Author Topic: The Siglent SDG2042X Thread (Read 566647 times)

Blue · « **Reply #1350 on:** March 31, 2021, 06:55:41 am »

Hi Tautech,

Due to the location of the components and their shielding against airflow, there is hardly any airflow over these components to make their temperature go down substantially.
Their surface area is extremely small (no cooling block on them).

Why don't you switch one SDG2000X on permanently for a minimum of 3 months to test for any degradation?

rf-loop · « **Reply #1351 on:** March 31, 2021, 08:03:07 am »

Quote from: Blue on March 31, 2021, 06:55:41 am

Hi Tautech,

Due to the location of the components and their shielding against airflow, there is hardly any airflow over these components to make their temperature go down substantially.
Their surface area is extremely small (no cooling block on them).

Why don't you switch one SDG2000X on permanently for a minimum of 3 months to test for any degradation?

You can switch it on after it arrived and keep 3 years continuously on, only care that room temp stay uder 30 ºC <90% humidity and up to 40 ºC < 50% humidity, naturally with original fan. If it fails there is 3 year warranty.

Who can guess how much Siglent have warranty cases if it fails due to thermal design. It is designed in south China where is quite hot (and humid), subtropic climate. This model is sold years around of world and this model have been quite popular.
If there is really some thermal design problem what leads to extra failures we have heard bout it and lot.

Blue · « **Reply #1352 on:** March 31, 2021, 08:57:06 am »

@rf-loop

Perhaps you are right - time will tell.
But I think most hobbyists are buying this device and do not have them on 24/7. So there will be hardly any reports. And the amount of instruments being sold is I quess in my country of 17M inhabitants around 2-3 per month. Weibull analysis is due to the low amount of instruments in the field questionable.

Therefore I ask others - who have this instrument - to check if they also have these components at 170 Celcius. Do you have this instrument rf-loop?

And if you are a enigneer, you will be aware that most components have rating of 1000h @85C. In other words you have to design in such a way that components do not fail due to thermal issues.

And in respect to the ambient temperature, there is hardly any difference between (170-20) = 150C or (170-30) = 140C. Besides, in China they use a lot of airco's removing water vapour from the atmosphere and reducing the ambient temps to 'normal' levels.

exe · « **Reply #1353 on:** March 31, 2021, 10:40:29 am »

Quote from: Blue on March 31, 2021, 08:57:06 am

And in respect to the ambient temperature, there is hardly any difference between (170-20) = 150C or (170-30) = 140C.

Idk, for electrolytics every 10C shortens life twice. Also, I'm not sure about the pcb. But I do know that my board house (jlcpcb) offers several pcb materials with different glass transition temperature. What is the max temp for "normal" fr-4?

Anyway, I agree I didn't hear much about failures of this model.

2N3055 · « **Reply #1354 on:** March 31, 2021, 11:47:54 am »

Quote from: exe on March 31, 2021, 10:40:29 am

Quote from: Blue on March 31, 2021, 08:57:06 am
And in respect to the ambient temperature, there is hardly any difference between (170-20) = 150C or (170-30) = 140C.

Idk, for electrolytics every 10C shortens life twice. Also, I'm not sure about the pcb. But I do know that my board house (jlcpcb) offers several pcb materials with different glass transition temperature. What is the max temp for "normal" fr-4?

Anyway, I agree I didn't hear much about failures of this model.

Glass transition temp is not important. Like Thomas rightly said, temps over 120-130 °C will contribute to epoxy oxidation and cause yellowing. Also there is a greater chance of delamination of copper from FR4. I personally don't like SMD power resistors for that reason, a nice TH power resistor mounted off the board is better choice for long term.
Although, there is no need for such high dissipation. If it's a bleeder circuit, it is drawing too much power for nothing.
If it's a standby, same, or should be replaced with physical power button. I personally like physical power switches, and am willing to trade for few more seconds of boot time...
Most of the time, it is not even that device is in some kind of standby, just someone thought that soft power switch is cool because it enables stupid blinking power switch. Of course, blue light and too damn strong one...

Qw3rtzuiop · « **Reply #1355 on:** March 31, 2021, 01:03:39 pm »

I have an unmodified unit. I measured 105° at the 20k Resistors with a cheap IR Thermometer. Considering that i wasnt able to get the thermometer at the correct distance the actual value is higher because i also measured parts of the colder pcb.

My unit is from mid 2017. There are no discolorations so far.

JohnG · « **Reply #1356 on:** March 31, 2021, 01:57:20 pm »

Quote from: Blue on March 31, 2021, 06:37:20 am

No difference measured with or without plastic cover (Its not possible to slide the metal cover back without slicing the thermocouple)
Hardly any airflow over the components due to their location and shielding.

I've worked on a number of projects where thermal management using airflow was an essential part of the design. Removal of a cover has a dramatic affect on airflow in every case I have ever seen. The cover forces the air to go where it normally would not. Under such circumstances, the air will usually form eddies that can cool areas without direct airflow. Can you see the airflow, or have otherwise measured it? That's what thermal engineers do, with a smoke generators. If they work for a rich company, they use a lidar to measure particle velocity, even. If you don't measure it with the cover on, it is a pure guess as to what is actually going on.

If you have changed the fan or otherwise lowered the airflow, stuff will get hotter. If the parts get that hot with the original fan and the cover on, you have a case for a design flaw. Otherwise, you have no case.

For some points of reference, a lot of hi-rel stuff limits temps to 100-105C for most parts that are rated to operate at 150C. It's usually the semiconductors and the resistors that can get this hot. Consumer stuff will often run around 110-120C worst case, i.e. wall-warts and such.

John

Cheers,
John

TurboTom · « **Reply #1357 on:** March 31, 2021, 06:15:19 pm »

So these resistors are probably built-in smoke generators to identify the eddies...

Blue · « **Reply #1358 on:** March 31, 2021, 07:07:23 pm »

Hahaha,

More seriously, do you know which notified body accepted the CE mark?

2N3055 · « **Reply #1359 on:** March 31, 2021, 07:24:11 pm »

Quote from: TurboTom on March 31, 2021, 06:15:19 pm

So these resistors are probably built-in smoke generators to identify the eddies...

LOL!!

ebastler · « **Reply #1360 on:** March 31, 2021, 07:29:37 pm »

Quote from: Blue on March 31, 2021, 07:07:23 pm

More seriously, do you know which notified body accepted the CE mark?

Why would a notified body have to accept a CE mark?

Blue · « **Reply #1361 on:** March 31, 2021, 07:38:59 pm »

have a look here:
https://wpo-altertechnology.com/notified-body-for-ce-marking/
https://en.wikipedia.org/wiki/CE_marking

nctnico · « **Reply #1362 on:** March 31, 2021, 07:43:47 pm »

Quote from: Blue on March 31, 2021, 07:07:23 pm

More seriously, do you know which notified body accepted the CE mark?

Doesn't matter. Has been debunked a long time ago.

Blue · « **Reply #1363 on:** March 31, 2021, 07:46:56 pm »

Quote from: nctnico on March 31, 2021, 07:43:47 pm

Doesn't matter. Has been debunked a long time ago.

I do not understand, can you elaborate....

ebastler · « **Reply #1364 on:** March 31, 2021, 08:07:02 pm »

Quote from: Blue on March 31, 2021, 07:38:59 pm

have a look here:
https://wpo-altertechnology.com/notified-body-for-ce-marking/
https://en.wikipedia.org/wiki/CE_marking

So, in your opinion, which of the special product categories which require the involvement of a notified body does the SDG2042X fall into? Implantable medical devices maybe?

https://ec.europa.eu/growth/single-market/ce-marking/manufacturers_en

Blue · « **Reply #1365 on:** March 31, 2021, 08:17:37 pm »

Measuring Instruments

I'll figure it out.
Power resistors in the powersupply @ 170C is a fire harzard. Siglent put a CE mark on the instrument.
Something is gonna be smelly....

tautech · « **Reply #1366 on:** March 31, 2021, 08:25:15 pm »

Quote from: Blue on March 31, 2021, 08:17:37 pm

Measuring Instruments

I'll figure it out.
Power resistors in the powersupply @ 170C is a fire harzard. Siglent put a CE mark on the instrument.
Something is gonna be smelly....

A data point a few posts earlier:

Quote from: Qw3rtzuiop on March 31, 2021, 01:03:39 pm

I have an unmodified unit. I measured 105° at the 20k Resistors with a cheap IR Thermometer. Considering that i wasnt able to get the thermometer at the correct distance the actual value is higher because i also measured parts of the colder pcb.

My unit is from mid 2017. There are no discolorations so far.

Maybe you should have your IR camera checked for accuracy ?

Blue · « **Reply #1367 on:** March 31, 2021, 08:30:26 pm »

Upto now, there have been 3 measurements of reporting of extremely hot resistors. By using a thermal camera, thermo couple and IR thermometer.
Humm, seems to be a systematic issue. No need to question these results.

If you have one of these instruments, please check the Siglent with it and report back here. Many thanks.

tautech · « **Reply #1368 on:** March 31, 2021, 08:39:28 pm »

Quote from: Blue on March 31, 2021, 08:30:26 pm

Upto now, there have been 3 measurements of reporting of extremely hot resistors. By using a thermal camera, thermo couple and IR thermometer.
Humm, seems to be a systematic issue. No need to question these results.

If you have one of these instruments, please check the Siglent with it and report back here. Many thanks.

I beg to differ....a 70C variance in reported results is not insignificant coupled with rf-loop's and my knowledge of zero SDG2kX PSU failures.

If there was a problem with these units he and I would know about it.....I'll leave the exercise for you to establish who rf-loop really is and why he might have considerable knowledge of SDG AWG's.

Blue · « **Reply #1369 on:** March 31, 2021, 09:16:07 pm »

Perhaps you or rf-loop could post some flir images showing temperatures that are not regarded as a fire hazard? ie way less than 85C.
I have no clue who rf-loop really is apart from that he writes in proper English and has a Chinese flag in his posts. Maybe the CEO of Siglent?

I really like the instrument, it has lots of capabilities. The fan just made a hell of a noise and therefore I opened it and found out that something is not good.

Great respect will be shown to those who are able to make the instrument better. Perilous defending or nullifying the flaws could backfire.

JohnG · « **Reply #1370 on:** March 31, 2021, 10:19:03 pm »

Quote from: Blue on March 31, 2021, 09:16:07 pm

Great respect will be shown to those who are able to make the instrument better. Perilous defending or nullifying the flaws could backfire.

Even if there was a real problem, you have not shown that it occurs under normal operating conditions on an unmodified unit. Having the cover off counts as modified. Changing the fan counts as modified. Note that I am not defending anything. I am only pointing out improper technique. If you do the measurement properly on an original system, and you still measure a temperature that is excessive either safety-wise, exceeding component ratings, or from a reliability analysis (not opinion about what's too hot), you have a solid case.

To do this, you can feed the thermocouple or other temperature sensor through the vent holes in the side of the case, leaving enough excess so that you can attach the tip of the thermocouple to the point of interest, and still have enough slack to put the cover back on. Before putting the cover back on, replace the original fan. Then do the measurement. Be careful, because the thermocouple may be at or near line voltage.

You may go to any knowledgeable certification body that you wish. If you take the temperature measurement with the case off, it will be discounted. If you have modified the device in any way, you will be discounted.

But don't take my word for it. Try it yourself!

As an aside, an 85C temperature hardly counts as a fire hazard. Some electronics have parts that routinely operate at 120C or higher. Go look inside some old CFLs if you want to see scary charred PCBs with crystallized solder joints, still operating. Heck, I worked on an electronic cooktop controller in a prior job that had to work in an ambient environment of 85C because some customers would install it in a cabinet right over the oven. The coolest parts were 85C. This was UL certified, and UL is strict and they don't take your word for it.

Cheers,
John

Blue · « **Reply #1371 on:** March 31, 2021, 10:35:39 pm »

Dear John,

You are absolutely right.

I did measure it with the thermocouple in place with a plastic sheet over the box and without plastic sheet. I still got a temperature above 120C in both cases. Unfortunately, the thermocouple was not affixed perfectly to the resistor - there was a small airgap between the resistor and the thermocouple.
This was due to the lack of stickeyness of the kapton tape (old tape). Therefore the reported temperature was too low! I deem the flir readings of 170C more accurate. Perhaps the closed lid will drop the real temperature a bit (5-10C?), but the that's impossible to measure with a flir camera since it does not look through metal. Because the resistors have a very small surface area, most of the energy will be radiated instead of being removed by air convection.

No difference in temperature was observed with or without this plastic cover with the thermocouple. That was expected since the airflow is not directed over these resistors. If you have one of these units, you can confirm it for yourself. Or just study the video by Dave. He also made a remark about the airflow.

TurboTom · « **Reply #1372 on:** April 01, 2021, 12:06:50 am »

Took some thermographs on my SDG6000X and could to some extent verify Blue's findings (~150°C at the resistors with an open casing). Yet, I doubt that it's so severe when the machine is closed. These temperatures would cause some discoloration of the PCB in relatively short time, and there's just nothing like that to be found in my instrument.

But I noticed another peculiarity: there's a hot MLCC in my PSU! I guess that's a candidate for a replacement really soon... See attached report for more information.

Edit: Some more testing on that issue over in the SDG6000X thread.

nez · « **Reply #1373 on:** April 01, 2021, 12:20:19 am »

I think that even with the proper setup as JohnG described, taping a thermocouple on would affect the temperatures negatively as well. It would likely directly block the airflow on the area of the component hidden under the tape like a miniature tent (or like wearing a wind-breaker outside in a breeze), degrading cooling ability.

This way of testing also assumes that the part of the tape that's flush on the component (rather than the 'tent' portion) has perfect thermal conductivity and doesn't insulate the part. There is a lot of variation between heatsink thermal pastes' qualities, for example, that makes a difference with cooling hot parts.

Maybe I'm wrong.

That said, I don't know a great way to get a true reading on the temp without a professional setup that avoids these pitfalls (calibrated thermal camera built into the side/top of a custom case?).

At a different level, important aspects to analyze are the airflow and static pressure specs of the stock fan vs. replacement fan (Edit: Assuming case completely closed, of course, since any additional gaps between panels would affect airflow).

Quote from: Blue on March 29, 2021, 06:01:13 am

I installed the Noctua NF-A6x25 FLX and it has two working fan settings. I used the lowest one and it works great!

I very much share your desire for low noise! However, we must still take care to compare the noise-lowering solution to the original design specs for cooling ability.

https://noctua.at/en/nf-a6x25-flx/specification

Using the medium or low Noctua fan speeds using one of the adapters, the airflow and static pressure might be too low compared to the stock fan.

My understanding is that in tight spaces with lots of air resistance, static pressure can be particularly important.

@Blue
Can you share a pic of the stock fan to see its model number?

adamgreig · « **Reply #1374 on:** April 01, 2021, 12:35:58 am »

For another data point, I measured my SDG2042X. With the lid open I get the same thermal camera readings as everyone else, around 150C on that resistor. I also note very minimal discoloration around it - some oxidation on the solder joint compared to other components but no discoloration to the resistor or PCB despite regular (albeit not continuous) operation since 2017.

I then attached a thermocouple to the resistor with a bunch of kapton tape and took measurements in the lid on and lid off position, in both cases waiting about 10min by which time readings had equilibrated. As has already been commented it's very tricky to get a good contact between the thermocouple and the resistor, and in the lid-off position I read around 98C instead of the 150C reported by the IR camera. Could be a mix of IR calibration and poor thermocouple contact. In any event I was able to close the lid without disturbing the thermocouple and the temperature dropped to and remained at around 78C over the next five minutes or so.

So, while it's definitely running hot, and I can't say exactly how hot, it seems to me that closing the lid does have a significant impact on the temperature.

Quote

But I noticed another peculiarity: there's a hot MLCC in my PSU! I guess that's a candidate for a replacement really soon...

I didn't measure it specifically but didn't notice this MLCC being hot on my unit.


EEVblog Main Site	EEVblog on Youtube	EEVblog on Twitter	EEVblog on Facebook	EEVblog on Odysee

Author Topic: The Siglent SDG2042X Thread (Read 566647 times)

Share me