EEVblog #829 - Siglent SDM3055 Bench Multimeter Teardown

[EEVblog]

Inside the Siglent SDM3055 5.5 digit bench multimeter
http://amzn.to/1lBAD7b

[Muttley Snickers]

Apparently the expansion port was intended to be used for a scanner option for the SDM-3055S model which they did away with, the SDM-3055A GPIB model is supplied with a USB/ GPIB dongle and not a rear port which was shown in some early pictures.

https://www.eevblog.com/forum/testgear/siglent-new-bench-dmm-sdm3055/msg666922/#msg666922

[plexus]

So wait a minute. on one hand you say that its cheaper than the Agilent but it's not as well made as the Agilent. what do you want? you want these things to be built as well as higher priced models but you want them to be the same price? care to provide solutions to how this can be done? I can understand if you explain how they could increase the quality for the same price but you don't do that. you just say they aren't built as well as more expensive units as if they can do better. how? I don't understand the complaining on cheaper quality on a cheaper instrument but if you can tell us how they could provide higher quality for the same price that would be great.

[TheSteve]

So wait a minute. on one hand you say that its cheaper than the Agilent but it's not as well made as the Agilent. what do you want? you want these things to be built as well as higher priced models but you want them to be the same price? care to provide solutions to how this can be done? I can understand if you explain how they could increase the quality for the same price but you don't do that. you just say they aren't built as well as more expensive units as if they can do better. how? I don't understand the complaining on cheaper quality on a cheaper instrument but if you can tell us how they could provide higher quality for the same price that would be great.

The engineering design cost of the meter in China is probably 10% of what it would be if it was designed in the US - so they darn well better be able to sell it for less, even if the quality was the same.

[Armxnian]

So wait a minute. on one hand you say that its cheaper than the Agilent but it's not as well made as the Agilent. what do you want? you want these things to be built as well as higher priced models but you want them to be the same price? care to provide solutions to how this can be done? I can understand if you explain how they could increase the quality for the same price but you don't do that. you just say they aren't built as well as more expensive units as if they can do better. how? I don't understand the complaining on cheaper quality on a cheaper instrument but if you can tell us how they could provide higher quality for the same price that would be great.

Build quality is not the only reason the Agilent costs more. It has better specs, operational firmware, better support, and the name Agilent on it.

[Hydrawerk]

You mean Keysight.

[Armxnian]

You mean Keysight.

No, I mean Agilent, because that is the branding Dave's unit has on it.

[rdl]

I wonder how old Dave's meter is? I thought they claimed to have fixed the rust problem, but I saw it as soon as the cover slid off. The power supply in the last show was new and it had rust also. They are at serious risk of becoming a laughingstock (if not already) unless they fix that. What happens when particles of rust start floating around inside a so-called precision meter?

[lukier]

It seems that the case and the digital mains earth referenced logic (ARM CPU) is shared with SDG2122X. Maybe the case will be used in something else, therefore two holes for fans. Nice design from SKU and software development point of view.

For the price the hardware seems to be OK, rusty case & crusty soldering, but decent parts, input protection, guard traces etc. SDG2122X was OK too. Both much better than the PSU you showed recently. I suppose this must have been a different design team

I wonder why discrete 24 bit ADC. Nice chip, but these chips can be quite expensive. Since they have Lattice CPLD on board and using ADG switches elsewhere, wouldn't it be cheaper to do multi-slope instead?

[SeanB]

Rust problem is easy to fix, just take all the metalwork after the punching process, dip in a pickling bath, rinse then electrogalvanise then etch again and powder coat. Even a short pickle and electrogalv operation after the painting on complete cases would put a protective film on the cut edges.

More expensive would be a tumbling operation in a large drum to get the sharp edges off first.

[EEVblog]

So wait a minute. on one hand you say that its cheaper than the Agilent but it's not as well made as the Agilent.

Correct, I said it's cheaper than the Agilent and isn't built as well as the Agilent. I used the term Apples and Oranges comparison.
You then go on to make a point as if what I've said is somehow contradictory?

what do you want? you want these things to be built as well as higher priced models but you want them to be the same price?

I neither said nor implied such a thing.

[EEVblog]

I wonder why discrete 24 bit ADC. Nice chip, but these chips can be quite expensive. Since they have Lattice CPLD on board and using ADG switches elsewhere, wouldn't it be cheaper to do multi-slope instead?

Maybe, but you have the dick around getting it right. The 24 bit ADC just works and has known parameters.
Keysight have spent decades refining their own dual slope technique.

[TheSteve]

It's another teardown that makes me want the product less instead of more...

[tautech]

Thanks Dave.
Siglent SDM3055 thread:
https://www.eevblog.com/forum/testgear/siglent-new-bench-dmm-sdm3055/

[mikeselectricstuff]

I think you were way too hard on the build quality.
Surface rust, presumably a by-product of laser or waterjet cutting, is a non-issue as long as there isn't powdery rust falling off anywhere.
Airflow path looked perfectly reasonable to get a more stable temperature - probably not essential but for a DMM, stable temperature can't hurt.
The holes spread out around the divider looked like they should give a reasonable airflow over most of the board.
I see no problem with spade connector onto a flap in the metal, as long as it's not rusty and thickness is OK, and it's arguably safer as screws can come loose.

As for caps - has anyone actually done any proper tests on lesser brands ? Just because you've never heard of them doesn't mean they're not perfectly adequate for the job.
And of course they're very easy to replace if they do eventually die, so not really a big deal, less so where they're not running in a hot environment.




 

[funkyant]

So wait a minute. on one hand you say that its cheaper than the Agilent but it's not as well made as the Agilent.

Correct, I said it's cheaper than the Agilent and isn't built as well as the Agilent. I used the term Apples and Oranges comparison.
You then go on to make a point as if what I've said is somehow contradictory?

what do you want? you want these things to be built as well as higher priced models but you want them to be the same price?

I neither said nor implied such a thing.

When they do this:



It's fair game to be comparing the two, regardless of price point.

If they don't want the comparison made, they should design an original front panel.

[Wytnucls]

Lelon has four factories. One in Taiwan, two in China, and one in Malaysia (joint-venture business with Elna Japan). Total capacity is 820 million pieces per month and the company has 3000 employees.
Joint venture with Elna points to an acceptable product quality, I would presume.

[SeanB]

The internal case is punched with a die, which leaves those sharp edges. They probably also reuse the same case parts in other equipment, so you will find other Siglent equipment in the same form factor case will have the same chassis construction, and with the same mounting holes and cut out portions, or a majority, you can use a multipart die with removable punches in part to do special cutouts, but it is often cheaper to punch them all and simply not use those you do not need in the particular model. That explains the fan cutout in the 2 sides, they are a mirror image, simply turning the sheet over to do the end bends  on a separate set of dies. You can see the draw marks on the edges of the metalwork.

Centre is done with a 3 step process, first punches out the holes and the openings, leaving the holes with shaped edges for the second stage of punching them down 90 degrees, then the third stage rolls it over to leave the smooth edge. they then either spot weld or form a rivet ( did not look too closely to see the method of joining the panels, but this is the most common method that does not need any extra material or parts) to hold the panels together.

Ideally they want to tumble the panels after punching in an abrasive pumice bed, to dull the sharp edges, then wash, pickle in acid, wash again then thin electrogalvanise to coat the edges exposed in the forming. Probably the pickle and electrogalv would be done on the finished parts ( less the bright coated inserts) to have a nice internal chassis with smooth edges for assembly, pretty much all the final assemblers would have cuts from those edges.

Wonder if you ran a damp swab over the edges, put it in a PCR reactor then did DNA analysis if you could trace who was the last person to bleed over the unit.

As to the Lelon capacitors, in this use they really are only mildly stressed, as regular electrolytics in linear supplied have used non low ESR for decades, so they should have a good life. On the SMPS side with high frequency current heating they might have a shorter life, but here the ESR can ride quite high before it will become a failure. You see many old power supplies running fine with capacitors with ESR in the range of 1-5R and only thing is increased ripple into a linear regulator.

[Lightages]

I thought that it is OK build for the price. Except for the "rust" I don't see much to complain about. I might buy one if I were in the market for a 5.5 benchtop, especially since they sped it up with a firmware update.

[Groucho2005]

Maybe it's really humid in Dave's lab... 

[kalleboo]

Maybe it's really humid in Dave's lab... 

Clearly Dave has been slobbering all over the equipment!

[bookaboo]

Has to be said Dave is certainly not ever going to be influenced to give positively biased reviews... even if the CEO flies in personally and leaves a bunch of swag. Fair play for pointing out every niggle.

One thing that obviously struck me was the similarity to the Agilent/Keysight and it's implications for the future. If Fluke can get a batch of meters stopped at customs and banned for literally "being yellow" how and why are Keysight letting a competitor (albeit a lower end competitor) get a foothold using such a blatant interface rip off?
Don't get me wrong, I'm all for competition and I shed no tears for large corporations but this unit is just asking for a copyright infringement lawsuit!

[Kleinstein]

Copyright does only applys to really 100% copy - it does not protect things like the look and feel of the software or the arangement of the keys. In some countries there may be a kind of "small" patend to protect a design  (e.g. Gebrauchsmusterschutz in Germany), but this is expensive and only works locally.  So even Samsung got around the round corners of the IPad.

It's true that the design is rather close, but this might help users to switch in one way or the other - often user start with a low cost and upgrade. So even Keysight might profit from the similarity even if they don't like it.

One thing that did not look good to me is the rather close spacing from the guard lines to the reistors strings made for high voltage rating. If the guard is so close there is not much sense to the long string and the cutouts.

With the fan I wonder if it is temperature controlled ? At moderate temperatures I don't think the fan is really needed. It may be at 40 C ambient. So I would consider a fan just for the summer a good idea.

[nctnico]

I think you were way too hard on the build quality.
Surface rust, presumably a by-product of laser or waterjet cutting, is a non-issue as long as there isn't powdery rust falling off anywhere.
Airflow path looked perfectly reasonable to get a more stable temperature - probably not essential but for a DMM, stable temperature can't hurt.
The holes spread out around the divider looked like they should give a reasonable airflow over most of the board.
I see no problem with spade connector onto a flap in the metal, as long as it's not rusty and thickness is OK, and it's arguably safer as screws can come loose.

I agree. You only need a little bit of airflow move a decent amount of heat.
Also the remark about the front/display cable running through the transformer section makes no sense IMHO. I would not want such a cable anywhere near a potentially sensitive analog board so I'd consider it a plus they went through the effort of putting that cable on the other side of the shield and wrapping it with an insulator. Same goes for the remarks about the holes for the fans. This clearly is a standard casing for Siglent which they use for many of their devices (look at the SDG1000 series for example). What is wrong with that? Only a complete fool designs a case from scratch for every product. HP has been using the same aluminium front, rear and frames for decades.

[jitter]

Lelon has four factories. One in Taiwan, two in China, and one in Malaysia (joint-venture business with Elna Japan). Total capacity is 820 million pieces per month and the company has 3000 employees.
Joint venture with Elna points to an acceptable product quality, I would presume.

I can tell you with 100% certainty that Lelon caps are being used in industrial procucts that are designed AND manufactured over here. But having said that, it's much more likely that one will find Nichicon, Rubycon, Panasonic, Nippon/United Chemicon and last but not least Vishay/BC (formerly Philips) in those kinds of products.

With respect to the remark on the rear terminals on Keysight bench meters: imagine you're doing my work, i.e. testing and calibrating measuring instruments. Then having both front and rear terminals is very useful. On the front I connect standard testleads with clips that connect to test points on pcbs. On the rear I connect a coax through an adapter that connects to the output BNCs of the instruments under test.
The procedure requires alternate use of clips and BNCs, and a push of a button is much quicker than swapping leads all the time. Granted, for "normal" use those rear terminals are probably not used much.