I was common for old SMPS designs to require 5% minimum preload to get 0.1% load regulation error. your design seems to be lacking the dynamic range on current & voltage error sense to maintain low errorwith no load.
For some reason when I put a load of 5W the voltage is 23.7V and it regulated out the whole input voltage (170-400V).
When I put the load to 10W the output drops to 23.3V and it regulated at that through the whole input range
What could be causing this output drop? Feedback?
Depends on where you are measuring the o/p volts if you measure at the same point that the volt control loop picks off the o/p voltage the change will be smaller, if you measure at the load at the end of some wires you will see a droop similar to the one you are seeing...
I'm a bit confused by the term "5W load". One could imagine it is an intelligent load to draw 5W independent of voltage.
I never heared of such a constant 5W load.
But I assume the OP..
* either uses an ohmic load (with a fix ohmic value, lets say 115 ohms)
* or a constant current as load (lets say 200mA)
Depends on where you are measuring the o/p volts if you measure at the same point that the volt control loop picks off the o/p voltage the change will be smaller
Would it be by a significant amount if they were on the same rail as picture below, with the voltage loop picking off from the electrolytic point, and the load being connected across the output?
What i meant is that I have used and ohmic load and measured the output voltage, then worked out roughly how much power it is drawing, not specifically put a "5W" on the output.
Would it be by a significant amount if they were on the same rail as picture below, with the voltage loop picking off from the electrolytic point, and the load being connected across the output?
I presume you have a digital voltmeter. As far as it refers to possible voltage drops, why don't you answer the question yourself by measuring voltages in the circuit operating under load?
Quite generally speaking, there are different possible reasons why the circuit doesn't regulate correctly.
- The feedback doesn't sense the output voltage correctly (e.g. is affected by voltage drops)
- Your own measurement is flawed
- The regulator circuit doesn't work as expected (e.g. not regulate the average value due to inappropriate circuit design)
In latter regard, a statement like "I have a design that is very similar to the one in the demo circuit " is problematic. Nobody knows how the actual circuit is different from the original design and if the change may bring up the problem. Without exact schematics, the only serious answer is:
Find how the circuit works and what causes the problem!