How to calibrate Siglent SPD1000X / SPD3303X / SPD3303X-E series power supplies?

[oz2cpu]

this is my first time using the record to usb stick feature :-)
I am sorry but it is a bit confusing,
I dont know if it is running or not, how to start or stop..
+2V -2V squarewave 0.1Hz

Mode,Time(s),Value
DCV,0.1224640012,-1.980109E+00
DCV,0.3224929869,-1.980099E+00
DCV,0.5225350261,-1.980221E+00
DCV,0.7225729823,-1.980194E+00
DCV,0.9232320189,-1.980166E+00
DCV,1.1231570244,-1.675675E+00
DCV,1.3231680393,4.330368E-01
DCV,1.5232470036,1.899562E+00
DCV,1.7232489586,1.979141E+00
DCV,1.9232670069,1.979064E+00
DCV,2.1233971119,1.979092E+00
DCV,2.3234310150,1.979073E+00
DCV,2.5233540535,1.979068E+00
DCV,2.7234768867,1.979066E+00
DCV,2.9235060215,1.979098E+00
DCV,3.1234118938,1.979249E+00
DCV,3.3235540390,1.979335E+00
DCV,3.5234611034,1.979267E+00
DCV,3.7234780788,1.979239E+00
DCV,3.9236209393,1.979196E+00
DCV,4.1235442162,1.979189E+00
DCV,4.3235640526,1.979174E+00
DCV,4.5237011909,1.979143E+00
DCV,4.7236280441,1.979153E+00
DCV,4.9236412048,1.979187E+00
DCV,5.1237788200,1.979096E+00
DCV,5.3237118721,1.979051E+00
DCV,5.5237450600,1.978988E+00
DCV,5.7238807678,1.978893E+00
DCV,5.9238147736,1.978840E+00
DCV,6.1238241196,1.857547E+00
DCV,6.5433020592,-1.980423E+00
DCV,7.4235730171,-1.980176E+00
DCV,7.6235952377,-1.980146E+00
DCV,7.8236031532,-1.980107E+00
DCV,8.0236463547,-1.980106E+00
DCV,8.2237596512,-1.980206E+00
DCV,8.4237136841,-1.980117E+00
DCV,8.6238460541,-1.980151E+00
DCV,8.8238334656,-1.980220E+00
DCV,9.0238161087,-1.980155E+00
DCV,9.2239456177,-1.980105E+00
DCV,9.4238948822,-1.980061E+00
DCV,9.6239099503,-1.980061E+00
DCV,9.8240585327,-1.980089E+00
DCV,10.0240192413,-1.980048E+00
DCV,10.2240056992,-1.980011E+00
DCV,10.4241542816,-1.980070E+00
DCV,10.6240949631,-1.980055E+00
DCV,10.8252677917,-1.980056E+00
DCV,11.0251789093,-1.974333E+00
DCV,11.2252397537,-1.307193E+00
DCV,11.4253664017,1.053806E+00
DCV,11.6252880096,1.962226E+00
DCV,11.8253135681,1.979114E+00
DCV,12.0254421234,1.978924E+00
DCV,12.2253580093,1.978975E+00
DCV,12.4253692627,1.979014E+00
DCV,12.6254377365,1.979111E+00
DCV,12.8254461288,1.979105E+00
DCV,13.0254764557,1.979070E+00
DCV,13.2255163193,1.978966E+00
DCV,13.4255256653,1.978929E+00
DCV,13.6255464554,1.978772E+00
DCV,13.8255786896,1.978756E+00
DCV,14.0256805420,1.978708E+00
DCV,14.2256231308,1.978740E+00
DCV,14.4257678986,1.978772E+00
DCV,14.6257085800,1.978803E+00
DCV,14.8257331848,1.978868E+00
DCV,15.0258588791,1.978925E+00
DCV,15.2258081436,1.978895E+00
DCV,15.4258232117,1.978820E+00
DCV,15.6259517670,1.978818E+00
DCV,15.8258924484,1.978791E+00
DCV,16.0259170532,1.978506E+00
DCV,16.2260551453,1.645835E+00
DCV,16.4259834290,-5.011422E-01
DCV,16.6259975433,-1.910020E+00
DCV,16.8260555267,-1.979830E+00
DCV,17.0260810852,-1.979768E+00
DCV,17.2261085510,-1.979764E+00
DCV,17.4261226654,-1.979775E+00
DCV,17.6261367798,-1.979843E+00
DCV,17.8267440796,-1.979848E+00
DCV,18.0266151428,-1.979829E+00
DCV,18.2266178131,-1.979847E+00
DCV,18.4267959595,-1.979837E+00
DCV,18.6266841888,-1.979698E+00
DCV,18.8266925812,-1.979707E+00
DCV,19.0267677307,-1.979766E+00
DCV,19.2267646790,-1.979769E+00
DCV,19.4267768860,-1.979799E+00
DCV,19.6269245148,-1.979892E+00
DCV,19.8269100189,-1.979952E+00
DCV,20.0268516541,-1.979956E+00
DCV,20.2270088196,-1.979872E+00
DCV,20.4269199371,-1.979882E+00
DCV,20.6269474030,-1.980007E+00
DCV,20.8270702362,-1.979978E+00
DCV,21.0269870758,-1.845071E+00
DCV,21.4470748901,1.978747E+00
DCV,22.3273601532,1.978798E+00
DCV,22.5275154114,1.978814E+00
DCV,22.7274436951,1.978876E+00
DCV,22.9274597168,1.978843E+00
DCV,23.1276111603,1.978771E+00
DCV,23.3275279999,1.978939E+00
DCV,23.5275440216,1.978910E+00
DCV,23.7276000977,1.978854E+00
DCV,23.9276371002,1.978866E+00
DCV,24.1276683807,1.978937E+00
DCV,24.3277091980,1.978940E+00
DCV,24.5277252197,1.978922E+00
DCV,24.7277297974,1.978934E+00
DCV,24.9277763367,1.979013E+00
DCV,25.1277961731,1.978926E+00
DCV,25.3278236389,1.978762E+00
DCV,25.5279483795,1.978718E+00
DCV,25.7278919220,1.978830E+00
DCV,25.9293479919,1.979036E+00
DCV,26.1293487549,1.846035E+00
DCV,26.5514125824,-1.980106E+00
DCV,27.4316883087,-1.979922E+00
DCV,27.6341247559,-1.979897E+00
DCV,27.8318672180,-1.979915E+00
DCV,28.0318107605,-1.979855E+00
DCV,28.2319431305,-1.979865E+00
DCV,28.4318809509,-1.979834E+00
DCV,28.6319141388,-1.979810E+00
DCV,28.8320484161,-1.979816E+00
DCV,29.0319805145,-1.979913E+00
DCV,29.2320022583,-1.980012E+00
DCV,29.4321346283,-1.979985E+00
DCV,29.6320552826,-1.979882E+00
DCV,29.8320922852,-1.979778E+00

is this what you want ?
also how about all sorts of settings ?
fast, med, slow ?
auto range ?
trigger auto, and delay auto, samples / trigger ??
here is a new file, much longer recording,
webx dot dk / siglent / csv_data_1.csv

[Bad_Driver]

This looks much better as what is possible with the SDM3065! Every 0.2 sec a reading!
Thanks for your efforts!

Can you please try again with lower DMM resolution to see what will be the fastest acquisition frequency?

[Bad_Driver]

Sorry, didn‘t answer your question.
No auto range, fast reading, trigger auto, trigger delay=0 -> let‘s see what is possible.

[oz2cpu]

exactly.. got it
please reload the link,
just added 3 new files, about 5 mins, 5V dc, in slow, medium, fast.
the fun thing is all modes save a lot of digits into the file ..

[Bad_Driver]

I got the file. I will have a look tomorrow and do some comparisons with the 3065X.

[Bad_Driver]

The SDM3055X behaves as expected! Very constant readings, in slow mode 5 Hz, in fast mode about 72 Hz!

So I'm pretty happy to spend 800 € for the SDM3065X with very weird and crappy acquisition behavior.   

I try to compare your readings with the SDM3065 for the SDM threat, hope that Siglent takes care of this issue (Tautech where are you??).

I did a FFT with your slow readings, I assume this is not the SPD3303X??? (X = frequency/Hz)

[oz2cpu]

>I did a FFT with your slow readings, I assume this is not the SPD3303X??? (X = frequency/Hz)

nope it is the oldie Oltronix_stabpac_mca_5-3 the most stable voltage i got :-)

I can make a slow-mode file SPD3303X tomorrow, just booted it up
I let the voltmeter stay on, all night too, this way they will be as stable as possible,
how long file do you think I need to record ?
only slow ? or also the other modes ?

[Bad_Driver]

Plz try to make a file with 4096 readings. This is the limit for EXCEL FFT.

[oz2cpu]

so this is the plan, agree ?

DC source : Siglent  SPD3303X set for 5.000 no load, been powered on for 24hrs at least.
slow 4096 points needed, 4096/5 = 819 sec, 13,6 mins
medium 4096 points needed, 4096/50 = 82 sec, 1,36 mins
fast 4096 points needed, 4096/92 = 44 sec, 0.74 mins

i add atleast a few secs, so we are sure you get over 4096, maybe it is a good idea to take double the length, and perform the test two times ?
see if results align ?

PS: how do i get the excel fft ? please email it to me, my email is my user name at gmail so that is easy to remember :-)
note : i have a hard time importing this CSV format, my excel dont accept the format.

[Bad_Driver]

Follow this advice to get exact 4096 readings to usb. For best resolution 2 volts DC I think.
https://siglentna.com/operating-tip/sample-and-store-sdm3000/

For FFT I send you the information via mail tomorrow.

Edit: Please adopt the advice so, that datas are written direct to USB-disk, internal recordings will not have any time stamp. 
Very old failure (I can‘t beliefe that this shall be a feature) in the firmware.

[oz2cpu]

here we go, refresh the link

I added slow, medium, fast,
2V and 5V same SPD3303X
the record length should be over double the min required, this means it is possible to make two fft pr test, and see if results are the same

[Bad_Driver]

I have started with the analysis but no results before Friday. Tomorrow I get my (3rd) vaccination.

[Bad_Driver]

With help of oz2cpu data acquisitions I calculated 2 new FFTs of the SPD3303's output oscillations.
oz2cpu and I recorded with our DMMs (oz2cpu: SDM3055X, medium speed; me SDM3065X, 0.05 PLC) after warming up DMM and
SPD for more than 24 hours more than 4000 values for further FFT with EXCEL.

I discovered some interesting findings for the SDM3055X and 3065X but I will them post in the Siglent SDM topic later.
Long story short: SDM3055X behaves much better with recordings (with regards to sampling frequency accuracy) but
not perfect, SDM3065X is (big surprise!) only useable with high sampling frequencies up to 600...2000 Hz.
This is the reason for the fewer (smoother looking) data's in my measurement, there is no additional information in the spectrum above 20 Hz (up to 1000 Hz).

Attached you see both FFTs with nearly the same result. There are not only some oscillations in the low-Hz range but up to 18 Hz as well with the SPD3303X. My former low-f FFT:
https://www.eevblog.com/forum/testgear/how-to-calibrate-siglent-spd1000x-spd3303x-spd3303x-e-series-power-supplies/msg3916976/#msg3916976

[mawyatt]

Good news on the rogue SDP3303X-E that had the current offset mentioned much earlier. We sent this to Siglent NA before the Holidays and just received the repaired unit. After 1/2 hour warm up, ran some quick tests for output voltage and current limit. Everything is Spot On, even lower current limits below 20ma, so we can now put this back in service with the intended use

Don't think we'll be attempting a self calibration on this unit tho 

Thanks Siglent NA (Jason)

Best,

[mawyatt]

With help of oz2cpu data acquisitions I calculated 2 new FFTs of the SPD3303's output oscillations.
oz2cpu and I recorded with our DMMs (oz2cpu: SDM3055X, medium speed; me SDM3065X, 0.05 PLC) after warming up DMM and
SPD for more than 24 hours more than 4000 values for further FFT with EXCEL.

I discovered some interesting findings for the SDM3055X and 3065X but I will them post in the Siglent SDM topic later.
Long story short: SDM3055X behaves much better with recordings (with regards to sampling frequency accuracy) but
not perfect, SDM3065X is (big surprise!) only useable with high sampling frequencies up to 600...2000 Hz.
This is the reason for the fewer (smoother looking) data's in my measurement, there is no additional information in the spectrum above 20 Hz (up to 1000 Hz).

Attached you see both FFTs with nearly the same result. There are not only some oscillations in the low-Hz range but up to 18 Hz as well with the SPD3303X. My former low-f FFT:
https://www.eevblog.com/forum/testgear/how-to-calibrate-siglent-spd1000x-spd3303x-spd3303x-e-series-power-supplies/msg3916976/#msg3916976

Wonder if this is something to do with the feedback loop that uses the ADC readings to correct the output, assuming Siglent has incorporated this type of error correction loop? If they did, then the cycle time between read and correction might be within the FFT spectrum you've shown.

Do recall Dave mentioning a high resolution SD ADC chip in his teardown video, these tend to be slow and their sample rate would fit the FFT artifacts. Would be interesting to hear from someone that has the schematic, or knows how this PS is implemented.

Best,

[blurpy]

Good news on the rogue SDP3303X-E that had the current offset mentioned much earlier. We sent this to Siglent NA before the Holidays and just received the repaired unit.

Really nice! Any idea what they did to fix it? Clear an eeprom and recalibrate perhaps?

[Bad_Driver]

Had a short look into Dave's video. The ADC is a AD7792 (3-channel, 16/24 bit) and the reference is a ADR03.
If they have done the regulation feedback loop in "software" than they should be able to fix it with a firmware update.

[mawyatt]

Good news on the rogue SDP3303X-E that had the current offset mentioned much earlier. We sent this to Siglent NA before the Holidays and just received the repaired unit.

Really nice! Any idea what they did to fix it? Clear an eeprom and recalibrate perhaps?

Agree, really nice they did this under warranty since I had "augmented" the X-E to an E. Our only cost ($50) was shipping to Siglent NA

No idea what they did.

Best,

[mawyatt]

Had a short look into Dave's video. The ADC is a AD7792 (3-channel, 16/24 bit) and the reference is a ADR03.
If they have done the regulation feedback loop in "software" than they should be able to fix it with a firmware update.

Siglent may have implemented a "Dual Loop" with a continuous feedback loop and another feedback loop formed with the ADC where this becomes more of an output voltage update and now behaves as a sampled data feedback in parallel with the continuous loop.

Agree the loop formed with the ADC should be affected by firmware.

Best,

[Bad_Driver]

I found some time to setup my high-end audio interface (Terratec DMX6 Fire USB) to measure the noise spectrum above 10 Hz of the SPD3303X.
(it's a night mare since all the drivers are from the good old XP/win7...8 decade)

Attached the results. Graph 1 is channel 1 @ 2 volts, no load. Graph 2 is channel 2 @ 2 volts, no load.
Nearly identical. Above 3 kHz is nothing of interest in the spectrum ( I did it up to 90 kHz with "AudioTester 3.0; 192 kHz sampling rate, 24 bits).

Please note that the levels of the spectrum are in calibrated "dBV". Overall the SPD is in specs. There is no dependency to the level of output voltage.
End of the week I can try it under load.

[oz2cpu]

I am so happy you look into the odd regulation loop noise,
if this is a sw or hw issue, i dont know,

I have been "playing" with SPD3303 today, i did find optocoupler drivers on both sides, running 5 times
hotter, than all other opto driver transistors, so i figure they do run some kind of hi speed link there too,

since I keept remembering the too hot parts, in the EEVblog 828 see at 26 mins
and i run my SPD3303 most of the time in week long experiments
i was looking for a way to , fix the hot parts, and to dimm the display (mine are very bright)

I gave up on the display dimm, looks like impossible to do, i digged all the way to the display it self,
there is just no way to get access to the led wires and put in a resistor.

I did however fix the too hot parts, since i fixed both psu in one go, made the video during the 2nd psu fix,
so i knew what to do fast and easy.
it is all explained here:

https://youtu.be/nvGxl44cNno

incase anyone want access to all my hires pictures of all details inside SPD3303
get the 490MB zip here
webx.dk/siglent/SPD3303PICS.ZIP

[Bad_Driver]

During the last firmware update I learned that there is a new hardware revision of the SPD3000.
(firmware update was only for new hardware release)
Has someone an idea what was improved?

[BreakingOhmsLaw]

During the last firmware update I learned that there is a new hardware revision of the SPD3000.
(firmware update was only for new hardware release)
Has someone an idea what was improved?

The SPD3303X(-E) had an issue with a bridge rectifier overheating, causing the unit to reset frequently under load. The issue is common. So far, Siglent has been repairing them under warranty as far as users here report. Funny enough, in Dave's teardown from years ago, he already noticed that hot rectifier.
https://www.eevblog.com/forum/repair/siglent-spd3303x-e-rebooting/msg4340536/#msg4340536

[oz2cpu]

you might want to see this one
https://youtu.be/nvGxl44cNno

it is about how to modify the psu, so its life time will be much longer

[Peter_O]

Never tried to calibrate mine.

acc. to siglenteu the 3.0 FW is the newest one: SPD3303X Firmware_V1.01.01.03.10 (Release Date 09.02.22 )

https://www.siglenteu.com/service-and-support/firmware-software/dc-power-supplies/#spd3303xx-e-series