KORAD KA3005P Schematic, Design Details, Teardown, Infos

[@rubenuspv]

This PS is old, it's 2013 or older, maybe at that time the dacs weren't as cheap as now, now you can get those under $1usd in single unit quantities, maybe even weren't available in China at all at that time. Now there are some switching power supplies from China that includes 16bit TI dacs. Great vid, subscribed.

[@KonstantinGrigoriadis]

Hi , Thankbyou, the spoken dac is from the early 2000’s was just an example, the developers would have used a 12bit dac made for the asian Market like the Nuvoton mcu’s.

[@AnnaVannieuwenhuyse]

@@KonstantinGrigoriadis most of the nuvoton mcu's used for this have the ability to put out 12bit on their own DACs. They very likely ended up going the way you can see here due to either preexisting designs that were retrofitted with a newer MCU, or the lack of R&D money, or some strange limitation that is due to architectural differences between the Nuvoton and whatever the Nuvoton ripped off. That latter issue can make them behave weird, and it cripples the GigaDevice ADC's unless you hack in workarounds or code with their own API for an example. On another note, the shift registers you noted are sub-cent parts at times. The DAC's with sufficient performance can definitely reach tens of cents.

[@KonstantinGrigoriadis]

Hi Anna, probably they just used a dac design the had used before and they had the code for it proven and tested. Even if the shift registers are dirt cheap, using more than 60 components to build just 2 12 bit dacs is ridiculus , this makes the design susceptible to noise for no real benefit.

[@AnnaVannieuwenhuyse]

@@KonstantinGrigoriadis I suspected as much!

[@odieadog4086]

The way how output transistors (power board Q1, Q2) are controlled by main board seems fascinating. The main board's supply seems to bootstrap itself to float on top of (power board) Q3's base by controlling the current through (main board's) Q4.

[@AnnaVannieuwenhuyse]

Nuvoton sounds quite foreign for many of us, but it's a very popular brand because they have such a wide variety of chips they sell at large volume that you can both get a chip ideally suited for your use & get it cheap. It's a spin-off of Winbond, the manufacturer we most often see making motherboard IC's and EEPROM/Flash or RAM

[@KonstantinGrigoriadis]

Hi Anna, Thankyou!

[@cristianstoica4544]

The rationale seems on point. Replacing a proven firmware with parallel drive to a book DAC with a serial I2C drive could have been the difference between a working product and a warranty neightmare. I can imagine the discussions between the engineers and the product manager

[@giorgostexnikos6976]

excellent analysis

[@KonstantinGrigoriadis]

Thankyou !

[@danielleblanc5923]

Thanks for the very nice and detailed walk through. To answer the question "why an externa EEPROM ?" I think it takes quite a beating even under regular use. The settings memory is dead in mine, so I have to replace this EEPROM. The issue shows up this way: - If at startup, the values show up correctly, the power supply will work normally until next power up. - If at startup, the values do NOT show up correctly, the power supply will continue displaying garbled data no matter what operation you perform and lock up within 20s I can not confirm this as I haven't had time to perform the swap yet. I have to be absolutely sure to read the old data in case there would be calibration data in there.

[@KonstantinGrigoriadis]

Hi Daniel, Thankyou! Yes Probably its easier to change the EEPROm than the MCU if the EEPROM Memory of it fails, Of course they could optimize when to absolute necessary to Write, but its easier and safer like that! hope there are no cal data there!

[@absurdengineering]

They could have made a DAC with excellent linearity by filtering down slow PWM. It’s one of the approaches used in the design of metrology-grade voltage sources. The resistor array is cute looking though.

[@KonstantinGrigoriadis]

Hi, yes it looks interesting of course :-)

[@stevenbliss989]

You could MOD the D/A using an board with 3 x 74HC165's (for 24 bit to serial convert, a MCU (8-pin Sop-8 ) like ATTiny13A/85, PIC12F629/675, or if feeling adventurous a PMS150C (the 3c micro), plus whatever 12-bit DAC(s) to taste and then do a lot of wiring (only a few track will need to be cut on the PSU). WHAM BAM, THANK YOU MAMA and you have no more magnetic influence issues!

[@stevenbliss989]

I have a KORAD copy brand PSU, so I might try this in the future (far future - health issues), just for fun! :)

[@stevenbliss989]

I just noticed that they are using 3 x 74HC595's, so a smarter idea would be to use a MCU serial in port ad do away with ALL THOSE WIRES! Feel dumb for not noticing! :(

[@stevenbliss989]

This would of course require VERY careful MCU selection to make the serial in work properly.

[@grahamhall2662]

Hello I have just bought the 10 amp version of this PSU, it is a beast. It has a much beefier fan heatsink arrangement and transformer , but otherwise very similar. Thank you for posting circuit diagram, never know when I might need it? Great analysis too, pity about the vast resistor array, I did wonder of it's purpose. I have learned a lot from your post, Regards Graham

[@KonstantinGrigoriadis]

Hi Graham, Thankyou ! Cheers, Konstantin

[@sokoloft3]

Thank you for the video. Probably going to pick one of these up. They seem more quality made vs the ones you can buy for $40 less. Hanmatek branded, switching supplies. I don't need the serial communication, so I'll go for the D variant.

[@KonstantinGrigoriadis]

Hello, I have it since 2015 and works fine :-)

[@sokoloft3]

@@KonstantinGrigoriadis I hope mine lasts just as long. Only using it for voltage injection, finding shorts in macbook logic boards.

[@laszlokovacs402]

I wonder how the output voltage of those DAC's get calibrated ?

[@KonstantinGrigoriadis]

Hi Laszlo, good Question. In this Design there is only doing it by software either by : 1) adding a correction list(Array) over the DAC Range by sampling a number of Points in the Line and correcting the Output accordingly. 2) or by reading back the output of the 2 Dacs and correcting the Output accordingly. (by dont see anything in the Schematics for that) This is one further Disadvantage of this Discrete Design Because of the Resistor Value Errors and thermal Drift. Even the Reference Voltage for the R2R Ladder is from a simple LDO.

[@therealspixycat]

That N054LAN has no digital to analogue converter. Plus they probably use analog comparators to monitor both current and voltage output so no constant looping is required

[@KonstantinGrigoriadis]

Hello, yes, the mcu is not in the regulation loop, to slow and to Unpredictable, just for setting and showing, typical digital controlled psu design, just too diskrete

[@therealspixycat]

@KonstantinGrigoriadis the analog comp can generate an interrupt so you would expect instant switch off of any further over current or over voltage. Very interesting to see this digital controlled psu which gives a very rough idea how the main features are accomplished. I am considering building a prototype with a stm32 and ie a uc3843 or similar chip but it has been very hard to find any information or design up until I saw your video. Thanks!

[@therealspixycat]

​@KonstantinGrigoriadis and after going through your video (several times...) it is not clear to me how they actually regulate the selected voltage or current control. There is only a kind of emergency voltage and current control using opamps so that in case of the mcu crash or freeze it won't go into a total runaway. Or or is the regulation done when the over voltage or over current opamp comparators go high that the 2 parallel transistors in the power supply will be disabled???

[@KonstantinGrigoriadis]

​​⁠hello, even with interrupts this is to slow and unpredictable for analog regulation, I do alot with mcus hard and software, there are nested interrupts, etc. I never have seen (and probably never will see) ca mcu control the analog loop except for testing with really high speed mcu‘s that break the bank :-) massively !

[@KonstantinGrigoriadis]

The 2 opamps with the ored diodes at the outputsare the regulation, its a typical u/i analog design regulating the current fed to the driver transistor of the power board, ocp and ovp hare handled from different opamps

[@pomonabill220]

I have one of those supplies and reverse engineered it several years ago. I too was surprised that they used such an old design for the A/D. Other than the lazy design, it has worked for me! One thing I did not like is the way the wires are routed between the power transformer to the back power board. The sheet metal baffle is SHARP and can cut the wires, so I re-routed the wires and cleaned up the edge of the sheet metal so it wasn't sharp, and put a piece of tape on the edge. Where did you find the schematic?

[@KonstantinGrigoriadis]

Hi, yes the metal sheet in the middle is sharp and can damage the cables, therefore i put the cables in a Protective sleeve. I found the schematics in the internet, cheers, Konstantin

[@nixxonnor]

I have the same power supply (branded Velleman LABPS3005D) One annoying thing with it is it rounds up the voltage shown on the display when the voltage is a little below the set voltage. When charging a LiPo battery with settings 4.2 V and 1 A, the voltage is shown correctly for a while as the voltage rises towards 4.2V. But around 4,10V (or thereabouts) the voltage suddenly jumps up to 4,20 on the display when the actual voltage is lower. The PSU nevertheless delivers the full 1 A until the voltage actually reaches 4,20V. Why did they do it like that? It is silly. Is it the same bug on the Korad?

[@KonstantinGrigoriadis]

Hi, I never used the KORAD as a charger, normally this should not happen bcs as seen in the schematics the Regulation Loop is purely Analog and not through the MCU. Seems to be only a Display Issue the has to do with the Internal 12Bit ADC or the code driving it, when the jump occurs, did you measure the real Output Voltage ?

[@hrandib]

Can't confirm. My Korad KA3005D behaves correctly in such a case

[@KonstantinGrigoriadis]

Hello, In what case ? Inductive loads or High EMC Pulses ?

[@hrandib]

@@KonstantinGrigoriadis I answered to the upper message. No issue with charging of li-ion batteries

[@KonstantinGrigoriadis]

​@@hrandib Ah, OK Thankyou!

[@josesalierno7177]

muy buena explicacion una pregunta que diferencia tiene esta fuente con las que son copias de la misma es el traformador o que es

[@KonstantinGrigoriadis]

Hello Jose, Thankyou, for what I know, there is no real Difference to the other Differently name Branded Versions of the KORAD except of the Name. The others ones are not Copies but Rebranded OEM Versions of it. Cheers, Konstantin

[@Ahmedmohamed-rs8no]

Thank you for the video but I have a question please if I set the current to 5amp and I cross the wire doesn’t give me the full 5amp it give me around 1.5amp I changed the tl082 but no change any advice please where the problem could be thank you for your time 🙏

[@KonstantinGrigoriadis]

Hello Ahmed, Thankyou! The 1.5A you see it on the Korad Display or you use an external Ampermeter ? If you see it only on the Korad Display, put in Series an External Amperemeter in Series with the Load and check what it shows.

[@Ahmedmohamed-rs8no]

@@KonstantinGrigoriadis thank you for the reply sir I see it on korad display and I also tried external ampermeter and I still get the 1.3 1.5 amp max when I close circuit

[@Ahmedmohamed-rs8no]

@@KonstantinGrigoriadis I found the problem sir was one of the d1047 was faulty I replace it and the current works good.now thank you sir

[@KonstantinGrigoriadis]

Very Good, You are Welcome.

[@paulcohen1555]

At 18:25 "current cathode" >> "common cathode"

[@RobertHammersteinArt]

Hi, I found Sunshine P3005A which looks visually the same (except USB). Do you think its safe to buy :) ?

[@KonstantinGrigoriadis]

Hi Robert, KORAD is the manufacturer of those psu’s, you will see them under many different OEM names, they are basically the same, unless that some OEM habe requested slightly differences.(specs and usage Eventually) .

[@inductr]

Thanks for the upload and greetings! Are the resistor array maybe old design they just kept?

[@KonstantinGrigoriadis]

Hello Arve, Thankyou, discrete R2R Dac (except in high end Audio) is gone since the 80‘s , the only reason I can Imagine is that they had the code ready for a proven working Design and used it, to speed up development.

[@inductr]

@@KonstantinGrigoriadis yes makes sense. No one had to hand solder it😀😀 just computer design and pcb robots.

[@pldaniels]

Wonder if it comes down to cheapness, 5c for all those resistors vs cost of the DAC? At least too you'd never have trouble of not being able to source the DAC :)

[@KonstantinGrigoriadis]

The 52 resistors the 3 74hc595 and a dual opamp, I dont think thats really cheaper than the 2 dacs i mentioned. But wait, no need to use big brand dacs, they can use cheap dacs made for the asian market, like the nuvoton mcu’s

[@KonstantinGrigoriadis]

BTW Paul, if you want the Schematics, send me an Email, Friday this Week when I am back i can send it to you

[@ghlscitel6714]

I like your presentation. Where did you get the schematics from?

[@KonstantinGrigoriadis]

Thankyou ! The schematics are from a private CN forum.

[@TheOliverHB]

Where did you get the schematics from? I am currently looking for those of a Stamos S-LS-28, which seems to be a KORAD KD3305. Any clue ?

[@KonstantinGrigoriadis]

Hello, from a non public Chinese forum, but now they are on the internet too available. Reg. The Stamos I have no info's, but seems to be just a rebranded Korad 3005. Open the stamos and check if the internals match the schematics

[@stevenbliss989]

U5b & u7B pin 5 should be tied to ground with a 10nF cap. The op-amps have an UNBALANCE frequency response between the two inputs - A BIG NONO!!!!!!!!!!!! It likely will help with your "crazy" response to TESLA coil issue.

[@KonstantinGrigoriadis]

Hi Steven, Thankyou !

[@TheStuffMade]

Does anyone know if it's possible to buy a new main board? The microcontroller on mine died.

[@KonstantinGrigoriadis]

Hi, i have not seen one yet being sold , cheers, Konstantin

[@zxrjimmy1]

I have 3005p but suddenly no output voltage, turns on fine, set voltage OK, when I press on I get nothing on output, any ideas what could be wrong

[@KonstantinGrigoriadis]

Difficult to say, can be many, check the +/-12V Power Rails that power the Analog Circuitry, further Transistors Q4 on the Main Board and Q3 on the Power Board

[@zxrjimmy1]

@@KonstantinGrigoriadis changed both power transistors and relays, still same issue

[@KonstantinGrigoriadis]

Check the driver transistors and the +-12v rails as i showed on the schematics

[@KonstantinGrigoriadis]

The relays are not the issue, max that the relay drivets the one uln2003 is a problem.

[@senthilkumars6045]

Where did you get schematic please share link

[@KonstantinGrigoriadis]

Hello,send me a Mail to send you the files. Cheers, Konstantin

[@truongtrongtran8520]

can you share schematic pdf? thank you.

[@KonstantinGrigoriadis]

Hello, Yes, send me a mail to reply you.Cheers, Konstantin

[@seoseo4017]

Can we increase amp of this psu .i have 5 amp version

[@KonstantinGrigoriadis]

Hello, no bcs of many reasons, 1) the software does not support more, 2) the transformer is for 5a, 3) the power stage is not designed for higher currents, Cheers, Konstantin

[@seoseo4017]

@@KonstantinGrigoriadis thank you for explaining. ...

[@юрийсергованцев-ь2у]

Посмотри PSL-3604

[@KonstantinGrigoriadis]

Hi, thankyou I know this Design, for some Years, I like it and his other Designs, Cheers, Konstantin

[@stevenbliss989]

Assuming your schematic is complete, the MCU (Novatron) is in the feedback path to control voltage/current, or am I missing something. If true then a much faster, even DSP like MCU should have been used.

[@stevenbliss989]

btw PIC and TI have several AP's on this.

[@stevenbliss989]

I am BLIND again, it is done all analog! Open your eyes Steven! :)

[@KonstantinGrigoriadis]

No, the Regulation is Analog, the MCU is not in the Regulation Path