Buying test equipment at work? How low can we go?

[Jolle]

Hi

I changed workplace two months ago, and since my new place is a small business the gear they had lying around which I have been given is... what should i say... dated...
They are aware of this, so when I asked to buy a new oscilloscope (with four channels and a colour screen, fancy I know  ) they didn't hesitate to say yes. Therefore I have been looking for different options to see what is available and so far I have been looking at Siglent SDS2000x Plus and GW Instek MDO2202EG, as they both seems to have the functionality I need (even though they are inexpensive).
At my last job I had a keysight DSXO3024A which I liked, but when I look at the price compared to the two others I find it really hard to justify the difference - especially since many of the keysight functions is locked. And would rather spend some money on more test equipment.

This is not a direct question about the two scopes, but rather a question about what is important when buying gear for work? At work time is money, so I believe test equipment which have a poor "price/performance on datasheet" ratio can be justified by things like good support, confidence in the mesaurements and the ease of operation.
But at the moment the amount of cheap and useful instruments is really impressive with things like migsigs current and differential probes, NanoVNA, component testers, cheap multimeters etc. you can get fun useful instruments at a fraction of the price of the professional equipment. But where should one draw the line between hobby equipment and equipment being suitable for use at work?
Is it only a matter of being aware of the specifications regarding ones need? is it a must that the equipment can be calibrated? Should one avoid the cheap fun stuff because it is too unreliable and we don't have a track record of their behavior on the long time scale - which could end up being a lot more expensive for a company than the cost of well tested equipment?

Any thoughts on this subject?

(Sorry, but just a small question regarding the two oscilloscopes: Having read about the GW instek I am really impressed by the amount of functionality it has implemented and actually find it hard to see what the siglent offers besides a larger screen with touch capability - so in case I have missed something please enlighten me on the siglents capability)

[nctnico]

For professional use get equipment with mature firmware. Second guessing your measurement results is the worst thing that can happen (one of the reasons I spend $3k on a not extremely special function generator recently). But it also depends on the complexity of a piece of equipment. There is not much that can go wrong with a relatively simple DMM or current probe for example. Another thing to look for is equipment which supports getting work done quickly with least fiddling of knobs. Workflow is everything. A simple bench DMM can be easier to work with compared to one with high precission and lots of bell & whistles you don't need for 99% of the measurement tasks.

[bdunham7]

I'll agree that you probably want a 'mature' product, one that has been out a while so that the bugs have been found and corrected and so that you can get well-informed opinions from other users.

However, when you are looking at 'A' and 'B' brands and factoring in the value of support, make sure you will actually get the support if needed.  Check with other users in your same area for opinions on this.  The support you get is highly dependent on where you are and in many cases, who you buy the unit from.

[uski]

It is worth establishing a good relationship with the reps of the equipment manufacturers early on.
In the past, I had a problem with EMI certiifcation and the Rohde&Schwarz local rep did not hesitate and lended me the equipment I needed to find the issue and fix it. I didn't have to purchase it. I really, really appreciated it. It really helped us and we did not have the budget to purchase the equipment anyway. Having the contact info of these people in your address book is going to help you in the long term, so when you are doing your research, give them a call, tell them your needs, and give them a reasonable chance to compete.

Also, just a small tip : my policy is to never disclose pricing from one competitor to the other one. I am up front about this and it actually helps me get even better deals than if I was showing everyone else the "best quote".
This is also valid here, I am never going to disclose some crazy deals that I got for some equipment here on EEVblog or anywhere else. But all I can tell is that if you negotiate properly you can get double digit off MSRP easily.

[uski]

I would add to blueskull's post that it is very important to understand the limitations of your instruments. I did use a cheap USB VNA at work to do antenna tuning and had no issue passing certification testing that involved checking the antenna performance. But that may or may not be possible, and if you are having issues, you need to quickly know where is the weak link in your metrology chain.

It also depends of the budget you have. In my previous company, we really did not have a lot of money to spend on this so I was very conservative. In my current workplace, we have higher budgets so it is less of an issue.

[Miti]

I think you should start with the budget and buy quality equipment within your budget, rather than squeezing a lot of junk. Prioritize what you need first and buy more later when the need arises. Sure I can be cheap at home but equipment that can’t do the job or craps up at work, doesn’t make a good impression.

[Fungus]

Is it only a matter of being aware of the specifications regarding ones need?

No.

is it a must that the equipment can be calibrated?

Not from a technical point of view, no.

If there's any chance at all of ending up in a lawsuit over a published figure then it's good to have an independent paper trail for all your devices.

Should one avoid the cheap fun stuff because it is too unreliable and we don't have a track record of their behavior on the long time scale - which could end up being a lot more expensive for a company than the cost of well tested equipment?

I'd say "no". Cheap hardware is just as reliable. You should be much more worried about firmware than hardware.

(and there's horror stories at both ends of the scale)

Any thoughts on this subject?

As a small business the single most important consideration might be "how close is the local dealer if this thing breaks in the middle of an important job"?

OTOH you can keep a spare around, or if it's a place with several instruments you might be able to grab another one and that's not very importantl. It's all up to you.

(Sorry, but just a small question regarding the two oscilloscopes: Having read about the GW instek I am really impressed by the amount of functionality it has implemented and actually find it hard to see what the siglent offers besides a larger screen with touch capability - so in case I have missed something please enlighten me on the siglents capability)

You may be right but I'm sure the resident Siglent salesman will be along any minute now to do his sales pitch.

[Berni]

Yes you get what you pay for with test equipment. I just had an adventure with that

Recently i bought a programmable PSU at work for cycling some batteries. It was a cheep Tenma 30V 5A one from Farnell (This is chinese gear re-badgeed as store brand) since it was cheep and i could get it next day delivery from Farnell. I just need to shove a few amps into a battery, i don't need a fancy PSU. Well... i ended up regretting that decision multiple times.

The thing arrived next day in the post as promised. Looked pretty decent and all. Connecting it via USB shows up as a serial port as a lot of these simple programmable instruments do. So i write my automation software in C# and its running fine, but then at random moments the voltage and current readings from the PSU got all mixed up. Hunted around in the software and could not find the cause, it simply looked like it starts returning wrong data. Then i managed to manually reproduce the problem by sending it commands KeysightIO, i ended up installing a different VISA library, still problem was there. Running out of options i noticed the PSU also had a Dsub9 on it, so i connected to that using a USB to RS232 adapter and suddenly all the problems disapeared. There goes 1 day of banging my head.

Okay so the system is running for a few days now, smooth sailing, not a single problem. So i connect it to a bigger battery and let it do its thing on it. Then while at home i decide to check in on it and remote desktop into it, noticing that the PSU is reporting 6A output even tho current limit for charging was set to 3A while its only a 5A PSU. Oh shit shit shit, this is not good . That lead acid battery connected to it was at almost 16V or something by now. I started sending it manual commands and turned out i could disable the PSU output, but this just reduced the current being pushed into the battery down to about 1.5A (While PSU was reporting 0A) because the relays inside the PSU switched to a lower transformer tap but the output pass transistors are blown and shorted. So to fix it i turned on the electronic load to sink away all that current so that the situation was stabilized until i got to work on Monday and unplugged it.

Thank god i was not using this to cycle a lithium battery.

[CatalinaWOW]

You haven't said what your business is, or what you will be doing, which are the most important pieces of information regarding the test equipment you need.

If you are a calibration lab, or doing EMI certification or similar work you need high quality instruments with a track record.  They impact directly on the credibility of your work.

If you are a small house designing custom boxes to solve small business control or automation problems you can probably get along just fine with the cheapest, nastiest pieces of gear that let you work efficiently (your time is money comment is apt).  But even awkward and slow user interfaces may not matter if you spend days or weeks at a time without changing setups, which can often happen.

These are just two examples.  There are a myriad of points between and beyond.  The difference between successful businesses and less successful ones is the care with which they allocate resources.  But time is money here too.  It is often better to make a sub-optimal decision than to spend a great deal of time refining an analysis beyond the quality of the available data or spending a lot of time improving the data quality.

[tautech]

(Sorry, but just a small question regarding the two oscilloscopes: Having read about the GW instek I am really impressed by the amount of functionality it has implemented and actually find it hard to see what the siglent offers besides a larger screen with touch capability - so in case I have missed something please enlighten me on the siglents capability)

You may be right but I'm sure the resident Siglent salesman will be along any minute now to do his sales pitch.

Don't need to, if engineers can't distinguish the features between 2 instruments then you're in the wrong job.

[CatalinaWOW]

Yes you get what you pay for with test equipment. I just had an adventure with that

Recently i bought a programmable PSU at work for cycling some batteries. It was a cheep Tenma 30V 5A one from Farnell (This is chinese gear re-badgeed as store brand) since it was cheep and i could get it next day delivery from Farnell. I just need to shove a few amps into a battery, i don't need a fancy PSU. Well... i ended up regretting that decision multiple times.

The thing arrived next day in the post as promised. Looked pretty decent and all. Connecting it via USB shows up as a serial port as a lot of these simple programmable instruments do. So i write my automation software in C# and its running fine, but then at random moments the voltage and current readings from the PSU got all mixed up. Hunted around in the software and could not find the cause, it simply looked like it starts returning wrong data. Then i managed to manually reproduce the problem by sending it commands KeysightIO, i ended up installing a different VISA library, still problem was there. Running out of options i noticed the PSU also had a Dsub9 on it, so i connected to that using a USB to RS232 adapter and suddenly all the problems disapeared. There goes 1 day of banging my head.

Okay so the system is running for a few days now, smooth sailing, not a single problem. So i connect it to a bigger battery and let it do its thing on it. Then while at home i decide to check in on it and remote desktop into it, noticing that the PSU is reporting 6A output even tho current limit for charging was set to 3A while its only a 5A PSU. Oh shit shit shit, this is not good . That lead acid battery connected to it was at almost 16V or something by now. I started sending it manual commands and turned out i could disable the PSU output, but this just reduced the current being pushed into the battery down to about 1.5A (While PSU was reporting 0A) because the relays inside the PSU switched to a lower transformer tap but the output pass transistors are blown and shorted. So to fix it i turned on the electronic load to sink away all that current so that the situation was stabilized until i got to work on Monday and unplugged it.

Thank god i was not using this to cycle a lithium battery.

I just want to understand your comment and recommendation.

1st.  The device didn't work with KeysiteIO and two somebodies VISA libraries.  Did they advertise that it would?  Is your recommendation to stick with a vendor you have used in the past so you don't waste time learning new software interfaces, or is it something else.

2nd.  Are you asserting that this was a random hardware failure due to low quality?  Or that it was a line spike or other unforeseen overstress that a name brand unit would be more robust against (possibly true, but no guarantees)?  Any possibility you operated it outside of its safe operating region?  Or are you asserting (incorrectly) that big name brands will operate with any combination of voltage and current within each of those limits?   Or asserting (more correctly, but not totally so) that name brand units will at least shut down without damage when the go outside their operating limits.

3rd.  Are you asserting that a name brand would have some form of continuous BIT that would have detected a hardware failure and shut down?  Again this isn't generally true.

I agree that your experience was discouraging and even terrifying, but I am not sure you couldn't have had the same ride with a high dollar unit.  With the possible exception that a higher dollar unit that advertised VISA compliance would probably deliver.

[Berni]

I just want to understand your comment and recommendation.

1st.  The device didn't work with KeysiteIO and two somebodies VISA libraries.  Did they advertise that it would?  Is your recommendation to stick with a vendor you have used in the past so you don't waste time learning new software interfaces, or is it something else.

The VISA library merely provides a standard API for talking to test equipment. You just tell it the equipment address and the VISA library does what it takes to get the data to the instrument, be it via USB, RS232, GPIB, Ethernet, PCIe...etc You can use any VISA library with any instrument. Be it the original National Instruments VISA, KeysightIO, Rigol UltraSigma...etc I just prefer KeysightIO.

The problem is that the USB interface on that Tenma PSU was unreliable and i ended up hunting all over the place where the occasional glitches came from. If you send the PSU a "Read voltage" command then the next response should be the voltage, and it was 99% of the time, but the 1% of the time it would start returning whatever was previously asked from it. The BK Precision electronic load is connected to the same PC over USB and i have never seen it glitch even once. The fact that simply switching to a USB-RS232 adapter and talking RS232 to the PSU made the problem go away shows that it was not my software at fault.

The USB port on that Tenma PSU simply is unreliable with prolonged use for some reason.

2nd.  Are you asserting that this was a random hardware failure due to low quality?  Or that it was a line spike or other unforeseen overstress that a name brand unit would be more robust against (possibly true, but no guarantees)?  Any possibility you operated it outside of its safe operating region?  Or are you asserting (incorrectly) that big name brands will operate with any combination of voltage and current within each of those limits?   Or asserting (more correctly, but not totally so) that name brand units will at least shut down without damage when the go outside their operating limits.

The Tenma PSU was rated for 30V 5A. I have read the manual and there is NO mention of a power limit or anything, you can pull 30V 5A simultaneously for an indefinite amount of time. The PSU was running in a room temperature environment with no air flow obstructions, there was a diode protecting it from backflow from the battery. It is a transformer linear PSU while the same mains circuit has plenty of smaller switchmode supplies that die 10x sooner from a mains voltage spike. So i believe i have never exceeded any of its specifications.

When it happened the PSU was idle for >1hour before told to supply 14V at 3A. It did that but then after about 1000 seconds according to the logs the output current started drifting out of calibration over one or two seconds, then went erratic and then shot up to 6A and stayed there as the PSU died. Last time i checked the fan was working fine and even if it died there should be a thermal shutdown because the fan is temperature controlled so there is a temperature sensor on the heatsink.

3rd.  Are you asserting that a name brand would have some form of continuous BIT that would have detected a hardware failure and shut down?  Again this isn't generally true.

I agree that your experience was discouraging and even terrifying, but I am not sure you couldn't have had the same ride with a high dollar unit.  With the possible exception that a higher dollar unit that advertised VISA compliance would probably deliver.

Well upon googling it i found out that Dave also blew up this same PSU branded as a "Korad" in one of his review eevblog videos by shorting its output erratically.

We bought another Farnell rebranded switching lab PSU (10A thing you see in ElectroBooms videos) and we blew it up the same day by also simply shorting its output when set to a high enough current. Once we got the replacement we found its such a EMI firehose that it might as well be sold as a radio jammer, we just gave it to the CEO to do what he wants with it just to get this radio transmitter out of the lab.In another cheap rebranded PSU from Farnell the X capacitors in the mains filter suddenly caught fire even tho those ware supposed to be self healing X class caps.

One of the Rigol DP832 PSUs suffers from an occasional random reboot, likely due to the 5V regulator overheating problem. These same PSUs also stop responding on LAN if it sees a DDNS packet on the network so we had to isolate it behind a layer 4 switch.

At home i mostly have name brand gear, HP/Agilent/Keysight, Keithley, Tektronix... etc with a sprinkle of some other old stuff like Gigatronix, Wavecrest etc.. Most of it used, some of it new. And i have never had any similar glicthy issues with the other gear. Done SCPI remote control on them via USB, LAN, GPIB, had LAN-GPIB bridges involved etc... running with KeysightIO to talk to all these various brands of gear and it never had any weird intermittent problems.

[tautech]

Sorry, but just a small question regarding the two oscilloscopes: Having read about the GW instek I am really impressed by the amount of functionality it has implemented and actually find it hard to see what the siglent offers besides a larger screen with touch capability - so in case I have missed something please enlighten me on the Siglents capability

Really ?
$ vs capability one blows the other away.
Maybe it's best you have a good study of the datasheets:
https://int.siglent.com/u_file/document/SDS2000X%20Plus_Datasheet_DS0102XP_E01A.pdf
https://www.testequipmentdepot.com/instek/pdf/mdo-2000egx_datasheet.pdf

Come back with any specific questions if you must.

[Fungus]

You haven't said what your business is, or what you will be doing, which are the most important pieces of information regarding the test equipment you need.

If you are a small house designing custom boxes to solve small business control or automation problems you can probably get along just fine with the cheapest, nastiest pieces of gear that let you work efficiently (your time is money comment is apt).  But even awkward and slow user interfaces may not matter if you spend days or weeks at a time without changing setups, which can often happen.

This.

[jogri]

I tend to follow two rules:

1) Can it be used if the task suddenly becomes more demanding? In other words, are you already pushing it to its design limit or is there some headroom for a different setup down the line (in the form of a higher bandwith/sampling rate/more digits etc)? Better to buy one more slightly expensive unit now than to buy a second, better one later.

2) Can it be used by the biggest moron in the room? Remember, it's not your personal equipment, so somebody else might get it later. Spending more bucks on an instrument with an intuitive/easy to use interface will save you money in the long run as working time isn't free (and even minutes of working time can quickly add up if you have to use it daily)... (This question can also be important for safety: What happens if someone screws up royally? Can it protect itself or will it get destroyed? Two cheaper instruments will cost more than an expensive one that only needs a small fix)

[TimNJ]

My two cents:

Obviously if there's a certain piece of high end test equipment that you absolutely need, there's no getting around that. You gotta do what you gotta do. Goes without saying.

Depends on how your management is, but usually they like when you save them money.  By that I mean: You buy the $1,000 scope and not the $5,000 scope. If your lab is outdated in general, it's probably more beneficial to get a lot of mid-range stuff, instead of a few pieces of high end stuff, given that your management doesn't want to pay for a lot of high end stuff. If you can show that you are actively thinking about not wasting their money, and show that it will make the engineering department more productive, they'll probably be okay with it.

[David Hess]

At work time is money, so I believe test equipment which have a poor "price/performance on datasheet" ratio can be justified by things like good support, confidence in the mesaurements and the ease of operation.

That is how I would think of it.  Any money saved by buying low end test gear will be lost in wasted time and effort.  I would definately avoid Rigol, Siglent, and GWI for only this reason.

[Fungus]

That is how I would think of it.  Any money saved by buying low end test gear will be lost in wasted time and effort.  I would definately avoid Rigol, Siglent, and GWI for only this reason.

Can you give an example of when this might happen...?

(or examples of when it's happened to you in the past)

The two 'scopes mentioned by the OP seem like very capable devices.

and (b) At what price level would you consider an oscilloscope to guaranteed to never waste any time or effort? What brand?

[tkamiya]

When it comes to scopes, I have Siglent, Rigol, HP, and Tek.  First two are low end products in their respective line up and last two are older but higher end (at the time of production) types.

What doesn't show up in spec sheet is things like how many button push does it take to change AC couple to DC couple, enable bandwidth, change trigger types and levels, and turn this feature or that feature on and off.  If you must use scope at work and your time is money, then higher-end scope will justify itself in no time.  Believing it is set one way and having it wrong can be a very costly mistake, too.

In last few years, Chinese makes have come leaps and bounds into main stream.  Each of them make low end to high end.  Rumor has it, Tek's low end scopes are actually made by one of them. 

[tautech]

Tek and HPAK scopes have been made in Asia for a decade or more.....old news.....but the prices they charged back then compared to what we can get now for the same money is the real advancement.
The SH TDS2012B I got some 10 years back was right up there then in capability however it's abysmal compared to what's available now.

[Calvin]

Hi,

with all due respect to David Hess and tkamiya, but you simply can´t generalize in this way.
At work we have scopes from LeCroy R&S and Siglent.
The LeCroy (a Wavesomething) is nice to work with due to its large touch-display and sensible GUI.
The R&S is a handheld ScopeRider and nice for its 10Bit Hardware and isolated inputs.
The GUI was seriously buggy till the latest update (channel inverting hid in the depths of a mathematics function  , killed two USB sticks whilst trying to update, due to a failure of R&S in the USB programming, etc. etc.  ).
But the worst is sticking rubber buttons which you almost can´t figure out by sight due to their minimal stroke.
Almost every second push on a button leads to some weird behaviour or kind of hang-up.
R&S didn´t offer any solution or compensation other than two weeks of free testing of two options. 
The Siglent is the smallest and by far cheapest of the three, but offers the most functionality.
Handling is easy and intuitive and most of the time its the scope we get results on the display the fastest.
And so far we have no reason to distrust the results it offers more than the results we are getting from the LeCroy or the R&S.
Maybe its just the exemption from the rule, but I think, the times when this rule applied have passed.

regards
Calvin

[tkamiya]

I have zero experience in R&S or LeCroy.  My opinion is based on my own limited experiences.

I tend to stick with ones that I can get service information.  Siglent was a gift I received.  Rigol was given to me because it didn't work.  (easy fix)

Tautech...  I know.  I once worked for Sony way back when I was in Japan.

[David Hess]

That is how I would think of it.  Any money saved by buying low end test gear will be lost in wasted time and effort.  I would definately avoid Rigol, Siglent, and GWI for only this reason.

Can you give an example of when this might happen...?

The change in bandwidth versus attenuation of Rigol's DSOs comes to mind; it is very frustrating to have measurements change with attenuation.  Their automatic measurements of the display record also cause inconsistent measurements.

(or examples of when it's happened to you in the past)

I have had problems in the past with multimeters which change input resistance with range.  What do you do when one range says 100.00 and another says 101.0?  That may not seem like much of an error but it absolutely is.

(b) At what price level would you consider an oscilloscope to guaranteed to never waste any time or effort? What brand?

It is more of a matter of trust.  I would trust Keysight, Rohde & Schwarz, and Tektronix.  Having dealt with Rigol in the past, I do not trust them at all.

One thing I look for is detailed documentation on instrument operation so it is possible to understand what is going on.  Multimeters should specify input resistance and tolerance for every range.  DSOs should specify how automatic measurements are made.

[nctnico]

And so far we have no reason to distrust the results it offers more than the results we are getting from the LeCroy or the R&S.
Maybe its just the exemption from the rule, but I think, the times when this rule applied have passed.

A few devices are just a small sample. The difference often comes out in the corner cases (where you really get close the limits of an instrument).

[CatalinaWOW]

I have to mention another obvious point.  We all agree that speed in use is a valuable asset.  What is less obvious is that a bad UI that you know thoroughly is faster than a good UI that is strange to you.  At least during the learning period.  And in many cases a "bad" UI isn't intrinsically bad, just not familiar to the evaluator.  A great example of this is the "Ribbon" interface introduced about a decade ago by Microsoft.  Actively disliked by most who were deeply trained in the older interface, the newer interface has objective evidence of its superiority.  In spite of that I still prefer the older interface.

The differences in UI can also be very task specific.  With one awkward and slow for one class of tasks, and much better on others and visa versa.

[Fungus]

A great example of this is the "Ribbon" interface introduced about a decade ago by Microsoft.  Actively disliked by most who were deeply trained in the older interface, the newer interface has objective evidence of its superiority.  In spite of that I still prefer the older interface.

The ribbon is slower to find things. It's too verbose, not as discoverable, all the mouse movements are bigger and need more clicks. You can tell they designed it around people who've never used a computer when they should really have designed it around people who actually use computers.

It's cutting off your nose to spite your face. The beginners can learn and get up to speed, the people who use it daily will be pissed off forever.

IMHO, YMMV.

[Fungus]

Can you give an example of when this might happen...?

Their automatic measurements of the display record also cause inconsistent measurements.

You had to go down to $350 Rigols to see that these days and once you're aware that it exists then it's much less of a problem.

(we're back to "know how it works" again, and I'm not sure if there's a price level where that stops being true)

I have had problems in the past with multimeters which change input resistance with range.  What do you do when one range says 100.00 and another says 101.0?  That may not seem like much of an error but it absolutely is.

Sure, but I've got cheap-ass multimeters that don't do that.

AFAICT there's real stinkers from all manufacturers at all levels. Spending more money is no guarantee in itself.

[David Hess]

Sure, but I've got cheap-ass multimeters that don't do that.

But how do you determine that before buying and testing?  Cheap test instruments come with incomplete documentation and specifications.

[nctnico]

But how do you determine that before buying and testing?  Cheap test instruments come with incomplete documentation and specifications.

That's why we have this discussion board.

Exactly. Investing some time into a cheaper instrument can be worthwhile.

[Fungus]

But how do you determine that before buying and testing?  Cheap test instruments come with incomplete documentation and specifications.

That's why we have this discussion board.

Yep. There's plenty of horror stories on the high end, too. Spending lots of money is no guarantee.

(and cheap stuff tends to be more up to date, using new tech...)