New Lab Recommendations

[ophidian]

Hi all,

I'd like some advice on test equipment.  I have two goals.  The first is for my current job: I'm working remotely as a firmware engineer for a startup creating some Zigbee devices.  I've spent about a decade working for various big companies with big labs and endless tools, but for the moment I'm on my own.  When we build a new board, I am the first to bring it up and diagnose any problems.  I work closely with a rotating cast of hardware engineers, who themselves frequently don't have much of a lab.  I'd like to buy a few things to help me do a better job for this and any future employers (at least until I get back to a "real job").

My second goal is to get deeper into electronics.  I have a computer engineering degree, which for me meant that I took an introductory circuit analysis courses, then more or less jumped into logic gates, computer architecture and progressively higher-level topics.  So the analog world is a bit of a mystery to me.  To help with that, I bought the Art of Electronics (including the lab course).  My plan is to work through all of the analog labs as if I know nothing, and then pick a few selected digital topics that I haven't done in awhile.

The test equipment I currently have:

• an Craftsman 82040 multimeter, given as a gift in 2002 or 2003 or so
• a cheap $10 multimeter, bought recently just for measuring two voltages at once
• Dave's µCurrent Gold, which I have used to help with measuring power usage for our devices.  I've had a lot of trouble getting stable < 100 mV readings from my Craftsman meter, though.
• a Salea Logic Pro 16, which I use for decoding digital signals...and as a makeshift oscilliscope.  My experience "muddling through" with its analog traces is what has made me want to buy a real scope.

The target price range is under USD $3K or so. 

Here's what I've come up with so far:

Oscilliscope: Rigol DS1104Z

AoE recommended the DS1054Z, which also seems popular here. 

The most complicated signals I would want to visualize right now are from a 10 MHz SPI bus.  The Salea Logic does a fine job of decoding it for me--when the SPI bus is working at all.  The scope would help diagnose what's wrong when it's not working.  From reading other threads, the DS1104Z should be perfectly capable of that. 

Power Supply: GW-Instek GPD-3303D

AoE didn't provide much guidance here.  I don't believe I have any need to go over 3A for the low-power microcontroller designs I am likely to be working with.  Forum threads seem to approve of three-output PSUs.  I believe I can also connect this one to my PC and perform voltage sweeps, which I have wanted to do recently.  I looked at Rigol's offerings, but the front panel of this Instek looks more like the PSUs I used to use in college (mid-90s Agilents, probably) so it feels very familiar.

Waveform Generator: GW-Instek AFG-2012

AoE recommended a B&K 4017A that seems more expensive with fewer features.  I don't expect to need this for work, but I will need it for the labs.

Multimeter: Keysight U1252B

My immediate need is to log readings a PC, which my old Craftsman cannot do at all.  I am aware I'll need to buy the optional cables, etc.  I'm not sure which brands have the best, most easy-to-use PC interfaces, though.

This meter is probably overkill for my needs, but I would like a well-built, safe meter that I can hang onto and use for many years, no matter what I might find myself doing.  I also considered one of the Flukes in the same price range, like the 287 or 289 but I just can't shake the feeling that I'm paying more for less with Fluke, and Keysight seems like a perfectly trustworthy brand.

I'm also curious where is the best place to actually buy all this equipment for a US customer?

I will be grateful for any advice or suggestions! 

[ebclr]

I would suggest a complementary multimeter fluke87 or 34461 ( if you have money or that ) and a rigol DS1054, also a hot air station

[Mr. Scram]

Note that the resolution of the D model power supply is lower than it is with the similar S model. You may want to look into the price difference and see whether upgrading is worth it to you. Also note that the 3 channel model appears to have variants that can control the third channel from software, but commonly seems to have an unregulated third channel. The four channel model should be regulated on all channels.

Your set is remarkably similar to what I've been looking into recently, so I'm familiar with at least some of the pitfalls

[stmdude]

I'm going to make an assumption that you're working with battery-powered devices here (deduced from Zigbee and the uCurrent).

The power-supply should be good enough. I still use similar ones at work, and they get the job done.

As for a waveform generator. You could look into getting a Rigol 1074z-s instead of the 1054Z, as it has a built-in 2-channel signal generator.
(However, think about what you would need it for. I pretty much never use the one built into my Agilent MSO-X 2000)

I don't really like the protocol decoding in the Rigol 1054Z. I'm assuming it's software based, as you need to zoom _just right_ for it to be able to decode the signal. I'd recommend you using the Saleae for that, and the scope to inspect the levels and edges of the waveforms, just like you're planning.

The uCurrent is a fine device for measuring low currents, but for devices that do the whole sleep->wakeup->sleep thing, you need a fairly high temporal resolution as well (samples/second), which most multimeters don't do, and the scopes doesn't have the accuracy/range for.
For this, I use an nRF52 Development kit, paired with the Nordic Power Profiler Kit. Right now, the only thing that would come close to it's performance/usability are the high-end Agilent power-analyzers.
Considering what it does, it's a bargain at ~$100.

If you're going to be working with batteries, I'd recommend having at least one multimeter that you can do datalogging on. (Typical use-case, measuring Vbat over time). You seem to have thought about this already though.

How are you setup for soldering?  If you're going to do the initial bringup of hardware, you would likely need to do small reworks and/or attach leads to probe signals, so some kind of soldering setup is probably needed.

[nanofrog]

The Rigol DS1054Z can be unlocked to an 1104Z (increased BW, double the memory, and all the software decode options). Very easy to do as well. 

That particular PSU would do just fine and likely the wave gen as well (not used that model). Siglent's got some decent offerings as well, and then there's the used market.

The U1252B is an excellent DMM IME, as are various models from Brymen, such as the BM869S.

PM sent.

[EEVblog]

What's your budget?
You live in the US, so you have the gold mine of ebay used test gear available.
You can set up an advanced lab for not a huge amount of money.

[tooki]

Overall, sounds like a good plan, except that despite it being an accurate, good-quality, and safe meter, I wouldn't recommend the U1252B as your only multimeter, because its continuity tester is maddeningly slow. That said, you will always need more meters (it's not uncommon to need 4, to monitor volts and amps each on the input and output of something), so you might consider getting the U1252B now, and then keeping an eye out on eBay for a used Fluke 87V at a good price.

[Neomys Sapiens]

In my not-so-humble opinion, the Agilent multimeter is the only acceptable choice of the selection. As said above, it should be no problem in the US to get pre-owned quality test gear instead of GW-Instek and Rigol. If you intend to do serious analogue circuitry, make sure that your power supply is a linear one and that it is able to produce a tracking symmetric output. It does not need do be big, and if you also want a big one, that can be a switcher/hybrid.
Scope: I'm more than biased here, but I recommend Tektronix and definitely nothing else.
For a standard function generator, make your choice from the available ones from Tek, HP, Wavetek, Philips, BK...
(HP3312A is for example a totally reliable instrument, some Waveteks offer a frequency counter etc.)
Consider to include a LCR-meter in your selection.
Also, look for one or better two resistor decades of reasonable quality. Add a C decade as soon as possible. L can wait some longer.
Do not pinch pennies on the auxiliary items, like probes, adaptors, test leads, and clips/grabbers. Low quality, unreliable connections can drive you crazy and cost you lots of time.

[Mr. Scram]

In my not-so-humble opinion, the Agilent multimeter is the only acceptable choice of the selection. As said above, it should be no problem in the US to get pre-owned quality test gear instead of GW-Instek and Rigol. If you intend to do serious analogue circuitry, make sure that your power supply is a linear one and that it is able to produce a tracking symmetric output. It does not need do be big, and if you also want a big one, that can be a switcher/hybrid.
Scope: I'm more than biased here, but I recommend Tektronix and definitely nothing else.
For a standard function generator, make your choice from the available ones from Tek, HP, Wavetek, Philips, BK...
(HP3312A is for example a totally reliable instrument, some Waveteks offer a frequency counter etc.)
Consider to include a LCR-meter in your selection.
Also, look for one or better two resistor decades of reasonable quality. Add a C decade as soon as possible. L can wait some longer.
Do not pinch pennies on the auxiliary items, like probes, adaptors, test leads, and clips/grabbers. Low quality, unreliable connections can drive you crazy and cost you lots of time.

Why do you feel the HP and Agilent gear is the only acceptable choice? Buying second hand equipment comes with its own risks and problems and you pay a premium for the name printed in the front panel. On the other hand, a brand like GW-Instek has been around for quite a while now, and is generally well regarded. Some would put Rigol in the same category, though they haven't been around quite that long.

If you want to have dead reliable equipment, you'd might look at the known brands, but even those aren't bulletproof. Some people would argue that a company being owned by Danaher is not a good thing. I also think some people have a thing or two to say about the current Tektronix oscilloscopes.

[Neomys Sapiens]

Well, it was clearly marked as my personal opinion on this matters and even explicitely so, where it recommended one over the others of the pack. Beyond that, I don't really see the need to justify my recommendations, as it was clear friom the outset that others will give different ones.
I just can say that quality industrial products from well established sources have very seldomly let me down.
In my personal own lab, scopes are Tek, RF gear is R&S, HP, Marconi, Power Supplies HP, Oltronix, R&S, Heinzinger and so on, even if this means paying second hand the same price of a shiny new one from XYZ or making do with one generation older. Even given the close to identical makeup of some instruments I will maintain this. Beyond the purely technical reasons, there are other ones coming from esthetical, political and personal realms that I do not want to elaborate about.

[Mr. Scram]

Well, it was clearly marked as my personal opinion on this matters and even explicitely so, where it recommended one over the others of the pack. Beyond that, I don't really see the need to justify my recommendations, as it was clear friom the outset that others will give different ones.
I just can say that quality industrial products from well established sources have very seldomly let me down.
In my personal own lab, scopes are Tek, RF gear is R&S, HP, Marconi, Power Supplies HP, Oltronix, R&S, Heinzinger and so on, even if this means paying second hand the same price of a shiny new one from XYZ or making do with one generation older. Even given the close to identical makeup of some instruments I will maintain this. Beyond the purely technical reasons, there are other ones coming from esthetical, political and personal realms that I do not want to elaborate about.

Typically, people have a motivation for their preferences and these can be very interesting to other people. A lot people on here have extensive experience with various machines and weighed the pros and cons carefully. If the preference is what it is "because", it is of no use to anyone. However, most people on here have quite well examined points of view, which is why I asked.

A personal opinion is not synonymous to baseless rambling

[ophidian]

The uCurrent is a fine device for measuring low currents, but for devices that do the whole sleep->wakeup->sleep thing, you need a fairly high temporal resolution as well (samples/second), which most multimeters don't do, and the scopes doesn't have the accuracy/range for.
For this, I use an nRF52 Development kit, paired with the Nordic Power Profiler Kit. Right now, the only thing that would come close to it's performance/usability are the high-end Agilent power-analyzers.
Considering what it does, it's a bargain at ~$100.

Yes, these are battery operated devices.  We're using a Silicon Labs EFR32.  Their development kit does include a power analysis tool, though it is difficult to get it to work with external boards.  So far I've just been measuring the peaks using the uCurrent and the Salea Logic's analog mode during Zigbee transmissions, and my readings seem consistent with the data sheets.  For the low power states, I essentially have been using special code to keep it at various states for long periods to measure current usage during each.  Other code helps track the amount of time we spend in each mode.  I then plug all the data into into a spreadsheet to estimate total usage. 

As for a waveform generator. You could look into getting a Rigol 1074z-s instead of the 1054Z, as it has a built-in 2-channel signal generator.
(However, think about what you would need it for. I pretty much never use the one built into my Agilent MSO-X 2000)

The sense I got from some of the forum discussions is that it's just easier to use the standalone device. 

I don't really like the protocol decoding in the Rigol 1054Z. I'm assuming it's software based, as you need to zoom _just right_ for it to be able to decode the signal. I'd recommend you using the Saleae for that, and the scope to inspect the levels and edges of the waveforms, just like you're planning.

I've been very pleased with the Salea for decoding.  I feel like browsing through protocol events is a task best suited to a PC.  I don't much like the Logic 1.2 software otherwise, though; it seems buggy (sometimes I can use the mouse to see the voltage at a point in analog mode, sometimes not) and there's no math operations available. 

How are you setup for soldering?  If you're going to do the initial bringup of hardware, you would likely need to do small reworks and/or attach leads to probe signals, so some kind of soldering setup is probably needed.

I bought a Hakko FX888D early on to remove an unnecessary inverter from an ADC's reset line.  I still haven't mastered adding wires to our test points, though, which is the main soldering task I've had to do; the solder just doesn't seem to want to stay on the wire.  This was actually my first experience soldering anything--somehow, in school I made it through numerous labs without having to do any soldering!  The EEVBlog video tutorials have been very helpful, though.

Thanks for your comments.

[ophidian]

What's your budget?
You live in the US, so you have the gold mine of ebay used test gear available.
You can set up an advanced lab for not a huge amount of money.

I'm hoping to land not much more than US $1000; at US $2000 I start getting anxious, and at USD $3000 my wife starts getting anxious--so that's about my limit.

I've browsed eBay a bit, but I find myself paralyzed by the sheer number of options when I don't have the experience to judge any of them. 

One thing I tried to do was find items that looked familiar.  For example, I'm pretty sure I used the HP E3631A power supply or something very much like it back in college (class of '04), but I don't know enough to say what would make it superior or inferior to a contemporary Rigol or Instek PSU, and at $500 on eBay it's in a similar price range. 

I did find a lot of well-regarded multimeters on eBay, so I'll probably pick up an extra as several folks have suggested.

Thanks!

[ophidian]

Note that the resolution of the D model power supply is lower than it is with the similar S model. You may want to look into the price difference and see whether upgrading is worth it to you. Also note that the 3 channel model appears to have variants that can control the third channel from software, but commonly seems to have an unregulated third channel. The four channel model should be regulated on all channels.

Your set is remarkably similar to what I've been looking into recently, so I'm familiar with at least some of the pitfalls

Wow, I totally missed the difference in resolution.  Thanks for the catch.  I just browsed the 3303 manual and didn't see any commands for controlling the 3rd output, even though the description on the website seems to imply all three channels can be controlled remotely.  But if it's a fixed 2.5/3.3/5 V perhaps I wouldn't really want to anyway.

Thanks!

[iainwhite]

If you do decide on an U1252B, check out this thread (particularly pages 22-23):
https://www.eevblog.com/forum/testgear/bunch-of-keysight-u1252b-meters-around/msg1270322/#msg1270322

A number of forum members (including me) bought them from eBay seller Nmori87
He has recently joined this forum and has offered to sell direct to forum members.

[Mr. Scram]

Wow, I totally missed the difference in resolution.  Thanks for the catch.  I just browsed the 3303 manual and didn't see any commands for controlling the 3rd output, even though the description on the website seems to imply all three channels can be controlled remotely.  But if it's a fixed 2.5/3.3/5 V perhaps I wouldn't really want to anyway.

Thanks!

No problem, that one nearly caught me too. Having an extra digit on the power supply can be nice. In many cases, it's the difference between needing another multimeter, or even two, or just reading the numbers off the power supply.

There appear to be three channel models with the triple voltage selector switch, and ones without. I assume the latter may be the result of the four channel model being introduced, making it easier for them to just clip a channel off it and sell it as a three channel, but I'm not sure. I do know they re-jigged the front panel to accommodate the extra fourth channel. It would have been easier if they would at least have given the various models different submodel or version numbers, but they haven't.

[ophidian]

As others have said, get a 1054Z, its firmware can be cracked and get free 1104Z features.
For the wave gen, I never found a good use of them. When I do need one, my Analog Discovery usually serve me well, as an electronics Swiss knife.
I would get a second meter, preferably a desktop meter or UniT 181 or this sort of thing -- can be stationed, with PC connection.
Then I would replace the 1252B with an 87V.

Why the 87V over the 1252B?  Looking at that gigantic multimeter spreadsheet, nothing is jumping out at me.  Is it just that Fluke is built better?

I hadn't considered the Analog Discovery.  Thanks for the tips.

[tooki]

As others have said, get a 1054Z, its firmware can be cracked and get free 1104Z features.
For the wave gen, I never found a good use of them. When I do need one, my Analog Discovery usually serve me well, as an electronics Swiss knife.
I would get a second meter, preferably a desktop meter or UniT 181 or this sort of thing -- can be stationed, with PC connection.
Then I would replace the 1252B with an 87V.

Why the 87V over the 1252B?  Looking at that gigantic multimeter spreadsheet, nothing is jumping out at me.  Is it just that Fluke is built better?

I already explained why in my comment above.

[ophidian]

Well, it was clearly marked as my personal opinion on this matters and even explicitely so, where it recommended one over the others of the pack. Beyond that, I don't really see the need to justify my recommendations, as it was clear friom the outset that others will give different ones.
I just can say that quality industrial products from well established sources have very seldomly let me down.
In my personal own lab, scopes are Tek, RF gear is R&S, HP, Marconi, Power Supplies HP, Oltronix, R&S, Heinzinger and so on, even if this means paying second hand the same price of a shiny new one from XYZ or making do with one generation older. Even given the close to identical makeup of some instruments I will maintain this. Beyond the purely technical reasons, there are other ones coming from esthetical, political and personal realms that I do not want to elaborate about.

I'd be perfectly happy to make do with a generation-older name-brand scope, but in browsing four-channel oscilloscopes in the $1000 range on eBay, the majority seem to be > 20 years old, sometimes far older.  I did briefly get excited when I saw an Agilent 1024A for around $800, but I quickly learned from the forum that it's just a re-branded Rigol .   

I guess I'm having trouble seeing eBay as the gold mine that others describe.  Am I doing something wrong?  Maybe it's just more of a gold mine for the $3000-$5000 budget.

[ophidian]

As others have said, get a 1054Z, its firmware can be cracked and get free 1104Z features.
For the wave gen, I never found a good use of them. When I do need one, my Analog Discovery usually serve me well, as an electronics Swiss knife.
I would get a second meter, preferably a desktop meter or UniT 181 or this sort of thing -- can be stationed, with PC connection.
Then I would replace the 1252B with an 87V.

Why the 87V over the 1252B?  Looking at that gigantic multimeter spreadsheet, nothing is jumping out at me.  Is it just that Fluke is built better?

I hadn't considered the Analog Discovery.  Thanks for the tips.

I have a U1461A, which presumably shares firmware building blocks with its little brothers. So far, my U1461A has these firmware bugs:
1. Input mux not working, reading 0V all the time in voltage mode, some low resistance (couple 10 Ks) in Ohm mode, reboot takes care of it.
2. Unit/mode display incorrect, in insulation test mode, uA and MOhm can't switch or doesn't reflect on screen after switching, reboot takes care if it.
3. Continuity buzzer bug. Latching mode behaves weird, direct mode works fine.

Mine doesn't suffer from RF interference bug, though.

On the other hand, my 289 has zero firmware bugs, besides slooooow boot up time and the "fuck you" warranty policy (input MOV broken, presumably corroded by leaky battery, neither battery supplier, in this case, Tektronix cal center, nor Fluke takes the loss, each blaming the other and all blaming me).

Well, that's depressing.  I've used so many Agilents in the past that I had it in my head that they were a reliable brand, and therefore Keysight would also be a reliable brand.  I guess I was wrong.  In that case, I think your 87V + UniT for logging is a smarter approach. 

Thanks.

[Mr. Scram]

Well, that's depressing.  I've used so many Agilents in the past that I had it in my head that they were a reliable brand, and therefore Keysight would also be a reliable brand.  I guess I was wrong.  In that case, I think your 87V + UniT for logging is a smarter approach. 

Thanks.

It is a reliable brand, but perfection is not something that realistically exists.

[iainwhite]

It is a reliable brand

I agree that Agilent/Keysight is a reliable brand. So is Fluke. I have both and I am happy with them.
I think UniT is a mixed bag - for example see Dave's video "EEVblog #712 - Uni-T UT71E Multimeter"

Looking at Blueskull's post, I see he suggests the Uni-T 181A. That is $350 on Amazon. For $400 on Amazon, you can get a Fluke 287.
For $180 you can get an Agilent U1252B from "nmori87".

I get why people like the Uni-T meters around the $60-100 price point but at the $300-400 point I would look elsewhere.

[Mr. Scram]

It is a reliable brand

I agree that Agilent/Keysight is a reliable brand. So is Fluke. I have both and I am happy with them.
I think UniT is a mixed bag - for example see Dave's video "EEVblog #712 - Uni-T UT71E Multimeter"

Looking at Blueskull's post, I see he suggests the Uni-T 181A. That is $350 on Amazon. For $400 on Amazon, you can get a Fluke 287.
For $180 you can get an Agilent U1252B from "nmori87".

I get why people like the Uni-T meters around the $60-100 price point but at the $300-400 point I would look elsewhere.

It depends on where you live. In the US, Agilent/Keysight and Fluke meters are relatively affordable, but they can be frightningly expensive elsewhere.

[iainwhite]

It depends on where you live. In the US, Agilent/Keysight and Fluke meters are relatively affordable, but they can be frightningly expensive elsewhere.

Totally agree.   My post was aimed at OP who looks to be located in the US.

[Mr. Scram]

Totally agree.   My post was aimed at OP who looks to be located in the US.

Yes, agreed. I understand that most of the advice is given for OP's situation. That's obviously sensible. I'm just trying to clarify why Blueskull, who's not in the US, may have made a different choice, because it's likely a fair number of passerby from different parts of the world will read this thread too. When a Fluke 289 costs double the number you mentioned, preferences obviously shift.