Good by quartz halogen bulbs, are we ready?

[Diyisfun]

If the lamp is marked 230V no.
If the lamp is marked 12v, there may still be stock for a while, but eventually its worth replacing it. You would need to find the transformer first though

Its 230v

 

[JohnW2]

... Examining the DC output from the Switch Mode current sources for noise / spikes with a digital ( sampling ) oscilloscope did not show any spikes, but when an analogue ( non sampling ) scope was used there was some indication of fast rise time. very short negative spikes at the switching frequency.

OK.

Fitting a ceramic plate capacitor across the DC output appeared to quench the spikes. Displays supplied from the modified supplies did not fade over time.

Fair enough.

At some time in the past the PCB of the supplies had been altered and this included changing the output smoothing capacitor from an axial type to a radial type. The radial had a significantly higher inductive impedance due to it's construction.

Was it ever demonstrated that changing back to an axial type made a difference? As I said, I found it hard to believe that, with either type of lead arrangement, there would not be enough inductance to have a significant reactance at VLF.

Kind Regards, John

 

[bernardgreen]

there would not be enough inductance to have a significant reactance at VLF.

It isn't VLF ( Very Low Frequency ) that is the problem, it is the very fast rising and falling edges that the reactance prevents from entering the capacitance

Was it ever demonstrated that changing back to an axial type made a difference?

Not as far as I am aware in the case I posted about. Some tests were made in the lab of a radio communications company and the inductive reactance was measured and it's effects were found to be less than significant when considering UHF sine waves inducing noise onto DC power PCB tracks and/or rack cables.

 

[JohnW2]

It isn't VLF ( Very Low Frequency ) that is the problem, it is the very fast rising and falling edges that the reactance prevents from entering the capacitance

Fair enough, but if the design is such that such "very fast rising and falling edges" are a known component of the waveform, then there surely should be a capacitor of an appropriate type and value (in addition to the electrolytic 'smoothing' the VLF) (or, as below, a ferrite bead) to deal with that, shouldn't there?

... Some tests were made in the lab of a radio communications company and the inductive reactance was measured and it's effects were found to be less than significant when considering UHF sine waves inducing noise onto DC power PCB tracks and/or rack cables.

From VLF to UHF is quite a conceptual jump ('sublime and ridiculous' come to mind) :- )

Are you suggesting that the "very fast rising and falling edges" (which you believe were fast enough for the leads of an electrolytic to have significant reactance) has faster rise/fall times than a UHF sine wave (when reactance was, you say, "less than significant")? In any event, if the answer were 'yes' (which I would doubt), then (a) would not other parts of the wiring/PCB track etc. have even more reactance than the electrolytic's leads, and (b) would not the answer be a ferrite bead or two, rather than a capacitor of any sort?

Kind Regards, John

 

[bernardgreen]

It is not the capacitors leads that contribute the majority of the inductance. It is the spiral wound foil electrodes in the capacitor.

In a radial capacitor there is one point of contact from each lead to the corresponding electrode so the length of the spiral forms a distributed inductor along the electrode. The current has to flow along the length of the foil

In an axial capacitor the entire length of the edges of the electrodes can be connected to the corresponding lead wire. The current only has to flow across the width of the electrode

 

[JohnW2]

In a radial capacitor there is one point of contact from each lead to the corresponding electrode so the length of the spiral forms a distributed inductor along the electrode. The current has to flow along the length of the foil ... In an axial capacitor the entire length of the edges of the electrodes can be connected to the corresponding lead wire. The current only has to flow across the width of the electrode

If that's true, I don't see why it needs to be, since it would seem trivial to make both in essentially the same way. At the most, 'converting' an axial one to radial (i.e. both leads at same end) would require just a wire the length of the capacitor but, if it was in an aluminium 'can', one wouldn't even need the wire, since one could use the can as the conductor to the 'other lead'.

Having said that, anyone who relies on an electrolytic capacitor (of any type) to filter 'spikes' with rise times of the order seen in a sinuoisoidal UHF signal, they are probably rather daft - as I said, one would surely add a capacitor of appropriate type/value and/or a ferrite bead to achieve that?

Kind Regards, John