qyuick question re dc converter

[njkmonty]

i  etched a little board from the schematic below , it got the desired 18v i was after.
However once i connected it the a 10 part transistor booster to test (which had worked previously perfectly on 9v) 
it has some weird noise.
if i place my hand near or over the power circuit it would change and when i physically touched it  it would slow to a motorboat slow weird noise?
identify by this the noise was affected by my hand around R2,
i by chance placed a 100nf cap parallel with it!  (R2)
its now dead quiet no noise .
i double checked all joints values and resoldered the joints.
Im grateful its working well, now
but, can anyone explain why this has happened???

[njkmonty]

.

[antonis]

Although I don't like chokes I'll let it be this time..

You've build a switching circuit with a fast Schottky diode - faster maybe than required..
Placement of 100nF cap across R2 "smooths" its switching behavior - it "softens" on-off turn time interval..
(you could have same result by placing the cap across D1 - it's frequently met on PS rectifier diodes..)

[njkmonty]

Thanks  heaps for the explanation !
That has helped a lot
I used 1n5817. However  their origins may be from eBay !
Their leads look very thin compared to the regular ones I've previously used.
I will have a squiz and see if swapping them out makes any effect!
Although the layout asks for the choke, I would love to learn about your opinion of why you don't like them! I've only used them a couple of times and would be interested to hear your thoughts
Cheers again

[Rob Strand]

Could be going unstable:
http://www.ti.com/lit/an/slva274a/slva274a.pdf

100nF seems a very large value to cure that even with the low valued feedback resistors.
1n to 10nF would be more ball-park.

[njkmonty]

cheers

[Rob Strand]

Although the layout asks for the choke, I would love to learn about your opinion of why you don't like them!

The shielded ones are far less problematic than the unshielded ones.

Shielded:




Unshielded:




Mixed: (so you can different types)

[njkmonty]

perhaps thats the issue!

the schematic asks for 1n5817, however i treid it with 1n4007 , but no change,
so i put the 1n5817's back in and tried a 1nf cap,   made no difference , so its back to 1n5817 and the 100nf  and quiet, i might order some shielded ones and see if any improvement

[Rob Strand]

perhaps thats the issue!

You might have more than one issue.

Did you try a 10nF?   

As you can see from the PDF I posted, working out he cap correctly is not a simple task.  It also depends on the ESR of the output cap.