Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Repair SMPS for cooktop

Status
Not open for further replies.

Cortezz

Junior Member level 3
Joined
Jan 9, 2020
Messages
30
Helped
0
Reputation
0
Reaction score
0
Trophy points
6
Activity points
332
I have a cooktop which SMPS is bad.
It should produce 24VDC but it only does 16V.
Here is a schematic with a very similar topology.
It works with a TOP254GN IC.

For first I tried to change 3 main components
as all other parts looks fine and as these 3
tend to fail on this panel as I read on forums:
- D6 Diode (original SMD mark: T4X5) -> replaced with 1n4148 kind of diode
- U1 TL431 (original SMD mark: 43A) -> replaced with a normal SMD 431 IC
- UFB1 Opto (original SMD mark: 181.002A) -> TLP185 and then HCPL-181-000E
But now the output voltages jumps so it's even worse when I started... :-(

My main question is:
Which part is now specific in such a design and which has to be chosen carefully?
My hint is the opto but I couldn't find the datasheet for the original one.
So I chose a standard one but I'm not sure about the CTR for example.

Could you give a hint or what are the methods of localizing such an error?

Thanks in advance for any help!
 

Attachments

  • smps_sch.jpg
    smps_sch.jpg
    65.5 KB · Views: 116

D6 may have to be a fast switching type, possibly a Schottky diode but a 1N4148 will probably do.
The actual output voltage is set by R5 and R20 but if I had to guess at the fault it would be either C20 or C35.
The Ic manufacturer recommends a PC817 optocoupler in that configuration. They are cheap and plentiful.

Brian.
 

R5 and R20 seemed OK (values: 89k and 10k, I checked them with desoldering them)
All the electrolytic cap has been replaced with brand new ones.
I ordered some new TOP254 ICs the will arrive next week.

The only blurry part for me is the feedback section based on the opto.
What colud be here the ciritcal parameter?
For example what kind of CTR should I look for at the opto selection?
They are from 50% to 600%. Or it's not relevant in this application..?
Or after replacing the opto is there some adjusments / fine tuning neccesary in the feedback loop?
Maybe with the resistors R32, R35, R122 and maybe C82?
 

I checked the controller IC datsheet (50 pages...) which I attched the
relevant sections into the first post but I'm still not perfectly
sure I understand the details about the feedback mechanism.
Here is the whole datasheet: mouser.com/datasheet/2/328/tophx_family_datasheet-1511305.pdf
If I get it, the PWM freq depends on the current fed into pin C via the opto transistor.
But it's not clear how exactly it should be designed as the current here is mixed with the supply current.
Is this an analog feedback or just a "2-state-mode" when output voltage exceeds a level?
Or the control current at pin "C" should be set precisly based on the CTR of the opto and the TL431 resistors?
Or how should I check that after replacing the TL431 and the opto-coupler that everything is OK?
 

Attachments

  • basic-functional-operation.PNG
    basic-functional-operation.PNG
    44.3 KB · Views: 77
  • block-diagram.PNG
    block-diagram.PNG
    136 KB · Views: 97
  • contol-current-sections.PNG
    contol-current-sections.PNG
    102.4 KB · Views: 87
  • control-currents-and-voltages.PNG
    control-currents-and-voltages.PNG
    181.7 KB · Views: 86
  • control-pin-current-vs-voltage.PNG
    control-pin-current-vs-voltage.PNG
    35.8 KB · Views: 80
  • control-pin-operation.png
    control-pin-operation.png
    390.9 KB · Views: 92
  • C-pin-description.PNG
    C-pin-description.PNG
    35.4 KB · Views: 75
  • error-amplifier-description.PNG
    error-amplifier-description.PNG
    69.5 KB · Views: 70
  • pwm-operation-by-control-current.PNG
    pwm-operation-by-control-current.PNG
    64.6 KB · Views: 84
  • multi-function-pin-schematic.PNG
    multi-function-pin-schematic.PNG
    71.5 KB · Views: 72
  • kexin-datasheet-control-currents.PNG
    kexin-datasheet-control-currents.PNG
    144.6 KB · Views: 78

Update: the replacement TOP254 has been arrived and it worked at the begining the 24V
come back but after a few minutes it was jumping again and even it went down to 0V.
Tonight I'll cut off the load from th SMPS and checkout the main parts.
I can't guess what can be the root cause as all the passive components looking good
(checked with a DMM) and all the active components (TOP254, TL431, Opto, Diode)
and the caps have been replaced...
Something collapses but it's strange as it was goind for ~15 minutes perfectly...
What can be the cause for such a collapse after so many replaced parts...?!
 

Hi,

All the electrolytic cap has been replaced with brand new ones.
"Brand new" does not necessarily mean that they are suitable. Voltage rating and capacitance rating is not enough information to buy such a capacitor.
As with all SMPS you need to choose capacitors suitable for high switching frequency and low ESR.
Don't use standard "mains frequency" electrolytics capacitors.

Klaus
 

Update: I cut off all the load on the secondary coils.
On the first turn on it was ok for a few minutes then is turns off...
After it gets warmed up (I don't know where...) this "won't turn on"
stablizes and sometimes after switching it off it turns on for a few sec.
I tried to change back the opto to the original one but it makes no diff.
Checking with the scope when it doesn't turn on the "C" leg is
in hiccup mode it's clearly jumping between the 4.8V and 5.8V.
The low ESR criteria, can this cause such an issue..?
 

I checked the PS in the morning and it was OK for more then 5 minutes (still withoud load)
so I could check all the voltages when it's working, I attached the voltage map.
And after a 5-10 minutes it went down. The primary 310VDC was ok but the switching was not working.
What should I check...? What can be causing this..? Something warms up and fails..?
 

Attachments

  • smps-sch-v2.png
    smps-sch-v2.png
    32.9 KB · Views: 91

Ehh, strange yesterday evening I get back to this project but now everything was stable...
There was't any turning down or instability. I even tried it with dummy load up to 500mA and it was solid as a rock.
I put iut back to the cooktop and so for everything is ok. I'm not sure wat was the last problem maybe it was just too late in the evening.
Anyways the last part I replaced between the surely bad state and from now was the original opto I put back.
It'd be great to know what exactly was the difference but I can't guess other the the CTR of the opto.
Now a real life test will be performed for days and I'll come back if anything is coming up.
 

Without knowing the specification of the original opto it's difficult to know if your replacement was suitable. They are however, very reliable. Typically they degrade in CTR over long periods but it is so small that it takes many years of operation to notice and in any case, being inside a feedback loop means and change in performance is cancelled until it becomes unsustainable.

I think the original device was probably a PC817 or equivalent as described in the IC manufacturers data sheet but possibly the one you replaced it with has digital characteristics or has a darlington output stage which can significantly change the way it works in a circuit of that type.

Brian.
 

BTW: I couldn't manage to precisely understand to design of the feedback part.
How it is done usually? What are the "corners", the criterias during a similar SMPS feedback design?
Is it voltage or current based? If the opto CTR can be vary in a large range how it's available to set these precisely?
This is a linear feedback or just a binary: if the output voltage is too much it is turning into high current mode and that's all?
Or the feedback current should be set up carefully for example to 2mA in OK condition and 4mA when the voltage is 26V..?
 

It uses voltage monitoring and linear feedback.

The principle is to convert the output voltage to a current through the opto-coupler LED. It could be done with a simple series resistor but it wouldn't be very sensitive because a large change in voltage would only produce a small change in current. The trick it uses is to use the TL431 to lift the bottom end of the LED to a stabilized voltage a little below the output voltage. The current through the LED is then determined by the difference in those voltages rather than the difference in output and ground, making it proportionately much bigger. The TL431 is a programmable voltage regulator with its cathode (top in the schematic) voltage being set by the voltage from R5 and R20.

The CTR across devices, even from the same production batch isn't very precisely controlled so the manufacturer will choose values supporting the worst case and also add on a safety margin. CTR does degrade slowly but to put it in perspective, with 100% on time at maximum rated current, it will drop by about 50% in around 75 years, not something most people worry about!

There are several types of optocoupler which is why you may have used one that wasn't suitable. Basically, they are a LED and a photosensor mounted so the sensor sees the LED light and it is all packaged in a lightproof plastic box to keep outside light out. The differences are in the sensor side, a 'normal' optocoupler has the light shining on the base junction of a single transistor, a high CTR type usually has two transistors wired so the second one works as an amplifier. A digital type has a threshold at which the output side abruptly switches from low to high rather than having a gradual change. It makes them more suitable for carrying isolated data streams but useless for applications where the output has to be proportional to the input.

Brian.
 

Thanks a lot Brian for the detailed explanation!
I guess I'll analyze this feedback mechanism in spice to understand it fully.

It's still not clear for example how the CTR is not so relevant if on the SMPS controller IC side the currents should be set up relatively precisly.

I chose a normal opto with 50-500% CTR. I should took a measerment from the original van regarding the CTR...
Or originally maybe it was a darlington type opto while I used a normal one as replacement..?
 

I tested it with spice (such a cool tool..! :)) and here are the results.
The simulation is a DC sweep type incrementing the output voltage from 0 to 26V.
I measured the voltage and the current on the TL431.
The results are certainly not so surprising as it is an adjustable zener.
But regarding this I dont see why the opto end its CTR could be an issue at all.
If I understand it correctly its rather a digital type of feedback as the TL acts very rapidly to a small voltage change.
So this system sets up itself to a specific output voltage and to set the feedback current to a specific value is not relevant at all.
Basically it works on "full power" under a certain output voltage level and turns off after it reaches a certain maximum level.
And between that few tenth of volts there is a linear region but it's not used at all.
Is this correct or am I missing something?
 

Attachments

  • schematic.png
    schematic.png
    5.1 KB · Views: 90
  • voltage.png
    voltage.png
    18.3 KB · Views: 87
  • current.png
    current.png
    20.5 KB · Views: 96

Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top