Output of Aurora 'constant voltage LED driver' (SMPSU)

[bernardgreen]

, without looking to see what one has actually got.

The average person who is buying lighting equipment does not have access to an oscilloscope

 

[JohnW2]

The average person who is buying lighting equipment does not have access to an oscilloscope

True - but, by the same token, the average person buying an 'LED driver' is going to be using it for supplying LED lamps/bulbs, in which case something supplying, say, 20-30V high frequency pulses (averaging to 12V), whilst not ideal, is probably essentially fit for their purpose.

As I've implied, unless I'm missing something, I'm not at all sure why Aurora are describing the output of this device as "12 V eff". How much cleaner would it need to be for them to call it "12V"? As I've said, many/most traditional PSUs will have more 50Hz ripple than the very small hf 'ripple' in this device.

[In fact, I'm about to do another experiment (something overlooked yesterday!) - so watch this space!]

Kind Regards, John

 

[bernardgreen]

say, 20-30V high frequency pulses (averaging to 12V), whilst not ideal, is probably essentially fit for their purpose.

The problems with pulses are

(1) the LED elements may be takein current pulses which could lead to degredation and / or early failure of the LED elements

0v--LED--LED--LED--Resistor--12 volt.

The LEDs will have about 3 volts per element, total 9 volts leaving 3 volts across the current limiting resistor. LED current = 3 divided by R

0v--LED--LED--LED--Resistor--24 volt.

The LEDs will have about 3 volts per element, total 9 volts leaving 15 volts across the current limiting resistor. LED current = 15 divided by R

So the 24 volt peaks will be driving 5 times the rated current through nthe LED elements/

 

[JohnW2]

The problems with pulses are ... the LED elements may be takein current pulses which could lead to degredation and / or early failure of the LED elements ....

Yes, I realise that, which is why I wrote "not ideal" but I wonder how much of a problem it actually is in practice.

Furthermore, if one uses a 'driver' from the same manufacturer as an LED light, and 'recommended' for use with that light, then I imagine that they would be obliged to quote a life expectancy of that light when used with their 'recommended' driver, wouldn't they?

Also, I must say that I wonder how common these 'high voltage pulse' devices actually are - although I'm often hearing about them, I don't think that I have ever personally met one. I'm not even sure why (if?) they are produced, since, on the face of it, I can't see why it would necessarily cost any more to have lower voltage pulses with a higher mark-space ratio.

Kind Regards, John

 

[JohnW2]

To get an idea of the size of the issue, leave the scope earth where it is and put the probe on the earth point as well.

Very interesting! Qualitatively similar (but now pretty symmetrical around 'zero'), but with a much higher frequency component (~ 30 Mhz) superimposed on the ~70 Hz 'pulses'. If I touch either or both sides of the PSU output, or even put my hand on its case, the amplitude drops by about 25-30% (all that with an ~360 mA load).

[In fact, I'm about to do another experiment (something overlooked yesterday!) - so watch this

I should have thought of this yesterday, in view of what happened when I 'touched things', but I think I now have even more evidence that what I'm seeing is essentially 'spurious' (and certainly of no consequence).

Having thought to try, I now find that if I separately and explicitly earth one side of the 12V output of the PSU (the same side that is already earthed via the scope), the highest frequency (~30 MHz) component vanishes, and what is left diminishes in amplitude by about 50%.

Kind Regards, John

 

[aptsys]

Further to the above, and as one might expect, when one increases the load (this is about 360 mA, i.e. about 4.3 W), things get somewhat 'cleaner' - that 'pulse' is around 30mV at its peak.

What I don't fully understand is that, although I don't think it's a problem as it is, I am not having much success in significantly 'suppressing' that pulse with RC filtering ....

View attachment 159186

View attachment 159185

Kind Regards, John

That looks like user error with your scope probes. That ringing isn't real.

 

[JohnW2]

That looks like user error with your scope probes. That ringing isn't real.

I'm not sure what sort of 'user error' you have in mind (there's not a lot of scope for much variation in what one does with scope probes) but, yes, as I said in my most recent post, it certainly does seem that much/most of what I'm seeing was essentially 'spurious', rather than 'real'.

I could understand much more easily if the scope (with a high input impedance) were being connected to a very high impedance source - but, in my words you quoted, I was talking about a situation in which there was a 33Ω resistor across the scope probes.

Kind Regards, John

 

[aptsys]

I'm not sure what sort of 'user error' you have in mind (there's not a lot of scope for much variation in what one does with scope probes) but, yes, as I said in my most recent post, it certainly does seem that much/most of what I'm seeing was essentially 'spurious', rather than 'real'.

I could understand much more easily if the scope (with a high input impedance) were being connected to a very high impedance source - but, in my words you quoted, I was talking about a situation in which there was a 33Ω resistor across the scope probes.

Kind Regards, John

The ground lead on a scope happens to be an excellent antenna and this is a characteristic ringing that you'll always see near to a potentially noisy circuit. You need to remove the ground lead from the probe use the little springy ground probe to keep the loop area to an absolute minimum.

 

[JohnW2]

The ground lead on a scope happens to be an excellent antenna and this is a characteristic ringing that you'll always see near to a potentially noisy circuit.

Maybe I'm being dim, but I still struggle to understand how that could/would work with a 33Ω resistor between the scope's input and it's 'ground lead'. The 'ground lead', by the way, is the sheath of coax.

What are you suggesting will be doing the 'ringing'?

You need to remove the ground lead from the probe use the little springy ground probe to keep the loop area to an absolute minimum.

Could you perhaps clarify that a bit?

Kind Regards, John

 

[aptsys]

Maybe I'm being dim, but I still struggle to understand how that could/would work with a 33Ω resistor between the scope's input and it's 'ground lead'. The 'ground lead', by the way, is the sheath of coax.

What are you suggesting will be doing the 'ringing'?
Could you perhaps clarify that a bit?

Kind Regards, John

It requires some understanding of RF, but think of it this way. A basic loop antenna is just a DC short, which is what you're thinking of. At high frequency it becomes a high impedance thus you're able to develop a voltage across what would appear to be a low impedance.

The ringing is caused by the LC tank circuit of your L (ground lead) and the C (scope probe).

 

[JohnW2]

It requires some understanding of RF, but think of it this way.

Ageing brain/memory apart, I can probably cope with that. I've been playing with RF circuits (and antennae), from 1.8 Mhz up to 1,300 Mhz from a time which was quite possibly before you were born (although they weren't "MHz in those days) - although I may be under-estimating your age

A basic loop antenna is just a DC short, which is what you're thinking of. At high frequency it becomes a high impedance thus you're able to develop a voltage across what would appear to be a low impedance. ... The ringing is caused by the LC tank circuit of your L (ground lead) and the C (scope probe).

Fair enough, and I suppose you must be right, although I am having some difficulties in getting the numbers to work.

I'm talking about ~600 mm of (undoubtedly low quality) coax. Regarded as a pair of conductors, using guesstimates of its LC characteristics and including the 25 pF alleged input capacitance of the scope, that seems to correspond to a resonant frequency of 50-100 MHz. Similarly, if one regards it as a quarter-wave coaxial resonator, it's intrinsic resonant frequency would be ~125 Mhz; that will be lowered by the input capacitance of the scope but, without doing some reading, I'm not sure how to calculate by how much.

Given that my guesstimates may be way off, that resonance might account for the highest frequency (~30 MHz) component of what I've been looking at. However, the 'main component' is an order of magnitude lower in frequency than that (~2.5 MHz), and I find that much more difficult to explain.

Anyway, the bottom line is that, if most/all of what i'm seeing is 'spurious', then the SMPSU I'm looking at appears to have a remarkably clean and stable output for something described as "12 V eff"!

Kind Regards, John