On 24/04/14 04:04, Uncle Peter wrote:
On Wed, 23 Apr 2014 13:08:37 +0100, Daniel
wrote:

On 23/04/14 08:02, Uncle Peter wrote:
On Thu, 17 Apr 2014 13:42:41 +0100, Daniel
wrote:

On 16/04/14 22:23, Uncle Peter wrote:
On Wed, 16 Apr 2014 13:01:00 +0100, Daniel
wrote:

On 16/04/14 06:11, Uncle Peter wrote:

My memory of teaching Power Supplies (about 15 years ago), the
constant
current was maintained but keeping the voltage developed across a
particular resistor constant, so, in theory, by increasing the
value of
the resistor, a lower current would develop the same voltage
across the
sensing resistor.

However, you would have to determine which resistor and which value
.....!

Thanks for the tip, I'll run it on a bench with a dummy load and test
some resistor voltages, then try changing one a little.

If you've got a circuit diagram of your power supply, and you can
determine which resister is the "sensing" resister, replace it (for the
time being) with a variable resistor/rheostat in the same ohmic
range as
the sensing resistor until you can determine what value resistor you
want to stick in.

This is a partial drawing of the circuit diagram of the power supply. I
can't see half of the circuit connections as they go underneath the
unknown microchip!
http://petersphotos.com/temp/PSU.jpg

O.K., that's more than I expected, thanks.

(I'm really guessing that the 280ohms you measured is the 282ohm
resistor next to it, but that's by-the-by)

No, there are two in series. They are marked as such. I didn't measure
any, except to determine which way round the figures or bands were
written when unclear.

The NPN transistor that you question mark as "Low Freq" ... I would be
expecting it to be as High Freq Switching Tranny, which would, somehow,
by "driven On/Off/On/Off by the IC in middle of the diag, causing an
On/Off/On/Off current to flow through the coil/Transformer, and these
pulses are transformer coupled across the coil/Transformer, to be
rectified by the diode and filtered by the capacitor and 4.7kohm
resistor across the output.

Are you sure this is a "constant current supply"?? Depending on the
arrangement of the pins of the coil, this could just be a fairly regular
variable voltage PS.

It's designed to have three CREE LEDs connected to its output in
series. If I connect different LEDs, or a resistive load like a torch
bulb, it behaves very differently (as though it's confused). It's not
behaving like I'd expect a constant current PSU to behave. It seems
designed specifically for LEDs and is changing according to what I
connect. For example when I connected three LEDs which have a lower
forward voltage to it then the expected CREEs, it gave out less volts
AND less current and lit them almost correctly. I would have expected
it to give more power until it hit either a voltage or current limit,
probably blowing the LEDs. Could it be "clever" like a NiMH charger and
is sensing how the LEDs are behaving? Open circuit it only gives out 1.5V.

Do you have access to an Ohmmeter?? If so, with the device switched off,
and unplugged, can you measure the resistance from each connected
terminal of the coil to each other terminal?? I'm really only interested
in connection/no connection, not the actual ohmic values.

This is the transformer close up:
http://petersphotos.com/temp/Transformer.jpg
This is the underside: http://petersphotos.com/temp/Transformer%20pins.jpg
The numbers on the photo correspond to the measurements below:
1-2=1.65 Mohms
1-3=open circuit
1-4=2.3 ohms
1-5=open circuit
1-6=open circuit
2-3=0.5 ohms
2-4=increasing reading, as though the ohmmeter is charging a capacitor
2-5=open circuit
2-6=open circuit
3-4=increasing reading, as though the ohmmeter is charging a capacitor
3-5=open circuit
3-6=open circuit
4-5=open circuit
4-6=open circuit
5-6=0.1 ohms

Are you right
in calling it a coil rather than me calling it a transformer!!

Yes, transformer. As in two coils together :-) I wrote coil for some
reason or other....

O.K., are you counting 1 through 6 starting at your drawings top left
pin on the coil and counting anti-clockwise??

If so, could it be that you are forward-biasing the diode and you're
charging the cap connected to pin two of the coil?? Try re-measuring
with ohmmeter connected other way around to reverse-bias the diode.

Probably same-same with pins 5 & 6. Measure both ways.

I'm guessing pins 2 and 3 are not actually connected to the other
transformer windings but may or may not be connected to each other via
another winding on the transformer (0.5 ohms). Are you sure pin 3 is not
connect to anything outside the transformer??

Pins 1 & 4 primary winding, pins 5 & 6 secondary winding plus, maybe,
pins 2 & 3 another secondary winding

If what I thought was a variable resistor is just another 280 ohm (or
1.8kohm) resister, is there actually a variable resistor anywhere in the
circuit??

Is there any Brand Name and model number which might help find a circuit
diagram somewhere online?? Any info on a back panel/whatever??

Daniel