 
|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
|
|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
In Reply to: Re: black gates in parallel posted by Robert Karl Stonjek on December 21, 2001 at 23:52:45:
Robert S. I am not going to go back to completely understand your original question, but this is how I look at things. First, there is the original input surge current to charge the cap. This can in fact be 100 or more amps, IF you do NOT use a transformer and have a large enough cap. In the old days, 20+ years ago, we would insert a power resistor in front of the cap and short the resistor out with a relay, after the cap is charged. Today, we use heat sensitive 'surgistors' to limit the original charge current, and only on our best designs, do we use a shorting relay like we did in the old days. Of course, any transformer will limit the peak current, both with its winding resistance and core saturation.
Of course, once the cap is charged up, then the cap doesn't have to conduct so strongly, unless you are shorting out the cap, once it is charged up. This could happen with a flashlamp load, but I doubt that an amp would be so difficult. The exact charge current through the caps, after first charge, is difficult to predict, but it could be measured easily enough by putting in a current sensing resistor (.01) ohms or so in series with the ground leg of the cap, and measuring it with a scope. I doubt that you will see too many amps, and they will be only there for a short part of the AC cycle, so even though the I(squared) R losses could be significant, at least the cap can rest for most of the cycle.
Follow Ups:
Thanks John. But aren't you talking about initial switch-on surge? Steve and I were discussing ripple current. This current continues all the time, not just at switch on. Or do you mean the relay rattles away 100 times per second? (sorry, 120 times per second in your neck of the woods)I must admit I haven't heard of 'heat sensitive 'surgistors' . Steve will be evoking the 'ignorant' word if I admit to too many of those :)
But I use NTC (Negative Temperature Coefficient) Thermistors in series with the transformer. Toroidal transformers are a little severe at switch-on.
I get the feeling you are referring to rectifying the entire 110V line with no transformer. I've often wondered if I could get away with ditching the transformer, but our supply voltage is 240V , so it isn't really an option.
But even with a rectified mains supply, the ripple current would be directly related to the capacitor's ESR and hence its ripple current specification.
Kind Regards,
Robert Karl StonjekPS I don’t think ESR has much to do with anything once the caps are matched to the transformer, but its still nice to know how it all fits together.
Also, I just tried to measure the switch on current (in my amp) and could only find slightly more current than the ripple current (1 amp). Seems the thermistor does its job!!
Double checked by testing the switch-on current at the active mains line near the thermistor ~ 3.5 amp.
Extra note:- I tested the voltage drop over the thermistor. When the amp is switched on I measure a 35~40. When the amp is running I see a slight ripple that seems to be exactly the same as the ripple I see when I connect the current probe to the amplifier’s secondary winding. The peak is 3 volts, peak to peak is 6 volts at 50Hz.
This is a rather exciting ‘discovery’. If the ripple voltage is responsible for capacitors heating up, and my thermistor has the previously undiscovered (by me) effect of ‘softening’ the ripple current, then it is no wonder my caps run cold and last so long. In effect, the thermistor is helping to protect the caps.
A relay across the thermistor would null this effect. Well there you go - that’s made my day!! :)
Robert S. I still think that you are looking at the problem differently than I do. Just because a cap has a high value capacitance and a low ESR doesn't mean that that it will pass high currents through it, once it is charged up. At this moment, my current amp has 66,000 uf/channel with about a 700VA transformer per channel charging up to +/- 90V and not overheating the caps. It is completely load dependent and the idle load is approximately 1A / channel. If my rectifier bridge shorted out and AC attempted to pass through the filter caps, they would probably overheat, but normally that is not the case. The caps just try to maintain a peak voltage after being drained for less than 10ms. The current through the caps is difficult to calculate, but it can be measured with a current probe or a very small resistor put in series with one of the caps. It would appear that this peak current is set mostly by the cap value, load current, and the peak limiting current of the power transformer. The lower the ESR of the cap, the less heat buildup due to I(squared)R losses in the cap. I have a amp coming within the week with twice these values, so I hope I'm correct. ;-)
John,
just so we know exactly what I'm on about I've recorded the waveforms I was referring to and will present them here.With the amplifier on and idling, the current probe clipped onto the Secondary winding I get this trace:-
The average current drawn is 480mA. This is what you would get if you just connected up a current probe or measured the voltage drop across a series resistor. But the Scopemeter shows the peak to peak transient current or 2.4 amp. The frequency is 50Hz, voltage is 56~8ACNow at the capacitor (between the bridge rectifier and capacitor) we get this trace:-
The average current drops to 340mA (though the voltage is now higher) and the peak to peak (now all negative for the negative rail cap) is 1.3Amps. The frequency is now 100Hz, double the secondary frequency. Voltage is 79VThe next trace shows us the current drawn at the primary of the transformer:-
1.24A Peak-to-Peak, 200mA average, 50Hz, 245.5V (morning surge)
Note the extra hump in the trace. This comes from the other transformer in circuit. The low voltage (+15, -15, +12, +12 unregulated) lines have their own transformer, bridge etc.The final trace shows the voltage across the NTC Thermistor.
The small positive going hump shows up as a negative going hump - didn't notice it 'till just now. The thermistor is in series, so the correct orientation of the probes is not immediately obvious. I should have been more careful.
The voltage is 6VRMS Peak-to-Peak and the average is just 1 Volt.
Note that quite substantial currents are being drawn when the amplifier is idling although the average current (the figure you gave, I think) is small. The same measurements on my main amps are double as there are two of the 15,000uF caps on each rail rather than just the one per rail as in my spare amp (which is for sale).
Your amplifier’s ripple current should be about twice as high as my big amp as you have twice the capacitance and approx. twice the average current (you said 1 amp, my big amplifier draws about 600mA (each - I have one amp for the left channel, one for the right channel)). But this depends on the ESR of your caps compared to mine (and so the ripple current spec and so the measured ripple current).
This should clear up any confusion. Within the simple current reading one might get by just measuring across a series resistor or using a simple current probe lurks the ripple current which is displayed clearly in the traces above.
Kind Regards,
Robert Karl StonjekPS isn't gif format wonderful? Only 2k per image, 1/10th of the jpg!
Good work, I'm too lazy to do it myself! Now, it would appear that we have to integrate the effective current over all time and not just on the peaks, to get the equivalent RMS current of the charging current. It would appear that we are only conduction about 15% of the time. So, I would take the peak current into consideration, but I would also know that the cap has over 5 times the charge time within the cycle, to dissipate the heat generated by the current peak. I would presume that you are in pretty good shape, relative to added heat.
Of course, adding a thermistor will reduce the peak current, but it will also effect the regulation of the power transformer as well. This is why we have chosen in our latest amp design to use the thermistors for initial turn-on, but then short them out with a relay after a few seconds.
Although I haven't explored the idea, one could use a fast blow fuse when a thermistor is in circuit rather than a slow blow with no thermistor. They do give some protection to power amplifier.To be honest, I've never encountered supply caps heating up in Audiophile power amps though I was aware of the possibility. There is so much genuine dynamic headroom in my amps that such considerations as ripple current can be ignored.
Even so, Steve has brought to our attention the possible relevance of such things when one is designing a power amplifier.
Kind Regards,
Robert Karl StonjekPS I measured well in excess of 110dB at the back of my room (only 4 meters) with a music signal, so I don't think power is really an issue.
PPS the above test was done with just one channel running!! :)
This post is made possible by the generous support of people like you and our sponsors:
Top of Page ]
[ Contact Us ]
[ Support/Wish List ]
[ Copyright © 2025, Audio Asylum, all rights reserved ]
