Up and Coming 6½ Digit bench meters in 2016 or Go for a Keysight 34401A ?

[diyaudio]

My end of year present will be arriving much later than I expected! as a result of our local Rand currency reducing itself to almost nothing. 

Anyway, I need a 6½ Digit meter for 2016 as I am starting to work and experiment with projects that calls for one, are there any up and coming 6½ Digit bench meters in 2016 I should be on the lookout for? or is the 34401A still the obvious choice.

I was considering a Rigol DM3068 , but it just lacks the Quality of the 34401A (Both PC software and Hardware) and im the owner of two Rigol instruments.
 

[Orange]

If you can afford, go for an Keysight 34461 or one other models from that series 34465 or 34467. I think a 34401A is nowadays more expensive than a 34461A.
I have both and like the display of the 34401A, it also starts up much quicker.

I would avoid the Rigol, they have not a good track record. 

[diyaudio]

I would strongly recommend DMM4050. This thing turns on/off within 0.5 second, and warms up quickly to get specified accuracy.
Besides, it comes with clean and simple VFD display, not the LCD rubbish, plus free GPIB.
Its specified accuracy in virtually all ranges are better than 344x1A, and are comparable to 34470A.
List price: $1450, but I believe if you wrote a letter to TEquipment, they can give you more than [eevblog coupon] off.
Last time I got 20% off on an item, so expect $1200-ish price.

Rigol is not a very good competitor in this case, as the price is only $200 cheaper than 34461A.
Consider K2000 instead, you can get a used one on eBay for $500, and the specs are better than K2100, comparable to or better in some terms than 34461A.

Thanks for the input.

I do like the LCD, but im not sure how the issue around boot time is an issue once the instrument has booted ? I will mostly use simply use the USB interface/scpi with Matlab, the legacy GPIB isnt really a requirement for me. 


Also note, the Keysight 34461A has a big display, the UX is clean and the layout works well, also Benchvue works very well.  so thats a big winner.  im almost 90% certain this is what I will be investing in.
thats $1,086.03, its Rand 24K, I can almost buy a small used car for that price here where I live.   


I do agree, the TEK has better specs in some respects.

http://www.testequipmentdepot.com/tektronix/digital-multimeter/digital-multimeter-dmm4050.htm

6.5, 0.0024% Accuracy, Bench Digital Multimeter
Product Features:

Measures Volts, Amps, Ohms, Frequency, Period, Continuity, Diode Test, Capacitance, & Temperature
6.5 Digit Resolution (24 PPM)
Basic VDC Accuracy of up to 0.0024% (1 yr.)
200 mV to 1000 V Voltage Range, with up to 10 ?V Resolution
200 ?A to 10 A Current Range, with up to 1 nA Resolution
200 ? to 100 M? Ohm Range, with up to 1 m? Resolution
CAT I 1000 V, CAT II 600 V
3-year Warranty
Available Functions & Features

Volts, Ohms, and Amps Measurements
Diode and Continuity Testing
Frequency Measurements
2×4 Ohms 4-wire Measurement Technique
Dedicated DC Leakage Current Measurements
Six Dedicated Buttons for Fast Access to Instrument Setups
Limit Compare Mode for Pass/Fail Testing
Connectivity

Front 2×4 Measurement Inputs
USB Host Port on Front Panel for Easy Storage of Measurement Data and Instrument Settings
RS-232 on Rear Panel for Quick PC Connectivity
Includes USB to RS-232 Interface Adapter Cable
Includes National Instrument’s LabVIEW SignalExpress™ TE Limited
Edition for Connecting Your Bench

[BloodyCactus]

the tek dmm4040/dmm4050 is just the fluke 8045/8046. I would rather see 'fluke' on my bench dmm than tek

I want to get the Iwatsu VOAC7602 if I can get it past the wife approval

[fanOfeeDIY]

I want to get the Iwatsu VOAC7602 if I can get it past the wife approval

 
That sounds like it is tougher than getting approval from your boss. 

[retrolefty]

thats $1,086.03, its Rand 24K, I can almost buy a small used car for that price here where I live.   

 You can get a $1K car even here in the US. It's just that meter will work a whole lot better then the car would. 

[EEVblog]

The 34401A is an ancient design. The 34461A is what you want.

[Macbeth]

Keithley 2015's are cheap enough (~$300) and come with THD audio DSP board that they would seem a no brainer for someone working with audio. However, the audio stuff is designed for the telecommunications industry rather than the audiophile stuff. Still it's worth considering just as a spare 6.5 digit meter anyway!

Definitely avoid the Rigol DM3000 series. Effectively totally unsupported, no firmware updates, loads of bugs that will never be fixed. (Though I am saying this as a 5.5 digit DM3058E owner. Siglent SDM3055 is a much better choice for anyone after a 5.5 digit, at least they actively support it.)

However Dave is right. If you want new then the 34461A is very nice indeed.

[linux-works]

The 34401A is an ancient design. The 34461A is what you want.

having just bought a used 34401a, I guess I'm biased

the used one was about $450 and it was in great condition.  its a proven time-tested design and its long manuf avail time says something (good) about it.  maybe the other will sell for a long time, but I don't think the 'fancy screen' TE boxes will have the same appeal as the minimal simple-style ones that are still sought after in the used market.

the new one has a fan.  34401 is fanless.  I hate fans in my HOME lab.  at work, maybe I would not care.  but it is a failure point and I'd prefer to avoid noise and failure points.

I have not used scpi yet for the 34401 but I expect it to be high quality and low bug count.  I can't say the same about any new bit of hardware or software from - well - ANYONE, these days.  no one cares anymore (company) and its all about ship-it-now and not about quality.  for that reason, I'll take an older design over a new one, most times.

if its my money and not the company spending it, I'll err on the side of classic test gear instead of brand new units.

[MarkF]

Definitely avoid the Rigol DM3000 series. Effectively totally unsupported, no firmware updates, loads of bugs that will never be fixed. (Though I am saying this as a 5.5 digit DM3058E owner. Siglent SDM3055 is a much better choice for anyone after a 5.5 digit, at least they actively support it.)

What are some of the bugs you are referring to?  This video on the DM3068 6 1/2 model indicates that it is capable of some measurements that other meters in his lab can not do.

[TheSteve]

I would hope nobody would even consider the Rigol - it costs way too much for what it is compared to 34461A etc.

[Macbeth]

What are some of the bugs you are referring to?  This video on the DM3068 6 1/2 model indicates that it is capable of some measurements that other meters in his lab can not do.

I only have experience of the DM3058.

* Those big trigger buttons with flashing lights? For reading auto hold? a facility that is nearly verbatim in the Rigol manual with the 34401A manual - doesn't work, not at all. No matter how I try and explain it to Rigol support they claim to not understand it and report not a single person has asked them about it. They have me dancing like a monkey before they will report the issue back to China. Because I guess they know China HQ abandoned these meters years ago. I'd explained how it doesn't work like a blue arsed fly, then suggested I could show them it working on my Keithley 2000 (which is very similar to how it works on 34401A), but they wanted me to film a vid of it not working on the Rigol and working with a 34401A which I don't have. I gave up.

* Agilent 34401A command mode - near useless on the 3058 but apparently does work on the 3068. 3058 returns corrupted strings. This can be sort of fixed by coding around the bug, but not if relying on third party libraries or 34401 code (which is what this command mode is for after all). This is a totally obvious and should be extremely easy to fix bug. I reported it over a year ago I think. Nothing has ever been done about it. Compare with Siglent who are listening to customers and releasing new firmwares regularly.

* The much advertised amazing user defined sensor mode, complete with cold junction compensation. Great for thermocouples! Only on the 3058 there is no way of reading these sensors over SCPI and so utterly useless. (I have since found a very convoluted way to do this with unpublished SCPI calls, but it also involves switching between the Rigol command and fecked up Agilent modes and is very poor). I think the 3068 may work because it has working Agilent mode and Agilent extended mode to cope with sensors. I guess Rigol just couldn't be arsed with Agilent extended mode on the 3058. I mean they can't be arsed getting the basic 34401 mode working so why bother?

* There is a piece of software called UltraSensor that apparently does allow you to mess with the sensor functions (at least to upload them over USB) but it uses low level undocumented data packets over USB-TMC, not SCPI. I'm not sure I should get into this but it is VERY EASY to brick your meter completely using this official Rigol Software. I had a hell of a lot of back and forth with Rigol EU on this one and was treated like an idiot, until I finally got the support guy to follow my simple "how to brick a Rigol DM3058 in 3 easy steps" procedure and it happened to him. I refused to return my meter to them and managed to recover the firmware myself using JTAG. It turns out Rigol EU or USA don't actually have a clue or the facilities to repair meters anyway - even JTAG recovery - they just get China to send them replacements it seems. Maybe board swaps at the most.

I've started threads on these subjects on EEVblog previously, but nobody seems interested. Rigol appear to have the same opinion. I guess most people who have a clue just send the meters back and go and purchase a proper Keysight or something and never touch Rigol again with a bargepole and so Rigol are left with the suckers who take their shit. 

[MarkF]

Thanks for the info.  I was interested in a bench meter for home use and the Rigol DM3058E would be at the top of my price range and needs.

[HighVoltage]

If money is of significant influence, get a 34401A used in good condition and you know, you have a rock solid instrument that also can be repaired, if needed. Or if you have the funds, get a Keysight 34461A, you will not regret it down the road.

[linux-works]

given the bug reports I am seeing, there is NO WAY I'd touch a china based meter if I needed any SOFTWARE (scpi) to be working.  the front panel may work (generally) but the eastern firms simply don't care much about software and as soon as they have your money, they don't care about you.

I would not say the US firms (few that are still around) care that much more, but at least there's more attention to detail at first ship.

this is, again, why I prefer things that are 20 years old or more.  back then, it wasn't nearly such a low-quality software (embedded) world.  companies seemed to care a bit more and not have the 'ship it fast, now!' mentality that entirely rules the industry, today.

really such a shame that there is no long term support from ANYTHING in china.  wonder if it will ever self-correct?  'made in japan' used to be a joke in the 60's and 70's and then it turned around 180 so that now, MIJ is a sign of top quality - so much so that no one can afford to have anything MADE in japan anymore.  but for a while, things were great and we enjoyed the attention to detail from japanese goods.

I'm not at all convinced china will ever fix themselves.  they show no signs of wanting to and the world is not really in a position to force them.  what is the world going to do - take their business elsewhere?  like, WHERE?  china owns us all, now, more or less ;(

anyway, I ignore china for all serious things, with very few exceptions.  sure, I have a rigol 4ch scope but I'm not using for real business.  if I needed mission-critical test gear, I'd never even consider rigol and its kind.

[diyaudio]

thanks, lots of valuable tips really appreciate the feedback guys!

question
why do people still use the old GPIB standard in new meters,why not make the switch? is it for legacy purpose? 

[diyaudio]

The 34401A is an ancient design. The 34461A is what you want.

[linux-works]

what's wrong with gpib?  if you have a test rig already working, why change?  its good to keep old protocols working, as long as they earn their keep.

if I can pick a faster, modern interconnect and protocol, great; but for times when I can't, I appreciate having my test gear speak both old and new forms of data i/o.  if there are multiple ways in, I consider that a huge plus.

[Kleinstein]

GPIB is rather outdated, but it's still present for backward compatibility. It just starts that lower cost instuments (e.g. keysight 34460) skip it for cost reasons.

Though USB has many features that could make it a good replacement for GPIB, but it also has issues (noise, real time requirements). So GPIB is not fully dead yet though the computers that had GPIB bulid in are hardly in use any more.
Unless you need it, I would not miss the GPIB interface, if something like USB or ethernet is there.

[diyaudio]

Thanks for the info.  I was interested in a bench meter for home use and the Rigol DM3058E would be at the top of my price range and needs.

if based on price of (http://www.tequipment.net/RigolDM3058.html) USD700 price tag. i would buy used gear 6.5 digit or maybe even 2x used 5.5 digit. on the flip side, maybe even 2 brymen.

sometimes you get lucky, and sometimes you buy new stuff like a Rigol and get the short end of the stick, in this situation ive been through all except the buying of second hand instruments, no way would I risk it again.


Fool me once, shame on you; fool me twice, shame on me 

[acbern]

Something that has always made me mad is that both the K2000 and the 34401 do not support PT100 measurements. Often need this, and its just SW anyway. Has eliminted those two for any future DMM purchases for me.

[bson]

I can't believe Rigol hasn't fixed that 34401A mode yet.  I mean, it's a selling point on their marketing collateral even - and it doesn't work!   

And how can they even ship something so non-functional?  We're not talking oddball corner cases - it flat out uses a completely different format for readings than Agilent!  Did the engineer who added this ever even try to use it?! 

[Macbeth]

Like I said, at least they can hide behind that one a bit as it needs someone to use the intricacies of SCPI.

But even the front panel doesn't work.

The biggest most highlighted buttons are the Trigger Run/Hold and Single buttons. Press Hold and it just starts flashing at you. It does actually do all the stuff required for a reading hold, like you can give it a percentage window for trigger and it switches to low impedance mode so when you remove the leads it will not false read. Problem is it then doesn't ACTUALLY HOLD ANYTHING! Even worse, there may be some scenario where you want it to only hold readings without removing leads. So as not to fill up it's memory buffer, say. Well it will do that - hold the reading on the display while leads connected, but it carries on triggering anyway and filling up its memory. Utterly useless.

As for bricking the meter in 3 easy steps.... 

ETA: I've actually come to the conclusion that it would be easier to reverse engineer the damn thing and fix it myself. After getting intimate with its Blackfin and flash memory using JTAG in order to fix my own meter I am halfway there 

[linux-works]

again, china companies 'think differently' than many of us, do.

what is ok for them to ship is quite far from what we would do if it were our company.

its the 'sell and run' mentality, as I call it.  once they have your money, repeat sales mean nothing to them.  there will be other suckers to buy their gear even if its just a single sale.  this is how they think and operate!  and its why they are the last region who will get my money if there are other options available to me, such as used HP or tek gear.

[bson]

The auto hold on the 34465A is glorious!  Yesterday I realized I had populated a bunch of 2.7nF caps with 2.2nF ones and I had printed the BOM so small to fit on one page I didn't immediately notice.  Doh.    (2.2nF were used elsewhere in a nearby I/V converter; hence the confusion.)  But to identity the offenders (0805 size) I used a pair of Pomona tweezer probes under the microscope to read them, while the caps were setting in solder paste ready to reflow.  One beep for each reading without taking my eyes off the microscope.  Then look up and check the table to determine which parts were wrong.  Which turned out to be ALL the 2.7nF ones were 2.2nF.  But then I didn't actually have any 2.7nF caps, at all, so left these in until I get 100 pc from Mouser.    (These are for a couple of Sallen-Key lowpass filters on a DAC, so the values actually do matter.) 

But just taking a bunch of readings and have them appear on-screen in a table is really sweet. 

Apart from price concerns I'd pick the 34461A over a 34401A any day.