Hantek PPS 2320A Tear-down and Repair Part 1 June 16, 2014 at 11:53 PM
I finally decided it was time to purchase a half decent lab power supply. After a bit of research and this great review by Hacked Gadgets I settled on the Hantek PPS-2320A 3 channel bench PSU.
Being one of the cheapest 3 channel supplies around, I knew there would be issues.. But hey! Things get boring without a challenge! Keep reading to hear the story..
I purchased the supply via one of my Chinese suppliers. The price was good and I had it air freighted along with a bunch of other stuff we had on order. Unfortunately the air couriers took their toll and the supply was dead on arrival… No problem, open her up and hope that whatever the issue is it’s repairable. Worst case would be something critical like the entire processor board being fried, that would be a huge issue.
After whipping the top off I discovered that the toroidal transformer (Nice! good sign for a decent supply) had broken off its mounting plate. Fortunately other than a bit of bent sheet metal this had dislodged a connector from one of the front panel boards. After reconnecting the connector the supply powered up and operated correctly (mostly… )!! 😀 😀 The story does not end here though! Pics at the bottom 0f my tear-down + repair of the bent sheet metal.
Shortly after getting the power-supply running I discovered another issue.. While Channel 1 works normally Channel 2 will not supply more than ~65mA of current, it seems to constant current limit at around this value no matter what. Even when the supply should be in constant voltage mode the current still saturates to ~65mA and both the C.C. + C.V. indicators are on. Interestingly the output voltage displays as whatever the set voltage is (as expected in constant voltage), yet when measured at the output terminals the voltage is lower and essentially whatever voltage achieves the constant current of ~65mA.. Time for some more trouble shooting! (video showing the problem embedded below.)
I decided a Tear-down and inspection was in order to see if the fault was due to anything obvious. The gallery at the bottom of this post contains my (admittedly bad sorry, was just snapping pics as I pulled it apart.) pics of the supply. Along the way I straightened out the sheet-metal that had been damaged in what must have been some very rough transport, I also found a broken 100nF cap, broken 1N4007 diode and two dry solder joints on two other diodes. Repairing these components didn’t solve the problem and the remainder of the construction didn’t seem too bad for a cheap supply, I didn’t find any dangerously bad construction (IMHO). I also swapped out the ch1 and output indicator LEDs for some blue ones to make the display a little more readable as I found the sea of read lights was a bit disorienting. Next up was figuring out the circuit in order to solve the channel 2 fault.
Video showing the problem
Inside the supply are 6 circuit boards:
- Input board with the rotary encoders and tact switches
- Display board with LED indicators and 7seg displays, all driven off a few I/O pins and a bunch of 595 shift registers
- A linear power board with two big heatsinks. This board takes the taps from the transformer (which is connected to the mains input via the power switch) and rectifies + regulates the 3 channels as well as powering the main control PCB. This board has no logic IC’s at all, only regulators, transistors and a small SMPS section that likely deals with powering the logic board.
- A logic board that controls the supply which I will talk about more in depth below.
- Opto-isolated USB-TTL serial board for USB comms with a PC
- What appears to be a small bodge board with two relays and a single control line from the logic board that switches in/out two resistors to the power board (I think..)
The Logic board has two C8051F410 micro-controllers as well as a bunch of TL082I Opamps. At first I assumed that one of the micros would be dealing with front panel inputs / USB and the display while the other would be controlling the 3 outputs, but on closer inspection I found that the display is connected to one of the micros while the switches + USB are connected to the other. I’m still not sure which channels are connected to which micros but I assume the designers are using these micros for their 12-bit DACs and other analogue circuitry to do all the control, this leads me to believe that each micro controls a channel and that they communicate with each-other somehow.
It seems to me that channel 2 has a problem with it’s current limit circuit so I started fault finding by tracing this circuit out. The output current appears to be measured by the large coiled wire current shunts, each of these has a sense line meandering into the mass of opamps on the control board. I assume that the current limit is implemented by some sort of comparator with a reference value which is likely created with one of the DACs on the micros, I mapped one of the DAC pins from the micro closest to the ch2 section but this lead directly to one of the large multi-pin connectors. There are two of these connectors that lead back to either side of the power board, it would seem that each connector controls one of the main output channels. Hmm.. This didn’t help as I was hoping all the current limiting would be performed on the logic board and I could trace the fault to an opamp or something. I’m not even sure if that line has anything to do with the current limit, it could even be the voltage reference for the voltage output!!
This is as far as I’ve come for the moment, ideally if I could get hold of a full circuit schematic that would make the problem much easier to trace. I’m going to post on some online forums as well as email Hantek and see if I can get some more hints where to look, but failing that I might take the time to draw out the entire circuit (ahhh don’t know where to find the time for that!! ). At least in the mean time I can use channel 1 & 3! If you have any ideas or resources, please feel free to let me know!!
Gallery of the Tear-down, Sheet metal repair and repair of some other misc components.