Analog Discovery 3 or budget bench setup

[hao004]

I am a sophomore EE student and I am interested in experimenting with electronics circuits, particularly amplifier circuits and applications. Should I get the Analog Discovery 3 with student discount ($250) or a budget bench setup consisting of Hantek DSO2D10 ($150, being aware of the potential freezing issues) and Korad KA3005D ($50 from taobao)? I would like to start small first with a budget of $250 and see what I am really interested in down the road.

[ataradov]

You would need a power supply regardless. AD3 power supply is very limited, I would not rely on it as a main supply.

Otherwise, AD3 is a good value. But if you can get AD2 cheaper, I would consider that too, there are not a whole lot of changes in AD3.

[tggzzz]

The Analog Discovery devices are very good for simple lab experiments, but I wouldn't use them for anything with a high voltage/current/power.

You may decide you also need a simple power supply, e.g. dual voltages and/or higher voltages.

Make sure you also understand the maths behind what you are building. When interviewing candidates I wanted to see (1) theoretical knowledge (2) practical demonstration of theory (3) going beyond the course's requirements. The last requires you to set difficult-but-achievable objectives, do them, be able to state what surprised you and what you would do differently next time.

[Antonio90]

[...]

Make sure you also understand the maths behind what you are building. When interviewing candidates I wanted to see (1) theoretical knowledge (2) practical demonstration of theory (3) going beyond the course's requirements. The last requires you to set difficult-but-achievable objectives, do them, be able to state what surprised you and what you would do differently next time.

I see a lot of emphasis on intuitive knowledge for beginners. I wasted a lot of time learning "intuitively" looking at videos, non-rigurous books or posts, etc. In my experience math modelling is the way of building theoretical intuition. Of course, you then need to apply it.
The Analog Discovery is a great device, and if you are into programming it can be expanded a lot. "The Stuff Made" youtube channel has a really good impedance analyzer design.
I really like my AD2, but you have to know its limitations. You can get quite far using it as an oscilloscope and signal generator (with the BNC breakout board), it has good XY mode, bode plots, good FFT, basic logic analyzer and digital pattern generator, and VERY limited bipolar PSU.
As much as I like it, there is no way around getting one or two DMMs and a standalone PSU.
EDITs: a few typos.

[tggzzz]

[...]

Make sure you also understand the maths behind what you are building. When interviewing candidates I wanted to see (1) theoretical knowledge (2) practical demonstration of theory (3) going beyond the course's requirements. The last requires you to set difficult-but-achievable objectives, do them, be able to state what surprised you and what you would do differently next time.

I see a lot of emphasis on intuitive knowledge for beginners. I wasted a lot of time learning "intuitively" looking at videos, non-rigurous books or posts, etc. In my experience math modelling is the way of building theoretical intuition. Of course, you then need to apply it.

Just so. Theory and practice are complementary, and both are required for useful intuition.

Theory without practice is mental masturbation.
Practice without theory is blind fumbling.
Both are necessary, neither is sufficient.

[RoGeorge]

For student work/learning, I'll recommend to buy the Analog Discovery.  It has all in one, oscilloscope, generator, power supply, but what's more important is that it has tons of free classes, experiments and tutorials, with software that can control all the instruments in an intuitive way.  Plus, it is small and portable, you can even take it with you at school if you want, find a laptop and experiment with other classmates.  When learning you don't want the hassle of extra cabling, learning SCPI or other cumbersome tool to run an experiment, and so on.  An analog discovery is very easy to learn and handle, and has a very versatile tool.



With classic/standalone instruments, you don't get any of the above advantages, particular with the very cheap instruments.  You want to focus on learning electronics, not on dealing with the pitfalls of very cheap/unreliable standalone instruments.  If you buy cheap standalone instruments, it will be a waste of money, because in that budget you can only find very unreliable instruments, so you'll end up buying again something more expensive.

There is no way to buy a good enough oscilloscope+generator+power supply for the money of an Analog Discovery, particularly the one with student discount.

An Analog Discovery is the best choice for learning, and it finds its use even in a fully equipped EE labs.  There are plenty of measurements and experiments that are much easier to run on an Analog Discovery than on standalone lab instruments.

Another alternative to Digilent's Analog Discovery might be the similar all-in-one learning kit from Analog Devices' "ADALM" series, I think it's called ADALM2000 https://www.analog.com/en/resources/evaluation-hardware-and-software/evaluation-boards-kits/adalm2000.html#eb-overview , not sure which one is the latest version.  ADALM2000 is also very good for learning electronics, has a lot of free tutorials/experiments/classes from the ADI website, but I think it is less widespread than Analog Discovery, not sure which wins.

I remember ADALM being cheaper than Analog Discovery, but I didn't check neither of the two protucts since some years ago, not sure which one would be better.  Both were designed mainly as all-in-one learning platforms for students.  Both have plenty of free learning classes, with experiments and examples designed for that particular kind of all-in-one EE lab.

[Fungus]

The Analog Discovery has a lot of limitations but it's a good all-round device. It will teach you a LOT more than a Hantek, especially with the signal generator.

Does that AD you're looking at come with the BNC breakout board? You need one of those.

(Also 3D-print the case for it - the BNC is a bit floppy, https://www.thingiverse.com/thing:3070779  )

Downsides: The AD software is powerful but can be a bit clunky. Physical knobs and controls are good. It's not great down at the mV range)

As others have noted, you'll still need a multimeter and power supply. You don't necessarily need a bench supply right away though, most things are either 12V, 5V, or 3.3V and you can use an old PC power supply from a dumpster for that. The only thing you won't get is current-limiting.

(nb. The ability to dial a voltage is maybe the least important feature of a bench power supply, current limiting is way more important)

[Fungus]

Another alternative to Digilent's Analog Discovery might be the similar all-in-one learning kit from Analog Devices' "ADALM" series,

I remember ADALM being cheaper than Analog Discovery, but I didn't check neither of the two protucts since some years ago, not sure which one would be better.  Both were designed mainly as all-in-one learning platforms for students.  Both have plenty of free learning classes, with experiments and examples designed for that particular kind of all-in-one EE lab.

The power of the Analog Discovery is in the software, not the device itself. I don't know how the ADALM measures up in that regard.

nb. You can use the Analog Discovery software for free with your PC sound card (minus the digital parts, obviously)

[Antonio90]

The ADALM2000 is quite a bit cheaper, and I think it is used on the "Real Analog" course. The software is miles behind of the Analog Discovery though.
At the end of the day they are just an FPGA with a bit of interfacing circuitry, the software is the real product there.

[tggzzz]

The power of the Analog Discovery is in the software, not the device itself.

That plus the ADC/DAC resolution, which is much better than low-end 8-bit scopes.

[Aldo22]

The Analog Discovery has a lot of limitations but it's a good all-round device. It will teach you a LOT more than a Hantek, especially with the signal generator.

The Hantek also has an AWG. 

But it is not really about the brand of the benchtop scope, but about
Standalone vs. PC based.
Bandwidth/sampling rate vs. vertical resolution.
1GS/s vs 125MS/s
150MHz vs 30MHz (with BNC)
8bit vs 14bit.

These are simply different types of devices, regardless of whether it is an expensive or cheap benchtop scope.

[nctnico]

[...]

Make sure you also understand the maths behind what you are building. When interviewing candidates I wanted to see (1) theoretical knowledge (2) practical demonstration of theory (3) going beyond the course's requirements. The last requires you to set difficult-but-achievable objectives, do them, be able to state what surprised you and what you would do differently next time.

I see a lot of emphasis on intuitive knowledge for beginners. I wasted a lot of time learning "intuitively" looking at videos, non-rigurous books or posts, etc. In my experience math modelling is the way of building theoretical intuition. Of course, you then need to apply it.

Very true! Once the quick calculations are done, simulating a circuit is also very insightful because it is much easier to see voltage and current waveforms than in a real circuit.

[RoGeorge]

Another alternative to Digilent's Analog Discovery might be the similar all-in-one learning kit from Analog Devices' "ADALM" series,

I remember ADALM being cheaper than Analog Discovery, but I didn't check neither of the two protucts since some years ago, not sure which one would be better.  Both were designed mainly as all-in-one learning platforms for students.  Both have plenty of free learning classes, with experiments and examples designed for that particular kind of all-in-one EE lab.

The power of the Analog Discovery is in the software, not the device itself. I don't know how the ADALM measures up in that regard.

nb. You can use the Analog Discovery software for free with your PC sound card (minus the digital parts, obviously)

A couple of years ago I've looked at both, and they were about the same.  I remember Analog Discovery to be richer in features, but both have all that is needed to run their classes and tutorials.  Both have additional functionalities like Spectrum Analyzer, Bode Plots, RLC measurement, DMM, etc.  Both have Python, C, LabVIEW, Octave/Matlab, etc. libraries for automation.

I do not own any of these, and only looked briefly at the software, out of curiosity.  Can not really say if it's a big difference in usability or not.  At a first look they seemed both OK.  I bet there are plenty of reviews out there to decide which one to buy.

I think both platforms have student discount, and sometimes promotions and/or discount coupons.  Analog Devices also sells a kit of components with a breadboard and wires (the kit is a separate buy) to use for their classes:  https://www.analog.com/en/resources/evaluation-hardware-and-software/evaluation-boards-kits/eval-adalp2000.html .  The kit has most representative types of opamps from Analog Devices, some transistors, sensors, breadboard wires and such, to complete the classes posted on their StudentZone:  https://www.analog.com/en/resources/analog-dialogue/studentzone/studentzone-june-2024.html (note there are many more other explained experiments).

I remember some similar components kit from Digilent, too, but I don't recall the name.

No matter which board, or kit or discrete components will be used, I would recommend first to do the labs/classes from own school, then from the manufacturer's website.  YouTube videos and other blogs might be more time consuming, and might be of a questionable quality.  As a beginner, it is hard to tell which resource worth the time, and which doesn't.

[tggzzz]

YouTube videos and other blogs might be more time consuming, and might be of a questionable quality.  As a beginner, it is hard to tell which resource worth the time, and which doesn't.

You are being generous. Most yooootooob videos are "talking heads" full of ums and ahs. They do little more than boost the ego of the creator. Very few have moving content that is necessary. Most would be better written as a blog. There are, of course, exceptions.

Major problem: you can't speedily determine which videos are worth your time (1%) and which are a waste of your life (99%). With written content it is easy to skip and speed read to find the good 1%.The

With so much material available now, determining which 1% to look at is a key skill.

[Aldo22]

If you buy cheap standalone instruments, it will be a waste of money, because in that budget you can only find very unreliable instruments, so you'll end up buying again something more expensive.

What is the difference between a reliable and an unreliable instrument?
What does that mean exactly? Can it be defined / quantified or is it more of a feeling?
How can I test if my scope is "unreliable"?
Thank you!

[Doctorandus_P]

I find the Analog Discovery far too expensive for what it delivers. For that price you can buy a real oscilloscope with GHz sample rate and two channels. It's power supply (max 5V) is also quite limiting for analog stuff. Something like 15V 200mA would have been much more suitable for generic analog stuff. I was once interested in it, but I would never pay over EUR100 for it. It may be a suitable thing for if you really like it or need it. Maybe you don't have room for bigger test equipment, or you need the PC connection.

For a power supply, if DIY has always been, and still is a good option if you are on the intersection of low budget and starting with electronics. DIY power supply can be as simple as an LM317 with a potentiometer, or design something with opamps and current amplification / buffer & limiting. This is a good opportunity to learn about control loops and stability.

If you're doing anything digital, then also buy one of those USD10 Logic analyzers (Saleaeaea clone) and use them with Sigrok / Pulseview. With 8 channels and up to 24Msps (for a few channels) you can look at anything from a few flipflops to I2C, SPI and around 100 other digital protocols.

[Doctorandus_P]

I think I like the "Expeyes". It's more limited than the analog discovery, but capable enough to learn about electronics and the design is (mostly?) open so you can use it as a basis to make your own variant. And with a price of USD 80 it's also almost cheap enough to buy it "just for fun".

https://expeyes.in/

https://www.tindie.com/products/CSparkResearch/expeyes-17-your-labhome/

Another affordable / fun instrument is the totem mini lab board:

https://totemmaker.net/product/mini-lab-labboard/

[tggzzz]

I find the Analog Discovery far too expensive for what it delivers. For that price you can buy a real oscilloscope with GHz sample rate and two channels. It's power supply (max 5V) is also quite limiting for analog stuff. Something like 15V 200mA would have been much more suitable for generic analog stuff. I was once interested in it, but I would never pay over EUR100 for it. It may be a suitable thing for if you really like it or need it. Maybe you don't have room for bigger test equipment, or you need the PC connection.

For a power supply, if DIY has always been, and still is a good option if you are on the intersection of low budget and starting with electronics. DIY power supply can be as simple as an LM317 with a potentiometer, or design something with opamps and current amplification / buffer & limiting. This is a good opportunity to learn about control loops and stability.

If you're doing anything digital, then also buy one of those USD10 Logic analyzers (Saleaeaea clone) and use them with Sigrok / Pulseview. With 8 channels and up to 24Msps (for a few channels) you can look at anything from a few flipflops to I2C, SPI and around 100 other digital protocols.

You seem to ignore the dual AWG and pattern generator, and the software that combines them to make spectrum analysers, network analysers, etc. And you also have the ability to script new tools.

Example: when playing around with an N-path filter, the only other instrument I used was a 15V PSU. To demonstrate the bandpass frequency response with a Q of 15000 and roll off rate of 140000dB/decade(!), I used the pattern generator to drive the switches the AWG+scope to sweep the frequency and generate the transfer function. If I FM modulated the input frequency by a few Hz, it was amusing to watch the IQ output rotate in one direction, stop, and rotate in the other direction.

When I first did that in 1980, it required a benchful of equipment, and the result was nowhere near as explicit.

[RoGeorge]

If you buy cheap standalone instruments, it will be a waste of money, because in that budget you can only find very unreliable instruments, so you'll end up buying again something more expensive.

What is the difference between a reliable and an unreliable instrument?
What does that mean exactly? Can it be defined / quantified or is it more of a feeling?
How can I test if my scope is "unreliable"?
Thank you!

The OP was mentioning an oscilloscope that sometimes freezes.  Most common problems are because of firmware bugs.  May be as well because bad design/manufacture of the hardware.  Usually the lower the price, the less quality is expected.  This is a thumb rule for any product.

Hard to quantify or to test that.  Usually time will tell.  I don't know what type/model/brand of instruments you were asking about how to test their reliability.  If yours don't show problems, that's great! 

[Aldo22]

The OP was mentioning an oscilloscope that sometimes freezes.  Most common problems are because of firmware bugs.  May be as well because bad design/manufacture of the hardware.  Usually the lower the price, the less quality is expected.  This is a thumb rule for any product.

Hard to quantify or to test that.  Usually time will tell.  I don't know what type/model/brand of instruments you were asking about how to test their reliability.  If yours don't show problems, that's great! 

I'm talking about the scope that the OP is mentioning.
With the current firmware, it does not freeze here.
The OP only knows about the problem from hearsay (or reading the forum), as he doesn't (yet) own the device.
Many people "know" a lot only from hearsay or they are only interested in bashing anyway.

I don't think general bashing of cheap devices is a good thing.
As with everything, a product can be good and reasonably reliable for a certain price without having all the features and performance of more expensive devices.
For many people, a Dacia Sandero is good enough. Why should you badmouth it? Because there are better cars? That doesn't make sense to me.

[egonotto]

Hello,

an Analog Discovery with a breadboard is a small laboratory that is very suitable for learning. An Analog Discovery 2 is cheaper and can do almost as much as the Analog Discovery 3.

Best regards
egonotto

[Smokey]

I think there is a difference between "... very suitable for learning..." and "bench setup to do actual engineering" that needs to be expanded on.

I've had a unit from the Analog Discovery1 days, and I agree it's "very suitable for learning".  But it's quite inappropriate for most actual work. 
If it's something you just need for a class, then go for it (short term).  If you don't plan on doing any real design work at the bench and just want to play, then go for it (short term).  If you plan on being an engineer that will spend any amount of time at the bench doing actual work (long term), then you can still get an Analog Discovery but expect to outgrow it essentially immediately. 

Unless you go straight into something specialized like RF, a basic bench setup of DHO800 + DP832 + DG1000Z (or Siglent equivalent), while not a prestigious as a Keysight or Tek is enough to be a real engineer for quite some time.

[Vertamps]

Buy used: Multi-meter, Oscilloscope, bench power supply.
Locally: FB market place or Craigslist, alternatively ebay is fantastic even with higher shipping prices these days.
You could get a cheap' new all in one Multi-meter/Oscilloscope, but it not as enjoyable and i don't really like those. Get a 4.5 digit bench meter, 5.5 digit if you want to spend the $100+. An old 1980's 3.5 digit Fluke hand held is fine as well, we use them at work. > Old Keithley  4.5-5.5 digit from $30 - $80. Hunt for a functional scope, $45-$100.  Extra cost are: Soldering Iron, Wire strippers/Cutters, tweezers. Pull electronics apart from the garbage if you don't want to order parts.

[ArdWar]

Analog Discovery are excellent tool for learning indeed. Especially its powerful charting and post processing. Some of them gives you capability only available on thousand bucks equipment^[1]. Heck, its decidedly weak supply can be a selling point for learners to limit them to something slightly stupid instead of something monumentally stupid.

However depends on your course/career path it can quickly get past its usefulness. One glaring point is its lack of robust protection. Probing around while it is direct connected to your PC/laptop can be unsettling sometimes. Not a problem if you keep doing low frequency small signal designs, but less useful when doing SMPS design for example.

[1] Often as an extra upgrade. Yes I'm still salty on why curve tracing didn't come standard on K2461

[Antonio90]

I think there is a difference between "... very suitable for learning..." and "bench setup to do actual engineering" that needs to be expanded on.

I've had a unit from the Analog Discovery1 days, and I agree it's "very suitable for learning".  But it's quite inappropriate for most actual work. 
If it's something you just need for a class, then go for it (short term).  If you don't plan on doing any real design work at the bench and just want to play, then go for it (short term).  If you plan on being an engineer that will spend any amount of time at the bench doing actual work (long term), then you can still get an Analog Discovery but expect to outgrow it essentially immediately. 

Unless you go straight into something specialized like RF, a basic bench setup of DHO800 + DP832 + DG1000Z (or Siglent equivalent), while not a prestigious as a Keysight or Tek is enough to be a real engineer for quite some time.

As a full bench setup, it really is only suitable for learning. The points, AFAIK, are (1) the OP is actually studying right now, and (2), the Analog Discovery keeps its usefulness even with a full bench.

Admittedly, it depends on the type of work you do, but as an example, I know of a design engineer currently working on EV battery charging circuitry (the circuitry that goes inside the car) and the tool he uses the most, by far, is the Analog Discovery. Not necessarily the Waveforms features, but the ones he programs.
Of course, they have oscilloscopes, expensive probes, PSUs (huge ones at that) and a lot of other equipment, but the little AD still gets a lot of use in a professional environment.

So, yeah, I don't think it's a bad option with the stated budget, as long as the OP is aware of the limitations. If he wants a full lab, he will have to buy it later while keeping an extremely versatile tool.