So, between the pcbs and the ADUM4166's that's almost $50 for those two components alone delivered to your door. Add a box, the other components, maybe some front and rear panels, and more shipping and it all adds up quickly. <sigh>

The USB isolator units I built didn't save me a penny compared to just using the eval board. Yeah, they're in attractive boxes with a finished look. And, they have some additional features that I wanted. None of that came for free.

Plus, I wanted to really do some experiments and push the limits of what could be done, at least by me. So, in that sense it was worth it. I validated some of the ideas I had about the limits of digital audio systems and proved to myself that a lot of the conventional dogma out there is more like dog-something-else. Most sane people would think that's crazy. In fact, I can point you to some web forums that have hundreds of people poised to tell you that the whole idea is crazy and anybody doing this is stupid. Or, worse. (I generally avoid those web forums and never post there.)

The HS02 doesn't do exactly what I want, but it seems to work just fine for measurements. At least the one I tried was OK. I bet there'll be a number of those available used from the people who want to try every gadget that comes along and will just have to move on from the HS02 to the next thing soon enough.

BTW, I'm certainly not trying to discourage you or naysay. I'm just trying to be realistic and give you the whole story as I found it, so that you can make your own decisions with your eyes open. If somebody can show me the error of my ways, I'd be very appreciative. Short of the kind of bashing found on those web forums, of course. :D

Also, you do know that there's a very nice audio analyzer suite for the AD2, right? It's limited to 14 bits of resolution, but that's actually pretty useful for a lot of work. AD2 Audio Analyzer Suite Info If you're trying to measure lower distortion in system components capable of better, then you'd probably want a QuantAsylum product or some combination that includes a Cosmos ADC. There's lots of solutions out there. Maybe too many to choose from.
 
Last edited:
I wasn't aware of the audio analyzer software/firmware written for the Digilent AD2. It looks like it duplicates some of the REW functionality. So that will be interesting to compare.
At the moment I'm still learning my way around REW and the Focusrite USB DAC+ADC combo. The noise floor seems to depend a lot on which USB socket I use to connect it to my computer.
This is about the best I am able to get out of that thing at the moment with a loopback connection, 93 dB SNR at 1 kHz, if I interpret the REW numbers correctly. I can get the noise maybe 1 dB lower if I increase the output volume to the maximum of 4.845 V, but then the distortion goes up significantly.
Focusrite Scarlett Solo loopback 3.37 V out min input gain (balanced).jpg


Regarding price-competitiveness of home built gear: as soon as you start figuring your own time in, everything becomes super expensive. I once grew potatoes in a garden plot: most expensive potatoes ever! And that was in the good years where I was able to harvest any.
 
With the Scarlett devices, it depends a lot on how you set them up, what gains you have set, and so on. Here's some measurements for the very similar 2i2 by a guy who certainly knows how to measure things. The DAC side appears to be the limiting factor for a loopback measurement.

PMA on 2i2's

Yeah, USB connections to computers often are a mixed bag. Some are better than others. Cables matter, too. As does the test set-up. Unfortunately.