I don't see anything obvious in the schematic I recommended you fit the preload resistor.

Would you confirm the specification and can you post the PCB layout. You mention "oscillation" if this is visible on the DVM then the LinkSwitch-TN may be in auto-restart. Do you have a scope which you can use to monitor the output voltage and post the image here?

Cheers

PI-Chekov

Hi PI-Chekov

Many thanks for the response.

I have sent the pcb artwork as PDF files for overlays, top and bottom layers, do you need it in another format? It contains the extra but unused components.

To recap, this circuit is simply a battery charging circuit. The controller schematic has been stripped down of the optional stuff which is not soldered on for the battery charger. The PCB has only the components mounted as in the schematics. The micro simply monitors the battery voltage to determine charging complete and switches the pass transistor Q1 via Q2 to turn off the AC to DC voltage generator using the LNK306.

My problem starts as soon as I turn on Q2 which turns on Q1, the regulator does not regulate and drops to about 2.7V (not around 5V) and keeps changing as if it is resetting (which is what you mentioned as well). This was tested with the micro and manually using a jumper wire without the micro. Even without the battery in place it does the same thing, drops to 2.7V.

The 1K5 load resistor is always in place otherwise the voltage rises very high with no load.

The circuit works perfectly on a vero/strip board. However one difference with the vero board design from the PCB schematic is that it does not have L2 and C7 and C6 is 47uF. I even removed these components from the populated PCB and shorted out L2 to complete the current path and it made no difference, the regulator stops regulating as soon as I turn on Q2.

I definitely feel that the layout is a problem because I made up the circuit on vero/strip board and it worked properly.

Does the layout make such a huge difference that the LNK306 will keep restarting?

Does using leaded resistors make a difference in the feedback components, because the PCB version i use 0805SMD resistors while the strip board version uses leaded type?

Many thanks for helping.

---- Priyend Somaroo c/o Vardaan Enterprises, www.vardaan.co.za for Electronic Research and Development in South Africa

Hi Priyend,

I'll take a look later today (on vacation this week) but a quick answer on the use of SMD vs leaded. This is not an issue - layout is important but this sort of problem doesn't sound like it is due to that.

Any chance of some waveforms of the output voltage - can you see if the supply is in auto-restart. From datasheet:

In the event of a fault condition such as output overload, output
short, or an open loop condition, LinkSwitch-TN enters into autorestart
operation. An internal counter clocked by the oscillator
gets reset every time the FB pin is pulled high. If the FB pin
is not pulled high for 50 ms, the power MOSFET switching is
disabled for 800 ms. The auto-restart alternately enables and
disables the switching of the power MOSFET until the fault
condition is removed.

Cheers

PI-Chekov

Hi PI-Checkov

I cannot capture the waveform because of the direct mains supply.

I have the Bitscope product BS100U but the Bitscope support people were not able to help me with connecting up the device safely (without possibly blowing it up) unless I use a 100X isolated probe which I don't have. I personally feel they are not familiar with direct mains circuits because they were "surprised" by the "dangerous" circuit.

I am pretty sure that if I use a laptop with only the battery supply, and the Bitscope powered by battery as well then I should be able to take the readings. Is this correct?

Is my thought process correct : Since the laptop and scope are both powered from their own battery supplies I can then attach the GND pin of my direct mains supply circuit to the scope without any hassles and risk of blowing or burning things up?

On another note... I hope you enjoy your vacation.
Cheers
----
Priyend Somaroo c/o Vardaan Enterprises, www.vardaan.co.za for Electronic Research and Development in South Africa

To recap, this circuit is simply a battery charging circuit. The controller schematic has been stripped down of the optional stuff which is not soldered on for the battery charger. The PCB has only the components mounted as in the schematics. The micro simply monitors the battery voltage to determine charging complete and switches the pass transistor Q1 via Q2 to turn off the AC to DC voltage generator using the LNK306 work from home

My problem starts as soon as I turn on Q2 which turns on Q1, the regulator does not regulate and drops to about 2.7V (not around 5V) and keeps changing as if it is resetting (which is what you mentioned as well). This was tested with the micro and manually using a jumper wire without the micro. Even without the battery in place it does the same thing, drops to 2.7V loans.

The 1K5 load resistor is always in place otherwise the voltage rises very high with no load.

The circuit works perfectly on a vero/strip board. However one difference with the vero board design from the PCB schematic is that it does not have L2 and C7 and C6 is 47uF debt consolidation. I even removed these components from the populated PCB and shorted out L2 to complete the current path and it made no difference, the regulator stops regulating as soon as I turn on Q2.