I use to work with sound cards as spectrum analyzer, since about 13 or 15 yrs ago, analysing engine knock sensor signals, while I tune the ignition advance time. Because the frequency of engine knock depends only on the piston bore diameter, and it's perfectly human audible, sound cards are perfect to use. And I can measure two channels at same time: engine phase and the knock sensor. So I can detect what's the real knock, in what cylinder(s) from some other noises like a bad adjusted valves, huge piston clearances and other engine noises. The other way is to use a simple audio amp and listen the signal generated by the sensor. It's obvious the hard way. Very hard.
Yeah, about that same time I was tape recording my chainsaw and playing it back to Windows 98 through Sound Recorder, then analyzing and displaying using Quick Basic. Looked just like a scope with 2 cursors for analysis. Very revealing I tell you. Also recorded firecrackers. It's not what you'd expect. :)
@@antoninperbosc1532 wow... I used to do this more then a decade ago. :P To do something like this now, and make a video, I will need a car/engine with a programmable engine management system (Programmable ECU) with a properly knock sensor, perhaps a good audio pre-amp, an adaptor cable (from the knock sensor to the mic/pre-amp input) and of course, the notebook with a sound card. And a car/engine dynamometer... If you are curious about the engine noises, I found a video from HPA ( th-cam.com/video/66okwgg6Rrw/w-d-xo.html ). I didn't found any kind of video with the same things that I used to do. Sorry. :(
Nice video and brings up some memories of the past. Back in 1997 I used a PC sound card to decode the DCC track signal of my model railroad and display the commands that are coming along the track in real time. The DCC signal is a pulse width modulated square wave with a maximum bit rate of about 16kbps, so it was quite easy to decode with 44.kHz sampling. As interface I used a resistor in each wire to the track and two antiparallel diodes to limit the input voltage on the microphone input. Software was written in Borland Pascal for Windows, running on a 486 with Windows 3.1. Tempi passati ;-)
@@sookmaideek not necessarily. Borland Turbo Pascal and Borland Pascal are different, where Borland Pascal was more high end, and contained more libraries out of the box.
the "loss" of your low-pass filter only comes from the limited low input impedance of your sound card. You can easily check it by using a filter with 1 kΩ and 160 nF. A passive RC-filter is lossless in the passband if driving a high-impedance input.
@@antoninperbosc1532 distortion is not related to the dampening of the signal. The RC filter is linear for a sine wave of constant frequency, even if it is connected to a low impedance input and does not cause distortion. Distortion is generally caused by non-linearities in an amplifier. Other problems are caused by the frequency dependence of the phase shift if one considers a mix of incoming frequencies like voice or music.
It is pretty easy to bypass the capacitor near the input or output jack on the soundcard to be able to input/output DC signals as well. I've done this both for using audio input as an oscilloscope, and for using the output as a signal generator for a laser galvanometer and it worked great for both.
@@iforth64 : AC coupling for soundcards (and a fair variety of other things) is the norm, because it reduces the chances both of ground loops, and of safety hazards.
@@iforth64 DC means V+ and Vzero with V+ will push the speaker forward then Vzero will return the speaker back to natural state. Then AC means V- will pull the speaker backward deeper and quicker to compare Vzero.
@@iforth64 In an amplifier you do not want to amplifie DC offsets since at high levels they will push the amplified signals to one of the powerrails. If that happens the peaks of the signals are cut off which causes extreme high distortion. You can not control if the applied input signal has a DC offset. But with a capacitor in series of the input you can block any DC offset preventing this problem.
Nicely done! I figured that this was possible back in the late 1990's but never found the time to explore using a sound card as an oscilloscope. It's great that other folks had the same idea but actually put in the time to bring it to reality.
I use a similar software (soundcard oscilloscope), its very easy and gives real oscilloscope like interface. I also use two silicon diodes (one forward, one reverse biased) in parallel to the potentiometer on the input signal and ground wire to bypass any AC voltage greater than 700 mV. It saves the soundcard from any accidental over voltage when the potentiometer is not correctly set.
Did the same . But added a 2 pole 6 way switch. It has a though , 50 R , 8 R , 600 R dummy load , a diode for AM demod , rf signal tracing & rectified output for a DVM . Signal tracing can be via the pc speakers if necessary & the microphone boost can up the gain . That's af or rf demodulated signal tracing . ( Noise can be reduced by subtracting the unused channel #2 from #1 ) . I almost fitted a calibrated RF wave absorption & notch filter . But the box wasn't big enough 🤦♂️ Note this is not isolated so not for high voltage or tube work !
Thanks for the contribution, I have a question: I have basic knowledge of electronics but I am actually a musician, as I have a 2-channel Apollo Twin interface like the focusrite, it also has 2 6.3mm inputs and a HI z input for guitars. I wanted to ask, would I also need the 20k potentiometer as a voltage divider for my Apollo Audio interface? Or is it just to protect the audio card that is built into the PC/Laptop? It would be good to use my Apollo as an external audio card to measure signals such as an oscilloscope, wave generator, etc. But I don't want to damage the inputs. Actually I want to calibrate my fostex R8 reel and it asks me to generate a 1Khz, -10dvu (0.3v)Signal waveform so I can adjust the input level. And measure with the millivoltmeter at the terminals of the 245mv channel for an optimal adjustment. Since I don't have the budget right now to buy an oscilloscope, tone generator, and my multimeter doesn't measure millivolts, so I'll get one I think that the voltages that I will handle are very low, so would that circuit with cables that has the 20k potentiometer be necessary? Or can I just use the inputs and outputs of my audio interface directly? Thanks and greetings from Germany By the way, the maximum entry level specifications of my Apollo say: +20.2 dbu which would be the same at 17.99Dvb = 7.93 Volts. Since the maximum output voltages of the Fostex are 0.3 V +/- 1DB then the voltages are very small. 🙂
A deep dive into existing 'stuff' right under my nose; presented fully as always. Amazing what free software is 'out there' as you show. A very helpful video. I, of course, have a wonderful oscilloscope but didn't even think I had one in my PC. Very enlightening.
I just remembered this one ! If the software has a audio analyser but you can't inject a sweeping signal . Try a simple wight noise generator injected into the amp . You will see the passband response on the analyser . It has limited uses with RF also , tune for maximum noise or work from a good audio stage into the radio / unit / DUT . Like passing a 100 khz square wave ( & overtones ) through a radio receiver . Old tricks we used when all we had was time , a box of scrap , war surplus & brass screws !
Using white noise as an input is a great possibility. Particularly with the new broadband SDR receivers. You immediately get the full frequency response.
@@AndreasSpiess Same here. I might not use this info today. But now I know where to look when I *do* need the info. I do need to do some audio experimenting later. Thank you.
White gloves and the tophat always make your videos feel special. It's like you're about to show us diamonds and artifacts. So classy! Also thank you for this. I've needed a way to analyze signals without a scope.
I actually did this years ago with my little netbook when I was putting a new stereo in my mum's car. I don't remember the programs I used but it was very handy to identify which wires went to which speakers.
Easiest thing to experiment with - is to connect transparent diode to line input. Now you will be able to receive IR signals from remote controls. If you write corresponding software of course :)
Nice one Andreas. If you learn something (which we may already know) it's (still) _your great way of presentation_ that makes it fun ... *Love That* 😘👍
Well done again Andreas! I was using something similar few years ago and not just on PC, but also using a tablet! It was working well, for what it is and mainly I had fun with it and so did my friends, for whom I made the cables for. Also old phones can be used to play with this kind of stuff, since the phone is old and not in use! I'm a Tracfone user so every year I get a new phone and my wife also, so at least 2 phones per year are dedicated for all kinds of projects, including Mooshimeter, Pokit Meter and few of my regular meters that have Bluetooth built in for remote measuring. Thank you again Andreas!
Excellent video and a very good introduction to sound cards and their benefits to signal processing. It's also a trip down memory lane... Many years ago (Pentium 3 and 4 era) I used a sound card to decode/encode tone signals in combination with radio equipment (CTCSS, SelCall tone sets like EEA, ZVEI I etc...). What a wonderfull time...
We still use the sound cards to decode signals in amateur radio. This is a cheap and high-quality procedure. I did not know that this trend started already long-time ago.
@@AndreasSpiess The encode/decode with soundcard project was my thesis project. I started with radio communication for my job in the early 90-ties. In that era our relay stations were connected with M1020 orM1040. I started a project (which got cancelled...) to interface radio's with ethernet, today known as RIPS (Radio over IP Streaming) to get rid of the expensive leased lines. A little later... the era of Tetra pushed back all the older equipment and projects to maintain/improve the analog network. I miss this older period somehow... I guess it's a little nostalgia.
@13:38 regarding disadvantage #4 that sound cards don't measure DC: For most soundcards that is true. However some soundcards (such as the MOTU 624) have DC-coupled analog outputs, meaning that those outputs can provide DC (and low frequency) output. Though I don't know if there are soundcards with DC-coupled inputs.
@@AndreasSpiess I was going to suggest that. ADCs are mostly DC-sampling devices and most sound-cards should be able to sample DC by shorting the input caps, unless high-pass filtering was done in the feedback or the DSP.
There's alot you can do with a PC but it is not realtime due to latencies in the operating system and USB protocol. This doesn't stop you from doing non-realtime (record and analyse) however for an oscilloscope realtime is needed for trigging. Only reason you may not want to use a PC is isolation - tube amps with hundreds of volts and PC are not a good combination without specialist hardware. A differential probe helps and could be used with a sound card - isolating and being safe. The only down side to a PC sound card is the lack of control over the processing - some operating systems perform conversion, and apply human ear perception equalisation etc. MacOS used to be a pain for that, the later operating system allows the developer better options to prevent this. I wrote a Bode plot AWG for my 1104x-e and currently writing a high resolution spectrum analyser (1Mpt and 1MSPS). You may want to use flattop filtering on the FFT as this provides the best measurement value.
Triggering does not need realtime. Just write the data into a ring buffer and evaluate any trigger condition at your time. This also enables pre-trigger, where you see some data leading to the trigger event.
What realtime is fast enough for you? It is easy to get ~100ms latency. ~10ms is harder, but possible too. ~1ms looks like maximum of what PC + sound card can achieve.
That's why you don't use WaveRT/DirectSound Drivers on Windows, or CoreAudio on MacOS. You have to use audio drivers which bypass any OS/Driver processing, putting the audio codec in a sort of "promiscuous mode". I don't know how that can be done in MacOS, but in Windows you use ASIO drivers for that. ASIO4ALL project provides a free of charge ASIO driver which works with the common Realtek codecs which don't provide ASIO by default.
15 yrs ago I used an old desktop sound card as an oscilloscope as well. To reverse engineer 1 wire 300 bps car head unit internal communication protocol between each module and it works very well.
I have tinkered with sound cards as oscilloscopes. I'd suggest connecting diodes in parallel to the 3.5mm mic input to clip their voltages to 0.7V to protect the sound card in case of input over voltage.
My sound card (SB Live!) survives even signals with 4 diodes protection (twice your level). I suspect that it may be capable to use even higher voltages.
I've used REW and a Dayton Audio UMM-6 USB calibration microphone to set up a couple church sound systems. That software is crazy powerful -- I've seen commercial solutions that aren't as comprehensive. Anyone interested in playing with sound systems, especially measuring speaker response, would do well with this combination. 👍️
@@AndreasSpiess Goodness... if it falls under the Acoustic domain, it's probably measurable. I won't claim to know everything it's capable of, though. 😁
Super cool! Very interesting in this topic. I've search for these setups years early, however, I did not able to fully setup a set of software and hardware in my computer. Thank you for sharing this!
I bought an Asus sound card recently for this purpose, in the manual it even shows you how to use ARTA to measure the specs of the mic input. My problem till now has been that all the sound cards I test inputs have been far from flat, some roll-off the bottom end, or top, or both. On some, response is all over the place. REW can calibrate a soundcard, it is great for removing a decibel here and there, but it can only do so much, you need somewhat decent hardware to begin with
Fantastic ! Will use this today; am building some ‘home brew’ filters for a (music instrument ) synthesiser so frequencies I’m interested in are all audio, yes can calculate it all but can’t beat empirical validation of your work.
I like adding small serial resistance (100 Ω) and two (or four) diodes as protection for overvoltage. It will be possible to put 10 V signal without damaging hardware with such setup. While low voltage signals will pass through almost not distorted.
I was thinking the same! but what about measuring mains voltages? Even if you trim / clamp down with diodes couldn't a single line be somewhat referenced to your soundcard and cause damage?
@@MattReb090 I'm not experienced enough to work with mains voltages, so I can only guess. I think transformers may help - secondary voltages should be isolated and proportional to primary. Also it is worth trying to make capacitive coupling. Just putting long wire next to mains line - making antenna in other words.
Some time ago, before I got a scope, I considered using one of my audio interfaces as a scope. The 24-bit resolution on the converters are exquisite, but the sample rate leaves much to be desired. My scope is 1 GSa/s, whereas my interface maxes at 192 kHz. (But I can measure up to 10 channels! Or if I use multiple interfaces, 34 channels at 96 kHz.) I think I did actually try this at one point, but I seem to remember there was a huge DC offset, and I was worried I'd blown something. They all have input caps, but... still scary. Never again. (Fortunately, they're all good.)
After watching the video in full, I take back my "never again" comment. I can see how valuable it'll be in the audio domain. As long as I know I won't cause problems for myself. Idiot-check with the scope or DMM, then detailed measurements with REW. 😁
I I've not checked all the comments but it it's important to to discuss the built-in hardware Nyquest fitters. The bandwidth of audio frequency oriented devices is much more limited as hinted at, but it is an unsumountable drawback if you want a more general purpose oscilloscope.
I have been using the free DAQARTA demo program which allows unlimited use of its oscilloscope and a few other functions, while its signal generator is limited to 30 times or 30 days use after which you need to purchase it or re-install the program for another 30 times/days of fully functioning demo, which includes its most important tool to me: an LCR/ESR meter which can measure as low as 1mΩ resistance, 1pF to at least 10,000uF and less than a uH resolution (I think as low as 1nH), though such low measurements are filled with noise, depending on the soundcard's S/N ratio. I use an Audiophile 2496 soundcard, so the noise is at least 10 times lower than on the standard cards.
@@AndreasSpiess Even with my 2496 sound card, actual usable resolution for inductivity is 1uH, with 0.1uH digit changing all the time and you sort of see which number is the most frequent. Below that, numbers are changing too much to be useful. All that is needed besides soundcard (and a cable) is a reference resistor and probes, and the ability to short out the reference resistor when calibrating the software. I plan to add an external output amplifier so that I can use a lower value REF resistor (improves S/N ratio) and a pre-amp at the probe (also to improve S/N ratio). One channel is used as an output, and 2 channels are used as an input - one takes signal before and the other after the REF resistor, comparing their phase shift and relative values to automatically give out results as capacitance or inductance, along with ESR and DF for caps and resistance and Q factor for inductors.
No one here seems to have mentioned USB isolators yet. Isolation on the USB side would seem a good general purpose approach that would be transferable between lots of use-cases where you want your precious PC connected to a bench of not-yet-proven project. A USB isolator must be a fairly tractable thing to implement, but I've done a load of searching on Amazon and not found anything that definitively performs this task. I'm sure they must exist and I'm just failing to find them. Thoughts ?
Right! I've always wondered about this, I had (sadly gone) a 192khz 24-bit sound card with 8 channels. No, it couldn't compete with a 100Mhz oscilloscope, but measured signals seemingly with incredible fidelity with very low SNR, onboard filtering etc. Certainly usable to observe behaviour of most MCUs
If you get a professional sound device (as opposed to an Ali-Express "professional(tm)" sound thing) you should be able to handle up to about 24Vpp of both input and output levels without distortion. Obviously if you are working with unknown-level signals it would be a good idea to build a +-15V clipper for the input signals.
@@AndreasSpiess The Scarlett qualifies as a professional device. Basically if it has an XLR jack and 6mm TRS jacks, it is likely to be +20dBu or better. It should be able to output similar levels into 600 ohms. Pretty much all of these devices run the internal opamps from +-15V, so subtract about 6V from that to get the max direct input level or max output level.
Unfortunately very few modern computers and external sound cards have stereo line input, a mono mic is predominant, which makes it difficult to input signals. My oldish PC is good in this respect. Two other free softwares are Digilent Waveforms (front end for their instruments but includes a sound card option) Soundcard Oscilloscope by Christian Zeitnitz - a bit dated interface but does the job reliably
You should look at more expensive sound cards. I have a thing sitting beside me with stereo mic inputs, and 8 channels of balanced line inputs and 8 channels of balanced line outputs. Of course it cost me a few hundred bucks.
@@alexandern8671 No it doesn't. The title of the video isn't "make a scope for free", it is "use your sound card as a scope". A $300 sound adapter is still a tenth the cost of a $3000 scope, which is about the minimum these days for a decent scope. And the scope won't have the gain dynamic range and noise floor of the sound card, nor is it likely to have FFT or spectrum analyzer capabilities, and may not have a way to record what it sees on the screen. And you can get that $300 sound card for maybe half that used on eBay. I've got both a scope and a cheap PC with a good sound adapter, and I use both. They have different strengths.
@@lwilton Agree with your point regarding the title of the video. However I do not think someone with a budget of $300 should invest in a sound card for the purpose of scoping some signals. Andreas has discussed the downsides - no DC, limited frequency range. If one has such a card already like you then of course it is nice to repurpose it though. I can think of some niche applications where the increased vertical resolution could make a big difference but they are hardly the mainstream. Can you please give some common use cases that would benefit from increased vertical resolution?
I think the point is that if you don't need it for high precision and rarely, it's probably good enough. If you're doing anything that requires serious precision, you'd probably want to just buy a quality bit of kit.
🙂 By the way, the maximum entry level specifications of my Apollo say: +20.2 dbu which would be the same at 17.99Dvb = 7.93 Volts. Since the maximum output voltages of the Fostex are 0.3 V +/- 1DB then the voltages are very small.
Thanks for your video! I think this is a great idea! I just wanted to say, that maybe you can use an optocoupler between the signal and the sound card's input. This would protect the sound card in a better way than the potentiometer.
Opto-isolators usually introduce a lot of non-linearities (as sdjhgfkshf mentioned). So this would disturb the signal. Maybe it would be better to use an external soundcard and a USB insulator.
The frequency sweep at 4:30 is -1.4 dB below max and in just one channel. No warning needed, commercials are louder. Some people have to turn up their sound so much because of weak sound often heard right here on TH-cam. Set up your soundcard playback level to mainstream sources not amateur streamers, then tell them to turn up their levels to standards. Such is the power of Audacity to scope audio streamed on the web. Wasapi mode instead of line in. Your voice track is -7.295 dB which is weak for pro production standards. Audacity has several tricks including spectrum display as well the details that I have gotten out of this video. It's just not realtime.
While audio card oscilloscope is handy, there's some caveats; the processors and / or ADCs, and DACs used by most sound cards tended to be slow so 48 kilohertzs bandwidth is usable at most, if that. However, it's cheaper to do it and get done quickly. And I strongly recommend that the line or microphone inputs be protected - the same way you would do when putting together a few more extra coaxial oscilloscope leads for the most expensive kit.
I was duing some experiments with audio signals and misusing a sound card. I was using just an ordinary integrated Realtek HD Audio. I've discovered that the sound chip is doing some funny stuff to the signal. Namely it seems it has some kind of autogain, which can't be controlled or turned off from the software. Also laptops usually have only a microphone input, which is more difficult to use for recording arbitrary audio signals. So it's a good idea to carefully test your equipment, if it's suitable for the task.
That's true, especially for laptops that have limited I/O and usually combine headphone and mic together. The easiest and less risky way is to use an external USB audio adapter with proper line in/out. Worse case you kill the adapter but keep you computer safe.
Using an external sound card is what I recommended, too. They are no more very expensive. But I agree, there is always a possibility that the signal is distorted by the PC.
back around Windows-95 in some astronomy magazine. Someone taken a PC and a 8.5" LDA photo-sensor from a simple document scanner. Then relocating the sensor to point at the night sky and with a little sensor on/off timer was able to scan the stars and moon as the earth rotated giving a basic image for very little money. 🌠🌛 another neat project 🪐 thanks 🤩
For maybe 20 dollars you should be able to get an audio-quality stereo I2S ADC and DAC, feed them through fast optocouplers, and send that data to a USB MCU that can send it to your computer. Arguably there are some computers that will take digital audio signals like SPIDF and some ADC/DACs that can produce it such that you don’t even need the MCU. Then I’d have a 6+ resistor range switch followed by clipping diodes for the input, and a couple of different gain amplifier settings for the output. Maybe an input buffer amplifier running on some larger voltage rails, for input impedance purposes.
@@AndreasSpiess from what I hear the Arduino Pro Micro has a fair bit of code written for it out there, acting like a sound card is almost certainly one of those. If that bitrate isn’t fast enough then the same should apply to the STM32F103, though they’re a little tougher to get programmed. Another option is using an FT232 and SPI ADCs/DACs, not sure how much programming they need, but I’d guess it would be less than the STM32 and still give you full speed USB2.0. For USB3.0 purposes there’s the Cypress FX3, but from what I hear that’s a pain to program and it’s only really necessary for really high bitrates or SDR territory.
I've certainly tried stuff like this throughout the years... There are just times when u want a particular waveform or pattern of waveforms. However, I would highly HIGHLY recommend to not use any direct lines into the computer. Just get ANY gear which is external to your computer. Great for making amps & guitar pedals :) Also, make sure to get an audio amplifier for your output. Output chain looks like: PC>audio interface > amplifier Input chain looks like: custom opamp divider > audio interface > PC I feel like this is a second step in the path of learning EE. First step would be to get a signal generator + oscilloscope. Caz, how are u going to test ur custom opamp divider thingy if u don't have suitable equipment to begin with ? REW is awesome. An absolute staple in the kit...
It is certainly good to protect your PC. But for experimenting with audio signals I think this is a valid alternative. Not for an oscilloscope, of course...
@@mikejones-vd3fg Just to be more clear with what I wrote caz it see it's kinda hard to read. I highly recommend against using any direct connections to your motherboard. I have fried 2 & a usb dock. What happens is that device fails. For example: I was using ardruino plugged directly into motherboard via usb = It took years but eventually I burnt something by accidentally crossing wires for the slightest of time :((. Just a touch. Oops. Now all the front panel usb's do not work. Looking at the motherboard >> power goes into the usb IC, but its dead. I got smart and started using usb docks. Resulting in another dead usb. But at least it's a dock & can be replaced. It's an accident waiting to happen. This is mainly caz computer devices are not designed to handle over-voltages. They're designed to stabilize voltages. Specifically about a soundcard, 9v is way above the "line in" rating of 1.5v. I don't know how that would turn out. Sounds like a bad plan. You would have to know how the front end is designed. i wouldn't be surprised if putting any DC into the soundcard for a period of time would mess things up. At least if u were to buy a usb soundcard, then that usb device would fry & not your computer motherboard. In general AD/DA's are pretty sensitive in my experience. I've seen many die from normal use, let alone outside recommended ranges. If you think about it > the PSU on a computer maintains the voltages precisely. They will dump large amounts of current just to keep voltages correct. I think my recommendation of using an audio interface is good. Many can handle +4 signal levels :) I've used a ton of these and never had a problem. I still think it's wise to design an op amp divider/buffer so it ranges your input correctly. Also an opamp has a known current output. It's a good thing to learn. . . Don't forget/overlook that you can set the opamp gain to at any level you wish. Then you can see all the small signals when the input is off, like the noise floor/crosstalk/signal bleed/etc ... Another thing to know is oscilloscope probes are amazing. If you compare alligator clips to Oscope probes, the noise pickup is night and day. Noise&signals will go right into the wires of alligator clips. Not with Oscope probe wires. Have fun, don't burn the mobo like I did ;p
Java driver is limited by your windows setup of soundcard output profile, it defaults to 44100ksps. Many proper soundcards support 192 ksps which is why 96 kHz works with the other driver.
@@AndreasSpiess Sorry if this is a double posted . TH-cam has deleted 2 posts with the links 😒 You mix the scope input with a reference frequency . ( Basically it works like a mixer & the output is on a sampling capacitor ) . So you get a very high bandwidth scope with limited hardware . Early high bandwidth scope used the technique as the input is repetitive so it works even with tubes ! ( Tektronix is the first name that comes to mind ) A 100 Mhz sine wave on a 22 Khz pc scope or old o-scope is like magic . Until you're shown how it's done ! The best I can find on the net is a tdr head project “ 1N5711 Sampling Head ” but the modus operandi is the same .
One of my favorites of all your videos. I'm not a big radio guy, so those videos are of limited interest, though I usually watch them to observe your design and assembly techniques. This one I can put to immediate use. Already ordered a USB Sound Card. Great stuff. Thanks!
The Java sound device not being able to generate beyond 22 kHz points to it internally capping at CD sample quality, 44.1 kHz. Per the Nyquist-Shannon sampling theorem, the highest reproducible frequency is half the sampling rate used.
Optocouplers are not linear and would not help in this situation because they would disturb the signal. You can use an isolated USB connector to connect an external soundcard and it would work.
Interesting. There was a old project ELFRAD That was analysing earth signals down to wavelengths days long if I remember right (earthquake prediction, the moon even) This would probably be ideal for something like that exploring those signals.
Hah! I did a soundcard oscilloscope project ~28 years ago when I was in college. Now I design IC test equipment for a living and I've always thought that a soundcard front end would be a fun project.
Many years ago a friend and I used to maintain a microwave radio scheme which used an analogue multiplexer We challenged ourselves to decode the baseband with just a laptop and no hardware (except a cable). Cooledit on a Windows 98 laptop did the job...
@@AndreasSpiess thank you for the reply, this was an analogue multiplexer (Granger DTL-7300) the individual voice channels are SSB with carriers at 4,8,12, 16kHz and extending further up, these are then modulated onto a sub carrier for transmission over a microwave FM link. Fascinating things to play with! 73
I already have a pretty good scope, but I think I'll stick to physical circuits when it comes to signal generators. Having said that though, these programs do look usefull.
I use to work with sound cards as spectrum analyzer, since about 13 or 15 yrs ago, analysing engine knock sensor signals, while I tune the ignition advance time. Because the frequency of engine knock depends only on the piston bore diameter, and it's perfectly human audible, sound cards are perfect to use. And I can measure two channels at same time: engine phase and the knock sensor. So I can detect what's the real knock, in what cylinder(s) from some other noises like a bad adjusted valves, huge piston clearances and other engine noises. The other way is to use a simple audio amp and listen the signal generated by the sensor. It's obvious the hard way. Very hard.
Yeah, about that same time I was tape recording my chainsaw and playing it back to Windows 98 through Sound Recorder, then analyzing and displaying using Quick Basic. Looked just like a scope with 2 cursors for analysis. Very revealing I tell you. Also recorded firecrackers. It's not what you'd expect. :)
These are very good examples of how this technology can be used. Thank you for mentioning them!
Interesting mmm
@Chiavaloni Now you win the right to made us a video to show us this interesting howto !!! 😁
@@antoninperbosc1532 wow... I used to do this more then a decade ago. :P To do something like this now, and make a video, I will need a car/engine with a programmable engine management system (Programmable ECU) with a properly knock sensor, perhaps a good audio pre-amp, an adaptor cable (from the knock sensor to the mic/pre-amp input) and of course, the notebook with a sound card. And a car/engine dynamometer... If you are curious about the engine noises, I found a video from HPA ( th-cam.com/video/66okwgg6Rrw/w-d-xo.html ). I didn't found any kind of video with the same things that I used to do. Sorry. :(
Nice video and brings up some memories of the past. Back in 1997 I used a PC sound card to decode the DCC track signal of my model railroad and display the commands that are coming along the track in real time. The DCC signal is a pulse width modulated square wave with a maximum bit rate of about 16kbps, so it was quite easy to decode with 44.kHz sampling. As interface I used a resistor in each wire to the track and two antiparallel diodes to limit the input voltage on the microphone input. Software was written in Borland Pascal for Windows, running on a 486 with Windows 3.1. Tempi passati ;-)
turbo pascal
@@sookmaideek not necessarily. Borland Turbo Pascal and Borland Pascal are different, where Borland Pascal was more high end, and contained more libraries out of the box.
@@thorbjrnhellehaven5766 turbo better because faster and turbo
You were very early using this technology! And you stayed with the model railroads. Now with much more sophisticated gear, of course ;-)
Borland Pascal for Window, yes those were the days, ... ~1984
the "loss" of your low-pass filter only comes from the limited low input impedance of your sound card. You can easily check it by using a filter with 1 kΩ and 160 nF. A passive RC-filter is lossless in the passband if driving a high-impedance input.
That or a opamp buffer as medium between the sound card and the RC filter
I agree with both statements.
so puting an AOP in follower with hight impedance to "isolate" the DUT could avoid distorsion ?
@@antoninperbosc1532 distortion is not related to the dampening of the signal. The RC filter is linear for a sine wave of constant frequency, even if it is connected to a low impedance input and does not cause distortion. Distortion is generally caused by non-linearities in an amplifier. Other problems are caused by the frequency dependence of the phase shift if one considers a mix of incoming frequencies like voice or music.
It is pretty easy to bypass the capacitor near the input or output jack on the soundcard to be able to input/output DC signals as well. I've done this both for using audio input as an oscilloscope, and for using the output as a signal generator for a laser galvanometer and it worked great for both.
Good to know. Thank you for the tip!
The 'no-DC' issue I did not anticipate. Is there a technical reason why the soundcard is AC-coupled (e.g. related to it having a 18-20 bit ADC)?
@@iforth64 : AC coupling for soundcards (and a fair variety of other things) is the norm, because it reduces the chances both of ground loops, and of safety hazards.
@@iforth64 DC means V+ and Vzero with V+ will push the speaker forward then Vzero will return the speaker back to natural state. Then AC means V- will pull the speaker backward deeper and quicker to compare Vzero.
@@iforth64 In an amplifier you do not want to amplifie DC offsets since at high levels they will push the amplified signals to one of the powerrails. If that happens the peaks of the signals are cut off which causes extreme high distortion. You can not control if the applied input signal has a DC offset. But with a capacitor in series of the input you can block any DC offset preventing this problem.
Nicely done! I figured that this was possible back in the late 1990's but never found the time to explore using a sound card as an oscilloscope. It's great that other folks had the same idea but actually put in the time to bring it to reality.
Indeed, soundcards are high performance devices which are also used by software defined radios, for example. Just different software...
I use a similar software (soundcard oscilloscope), its very easy and gives real oscilloscope like interface. I also use two silicon diodes (one forward, one reverse biased) in parallel to the potentiometer on the input signal and ground wire to bypass any AC voltage greater than 700 mV. It saves the soundcard from any accidental over voltage when the potentiometer is not correctly set.
Good idea with the diodes. And I placed a link to the Soundcard Oscilloscope SW. Thank you!
Did the same .
But added a 2 pole 6 way switch.
It has a though , 50 R , 8 R , 600 R dummy load , a diode for AM demod , rf signal tracing & rectified output for a DVM .
Signal tracing can be via the pc speakers if necessary & the microphone boost can up the gain .
That's af or rf demodulated signal tracing .
( Noise can be reduced by subtracting the unused channel #2 from #1 ) .
I almost fitted a calibrated RF wave absorption & notch filter .
But the box wasn't big enough 🤦♂️
Note this is not isolated so not for high voltage or tube work !
Thanks for the contribution, I have a question:
I have basic knowledge of electronics but I am actually a musician, as I have a 2-channel Apollo Twin interface like the focusrite, it also has 2 6.3mm inputs and a HI z input for guitars.
I wanted to ask, would I also need the 20k potentiometer as a voltage divider for my Apollo Audio interface? Or is it just to protect the audio card that is built into the PC/Laptop?
It would be good to use my Apollo as an external audio card to measure signals such as an oscilloscope, wave generator, etc.
But I don't want to damage the inputs.
Actually I want to calibrate my fostex R8 reel and it asks me to generate a 1Khz, -10dvu (0.3v)Signal waveform so I can adjust the input level.
And measure with the millivoltmeter at the terminals of the 245mv channel for an optimal adjustment.
Since I don't have the budget right now to buy an oscilloscope, tone generator, and my multimeter doesn't measure millivolts, so I'll get one
I think that the voltages that I will handle are very low, so would that circuit with cables that has the 20k potentiometer be necessary? Or can I just use the inputs and outputs of my audio interface directly?
Thanks and greetings from Germany
By the way, the maximum entry level specifications of my Apollo say: +20.2 dbu which would be the same at 17.99Dvb = 7.93 Volts.
Since the maximum output voltages of the Fostex are 0.3 V +/- 1DB then the voltages are very small.
🙂
A deep dive into existing 'stuff' right under my nose; presented fully as always. Amazing what free software is 'out there' as you show. A very helpful video. I, of course, have a wonderful oscilloscope but didn't even think I had one in my PC. Very enlightening.
Glad you liked the content. Not a replacement for your oscilloscope, but maybe an estension...
I just remembered this one !
If the software has a audio analyser but you can't inject a sweeping signal .
Try a simple wight noise generator injected into the amp .
You will see the passband response on the analyser .
It has limited uses with RF also , tune for maximum noise or work from a good audio stage into the radio / unit / DUT .
Like passing a 100 khz square wave ( & overtones ) through a radio receiver .
Old tricks we used when all we had was time , a box of scrap , war surplus & brass screws !
Using white noise as an input is a great possibility. Particularly with the new broadband SDR receivers. You immediately get the full frequency response.
Yet another useful AND interesting video. Once again, you've saved me a week of research :-) Thank You!
Glad you liked the video!
@@AndreasSpiess Same here. I might not use this info today. But now I know where to look when I *do* need the info. I do need to do some audio experimenting later. Thank you.
White gloves and the tophat always make your videos feel special. It's like you're about to show us diamonds and artifacts. So classy!
Also thank you for this. I've needed a way to analyze signals without a scope.
Never looked at it this way. The hat and the gloves just happened...
I actually did this years ago with my little netbook when I was putting a new stereo in my mum's car. I don't remember the programs I used but it was very handy to identify which wires went to which speakers.
Cool idea to use it for such a project!
Easiest thing to experiment with - is to connect transparent diode to line input.
Now you will be able to receive IR signals from remote controls.
If you write corresponding software of course :)
Good idea. Most IR remote signals are around 38kHz. So you probably have to use a high sampling rate, I think.
@@AndreasSpiess it is possible with low sample rate too.
Nice one Andreas. If you learn something (which we may already know) it's (still) _your great way of presentation_ that makes it fun ... *Love That* 😘👍
Thank you for your kind words!
Well done again Andreas! I was using something similar few years ago and not just on PC, but also using a tablet! It was working well, for what it is and mainly I had fun with it and so did my friends, for whom I made the cables for. Also old phones can be used to play with this kind of stuff, since the phone is old and not in use! I'm a Tracfone user so every year I get a new phone and my wife also, so at least 2 phones per year are dedicated for all kinds of projects, including Mooshimeter, Pokit Meter and few of my regular meters that have Bluetooth built in for remote measuring. Thank you again Andreas!
You are right, Smartphones contain a lot of good technology and can be used also when they are old.
Excellent video and a very good introduction to sound cards and their benefits to signal processing. It's also a trip down memory lane... Many years ago (Pentium 3 and 4 era) I used a sound card to decode/encode tone signals in combination with radio equipment (CTCSS, SelCall tone sets like EEA, ZVEI I etc...). What a wonderfull time...
We still use the sound cards to decode signals in amateur radio. This is a cheap and high-quality procedure.
I did not know that this trend started already long-time ago.
@@AndreasSpiess The encode/decode with soundcard project was my thesis project. I started with radio communication for my job in the early 90-ties. In that era our relay stations were connected with M1020 orM1040. I started a project (which got cancelled...) to interface radio's with ethernet, today known as RIPS (Radio over IP Streaming) to get rid of the expensive leased lines. A little later... the era of Tetra pushed back all the older equipment and projects to maintain/improve the analog network. I miss this older period somehow... I guess it's a little nostalgia.
This week ive learned about how SDR works with your dual channel sound card thanks for this video!
@13:38 regarding disadvantage #4 that sound cards don't measure DC: For most soundcards that is true. However some soundcards (such as the MOTU 624) have DC-coupled analog outputs, meaning that those outputs can provide DC (and low frequency) output. Though I don't know if there are soundcards with DC-coupled inputs.
Thank you for the info. Another viewer also wrote that he removed/bridged the input capacitors on a "normal" Soundcard
@@AndreasSpiess I was going to suggest that. ADCs are mostly DC-sampling devices and most sound-cards should be able to sample DC by shorting the input caps, unless high-pass filtering was done in the feedback or the DSP.
My lab cat is very fascinated by todays presentation because of all those moving cursors :-)
I can imagine that this is interesting for a cat ;-)
you’re channel is a hidden gem 💎👍
I would hope it becomes a little less "Hidden" ;-)
Thank you!
There's alot you can do with a PC but it is not realtime due to latencies in the operating system and USB protocol. This doesn't stop you from doing non-realtime (record and analyse) however for an oscilloscope realtime is needed for trigging. Only reason you may not want to use a PC is isolation - tube amps with hundreds of volts and PC are not a good combination without specialist hardware. A differential probe helps and could be used with a sound card - isolating and being safe.
The only down side to a PC sound card is the lack of control over the processing - some operating systems perform conversion, and apply human ear perception equalisation etc. MacOS used to be a pain for that, the later operating system allows the developer better options to prevent this.
I wrote a Bode plot AWG for my 1104x-e and currently writing a high resolution spectrum analyser (1Mpt and 1MSPS). You may want to use flattop filtering on the FFT as this provides the best measurement value.
Triggering does not need realtime. Just write the data into a ring buffer and evaluate any trigger condition at your time. This also enables pre-trigger, where you see some data leading to the trigger event.
What realtime is fast enough for you?
It is easy to get ~100ms latency. ~10ms is harder, but possible too. ~1ms looks like maximum of what PC + sound card can achieve.
@
Nick K: I agree on the limitations you mention. Particularly I would not use it with tube amplifiers ;-) Fortunately, they are rare these days.
That's why you don't use WaveRT/DirectSound Drivers on Windows, or CoreAudio on MacOS. You have to use audio drivers which bypass any OS/Driver processing, putting the audio codec in a sort of "promiscuous mode". I don't know how that can be done in MacOS, but in Windows you use ASIO drivers for that. ASIO4ALL project provides a free of charge ASIO driver which works with the common Realtek codecs which don't provide ASIO by default.
Awesome video Andreas! And thanks for the tip to the kickstarter project you showed as well!
I thought the kickstarter was a good fit because it reduced the complexity...
If you build the "Artabox" for REW you can use it as very accurate LCR Meter.
Good to know. Thanks!
genius. I had been thinking about this a while back... You've gone well beyond mere thought!
awesome
Glad you like the idea!
Love that little hand pointing thing! Great job with the video.
Glad you enjoyed it!
I love how the sweep test just randomly interferes with the talking 😂
That one was easy...
Very GOOD, I am an audio projects addicted, I built my own Stereos and pre-amps. I will try this tool. Thanks
You will love it, I am sure!
15 yrs ago I used an old desktop sound card as an oscilloscope as well. To reverse engineer 1 wire 300 bps car head unit internal communication protocol between each module and it works very well.
So you were happy that the levels changed a lot and the signal behaved like AC... Today I suggest to use a cheap logic analyzer for digital signals.
Thank you Andreas. And thank you for not following the "Make a stupid shocked face thumbnail" trend on You tube!
You are welcome!
Andreas: Man you are our Gem.
Merci Veilmals fur das Lecture.
Gern geschehen!
I have tinkered with sound cards as oscilloscopes. I'd suggest connecting diodes in parallel to the 3.5mm mic input to clip their voltages to 0.7V to protect the sound card in case of input over voltage.
My sound card (SB Live!) survives even signals with 4 diodes protection (twice your level).
I suspect that it may be capable to use even higher voltages.
It is well possible that some cards support higher levels. My Focusrite seems to support 25Vpp. But only at the lowest gain...
Good idea.
I've used REW and a Dayton Audio UMM-6 USB calibration microphone to set up a couple church sound systems. That software is crazy powerful -- I've seen commercial solutions that aren't as comprehensive. Anyone interested in playing with sound systems, especially measuring speaker response, would do well with this combination. 👍️
Thank you for sharing your experience. So you understand much more about all the measurements possible with REW...
@@AndreasSpiess Goodness... if it falls under the Acoustic domain, it's probably measurable. I won't claim to know everything it's capable of, though. 😁
Very interesting and useful video Andreas! 😄
Thanks! 😃
Great Stuff ! Never have thought that you can use a soundcard for this, also not aware for the free software programs for this, thanks !
You are welcome!
Super cool! Very interesting in this topic. I've search for these setups years early, however, I did not able to fully setup a set of software and hardware in my computer. Thank you for sharing this!
Glad it was helpful!
Very good explanation...he gave a professional class...thanks and greetings from Paraguay
Glad you liked the video!
Also a VLF digitizer paired with SDR software for receiving and decoding VLF transmissions.
You are right: Many SDR receivers use the soundcard. And all digital modes in HAM radio, too.
Dear Andreas, you are one of my favorites.
You drove my cats crazy with the sweep sample🤬. Still love your channel... and the accent bdw.
Sorry for that!
Very interesting video. I was using number of "PC scope" like programs, but about REW and Artalabs didn't know. Thank You Andreas.
You are welcome!
Wonderful! That's going to be great stuff for experiments with my 12 yo in the workshop 👍
I hope she/he will like it. Good it is audio where kids probably can relate because they hear the sound.
I bought an Asus sound card recently for this purpose, in the manual it even shows you how to use ARTA to measure the specs of the mic input.
My problem till now has been that all the sound cards I test inputs have been far from flat, some roll-off the bottom end, or top, or both. On some, response is all over the place. REW can calibrate a soundcard, it is great for removing a decibel here and there, but it can only do so much, you need somewhat decent hardware to begin with
As I showed it does not work for DC as well as above 1/2 the sampling rate. This cannot be changed.
Fantastic ! Will use this today; am building some ‘home brew’ filters for a (music instrument ) synthesiser so frequencies I’m interested in are all audio, yes can calculate it all but can’t beat empirical validation of your work.
Checking if it works is always rewarding...
After watching every video of yours I learn something, and often a surprise--a bonus! ;)
Thank you for your kind words!
I like adding small serial resistance (100 Ω) and two (or four) diodes as protection for overvoltage.
It will be possible to put 10 V signal without damaging hardware with such setup.
While low voltage signals will pass through almost not distorted.
Good point. Thanks!
I was thinking the same! but what about measuring mains voltages? Even if you trim / clamp down with diodes couldn't a single line be somewhat referenced to your soundcard and cause damage?
@@MattReb090 I'm not experienced enough to work with mains voltages, so I can only guess. I think transformers may help - secondary voltages should be isolated and proportional to primary. Also it is worth trying to make capacitive coupling. Just putting long wire next to mains line - making antenna in other words.
Some time ago, before I got a scope, I considered using one of my audio interfaces as a scope. The 24-bit resolution on the converters are exquisite, but the sample rate leaves much to be desired. My scope is 1 GSa/s, whereas my interface maxes at 192 kHz. (But I can measure up to 10 channels! Or if I use multiple interfaces, 34 channels at 96 kHz.) I think I did actually try this at one point, but I seem to remember there was a huge DC offset, and I was worried I'd blown something. They all have input caps, but... still scary. Never again. (Fortunately, they're all good.)
After watching the video in full, I take back my "never again" comment. I can see how valuable it'll be in the audio domain. As long as I know I won't cause problems for myself. Idiot-check with the scope or DMM, then detailed measurements with REW. 😁
Indeed they should not be compared 1:1. Both have advantages and disadvantages.
I
I've not checked all the comments but it it's important to to discuss the built-in hardware Nyquest fitters. The bandwidth of audio frequency oriented devices is much more limited as hinted at, but it is an unsumountable drawback if you want a more general purpose oscilloscope.
I agree.
I have been using the free DAQARTA demo program which allows unlimited use of its oscilloscope and a few other functions, while its signal generator is limited to 30 times or 30 days use after which you need to purchase it or re-install the program for another 30 times/days of fully functioning demo, which includes its most important tool to me: an LCR/ESR meter which can measure as low as 1mΩ resistance, 1pF to at least 10,000uF and less than a uH resolution (I think as low as 1nH), though such low measurements are filled with noise, depending on the soundcard's S/N ratio.
I use an Audiophile 2496 soundcard, so the noise is at least 10 times lower than on the standard cards.
I did not know that they could do such precise measurements!
@@AndreasSpiess Even with my 2496 sound card, actual usable resolution for inductivity is 1uH, with 0.1uH digit changing all the time and you sort of see which number is the most frequent. Below that, numbers are changing too much to be useful.
All that is needed besides soundcard (and a cable) is a reference resistor and probes, and the ability to short out the reference resistor when calibrating the software.
I plan to add an external output amplifier so that I can use a lower value REF resistor (improves S/N ratio) and a pre-amp at the probe (also to improve S/N ratio).
One channel is used as an output, and 2 channels are used as an input - one takes signal before and the other after the REF resistor, comparing their phase shift and relative values to automatically give out results as capacitance or inductance, along with ESR and DF for caps and resistance and Q factor for inductors.
Very interesting and honest (about the limitations) video.
Glad you liked it!
No one here seems to have mentioned USB isolators yet. Isolation on the USB side would seem a good general purpose approach that would be transferable between lots of use-cases where you want your precious PC connected to a bench of not-yet-proven project. A USB isolator must be a fairly tractable thing to implement, but I've done a load of searching on Amazon and not found anything that definitively performs this task. I'm sure they must exist and I'm just failing to find them. Thoughts ?
A good idea. I added a link in the description.
Magnífico!!!! I never thought that could be done, thank you very much Andreas You are the best
You are welcome!
I was waiting for this video for so many years.....
So it is good it appeared now ;-) Glad I matched your wishes.
Its amazing how a Simple RC circuit is still usable.
It is physics which probably will not change soon ;-)
great and useful video. It's a pleasure to learn new things every day. Your channel is the best for DIY electronics.
Glad you like the channel!
Right! I've always wondered about this, I had (sadly gone) a 192khz 24-bit sound card with 8 channels. No, it couldn't compete with a 100Mhz oscilloscope, but measured signals seemingly with incredible fidelity with very low SNR, onboard filtering etc. Certainly usable to observe behaviour of most MCUs
Indeed, we should not compare it with an oscilloscope. But the SNR is truly much, much better...
I could not understand anything. But one thing I can say you all are very very intelligent and gifted people
Thank you for watching, even if the content was not very understandable!
If you get a professional sound device (as opposed to an Ali-Express "professional(tm)" sound thing) you should be able to handle up to about 24Vpp of both input and output levels without distortion. Obviously if you are working with unknown-level signals it would be a good idea to build a +-15V clipper for the input signals.
That is interesting! After this comment I looked up the specs of my Focusrite and it has a max input of 22dBu which s 27Vp-p at minimum gain.
@@AndreasSpiess The Scarlett qualifies as a professional device. Basically if it has an XLR jack and 6mm TRS jacks, it is likely to be +20dBu or better. It should be able to output similar levels into 600 ohms. Pretty much all of these devices run the internal opamps from +-15V, so subtract about 6V from that to get the max direct input level or max output level.
you can receive DCF77 and other long wave time signals with a soundcard and a long wire
Or Russian military signals at 18.1kHz.
Interesting. I never tried it.
Unfortunately very few modern computers and external sound cards have stereo line input, a mono mic is predominant, which makes it difficult to input signals. My oldish PC is good in this respect.
Two other free softwares are
Digilent Waveforms (front end for their instruments but includes a sound card option)
Soundcard Oscilloscope by Christian Zeitnitz - a bit dated interface but does the job reliably
You should look at more expensive sound cards. I have a thing sitting beside me with stereo mic inputs, and 8 channels of balanced line inputs and 8 channels of balanced line outputs. Of course it cost me a few hundred bucks.
@@lwilton ... which totally defeats the point of the original video
@@alexandern8671 No it doesn't. The title of the video isn't "make a scope for free", it is "use your sound card as a scope".
A $300 sound adapter is still a tenth the cost of a $3000 scope, which is about the minimum these days for a decent scope. And the scope won't have the gain dynamic range and noise floor of the sound card, nor is it likely to have FFT or spectrum analyzer capabilities, and may not have a way to record what it sees on the screen. And you can get that $300 sound card for maybe half that used on eBay. I've got both a scope and a cheap PC with a good sound adapter, and I use both. They have different strengths.
@@lwilton Agree with your point regarding the title of the video.
However I do not think someone with a budget of $300 should invest in a sound card for the purpose of scoping some signals. Andreas has discussed the downsides - no DC, limited frequency range. If one has such a card already like you then of course it is nice to repurpose it though. I can think of some niche applications where the increased vertical resolution could make a big difference but they are hardly the mainstream. Can you please give some common use cases that would benefit from increased vertical resolution?
@AlexanderN: Thank you for the links to the software packages.
RIP headphone users @ 4:22. Even on the speaker on my laptop it was really loud.
Sorry for that!
OMG. I never though this would be possible. THis is awsome!
Glad you liked it!
I'll stick with my scope for the measurements but having a programmable signal generatr could be extremely useful
I think the point is that if you don't need it for high precision and rarely, it's probably good enough. If you're doing anything that requires serious precision, you'd probably want to just buy a quality bit of kit.
get an analog electric analog computer. Best SG you will ever get.
@Frank: I agree it is not a replacement for an oscilloscope.
th-cam.com/video/SYBkL4Or2To/w-d-xo.html
Go to this link... You will find more close waveform on pc as you use Arduino as DSO
You can also use your soundcard to measure the thiele-small parameters of a speaker.
Good to know (I had to google "thiele-small parameters" ;-)
@@AndreasSpiess Interesting, hope we can see you talking about it.
I guess audio is analog! This is bring back memories of sound engineering when I was a kid
Yes, audio is analog. But here most of the tasks are done in the digital domain...
🙂
By the way, the maximum entry level specifications of my Apollo say: +20.2 dbu which would be the same at 17.99Dvb = 7.93 Volts.
Since the maximum output voltages of the Fostex are 0.3 V +/- 1DB then the voltages are very small.
Thank you for the additional info!
Thanks for your video! I think this is a great idea!
I just wanted to say, that maybe you can use an optocoupler between the signal and the sound card's input. This would protect the sound card in a better way than the potentiometer.
For high voltage digital signals it is good idea for sure.
But I doubt that optocoupler will preserve shape of more precise signal.
Opto-isolators usually introduce a lot of non-linearities (as
sdjhgfkshf mentioned). So this would disturb the signal. Maybe it would be better to use an external soundcard and a USB insulator.
@@AndreasSpiess : Yeah, I got an external USB soundcard for this very purpose (around $4, so cheap), though not a USB isolator.
Software defined radios also work with a pc soundcard
You are right. And all digital modes in HAM radio
The frequency sweep at 4:30 is -1.4 dB below max and in just one channel. No warning needed, commercials are louder. Some people have to turn up their sound so much because of weak sound often heard right here on TH-cam. Set up your soundcard playback level to mainstream sources not amateur streamers, then tell them to turn up their levels to standards. Such is the power of Audacity to scope audio streamed on the web. Wasapi mode instead of line in.
Your voice track is -7.295 dB which is weak for pro production standards. Audacity has several tricks including spectrum display as well the details that I have gotten out of this video. It's just not realtime.
Thanks for the info. I use Audacity, and I level the sound to around -2dB peak, not average. But I am not sure what TH-cam does with audio.
a small warning before the sweeptest would have been nice
He even talks while the sweep test lmao
Sorry for that!
I think Sdr radio has been using this for a while, it’s very clever
You are right, many SDR receivers use the soundcards for the I and Q signals.
Another excellent video. Thank you.
Glad you enjoyed it!
A very interesting video. I learned new things today. Thanks 👍
Glad you enjoyed it!
Very interesty video, great work
Glad you enjoyed it!
While audio card oscilloscope is handy, there's some caveats; the processors and / or ADCs, and DACs used by most sound cards tended to be slow so 48 kilohertzs bandwidth is usable at most, if that. However, it's cheaper to do it and get done quickly.
And I strongly recommend that the line or microphone inputs be protected - the same way you would do when putting together a few more extra coaxial oscilloscope leads for the most expensive kit.
Thank you for the additional info!
I was duing some experiments with audio signals and misusing a sound card. I was using just an ordinary integrated Realtek HD Audio. I've discovered that the sound chip is doing some funny stuff to the signal. Namely it seems it has some kind of autogain, which can't be controlled or turned off from the software. Also laptops usually have only a microphone input, which is more difficult to use for recording arbitrary audio signals. So it's a good idea to carefully test your equipment, if it's suitable for the task.
That's true, especially for laptops that have limited I/O and usually combine headphone and mic together. The easiest and less risky way is to use an external USB audio adapter with proper line in/out. Worse case you kill the adapter but keep you computer safe.
Using an external sound card is what I recommended, too. They are no more very expensive.
But I agree, there is always a possibility that the signal is distorted by the PC.
Amazing! Great video, thank you for sharing this!
Glad you enjoyed it!
back around Windows-95 in some astronomy magazine. Someone taken a PC and a 8.5" LDA photo-sensor from a simple document scanner. Then relocating the sensor to point at the night sky and with a little sensor on/off timer was able to scan the stars and moon as the earth rotated giving a basic image for very little money. 🌠🌛 another neat project 🪐 thanks 🤩
Interesting project! It is always interesting to repurpose consumer HW because it usually has an extremely good price-performance ratio
That is amazing, I'll check this later with my own hardware
Enjoy your experiments!
What an interesting idea! Just preparing to build a diy headphone amplifier, this could be very handy for testing and troubleshooting.
For sure you can test its frequency response and its distortion.
The sine sweep was painful with in-ear headphones! Please make sure to tune the noiselevels better
Sorry for that!
For maybe 20 dollars you should be able to get an audio-quality stereo I2S ADC and DAC, feed them through fast optocouplers, and send that data to a USB MCU that can send it to your computer. Arguably there are some computers that will take digital audio signals like SPIDF and some ADC/DACs that can produce it such that you don’t even need the MCU. Then I’d have a 6+ resistor range switch followed by clipping diodes for the input, and a couple of different gain amplifier settings for the output. Maybe an input buffer amplifier running on some larger voltage rails, for input impedance purposes.
I agree that this can be done. But you forgot an important ingredient: The software which also would have to be written.
@@AndreasSpiess from what I hear the Arduino Pro Micro has a fair bit of code written for it out there, acting like a sound card is almost certainly one of those. If that bitrate isn’t fast enough then the same should apply to the STM32F103, though they’re a little tougher to get programmed.
Another option is using an FT232 and SPI ADCs/DACs, not sure how much programming they need, but I’d guess it would be less than the STM32 and still give you full speed USB2.0.
For USB3.0 purposes there’s the Cypress FX3, but from what I hear that’s a pain to program and it’s only really necessary for really high bitrates or SDR territory.
I've certainly tried stuff like this throughout the years... There are just times when u want a particular waveform or pattern of waveforms. However, I would highly HIGHLY recommend to not use any direct lines into the computer. Just get ANY gear which is external to your computer. Great for making amps & guitar pedals :) Also, make sure to get an audio amplifier for your output.
Output chain looks like: PC>audio interface > amplifier
Input chain looks like: custom opamp divider > audio interface > PC
I feel like this is a second step in the path of learning EE. First step would be to get a signal generator + oscilloscope. Caz, how are u going to test ur custom opamp divider thingy if u don't have suitable equipment to begin with ?
REW is awesome. An absolute staple in the kit...
It is certainly good to protect your PC. But for experimenting with audio signals I think this is a valid alternative. Not for an oscilloscope, of course...
So probing around a breadboard distortion pedal circuit with a 9v supply would be unsafe with a soundcard oscilliscope?
@@mikejones-vd3fg Just to be more clear with what I wrote caz it see it's kinda hard to read. I highly recommend against using any direct connections to your motherboard. I have fried 2 & a usb dock. What happens is that device fails. For example: I was using ardruino plugged directly into motherboard via usb = It took years but eventually I burnt something by accidentally crossing wires for the slightest of time :((. Just a touch. Oops. Now all the front panel usb's do not work. Looking at the motherboard >> power goes into the usb IC, but its dead. I got smart and started using usb docks. Resulting in another dead usb. But at least it's a dock & can be replaced. It's an accident waiting to happen. This is mainly caz computer devices are not designed to handle over-voltages. They're designed to stabilize voltages.
Specifically about a soundcard, 9v is way above the "line in" rating of 1.5v. I don't know how that would turn out. Sounds like a bad plan. You would have to know how the front end is designed. i wouldn't be surprised if putting any DC into the soundcard for a period of time would mess things up. At least if u were to buy a usb soundcard, then that usb device would fry & not your computer motherboard. In general AD/DA's are pretty sensitive in my experience. I've seen many die from normal use, let alone outside recommended ranges.
If you think about it > the PSU on a computer maintains the voltages precisely. They will dump large amounts of current just to keep voltages correct.
I think my recommendation of using an audio interface is good. Many can handle +4 signal levels :) I've used a ton of these and never had a problem. I still think it's wise to design an op amp divider/buffer so it ranges your input correctly. Also an opamp has a known current output. It's a good thing to learn. . . Don't forget/overlook that you can set the opamp gain to at any level you wish. Then you can see all the small signals when the input is off, like the noise floor/crosstalk/signal bleed/etc ...
Another thing to know is oscilloscope probes are amazing. If you compare alligator clips to Oscope probes, the noise pickup is night and day. Noise&signals will go right into the wires of alligator clips. Not with Oscope probe wires.
Have fun, don't burn the mobo like I did ;p
Java driver is limited by your windows setup of soundcard output profile, it defaults to 44100ksps.
Many proper soundcards support 192 ksps which is why 96 kHz works with the other driver.
I do not know where the limitation comes from. You are right, as I showed, the proprietary drive went up to 192k.
I have been thinking about adding a sampler in-front of a pc scope .
Like the early scopes did .
PS
You can get a SINAD program for sound cards .
I assume the sound cards also have a sample and hold. What would be the advantage?
@@AndreasSpiess
Sorry if this is a double posted .
TH-cam has deleted 2 posts with the links 😒
You mix the scope input with a reference frequency .
( Basically it works like a mixer & the output is on a sampling capacitor ) .
So you get a very high bandwidth scope with limited hardware .
Early high bandwidth scope used the technique as the input is repetitive so it works even with tubes !
( Tektronix is the first name that comes to mind )
A 100 Mhz sine wave on a 22 Khz pc scope or old o-scope is like magic .
Until you're shown how it's done !
The best I can find on the net is a tdr head project
“ 1N5711 Sampling Head ” but the modus operandi is the same .
"#no #midroll #ad"
Sir you are a #legend
Herr Spiess hatte immer geniale Ideen
Danke! :-)
One of my favorites of all your videos. I'm not a big radio guy, so those videos are of limited interest, though I usually watch them to observe your design and assembly techniques. This one I can put to immediate use. Already ordered a USB Sound Card. Great stuff. Thanks!
Cool that you found something usable. I love wireless. This is probably why it appears quite often on the channel ;-)
Most all sounds cards can go to 320k in the settings menu
Thank you for the info!
Digilent’s Waveforms software is free and has a mode that connects to the PC sound card.
Thank you for the information!
The Java sound device not being able to generate beyond 22 kHz points to it internally capping at CD sample quality, 44.1 kHz. Per the Nyquist-Shannon sampling theorem, the highest reproducible frequency is half the sampling rate used.
I think that is right.
Nice work. Keep making great videos.
Thanks, will do!
Apps for iPhone and Android use the audio jack for sensor too.
Good to know! Thanks.
I have heard, but not tried it myself, that it is POSSIBLE to measure DC, if you short-circuit the input capacitors ...
You can try that and it should work. I certainly will not do it. Too complicated...
so we're looking at harmonics distortion something that i know of when it comes to sound design and Electronics
:-)
Great video. I though we had to put optocoupler to protect the soundcard?
I know many people use Audacity software as well
Optocouplers are not linear and would not help in this situation because they would disturb the signal. You can use an isolated USB connector to connect an external soundcard and it would work.
Interesting. There was a old project ELFRAD That was analysing earth signals down to wavelengths days long if I remember right (earthquake prediction, the moon even) This would probably be ideal for something like that exploring those signals.
Maybe. But then you would have to bridge the input capacitor because it blocks DC and low frequencies.
@@AndreasSpiess right, forgot! There would be a filter cutoff on the input.
Hah! I did a soundcard oscilloscope project ~28 years ago when I was in college. Now I design IC test equipment for a living and I've always thought that a soundcard front end would be a fun project.
Cool. You were very early!
This need to try....thank alot for video
You are welcome 😊
I wonder if we can use a soundcard with a down converter or upconverter as an SDR?
Many SDRs already use the soundcards... So it is possible
Many years ago a friend and I used to maintain a microwave radio scheme which used an analogue multiplexer We challenged ourselves to decode the baseband with just a laptop and no hardware (except a cable). Cooledit on a Windows 98 laptop did the job...
Indeed, many current SDRs use soundcards for demodulation. And of course all digital modes. But you were very early! 73 de HB9BLA.
@@AndreasSpiess thank you for the reply, this was an analogue multiplexer (Granger DTL-7300) the individual voice channels are SSB with carriers at 4,8,12, 16kHz and extending further up, these are then modulated onto a sub carrier for transmission over a microwave FM link. Fascinating things to play with! 73
Life-changing tip
Glad you liked the idea!
I already have a pretty good scope, but I think I'll stick to physical circuits when it comes to signal generators. Having said that though, these programs do look usefull.
I stick to my instruments, too. But not for audio analysis. Here, this PC software is much easier to use...