Check the LNK306D design to see if it will deliver the expected 320mA
Could you check the attached design to see if it will deliver the expected 320mA at effectively 12V. The V3P3 is 3.3V and is tied to neutral. The logic ground is generated using LT1964. The 3.3V and -8.7V are used to turn 12V relays on and off (8 of them).I have attached the PI expert file and the circuit diagram.
The LNK306D stops working when two relays are powered on. All relays on should equate to 280mA. The logic takes max 20mA.
Much appreciate your response.
Comments
Tried a resistive load this had the same effect.
Lowered the AC voltage(50hz) to 87VAC and was able to switch all relays on.
With the relays on I raised the voltage and they dropped out, IE supply shutdown, at 107VAC.
We need this to work for both US and UK voltage systems.
Thanks.
The only thing I can think of is noise injection that causes the chip to stop switching. You are going to have to check you layout and make sure that switching current paths are not too long
Also from your schematic I have the following comments
you seem to have too many capacitors in this circuit...
What is the purpose of C73, C77 and C78? Remove them
Make sure C76 is right at BP and S pins.
C75 is ok but what is the purpose of C74?
Try to follow the concept of layout around the LinkSwitch-TN part as shown in
http://www.powerint.com/sites/default/files/PDFFiles/der49.pdf
The purpose of C73,C74,C78 and C60 was I believe to reduce emmissions. These were on the original reference design from the meter chip manufacturer, although they were no fit components. With them removed it works ok over the full range. I did however try leaving some of them in and the only one's to make a difference were C73 and C60 which are across the diodes. I will leave out as suggested C77,C78,C73,C74, and C60 and see how EMC testing goes. The problem now is that at full load there is whistling, what is the best way of minimising this?. I believe it's due to the varying frequencies at full load where the fet is not always switched on to get the regulated output. Also what is the best way to minimise the temperature of C53( it is a low ESR cap).
Regards,
The audible noise is because of cycle skipping at low loads. The inductor is causing this audible noise due to magneto-striction. It vibrates and together with the PCB it causes audible noise. Are you using a off the shelf inductor?
About the cap getting hot - make sure that the ripple current rating of the cap is within the ripple current flowing through it.
The inductor is an off the shelf device (Panasonic ELC10D122E(Farnell part no 1749078)). This is within the resistance value required,and will fit in the very restricted space I have. Presumably I need to try different types of inductor to find the quietist?. I Note that you say noise at low loads, but I am experiecing the noise at full loads? It runs at a lower constant frequency at low loads and then as the load increases the frequecy increases then there is cycle skipping to generate the noise. The capacitor is a PANASONIC EEETP1E221AP(Farnell part no 1751942) and is within the ESR and ripple requirements.
Do you agree that the components are within spec? If not what would be the best alternative/s.
Thankyou for your quick response.
Regards.
The design looks good.
Have you tried powering on the circuit with a resistive load? That will give you an idea whether it is the inductive nature of the load thats causing the problem.
Also, is this problem occurring at a specific line voltage?