Analog Discovery 2 - Evaluating AC/DC Voltmeter & FFT Spectrum Analyzer For PC Based Audio Analysis

[@Bianca-zj8xo]

Hello, This is Bianca from Digilent. Thank you for the interest you are showing in this product. We saw your video and we wanted to say that is normal to have these results. The problem is real and affects the AC voltmeter, in Rev 2.5.4 and earlier of WafeForms. It does NOT affect the Spectrum Analyzer, the Network Analyzer, the Scope and FFT within the scope. The problem is solved in the Logger instrument. The Voltmeter instrument uses the same HW resources as the Scope. The RMS value is computed from the acquired samples. For Analog Discovery (2), the data buffer size is 8ksamples/channel. The Update rate is 1s, and the sampling rate is 8kSPS. The maximum frequency, for accurate AC RMS measurements is about Sampling Rate/4 = 2.048kHz (even at this frequency, the RMS value is computed from 4 samples /period, which is far from optimal). The lowest frequency for accurate AC RMS measurements is about Buffer size/Sampling Rate*4 = 8k/8kSPS*4 = 4Hz. (which is 1/(update rate)*4 ). This provides 4 periods of the input signal in the acquisition buffer (also not optimal). The problem is that the Voltmeter instrument always samples at the same frequency. Use the LOGGER instrument for this kind of measurements. The Data Logger instrument also performs RMS measurements, but allows setting the update rate. As described above, different update rates allow different ranges of frequency in the input signal. The Logger instrument displays the frequency range next to the Update rate control. The next versions of WaveForms will allow to select the update rate (and the resulting frequency range) in the Voltmeter instrument also. Thank you and best regards, Bianca Digilent

[@ElPasoTubeAmps]

Hi Blanca, Thank you for your kind reply and letting me know that I wasn't overlooking something simple and misrepresenting the AD2, AC voltmeter in my videos. As you can see from my KZbin postings, I am very pleased with the performance of the oscilloscope and signal generator, and I find the FFT Spectrum Analyzer usable into the MHz range. I am sure I will be using the AD2 in upcoming videos on audio devices I build and test and I look forward to upcoming versions of WaveForms. Again, thank you and I wish you and all the staff at Digilent the very best. Regards, David Beard

[@setitthen]

Bianca Peterlin with the logger instrument it would be of value if you were able to graph, frequency pulse width, duty cycle, a.c voltage etc over time like a histogram. For instance with the frequency you could look at the frequency stability of an oscillator over time. Just something to add. Thanks.

[@anks888]

There is some 3rd party software for audio analysis that works by stepping the scope and signal generator through different time bases. That method maximises the useful range of the device. It's a limitation really of the channels being designed as a scope primarily and the way it is sampling. The same issues come up with all scope FFT and they struggle with the low speed audio signals. As for the AC meter I've not looked at the spec but wouldn't be surprised if it meets them.

[@joeking5211]

Wow, it's 5 years on from this vid, but was just going to order one, but now wondering. Have they sorted this now do you know ?. Also, how does the Bode plots operate then, don't you have to measure AC ( to DC ) and convert to db for the Y axis, so that would mean no more than a few kHz ????, yet I think the AD2 Bode plots go to 10MHz, or maybe I have this all wrong. Would be grateful for you comment. I very nearly clicked the button to order the 'Ultimate Analog Discovery 2 Bundle' at quite high cost, so now in limbo. Thks for you vids on this product.

[@ElPasoTubeAmps]

Hi Joe, I don't know how to comment to help because I hardly use mine anymore. The model I have has a pretty decent FFT spectrum analyzer with, as you said, a 10 MHz bandwidth that I wanted for audio but settled on an audio specific SA that works great but has a horrible oscilloscope. Seems these PC based instruments aren't good at all things. I can see the AD2, or whatever version they make now, as really good for classroom demonstration. If you know what you are looking for and take care to set the device up right, it could be a jewel in a class room but for a general all-around lab instrument, I don't think so. I could be wrong and as can be seen in some of the comments, the AD2 people were quick to write back to me and defend themselves and show me how to get correct readings, I think..., on the AC scale, but that did not satisfy me that I have to know a special-trick to get accurate answers. It was pretty expensive when I bought it and I bought it expecting it to be that all-purpose lab instrument but, it did not turn out that way for me. Sorry I can't be of more help. I just noticed... see the comment just below this comment of mine from, Bianca.

[@jj74qformerlyjailbreak3]

Hey for 100 on FB. It was well worth it to me even with the BNC adapter.

[@behysun]

You forgot to move the jumper wire on breakout board to AC. That is why its readings are not well accurate.

[@ElPasoTubeAmps]

If you read the comment below from Digilent, the manufacturer of the Analog Discovery 2 just below your comment here, you will see the AC voltmeter has a known problem and they give a workaround. The AD-2 is a nice instrument for a classroom where you know the results you are looking for and there are a few other features about it, like the spectrum analyzer, and other features as she mentions, that are quite nice, but as an all-feature lab instrument, it can be misleading. They have come out just recently with the model AD-3 and I suspect they have resolved all of these problems.

[@bucyruserie1211]

Might be an obvious question, but do you need to move the switches or pins back to the original position (back to A/C ) on the fft like you showed in the beginning of the video?

[@ElPasoTubeAmps]

It should work OK either way except for possibly low frequencies with AC coupling. I have tried it with AC and DC coupling and with a direct wire, eliminating the probes and it is all the same.

[@JasonLeaman]

Are you still using this for working on amplifiers ? Do you recommend it still or ?

[@ElPasoTubeAmps]

I don't use if very often but I am glad I have it. It is like a Swiss army knife. You can do a million things with it but nothing really well. I see it as a nice teaching tool where you can set up situations and you know the outcome but if you use it as a laboratory instrument, you have to be very careful as you can get results that are inaccurate because of the way you have it set up or are using it. The scope and the FFT bandwidth are good enough for audio work if you are careful with it. I like the scope probes that came with it and have bought others like them. They are inexpensive and work pretty good.

[@kadosas]

To get rid of the glare you need a polarizing filter, sky filter or ND filter will not help. Like your videos!

[@ElPasoTubeAmps]

The ND filter didn't look like it was helping much from what I could see in the viewer. I will have to see what I can find in a polarizing filter. thanks.

[@ElPasoTubeAmps]

I finally looked up a52 mm polarizing filter on Ebay and ordered one. Thanks for the suggestion.

[@kadosas]

You are welcome. remember to remove both the skylight and ND, they only the light more light away. Question. What is your opinion about the tube 7189A?

[@electronicsoldandnew]

I've got a feeling the scope probe is messing with you on the ac voltmeter test.

[@ElPasoTubeAmps]

I wish that were true but I have tried it direct connect with a cable and the result is the same. Lots of good suggestions from all of you guys. I am going to continue looking into it and see if there is some calibration procedure that I missed and post my findings. Thanks.

[@harbselectronicslab3551]

Its not a skylight filter you need it's definately a polarising filter......it cuts out all the reflections of glass, water etc etc....... I use them every day in my job.....yep yep

[@DonDegidio]

Dave, Noticed you have the blue scope probe connected to the yellow input. Would that make any difference? 73 WB3BJU

[@ElPasoTubeAmps]

You sure have a sharp eye for detail. I didn't even notice that but the scope probes work the same in each channel. The colors just help a guy like me that has a hard enough time seeing anything anymore

[@agussugiono2362]

Have you ever tried using regular cable/lead instead of oscilloscope probe for AC Voltmeter?

[@ElPasoTubeAmps]

I have used regular coax cables and get the same results as using oscilloscope probes in the 1X position. The oscilloscope probes in the 1X mode are straight thru cables. The company that makes the AD2 acknowledges that the AC voltmeter is a problem but suggests other ways to get accurate AC volt readings. The AD2 is a nice instrument but it is primarily made for educational and instructional situations where you are instructed to set the instrument up in a certain manner and you will get the results expected. Using it as a laboratory grade instrument may result in answers that are very far off. I like the AD2 but it is best used after you know what the results should be and not as a first approach to solving measurement problems.

[@agussugiono2362]

@@ElPasoTubeAmps Ohh.. Ok. Thanks for the info. About oscilloscope probe, it's not just simple straight wire as far as i know... and it has resistance, about 100-200 ohm.

[@ElPasoTubeAmps]

@@agussugiono2362 I didn't know that about the 1X probes but I just measured two, one was the inexpensive ones that come with the AD2 and the other was a 1X Tektronix and the DCR was 247 ohms and 330 ohms respectively. I learn something every day from you guys. Thank you.

[@jp040759]

What is the AC voltmeter frequency range spec.? Every AC volt meter has a frequency range they guarantee accuracy on. It sure looks like this is rolling off after 1000 hz or so.

[@ElPasoTubeAmps]

It is rolling off bad as the frequency goes up. With a FFT that is usable and a scope that is good, it is hard to believe the voltmeter can be so wrong. I am reading all thoughts and will see what I can find. Thanks.

[@jp040759]

I have an old Fluke 8020A. Its AC volts specs roll off at 2khz. It states it is +-5% inaccurate at 5Khz. My Fluke 87 IV is good to 100Khz. AC Volt meters specs vary.

[@ElPasoTubeAmps]

I have been working with it tonight and it is interesting that my almost 50 year old Triplett 630 is quite accurate well past 20 KHz and even at 50 KHz 7.0 volts reads as 6.8 volts. Unbelievable that this otherwise very nice instrument is basically only accurate as a voltmeter at 60 Hz. The DC is perfect. It will read 5 volts as 5.00 from the Tek calibrator.

[@DAVIDGREGORYKERR]

the input OP amp chip might not have the gain at the higher frequencies, try 1024 bins instead of 16384 bins, using higher bins will make the FFT slower as it takes longer to calculate if you want 4096 bins would be your MAX, when I do a FFT at 32768 bins it does run very slowly so I stick to 4096 bins.

[@ElPasoTubeAmps]

Unfortunately, it is fixed at 4096 BINS. I can't change it to anything else. I agree that this would help and maybe a software upgrade in the (near) future will help. Thanks.

[@umajunkcollector]

I have an older PC scope, haven't messed around with it yet.

[@EJP286CRSKW]

Seems like the AC voltmeter is -3dB at 5kHz. Nothing about it in the specs. Did you try True RMS? Is there an LPF somewhere?

[@ElPasoTubeAmps]

I will search for some LPF setting. It does seem to work OK at all the lower frequencies. Thanks - I hope I can find something.

[@God-CDXX]

NICE AMP