24VDC/4A PSU repair help need

[Chriss]

Hi!
I have this psu on my mini cnc machine.
Unfortunately I realized a problem on my machine.
The drill motor is so to say "jogging" when I try to start the motor. Then I have to stop the motor and again try to start and so on so forth until it starts to spinning as it should.
The motor is a 24VDC motor. On a simple PCB fabricating the current never goes over 1A, just on the startup, then it will rise up to a bit over then 2A.
The PSU is a 24VDC/4A.

After checking around the control board of my cnc I found does the PSU is for some reason the problem. It will cut out the power from the motor and then power on and so every maybe 1sec and it looks like the motor is jogging. Start/Stop/Start/Stop....

I checked the whole PSU and didn't found any problem, it is a crap from china I'm realized that very quickly, but it was worked almost a year.
When it starts to deliver enough power to the motor and the motor spins up then there are no more problem,
even under heavy load.

What I suspect are 3 cap's 330uF/35V even if I checked them with ESR meter.
They are in parallel on the output of the PSU and feeds into the motor so to say.
They are for sure no some sort of filters etc. They must act as a buffer or so.

I think my PSU act's like if the output is in some short.
I checked the motor with another good PSU and it works like a charm.

If some caps is in some sort of short then it would cause a power fall out I suspect.

Can somebody prove my theory?
Or is there maybe something other in problem?

Check the pictures.
You will see on the PSU-PCB1.JPG left side are 3 caps, they are the cap's I think are bad.
The little chinese cut the wires first and then soldered them back to the pcb, they realized too late does one pad of the wire is nowhere connected .

Thank you very much for any suggestion.

[PKTKS]

Looks very simple
some kind or forward converter.

I would clean that crappy "stuff" over some components

and check the semiconductors in HOT working condition
they may be faulty - a very common issue in reactive loads

Paul

[Gyro]

It's a simple flyback converter.

It has obviously been 're-purposed' from some other application, but for a crappy Chinese PSU it actually looks surprisingly well implemented. At least they've kept the creepage / clearances acceptable (can't say for the transformer windings, obviously) and they've kept the thermistor, filter components and a 'real' fuse. The tacked-on mains wires need attention. 24V 4A looks wildly optimistic though.

Output capacitors would be a prime suspect, a total of 1000uF for 4A rating is pretty lean. Even though the motor average load is only around 1-2A, from stationary rotor startup, I bet it peaks briefly at >4A.

The quickest check would be to an a extra low ESR cap in parallel with the existing ones (as long as you can do it safely without exposed mains).

Adding a low value series resistor (in the 0.1 to 1R range) in series with the motor would reduce startup current spikes without significantly affecting normal running - it might get a bit more life out while you look for something better.

P.S. I can't see if those two blue EMC Y caps are actually connected to anything on the secondary side - there looks to be a pad and then a couple of chopped wires.

[mariush]

I would remove the metal heatsinks on the sides to make more room.
I would remove as much of that goop that's poured all over the components (that cream yellow stuff, scrape it off)... it's there only to prevent vibrations and mechanical shocks from moving parts out of the circuit board when it goes through reflow oven (during manufacturing) and during day to day operation.
Depending on the chemical composition, some of those goops can become slightly conductive after long periods of time, as they dry out. So you may have some resistors under that goop that have their resistance slightly changed due to the paste sitting on them, or some high voltage diode being partially on all the time etc...
Scrape it off.

With the heatsinks removed, you can check the parts (the mosfet or transistor that pulses the transformer on the input on one side) and I guess the two diodes that rectify the transformer output on the other heatsink.
Without heatsinks you could use continuity mode on your meter to check if they're shorted or not (though if you still have voltage on the output that's probably not the case)

Gently flex the parts a bit... it's rare but I've seen a pair of diodes in a TO-220 package that was cracked - when it got hot enough, the part would randomly fail.  Also had one of these parts (a transistor) where the lead was broken right where it entered the package (so it randomly broke contact, with heat expansion and cooling)

[Chriss]

I re-checked everything and didn't found any bad components.
I got a new idea why the psu is cycling under startup.
Maybe the TH1 (that is that one guy right up on the PSU-PCB1.JPG) it is directly connected to the main wire, and after that part it is feeding to the bridge rectifier diodes.
It should serve as a inrush current limiting device - NTC.
Maybe he is out of spec by time...

I can't approve the original value because I can't find any info about it.
The labeling on it is something like WTR 03D100.

I tested the PSU with halogen lamps in series and they started also pulsing until I turned them on and off several times but the main was not interrupted on the psu during the lamps on/off cycling by me.
Ones they started to shine everything was ok.

That look's to me like a protection of inrush current problem...

[cncjerry]

Keep in mind that stepper based CNC systems are very susceptible to noise.  If practical, I would switch to a linear supply.  I use large toroidal transformer supplies on my machines and they aren't all that expensive, at least not what you would think they would be.  So instead of trying to fix that one, I would just replace it and be done with it.  It must be defective anyway.

[Chriss]

Yes, you are probably 100% right, I think also like you.
I was just wondering what could cos this problem suddenly, almost after one year of working without any-mini problem...

I have a nice & good quality 30V/6A bench lab PSU and I can hook up the cnc to it to work with.
Maybe I will use it through that because I didn't use the cnc so often.

Thanks guys for suggestion and ideas.

I always like to hang out with good peoples.

My best regards.

[shakalnokturn]

Check the power components are tight enough on their heatsinks (nothing to do with your problem though)
Make sure you test the primary side electrolytics (4 in total), the ones right next to the heatsinks are the first suspects (you'll sometimes see a SMPS full of 105°C electrolytics with just one small 85°C one up against a hatsink...), don't forget the horizontal one hidden next to the optos.
Worth checking the large cap also they fail quite fast in some cheap PSU's.

[montarbo]

Check which one of these caps is for the supply voltage for PWM chip, most probably the little one beside the big green filter cap. The symptoms u r experiencing indicates the chip supply can is failing.
If not then some cold joints to search for.

[Chriss]

Ok, I didn't found the fault for now.
I checked all the casp and they all ok. I was a bit suspicious with the cap on the rectifier bridge in the primary side and I put a new cap just to be sure.

The problem is still the same.
I checked the diode and the FAT too on the heatsink. Also applied new thermal paste between the part and heatsink.

I looked into the datasheet of the PWM IC here: https://www.onsemi.com/pub/Collateral/NCP1200-D.PDF and I realized:
When the problem occur, the voltage on pin 5 of the IC falls down near to 6v. Better to say it is pulsing around 6-7v. When it is operating as it should then I have 12.9V VCC on the PWM IC.

The cap which is in series with the vcc pin of the IC is in good condition (C9 on page12 from the link).
I have some sort of feeling does the whole power source is collapsing which is delivering the power to the IC.
What I not really understood from the datasheet is the capability of how the IC sens a short and powers off if a shortening occurs in the circuit.

"TO PWM IC" I mean that is the winding which feeds the PWM ic circuit with power.

[montarbo]

Pin 5 is mosfet driver. Question is what voltages on VCC pin 6 do u measure during abnormal and normal functionality. The PWM ic in ur case get is supply through another winding than shown at datasheet which is the rectified by a single diode, sometimes two in series with resistor and a smoothing cap which C9 in data sheet stands for. Change the cap and the mentioned diode then report.   

[Chriss]

Oh!
Sorry! My fault in writing in my post before.
I interchanged the PIN5 and PIN6, I would write PIN6 but I wrote PIN5.
Yes, the PIN5 is the driver pin for the FAT.
Pin6 is the VCC pin.
Everything is the same as I wrote it above just not PIN5 but PIN6...
Sorry for this.
So, the values between the PWM IC VCC pin6 and GND pin4 are:
1. Good condition: 12.9V
2. When the unit is not working: fluctuate between 6V - 7.5V

[Chriss]

I think I figured out the problem in my PSU!

Here is a small description:
I found a 15v zener diode in the circuit which delivers the voltage to the VCC pin on the PWM IC.
It is connected on the small transistor base and act like a voltage regulator.

Whenever I hade the problem with my motor to spin up the base of the transistor was on 0V and no correct
Vcc voltage was delivered to the PWM IC.

I was checked with DMM the zener diode and I didn't noticed any problem, but I come up with an idea, if the ZD
fails in any way and he short out the base of the transistor to gnd then I would have 0V on the base.
So, I decide to change the ZD however.

I'm playing with my CNC for hours now and no one time comes up any problem with the drill motor starting.
Not a single time I had any issue since I changed the ZD.

The lines on the pic represents the current path of that suspicious circuit with the ZD.

However, I will playing tomorrow with the CNC machine again and will see if fails, but I think no.
Thank you for everybody for your time you took to write and help me to solve my problem.
I really appreciate that.

My best regards.