[Nerull]

Hey guys,

I'm looking for some equipment that'll prove useful when designing circuitry. When I say circuitry, I mean I/V converters, digital audio circuits, and whatnot (mostly audio). It's semi-serious, so I wouldn't mind investing in something that'll prove truly useful and worthwhile (not Kevin Gilmore extravagant though). I'm also thinking about other digital design (DSPs, microcontrollers, PLCs) so keep that in mind. I guess I need tools that'll aid in the design of these electronics, but I'm not sure which ones. I, of course, have a multimeter, breadboard, and the like, but I'm primarily looking for recommendations for variable power supplies, oscilloscopes, and I don't know what else. Basically a nice lab for useful measurements and design of digital / analog circuits.

The M-Audio Audiophile will be my way of doing spectrum analysis with THD+N and IMD measurements, as well as a signal generator for audio frequencies.
How does AMB go about measuring this? I know he uses a dummy load, but the maximum input voltage for the Audiophile is around 2Vrms (or somewhere there).

Another nice thing I've been looking at is the LNMP by tangent, which AMB used for his Sigma22 power supply.

My budget is quite flexible, and I'm looking for an investment sort of thing in electronics. Just not too expensive, like >$1200 for any one thing, but I guess it's negotiable...

Any other advice would be great!

~Thomas

 

[tangent]

Quote:

Originally Posted by Nerull /img/forum/go_quote.gif I'm also thinking about other digital design (DSPs, microcontrollers, PLCs)

You'll need a DSO, then. IMHO, the only thing at the low end worth talking about is a new Tek TDS 1000 or 2000 series unit. Off-brand and used DSOs are harder to trust than analog ones. There's just too much to go wrong in a DSO that you can't fix yourself.

Also budget for device programmers, compilers, etc.

Quote:

variable power supplies

Almost anything will work here. What you get at the high end are better protection features, better meters, more outputs, higher currents.... all useful things to have, but none of them absolutely essential. For experimenting, power is pretty much power.

Quote:

the maximum input voltage for the Audiophile is around 2Vrms (or somewhere there).

I'm not sure what your question is, exactly, but yes, it's a limit you have to live within. What makes the whole RMAA + sound card route worth taking is that at audio frequencies, you can do much better than a lot of high-end test equipment, because sound cards can focus on doing the low frequencies really really well. You have to spend thousands to get wideband instruments that have the same AF performance.

Quote:

Just not too expensive, like >$1200 for any one thing, but I guess it's negotiable...

Staying within that limit, here's my recommendation:

Tek TDS1002B scope (or 1012, or 2002)
Fluke 8845A bench DMM
B+K Precision 9130 bench supply

 

[error401]

Quote:

Originally Posted by tangent /img/forum/go_quote.gif You'll need a DSO, then. IMHO, the only thing at the low end worth talking about is a new Tek TDS 1000 or 2000 series unit. Off-brand and used DSOs are harder to trust than analog ones. There's just too much to go wrong in a DSO that you can't fix yourself.

Good advice. Doing digital with a CRO is tricky, and it's really difficult to analyze the real-world failure since it only happens once and the CRO won't capture it. You'll need to fake it by guessing at where the failure is and creating stable waveforms near it to see what's wrong with the timing. Not fun. If you're going to be doing digital, the advantages of the DSO definitely outweigh its pitfalls.

If you wanted to do this a bit cheaper, you could get a CRO and logic analyzer instead. A quick scan of the Bay and it looks like you can get a decent 100-200MHz logic analyzer for $400-500. Coupled with a fairly good used CRO you're looking at around $1000 for the set.

Quote:

Also budget for device programmers, compilers, etc.

Or choose devices that gcc can target and easy programming interfaces that a parallel port will drive

. The programmers are often worth it. The AVRISP for example is great value at $35.

Quote:

Almost anything will work here. What you get at the high end are better protection features, better meters, more outputs, higher currents.... all useful things to have, but none of them absolutely essential. For experimenting, power is pretty much power.

Oh, the joys of adjustable current limit. It's worth buying a proper bench supply for this alone. I'd say the supply is the last thing on the list, of major test gear that you'll need. You can build a perfectly fine supply without the fancy features for a few bucks, and it'll serve you well. Just be a bit more careful.

Just talking about this stuff makes me wish I still had access to a real electronics lab. I don't even have a scope at home

. I think I will soon buy a Spartan dev kit just to run this nifty, open source logic analyzer. Since I mostly do digital it'd be very useful...

 

[tangent]

Quote:

Originally Posted by error401 /img/forum/go_quote.gif Good advice. Doing digital with a CRO is tricky, and it's really difficult to analyze the real-world failure since it only happens once and the CRO won't capture it. You'll need to fake it by guessing at where the failure is and creating stable waveforms near it to see what's wrong with the timing. Not fun. If you're going to be doing digital, the advantages of the DSO definitely outweigh its pitfalls. If you wanted to do this a bit cheaper, you could get a CRO and logic analyzer instead. A quick scan of the Bay and it looks like you can get a decent 100-200MHz logic analyzer for $400-500. Coupled with a fairly good used CRO you're looking at around $1000 for the set.

This still only gets you part way down the path. It won't help you with serial protocols, for example.

Quote:

Oh, the joys of adjustable current limit. It's worth buying a proper bench supply for this alone.

Yes, I forgot to extoll the virtues of current limiting. It's saved circuits for me so many times it's not even funny.

Which reminds me: not all current limiters are made equal. The power supplies from Circuit Specialists claim to have it, but setting the limit is harder than it has to be. The UI on those things is a crock in general, for what that's worth. Don't get me wrong: they do the job. Just realize that there's a reason they're so cheap.

 

[error401]

Quote:

Originally Posted by tangent /img/forum/go_quote.gif This still only gets you part way down the path. It won't help you with serial protocols, for example.

Eh? That's what they're for (among other things). Modern ones will even decode them for you and flag violations. Or were you talking about the DSO pretending to be a logic analyzer?

 

[tangent]

I'm thinking of something very primitive: typically 8 wires plus ground, each connected to an LED or to a logic gate that gets displayed on a DSO like screen in parallel. I knew there were serial analyzers, but I didn't know you could get something decent for less than the cost of a DSO, which would be the point, yes?

 

[error401]

Quote:

Originally Posted by tangent /img/forum/go_quote.gif I'm thinking of something very primitive: typically 8 wires plus ground, each connected to an LED or to a logic gate that gets displayed on a DSO like screen in parallel.

An, I didn't know such a contraption would be referred to as a logic analyzer!

The typical device I'm talking about will be able to sample >30 channels at >10MHz, either synchronously or asynchronously. Triggering is pretty advanced, at the very least you can trigger on any combination of values, often multiple values; fancier ones can decode parallel and serial buses and trigger on those, or even trigger on a series of many states. They can usually be set to stop capture after a certain number of samples, and then be retriggered, so can capture multiple events in a short time. You'll get enough RAM to sample many signals for a large number of periods, since data can be much more efficiently stored than in a DSO, the memory lasts a lot longer. The downside of course it that they're purely digital, so are completely useless for analog. For digital though, they're the tool, alongside JTAG.

Like with o-scopes, PC-based ones are available. Generally a lot more viable than an oscilloscope on a PC. That Spartan 3 project for example, would cost around $175 and is apparently good up to 200MHz. It's probably some work to get it going and doesn't have the nice probes and all, but certainly more useful than a DSO in the digital world.

Unlike with oscopes, new gear is probably totally out of a DIYer's budget - Agilent's cheapest unit is over $7k, unless going for a USB option. There are some used units on the 'bay like I said though, like this. When looking at them, you'll see two sampling rates quoted as 'state' and 'timing'. State capture is synchronized to an external clock, while timing capture is asynchronous against an internal clock. State inputs are usually slower, but you usually get a lot more of them.

Oops, sorry for adding something to your shopping list