Voltlog #39 - 60W Electronic DC Load Reverse Engineering

[@Hugatry]

I finally decided to buy one of these, mainly because of info you've shared in your videos. I received it today and it surely is awesome device. There was one problem: Fan was always on (full speed). Thanks to values printed on the board and the schematic you shared, it took less than 5min to diagnose and fix the problem. Thermistor is 10kOhm version, instead of 50kOhm. Problem was easily fixed by reducing resistance of R14 to 2kOhm, making the voltage divider's ratio the 1:5, which it should have been in the first place. Thanks!

[@voltlog]

so it seems they are shipping a batch of these with the wrong thermistor value. Another viewer reported the same thing. Nice to hear you fixed it.

[@MaxAirGo]

I have one with fan always on, not sure if its full speed. I used dvm on ohms and put leads on thermistors, after meter settles, I get 7.24kOhms. How does your read? How did you find out its 10k? Is the resistor that needs to be replaced have 3 caps below it, 1 cap on left, 1 cap on right and 1 103 resistor on top?

[@Hugatry]

I desoldered one of the thermistor's wires from PCB and then measured resistance, which was little bit under 10kOhm. Now I measured resistance while both thermistor's wires are connected to the board and got same resistance as you. Yep, that description of resistor's surroundings sounds right. When looking at the main board, heatsink on top, crystal on bottom, there are three caps to right from microcontroller. The 10k Ohm (103) resistor that needs to be replaced with 2kOhm resistor is right above them.

[@MaxAirGo]

Thank you. I didn't have 2k so used a pair of 1k in serious as seen here s30.postimg.org/b87j7jln5/20170130_124832.jpg Now the fan stays off with power applied, turns on at slow speed when unit is set to run and speeds up as temp increases. When I turn the load off, fan stays on until temp goes below 43c. Measured the thermistor in circuit now and its showing 6.63kOhms after fix. I am so happy I got this fixed now thanks to you and voltlog. P.S. One other thing I noticed about my unit is the supplier only used 2 wrong size brass standoffs because the display board is at an angle. Also some units come with extra standoffs as legs, mine didn't, oh well.

[@voltlog]

from the picture your board doesn't seem to have the ZHIYU markings, could this be a clone? hence the issues?

[@ericksonengineering7011]

Nice work, and thanks much for your video. The 510K resistors are there to apply a small positive bias at the ADC input. The on-chip ADC plus any opamp offset might have a small negative bias, preventing the ADC from accurately measuring 0.0V. So the circuit applies a small positive bias and then subtracts the offset in software. Maybe it gets calibrated out and the offset calibration gets stored in EEPROM. The diode at the input of the op-amp is there to shut the op-amp fully off. Without it there may be a small positive offset applied to the opamp. Add the opamp offset, and you get a few mV error, which causes a significant offset current. A better way is to inject a small positive current via a resistor into the opamp - input. The R-C in the feedback loop is to compensate the loop. Power FETs have a large capacitive input (G-S) and a large feedback capacitance (D-G) which cause a pole in the loop. The R-C compensate for that. Cheers, Dave

[@voltlog]

thank you for the thorough explanation.

[@cmuller1441]

pin 17 is a shutdown control. when high the oa thinks there's a huge current flowing in the sense resistor and turn off mosfet. the main feeback loop for the oa is with the mosfet and the sense resistor. The problem is that it takes time to load the parasitic capacitor in the mosfet to activate it. The oa could then oscillate. The extra feedback is active for frequencies above 1/(2piRC). pin 19 indirectly detects that the voltage on the load is too low for the programmed current and the oa goes high to try to let current flow but the oa output is just maxed.

[@voltlog]

thank you for explaining this.

[@glennsprigg2378]

Thank you SO much for your usual in depth explanations !!!! You should feel PROUD of your knowledge and sharing !!! Glenn.

[@voltlog]

thanks!

[@peterdkay]

Excellent video and thanks for schematic. I purchased one of these as a battery capacity tester.

[@HowardBeachNYC9657]

Very nice, I just bought one because of your review. Thanks.

[@sureshlingabathina]

By watching your videos someone can design good products, excellent

[@davidkovacs2313]

Servus! Es war schöne Arbeit! Gratuliere, and Danke Dir, Dein Video ist sehr interessant! Beim 5:04 sieht man 510K und 10K Resistor (R5 und R3), die geben wenig Offset, damit man das im Software korrigieren kann, und so wird Offsetfehler von ADC vernachlässigt. Gleiche ist bei 6:21 R18 und R21. TL272 hat gross offset, mehr mV, und man weiss nicht, ob das negativ, oder positiv. So wird im positive Bereich verschoben, und Fehler kann korrigiert werden. Bei IC4 fehlt dieses teil, das ist das Grund, warum man nicht weniger mA einstellen kann. Dieses Fehler wird mit Software versteckt. Bei einem Elektrische Load braucht man nicht schnell reaktion von DAC (Strom referenze), deswegen könnte man das ganze 12-16bit verwenden, und bei 10A könnte man mit 12bit 2.44mA Auflösung auch erreichen, mit 16bit 0,15mA pro bit. Pin17, wenn das LOW ist, zieht R28 nach GND, und bei D3 wird umgefähr 0V, also Diode nicht leitet, mach gar nichts. Wenn Pin17 HIGH wird, dann D3 leitet, und beim Invertierende Eingang von OPAMP bekommt zirka 4.3V. Das bedeutet, dass dieses OPAMP sieht 4.3V beim Shunt Resistor, also OPAMP sieht ein riesiges gross Strom. Deswegen schaltet aus, und fliesst kein mehr Strom. So kann STM uC das ganze einschalten, oder ausschalten. Bei T1: Pin19 hat sicher PullUp Widerstand, also im normale Betrieb Pin19 ist HIGH. T1 wird dann Pin19 nach GND ziehen, von Opamp mindestens zirka 5-6V bekommt. Es kann nur dann passieren, wenn OUTPUT von IC4B arbeitet im höchsten Spannung Bereich, also beim FET kann nicht so viel Current fliessen, wie viel eingestellt wurde. Zb.: Wenn man dort zu wenig Voltage hat, und 0.2V zu wenig mit series Diode und 0.01ohm zu 10A, dann STM uC kriegt ein warnung. Oder wenn keine Spannung beim Eingang liegt, oder kontaktfehler gibt, usw. In dem Fall muss IC4 ausschalten. Beim 0V Eingang, IC4 wird für Gate von FET maximal Spannung geben, FET wird mit seine RDSON leiten. Wenn man in dem Sekundum zu Eingang ein Voltage gibt, es wird unendlich gross Strom fliessen, bis IC4 reagiert, und den FET ausschaltet. Und es dauert lange, vielleicht zu lange. So viel könnte ich dazugeben. Schönen Tag!

[@linuxcow9870]

Thank you for sharing , I now have a schematic for my fine DC Load

[@voltlog]

+Linux Cow yeah, that's the spirit :)

[@aaaooo8840]

Great work. Thank you for sharing schematic.

[@epiendless1128]

Thanks for this. It was very helpful when I wanted to tweak the accuracy. One minor error in the linked schematic - from the thumbnail I see that your unit is the same as mine, and R17 is 1k, not 10k.

[@lucianopalermo7097]

Hi, I'm unable to see which pins are used to read shunt voltage

[@liviupatty]

Hi, i have to say thanks for the video , i love your romanian accent because i'm romanian too . About calibration of this dummy load , i have one witch is about 37 mAmps off and i want to calibrate , i didn't find anything about this so i decide to email the seller and surprise hi sent me this : ZPB30A1校准 1.器材准备：12V基准电源（或者可调电源）；能够测试到5A以上的万用表；能提 供5-12V电压5A以上电流的电源；导线螺丝刀等。 2.进入校准模式：如下图这样在背面焊接一根短路线（旧版）： 新版本则是如下图： 然后保持安装ONOFF按键情况下给板子通电，一直到显示J0.0u再松开按键，即进入校准程序。 3.校准操作： J0.0u：将V+，V-，P+，P-四个输入口全部短路，按动ONOFF按键…… J12u：将V+和P+连接到12V基准电源正极，V-和P-连接到12V基准电源负极， 按动ONOFF按键…… J0.2A：电源串联电流表后按照正负极连接至P+和P-口，调整旋转编码器使得电 流为0.20A，保持此状态下按动ONOFF按键…… J5.0A：同上所述，调整至实际电流5.00A，保持状态下按动ONOFF按键…… Pxxx：该状态时候可以调整功率限制值，范围50-150（W），调整好后按动 ONOFF按键…… r xx：该状态时调整的参数是线电阻补偿值，单位为mΩ，例如当设置为10时， 则在电流为1A时候电压会补偿显示0.01V。裸板款最佳值为3 以上操作完成后如果参数认为有效，则会轮显6个校准记录参数后保存并退出，如果无效则只保存功率限制值和线电阻补偿值并退出。 注意：1. J0.2A和J5.0A时，可以通过按下编码开关同时拧的方式执行快速调整（不按 住则是微调） 2.程序能够记录校准运行次数，出厂默认都是1（最开始显示），当自行运行校准后， 则不再有免费保修。 I'm sure you will figure out the way to calibrate before me and you will teach us how to do . Thanks and sorry for bad English

[@Doctorbasss]

Hello Daniel, Please could you email me the calibration procedure you received from teh seller?

[@isprithul]

Can you please tell me the frequency of the PWM output on pin 18 of the microcontroller?

[@carysmith638]

I watched episode 38, your review and many thanks. I made a comment there (on #38) that I would like a lower minimum current than its 200ma. For small batteries 50ma would be good. Can this unit be modified to make it 50ma? Would changing the 10mohm shunt resistor work to do that? And then apply a scaling factor to the meter read out? I expect my unit to arrive (USA) in a few weeks. Thanks. Tom

[@voltlog]

+cary smith That's an interesting idea. I thinks is worth trying with a 0.1 ohm resistor, then everything on the amps meter display would be the actual measurement x10, so the actual range would be 20mA to 999mA. I will give it a try on mine if I can find a 0.1R shunt resistor in my lab and maybe post a video soon. Thanks!

[@voltlog]

+cary smith It's confirmed! It was so interesting that I immediately tested the idea and it works! A video will follow soon.

[@PaulJosephdeWerk]

I don't see the schematic portion of the satellite board. There are two switches there, the pushbutton you see, and also the rotary encoder (push it down to activate the switch). It would be nice to see how it interfaces with the main board.

[@MaxAirGo]

in description check out Andreas Spiess - Reverse engineering of the display part

[@Electester]

Could you tell me (because I bought an counterfeit one - with black microswitch) where is R15 and R14 on the board respectively to your diagram of "Fan PWM drive"... Because my Fan is running carelessly and still and I want to change the value of reistors. I have even one NTC thermistor (47k actually) on stock.

[@voltlog]

I don't remember their position but R15 has one end connected to pin 16 of the mcu, you can figure out the rest by probing on the pcb.

[@Electester]

VoltLog Ok. Doesn't really matter because first I changed the NTC from the 10k one to 47k and it works as it should. I bought the device on Banggood and it turned out as scam, unbelievable... So this is the tip for everyone who is looking to solve the "problem ". Now I want to improve only the current accuracy because when you e.g set the current to 600mA, there is actually flowing ~650mA. It's pretty big error about 8,33%. Accuracy is better with bigger currents.

[@julianbaugatz6967]

I am right when I think that the load current is not measured directly - is only set by the µC?

[@voltlog]

Not sure what you mean.. it has to measure load current in order to operate.

[@DeeegerD]

Ok OK! I bought one ;) If I need PC logging I can use my logging multimeter to capture the voltage of the source and with constant current and voltage cut off should do the job. How about a multiple shunt modification (10 milliohm, 100m, 1 ohm, 10 ohm to give ranges from .2mA to 9.99A)? Of course you would need ohm free connectors and switches :) (I think eBay has em lol) or tinker a lot. I still plan to build one of my own design (well borrowed and modified) with PC logging and more functions but this will do in the meantime. Thanks for your vids on this and all the work.

[@mauriceb]

Digger D I am looking forward to your Project. If I Can help you to design a Dummy load with pc logging, i would be happy to help you (if I Can)

[@crocellian2972]

Outstanding work young man. You have a great future if you stick to it. Read Horowitz & Hill.

[@voltlog]

thanks!

[@JS106351]

I've been watching more of your videos, and I have to say keep up the good work! I did a little research and the configuration of that op-amp is in is a "augmenting integrator". Have a look at this: www.globalspec.com/reference/53124/203279/5-5-integrators, it explains that it is a type of PI system used for closed loop feedback control for servos and the like. I hope this bit of info helps, but thank you for sharing.

[@voltlog]

+JS106351 yes it all makes sense, thank you for the info. I added a note on video saying that.

[@awesamhead]

In the schematic, I don't see how a load would be connected, it seems that the MOSFET drain and source are connected in parallel to the load, which I'm sure is wrong. Could you confirm?

[@voltlog]

This is the load, you are not connecting another load.

[@awesamhead]

I don't understand, the load should be connected to P+ & P-, or am I wrong? If that's the case, why is the MOSFET in parallel with P+ & P- ?

[@gregc2222]

Yes, you are wrong. The device being described in this video *is* the load. There is no external load. You connect a source of voltage: a power supply that you want to test under load or a battery for example, to the P+ and P- terminals. The current flows through the MOSFET and the R-SHUNT. The MOSFET and the R-SHUNT make up the load. The rest of all the circuitry on this board controls how much current is allowed to flow through those 2 components.

[@dennise4163]

Could someone please measure the voltage on their fan when they are testing a battery at 1 amp. My fan does not begin turning unless I physically move it first no matter what the current or how hot the heat sink becomes. It will continue turning after I move it and at 1 amp the fan measures 6.3 volts. I cannot troubleshoot my problem.

[@cmj20002]

Have you figured out a way to hack into it? Thanks for the video. I would like to know any special function that you figure out.

[@voltlog]

Voltlog #38, #39 and #41 show everything I know about this dummy load. The only hack I did is described in #41.

[@ufohunter3688]

Can you post your CAD schematic file? It doesn't matter what software it is. I can read all of them. I like to copy and paste portions of it into LTSpice, and don't have to draw the schematic from a PDF file, if I can help it. Thanks for another great video.

[@voltlog]

Sorry, I don't have those files anymore.

[@ufohunter3688]

Oh well, had to ask. No worries. Will draw it from scratch. I really enjoyed all your videos. I binge-watched all of them yesterday while in bed rest. Thank you for this generous service.

[@voltlog]

I like doing this videos, thank you for watching!

[@MonicaGarcia-cq3kp]

Excellent video. You can increase 60 Watts????? You can go beyond 60 Whats??????

[@voltlog]

+Monica Garcia You cannot go beyond 60W unless you modify the firmware and the hardware.

[@MonicaGarcia-cq3kp]

+VoltLog Can you help me? It is a project for school I want to 180 Watts Please Help me

[@MonicaGarcia-cq3kp]

+Monica Garcia Or some similar diagram 180 Whats me?

[@javierpallalorden]

Excellent!

[@voltlog]

+Javier Palla Lorden thanks!

[@felixcat4346]

It´s possible that they want pin 17 to be zero volts. Maybe the spec for the chip requires a low condition.

[@voltlog]

+Joseph Nicholas pin 17 would actually be pulled up to +0.6V, the forward voltage of D3(1n4148) in combination with 10K + 51R resistors in series to 12V. I think that pin is actually used as an output, but I might be wrong :)

[@stephaniemarkham5724]

+Simon Markham, I think that this is the run/stop control.

[@principecristiano]

I like Reverse Engineering, think about mounting PZEM-031 100V 20A LCD Panel Meter to this device

[@TheAmmoniacal]

Good job :)

[@voltlog]

+TheAmmoniacal thanks!

[@stevebliss6056]

WOW, you went to a lot of work there! Anyway, I suspect you already realised this by now, but for what it is worth: c18 + R23 are frequency compensation, good to keep things stable and prevent overshoot/undershoot and/or ringing on large fast step changes in control or load source values. Hook up an oscilloscope to the op-amp output and physically connect/disconnect the load source, then pull either c18 or r23 out of circuit and do it again.. You will see how "yucky" the control loop response becomes to step changes.

[@voltlog]

thanks!

[@Hasitier]

Great. If Mine arrives here from bangood i might to try finding out more about it 😀

[@voltlog]

+Michael Bernhardt they always arrive, sometimes just a bit late :)

[@Hasitier]

+VoltLog yesterday it did arrive but the rotary encoder was broken during delivery. So i have to try to get a new one and solder it it. At least it turns on but i cannot operate it

[@voltlog]

+Michael Bernhardt sorry to hear that but I assume it was packed with just some of that soft protection wrap which is really insufficient for a product this size. They should send these in cardboard boxes.

[@Hasitier]

+VoltLog you are absolutely right. Only 2 layers of bubble wrap. And when one piece of the ihrem than is rising up from the flat pcb as the rotary encoder does it will break for sure. Or as Dave would say "Murphy gets me every time" ;-) But i got the item refunded and will try to replace the encoder over easter. Should be a standard one 3 pin with 2 additional pins for pushbutton i think.

[@voltlog]

+Michael Bernhardt yes, it's a standard rotary encoder with switch, search for "rotary encoder" on ebay, the blue ones are everywhere.

[@Old1Papa]

Hello, one of my load always showed "Err6", so I examined the circuit. Result: the circuit diagram still lacked a module "VSENSE +12". This is evaluated at pin28 of the μC (transistor switches at 12V Pin28 to GND). This NPN (J3Y) was broken. Here is the extended circuit diagram: www.old-papa.eu/images/dummy-load-sch-new.pdf Revision: ZPB30A1 V2.7 1510 Old Papa

[@voltlog]

thank you for sharing!

[@ElectronicMarine]

I almost buy one of this... then I realised that is not suitable for what I need because i need big currents like 20A @12-24V... I guess ì'll build an analog one with an opamp and mosfet...

[@voltlog]

+ElectronicMarine I guess this 60W load is designed for battery and small power supply testing. It would be nice if the software would be open-source, then it would be easy to add bigger heatsink, adjust some parameters and re-compile.

[@ElectronicMarine]

yep it wold be nice to put another 5-6 mosfets on the output side also...

[@arduinohannover2286]

I've just done a full reverse engineering of the hardware: github.com/ArduinoHannover/ZPB30A1_Firmware So: - there is a dedicated voltage divider for the power supply - the zener diode is used to detect an overload - the diode at the opamp negative input is to enable or disable the load Open source software will be pushed as soon as my spare ICs arrive (don't want to brick my working one).

[@voltlog]

It's so nice that someone finally accepted the challenge. Good job man! I'll do a video on this once you have something that works.

[@arduinohannover2286]

But I'm wondering: how did they get a resolution that high (seems to be 12 bit)? There is only a 10 bit ADC in there, so over the full range of the dividers / opamp we get a resolution of 0.035V (35.6 / 1024). The original one can measure 0.01V without a problem and is still very accurate. ± 2 digits is not too bad, but it seems there is room to improve (as the original firmware can do it too).

[@voltlog]

I am not sure but they could theoretically increase the number of bits by oversampling & averaging.

[@arduinohannover2286]

Yep, just done that, so reading 4 x n times and dividing by n gives us 12 fake bits :-)

[@LAHUERTITADEGUILLE]

Hola muy buenas estoy con una huertita casera en casa le invitó a pasar por mi canal y si quiere suscribirte se agradece mucho desde la isla de Lanzarote en la huertita de GUILLE un cordial saludo gracias🌴🌱

[@Doctorbasss]

Look like some people built a FIRMWARE UPGRADE to this electronic load with OPEN SOURCE stuff.. interesting! github.com/ArduinoHannover/ZPB30A1_Firmware

[@voltlog]

cool!

[@mrlazda]

"Augmenting Integrator" is most commonly referred as Type 2a compensator in power electronic. Little correction FAN PWM DRIVE is not driving fan with PWM signal, that circuit is low-side-switch buck converter (composed of L1, D6, C20, Q3 (and R29)).

[@voltlog]

+Miloš Lazović thank you for the info. The circuit regarding the fan control was unclear to me but it makes sense now, I should have seen it as a buck converter while drawing the schematic. It makes sense to have a buck in there to vary the voltage across the fan and not just the duty cycle as that will not work very well with the fan internal brushless driver.