-
EEVblog 1382 - Keysight EDU34450A 5.5 digit Bench Multimeter TEARDOWN
Posted by EEVblog on 23 Mar, 2021 01:24 -
Teardown of the new Keysight EDU34450A 5.5 digit Bench Multimeter
-
why are the guard traces split?
-
why are the guard traces split?
Good question!
I'm not sure if I like this instrument.
PCB is not enig, just bare copper. Ref could be way better - 7ppm/C. Duhh.
Terminals are made from what? And terminals are obstructing my hand when I press buttons (rant of all instruments).
And for 25% more you can get a Keithley with way better specs, functions etc.
-
Maybe some of the traces are on inner-layers for protection from moisture or condensation?
The comb over comb looking thing.... is it for condensation detection? -
JH brand electrolytic capacitors are produced by a famous Chinese capacitor manufacturer Jianghai.
Jianghai and Hitachi jointly invested and established a company to produce electrolytic capacitors.
http://117.83.178.187/file/upload/2020/03/26/CD263.pdf
-
why are the guard traces split?
Not sure. They go to dedicated pins. -
Maybe some of the traces are on inner-layers for protection from moisture or condensation?
The comb over comb looking thing.... is it for condensation detection?
Exactly! I'm pretty sure it is, he should've tried shorting it out after powering it on just to see what it does.
-
From 6.36 on the video I noticed what looked like to be rust on the edges of the chassis .
Also Keysight put the model number on the screen just below the model number on the case. (20.00 on the video)
I will stick to my 34465A -
From 6.36 on the video I noticed what looked like to be rust on the edges of the chassis .
So they copied the Sigilent typical rust
Some rust is fast to come up with water based detergents used to clean the metal work.
There is plenty of spacing with the input protection - more than to expect from a Chinese handheld DMM with CAT 4 specs. Interesting to have the spark gaps to PE and not so much between the terminals. After all there essentially 4 terminals (the current input is kind of connected to COM) - so the ohms sense inputs seem to have no such protection (at least not so visible, may still have a PCB sparc gap) - just the long chain of resistors.
The custom chip nearly looks like a special DMM chip like found in handheld DMMs.
With a price point not that much lower than the 34460, it is a bit disappointing in some respects. At least the trend chart would be really nice to make use of the large screen. Also the acquisition mode with a max. speed of 1 SPS is rather slow - it looks like the meter should be capable of more (at least some 30/25 SPS for 1 PLC).
The Clicking from auto-ranging is a bit annoying, especially as it is quite fast. One can see a similar effect with other meters. AFAIR the HP3478 has a similar issue, though there it was not just mains hums, but the input bias slowly charging the input capacitance until the 10 M divider is engaged to discharge the cap. At least the clicking was slower. It may be hard to avoid this effect all together - the hum related part should be relatively easy to detect. -
This meter really looks way to crippled to justify that price point... Sure, they don't want to loose customers of the big toys to this cheaper alternative, but it just makes a pretty uninteresting product. If you can afford the little gap to the DMM6500 you definitly have a loooot more value for your money ;/
With that autoranging with open probes: If you go to lunch and accidentally leave the meter on the relays will be done before you get to dessert... -
Seems awful/pointless on many levels.
* Couldn't even hear the continuity buzzer on the video. Had to turn the volume up to max.
* Weird button labels: eg. "DCI" and "ACI" for current? Who calls it that?
Is it so that it matches Ohms law's V=IR (and if so, why isn't the Ohms button labelled "R"?)
* Is it really necessary to put "
2W" on the Ohms button instead of just ""?
(Took me ten seconds to figure out it wasn't "Watts" - "2 Watts or 4 Watts?" WTF? )
Would love to see it side by side with the more sensibly labelled $120 OWON in the other thread.
-
My crystal ball thinks that there will be a promotion soon: buy T&M for more than US$ 3.5k and get an EDU34450A for free.
-
Could that "touch sensor" be something to measure the board conductivity? Not sure why this would be as it is only 300 V CAT II, unless it is a hold over from another product.
-
I was looking for RMS converter chip but couldn't find any on the teardown pictures.
Then checked AC voltage measurement specification and I they seems bit low to me for this price range.
I know that all multimeters have wider tolerances on AC a ohmic measurement but this one seems to wide when spec-ed as 5 1/2 digits.
Overall impression is that this toy has features of 150$ handheld multimeter and rest of money are just to make it bigger so no one will steal it from classroom.
-
...rest of money are just to make it bigger so no one will steal it from classroom.
Also to impress the Dean when he visits the lab to see where all the money went.
-
That coil-y part could be connected to some testpoints to verify board plating thickness.
The more I hear about this meter, the more disappointed I am about it. I was considering it as a work-from-home multimeter. But no, sorry. They made all the efforts to make it unusable as an engineering tool.
That voltage reference doesn't really give me the confidence. With 8ppm tempco? At least it is easy to replace it with something more accurate.
The temperature measurement is useless. It can measure temperature, but not a PT100. Or a thermocouple. Only one type of NTC.
The analog frontend seems a lot noisier than the 6.5 digit meters. Pun intended, switching relays is not acceptable.
No trend chart. 5000 point memory. But surely, if it only takes 1 measurement you can just log, to the USB, right? Right? No.
NPLC cannot be set. I guess it would confuse students? Maybe the teacher can explain what NPLC is so they learn something.
The specification is all over the place. Basic DCV, the most accurate range is the 1V one. Yes, not the 10V one. I guess it has to do with the 1.25V reference voltage.
Tempco is 0.0015 + 0.0008 (0.0008 is 8 ppm. the reference voltage is 8 ppm). For resistance the tempco is 0.0080 + 0.0005. So they use a more precise reference than the voltage reference.
I'd like to know what happened here. Did they run out of the flash memory of the microcontroller? Did a marketing guy bring a pen, and cross out every second feature from the 34465A?
Seriously, I think this DMM could be vastly improved by spending some extra effort on the AFE, and the firmware. -
Why tell you the model number on the case and again of the screen? Kind of stupid.
-
With that autoranging with open probes: If you go to lunch and accidentally leave the meter on the relays will be done before you get to dessert...
It is not hat bad: the switching is usually not that fast and at some 1 switching per second the meter could last a few months, though maybe not the full warranty period.
With a meter of this grade it is not needed to keep it running 24/7 and if one is in the room one would normally notice.
The Owon meter is in a different league. It lower resolution, has less PC interface and no 4 wire Ohms. It still has a very nice look and may for many uses be a alternative, though not a direct competition.
A more direct competition is the Sigilent SDM3055 (with a reasonable low bug firmware ).
At least the EDU edition is more competitive priced than the old 34450 version.
For the RMS ranges, the RMS converter could be inside the DMM chip - at least I don't see much extra amplification part to use the µC internal ADC, which would be another low cost alternative.
No surprise that the 1 V range is the most accurate one - this range can go directly to the ADC. Chances are the ADC input range is some +- 1-2 V. The 10 V range would already go through the input divider. A high impedance 10 V range would take extra effort and this is not found in the 5.5 digit meter class. -
The exposed copper finger thing has 495 sq marked next to it, so it's probably a PCB capacitor used to calibrate out trace parasitics.
Humidty compensation ? Seems a stretch on something in this price range though -
I would not expet compensation for humidity effects - not even found at 8 digits.The exposed copper finger thing has 495 sq marked next to it, so it's probably a PCB capacitor used to calibrate out trace parasitics.
Humidty compensation ? Seems a stretch on something in this price range though
It may be a warning from too high humudity, which may be a safty hazzard - so maybe a carry over form an handheld CAT 4 design.
It could also be just a test structure to check board cleaning - though from Daves picture they could do better, especially around the shunts (does not really matter there) and more important the high voltage divider. -
This meter really looks way to crippled to justify that price point...
Unfortunately I'm beginning to believe this. I was really hoping for a good general purpose bench meter, not something half way between a high end hand held and a systems meter. The reality is it is hard to find a good bench meter for the repair bench that doesn't cost an arm and a leg.
In this sense a bench meter has to have a large digit display as the number one feature. The a reasonable range of features and performance. Honeslty I would have preferred one less digit and a $400 price tag.QuoteSure, they don't want to loose customers of the big toys to this cheaper alternative, but it just makes a pretty uninteresting product. If you can afford the little gap to the DMM6500 you definitly have a loooot more value for your money ;/
More like a worth less product. I mean honestly how many educational institutions will be buying this meter, especially on tight budgets. HPAK had two choices here, make a ≤ $400 meter or fully enable this meter as a replacement for an existing model. The high end can easily be managed by offering better specs and system meter features.
What bothers me here is that HPAK apparently already has the software done for many of the missing features. I hate to say this but they are effectively handing a very large market to the competition.QuoteWith that autoranging with open probes: If you go to lunch and accidentally leave the meter on the relays will be done before you get to dessert...
Obviously a partially tested meter.
Part of my disappointment comes from the reality that i really like a few of the concepts in the meter. The large display with big digits is very desirable. So far the actually panel lay out looks OK. But then you have a continuity function that you can't hear.
The B&K 2831E is a better meter at less than $400 for the educational market. Frankly it is a better meter for general use on most repair benches. for people that need the resolution and accuracy you would be going upscale by a significant amount so this EDU meter wouldn't come into consideration anyways.
So agree 100% big disappointment.
-
It is interesting that the similarly shaped OWON instrument is not listed on OWON's web site anymore. Hopefully that doesn't mean this HPAK meter will go down the same route. OWON has more conventionally shaped meters listed such as the XDM3041 which still are listed for far less than this so called EDU meter.
By the way I'm not one to expect HPAK to compete price wise with a Chinese company that is new (somewhat) to the market. I just don't see this meter as being remotely competitive with a lot of other established producers of test equipment. I'm just not seeing the value here.
As I mentioned elsewhere I would really like to see HPAK take a different approach to upgrading their EDU hardware. It is one thing to release early but when you have a habit of not adding to the initial release it just turns what is a poor value, into a extremely poor value over time. It is reasonable to think that EDU's will buy in bulk so something like the OWON XDM3041 is likely to end up costing far less than $400, which would result in buying power of twice as many instruments as the HPAK offering. Maybe that is what HPAK wants is the opportunity to offer bulk purchases at huge discounts. You know the ability to convince a buyer that they knocked off $300 is pretty delicious. -
My guess is that their business plan is to sell you the Keysight name with as little actual product as possible included in the deal. Sort of like promotional t-shirts. Anyone think the BOM on that thing exceeds $100? How low could you go in large quantities?
-
Another question is do student labs need 5.5 digit multimeter? We used venerable Mastech M830 in our physics labs, and it worked a charm. Typical labs don't need anything better. I'm not advocating for M830, of course, but if you are spending more than $100, you are probably just wasting money.
I can see there may be a need for improved performance in a research lab, but why bother with crippled EDU versions then? Just buy real equipment. -
The EDU meter is quite a bit (some $200) cheaper than the original 34450. In most aspects it is similar and the larger screen depends on the use. So it is not such a bad offer - still some extra for the KS logo and maybe some SW. Compared to the Sigilent SDM3055 it is still quite a bit more expensive, but at least there is a chance to get at least the main SW bugs fixed.
Hoping to see a trend chart function may be a bit too much - but the hardware should be sufficient.
At least it looks like still an original Agilent design and not some rebranded Owon, Uni-T or similar.
For some reasone the OWON XDM3051 is also quite expensive - so there seem to some magic at the 5.5 digit level (or some premium paid by the market).
Most student labs don't need 5.5 digts. It is more that one may want 1 µV resolution. From the logic there should not be that much extra costs for the 5 th digit unless it comes with extra accuracy.
So a 5 digit meter makes sense, but the performance of the 34450 / XDM3051 or SDM3055 is still a bit disapointing. They could as well habe 4 digits and a 10 mV range. I remember using an old nixi tube meter with 3.5 digits but a 2 mV range - would still be good enough in many cases.