Hi Everyone,
I recently bought this Multicomp Pro MP710503 PSU, it appears to be a Multicomp rebrand of Korad KA3005, but sold by CPC/Farnell in the UK.
Edit - this thread makes more sense if you can open the schematic added in post #3
It had a fault that upon switch on, you can set the voltage and current, but enabling the output it would immediately go into "Constant Current" mode, tripping the Over Current if set.
After fault finding it further, i think its a very unusual fault/failure mode which i havent read about on any of the eevBlog threads on this PSU. Although there are some mismatched component references i was about to use a KA3005 schematic from here to fault find it. It was the closest i could find.
I initially found that "OCP" was always high, which was triggering the uC to go into the Over Current Mode. When i traced this signal back, i found although there was no load current, U6A is slightly biased with a 1M resistor to +3.3VDC which gives a slightly positive output. This is then routed to the inverting input of U7B. So although there was no output current to the PSU, Pin 6 (inverting) was greater than pin 5 (Non inverting) and was causing Q2 to switch off, causing the OCP high state.
When i shorted OCP to ground (to force the PSU to run in normal mode) it did not have any current drive. I could find the Voltage reference (Pin 5 of U6B) but there was no Current voltage reference (U7A pin 3).
I cut the track leading to U7 Pin 3 and bridged it to the Voltage reference, this got the PSU working, however it meant that as you adjust the voltage reference, the current limit would also change. meaning at 31VDC you would get 5A, but at 0.1V you would get almost no amps at all.
This gave me confidence that the output drive circuitry and the pass transistors etc all worked fine. I did some loaded DC tests and for a solid half hour or so, there were no faults (other than the lack of current limit control).
Looking further into it, both Voltage and Current references are generated by 3x 8 Bit shift registers, data comes into the 74HC595, which is a Serial in, 8 bit parallel out Shift register, as the data gets clocked through, it alternates between Voltage reference and Current reference R2R ladders, starting at the LSB, which all makes sense, it allows both current and voltage references to "build up" nearly at the same time and from the least sig bits to gradually drive to the correct limits. the serial data gets passed through all 3 shift registers as there are 3 daisy chained. approx 24 bits, so if my understanding is correct 12 bits ADC resolution per reference.
However, what was happening was every single 74HC595 pin which was connected to the "current reference" R-2R ladder,
stayed at zero.So although a long string of serial data was getting fed into the "serial input" of the first Shift register, the voltage pins would generate the correct reference, being 2.745VDC for 30VDC output. but there was no current reference!
I lifted pin 14 of the first 74HC595 (data in) and using a 10K pull up, i was able to force both Current and Voltage references to their maximum. This pull up meant data in was always "1" and the Shift Register Clock from the uC would still be driving the correct number of pulses to latch the data through the shift registers. essentially filling both 12 bit R2R ladders with the maximum values.
In doing this i did infact get 3.3VDC on both current and voltage references. So this says to me that the R-2R ladders and the 74HC595's all work fine.
But for some odd reason, the uC is not setting the correct current reference value. As this stays at zero, it then means that the tiny bias voltages with U6/U7 will then force the uC to think there is a Over Current condition and trip out the PSU.
For the time being, i have bodged a dirty analog reference together, 1.5kOhm resistor, with an LM4040 2.5VDC reference, (approx 6mA bias current from +12VDC supply) results in about 4.9A Output from the PSU, however i would not recommend this as when the output is off, you still get upto 0.5A output current as the current reference is not inhibited (in normal operation the uC would set it to zero)
Its very odd as the serial data is one wire into the 74HC595 shift registers, i would expect either both references to work, or neither to work. The fact that the uC is only outputting the Voltage data, i wonder if there is a defect with the code (corrupted) or the uC itself. perhaps it is looking in the incorrect memory location for the target current ADC value. but in the mean time, the PSU is kind of working despite missing its over current protection.
I wonder if anyone else has seen this issue? the PSU itself is nearly new btw still with its Plastic coating on the front panel. Date codes indicate late 2020 early 2021 built unit.