Question about Slow charging capacitors for amplifier power supply

Discussion in 'Power Electronics' started by Nutball, Dec 8, 2016.

  1. Nutball

    Thread Starter New Member

    Jan 29, 2016
    23
    1
    I bought an amplifier circuit board, assembled but with no case or power supply, and I'm using the transformer, caps, and diodes out of a broken old amp. Something I've wanted to know how to do for a while even before this project is slow charging capacitors from the power supply while still having full unrestricted amps and volts available to the load. Ideally I want just a temporary slow charge circuit to avoid too big of a power surge when the device is powered on. I kept the switch and fuse that is connected to the transformer out of the old amp and I don't want to be blowing fuses.

    I'm not familiar enough with circuits and reading schematics to easily find the answer through research. I've spent a couple hours googling things and just can't find what I need. I've thought about using diodes, but that involves a voltage drop, I thought varistors could be useful to but I know little about them and their datasheets seem to only show surge protection specs instead of continuous use specs. I'm thinking a switch would work, but I don't know how well a mosfet would work for a delayed turn on to bypass the initial slow charge resistor and still let ripple current flow in and out.
     
  2. Nutball

    Thread Starter New Member

    Jan 29, 2016
    23
    1
    This is what I have salvaged off of the old amps so far for my project. I plan on rectifying the two 26v outputs on the transformer, then paralleling them. Would paralleling before or after the capacitors matter? It would be a great time and money saver if I could get all the parts I need off of my old amplifiers, so keep that in mind when suggesting components.

    Thanks
     
    ErnieM likes this.
  3. paulktreg

    AAC Fanatic!

    Jun 2, 2008
    766
    167
    If you are blowing the original fuse then you are going beyond the specification of the device?

    Why do you think you'll keep blowing fuses?

    What is the power rating of your amplifier and transformer?
     
  4. #12

    Expert

    Nov 30, 2010
    18,076
    9,691
    I don't think a fuse ever popped. I think this is another imaginary problem.
    "Hi. I never blew a fuse, so I want to complicate the circuit, just in case hundreds of millions of other similar circuits were the exception and mine is the rule."

    Is that what's going on here?
     
  5. Nutball

    Thread Starter New Member

    Jan 29, 2016
    23
    1
    For the most part. I read somewhere about surge current and blowing fuses because of the caps, which reminded be before I start I want to cover that potential problem. The amplifiers appeared to have extra components that probably did what I want to do now, but I didn't exactly understand what I saw and how to replicate it. If you think the surge current is not a problem, then I'll just build it without such a feature, but I'd still like to know how to temporarily reduce the current flow while the caps charge so I can use such a method on the big ESCs I use on my RC airplanes which currently use a charging wire with a resistor (to avoid a connector damaging spark), then you plug in the main power wire.
     
  6. #12

    Expert

    Nov 30, 2010
    18,076
    9,691
    Charging all 4 of your capacitors to 26 volts will require 0.49 amp seconds on the 26 volt side of the transformer.
    You would need a pretty small fuse to pop on half an amp second.
    You would shake your head at the number of these, "somebody said's" or, "I read it somewhere's" that arrive here.
    Then we whip the math on it and the "problem" disappears.

    Still, you want to do a, "just in case" on some RC airplane batteries.
    Hmmm...a PTC resistor might work, but you need solid numbers to select one.
    A resistor in series, with a relay that kicks in a second later would work.
    How about a P-channel mosfet?
    Use 1uf and 1 meg to get most of 1 second of delay.
    Without any voltages or currents given by you, nobody can name a part number. slow start.png
     
  7. dannyf

    Well-Known Member

    Sep 13, 2015
    2,196
    422
    Google soft start power supply.

    Generally not a problem.
     
  8. Nutball

    Thread Starter New Member

    Jan 29, 2016
    23
    1
    I saw somewhere where a PTC resistor was recommended, but I don't see how it would be very beneficial since it would start out just like the capacitors: low resistance. .5 amp seconds looks low, but I didn't know if it would start out at 50A for .01 seconds and if that is enough time and current to cause problems.

    One amplifier had a lot of relays, so maybe I can figure something out there. I don't know what the power ratings are on those, I wouldn't want to overheat one or do they have a built in resistor for current limiting?

    As for the ESCs, I usually use between 2500uf and 40v up to about 4500uf and 65v using rubycon zlh series caps.
     
  9. Nutball

    Thread Starter New Member

    Jan 29, 2016
    23
    1
    My edit time expired: I hope to get 5-10A 24v out of this power supply although I'll probably never use more than 3A just so you have an idea on the current I'll be using.

    I've thought about a mosfet for soft starting an ESC, but they use such a wide range of voltage, I don't know much, but I think I'd also need a little voltage regulator to go with the mosfet to avoid a problem where the capacitors have not discharged enough to turn off the mosfet before the battery gets plugged back in causing a huge spark from lets say a 40v difference. Otherwise the voltage range of the ESC might have to be reduced from 12-65v down to 40-65v for example.
     
  10. Nutball

    Thread Starter New Member

    Jan 29, 2016
    23
    1
    I looked up some info on relays, I think I could figure out a time delay relay circuit.

    For the ESCs, I guess some sort of non voltage dependent time delay circuit could be used to turn the mosfet on and off.
     
  11. Nutball

    Thread Starter New Member

    Jan 29, 2016
    23
    1
    I finished the power supply, it looks like junk and it probably is junk because when I powered it on I read I think over 60v. I'm not 100% sure because when I saw over 23v which it should have been I turned it off. Maybe I got some sort of voltage booster circuit by accident. I was so sure it would work. I had 2 diode rectifier bridges both read 23v, one on each 26v output of the transformer sharing the center tap. I then paralleled the 23vdc output ends and added capacitors and a half second delay solenoid switch to give full power to the caps. I don't want to turn it on again now to be sure of anything because I think it exceeded the 42v capacitors max voltage, maybe went as high as 64v which theother caps are only 63v, and I don't know if the solenoids got overloaded, but it sounded like they clicked off, so they might be fine, but they were using only a 35v capacitor for the delay.
     
  12. Nutball

    Thread Starter New Member

    Jan 29, 2016
    23
    1
    I powered it on without a load on the output end. Could the changing current draw involved with the delay circuit and just having big caps connected to an inductor (the transformer) without much of a current drain cause a DC switching boost converter effect? I keep thinking on it and I can't think of any way how two 23vdc sources being paralleled could cause at least double the voltage especially with filtering capacitors.

    I've always wanted to build a switching voltage booster, but didn't know enough to use a timer IC, and now here I may have just made a booster.
     
  13. Nutball

    Thread Starter New Member

    Jan 29, 2016
    23
    1
    If my power supply components are not damaged, is it safe to parallel (to handle more current) individual combination buck/boost regulators to regulate the voltage out of my power supply? I assume they would provide enough load to help keep the initial voltage spike down while of course giving me the output voltage I want. I could add a variable resistor to adjust until I get a sufficient load to prevent a spike that could damage the regulators

    Your thoughts?
     
  14. Nutball

    Thread Starter New Member

    Jan 29, 2016
    23
    1
    I tested my power supply again with a 40W light bulb attached, 26ohms I think, and I still got 64v after a second or less. I tried a motor with 4-5ohms when stalled and after a second or two of accelerating it reached 44v where I cut power. Point is I bet the regulators wouldn't help in this situation. oh well
     
  15. ronv

    AAC Fanatic!

    Nov 12, 2008
    3,768
    2,862
    You might have gotten the wrong 2 wires from the transformer. Measure the AC voltage before you hook it up to the diodes.
     
  16. Nutball

    Thread Starter New Member

    Jan 29, 2016
    23
    1
    I did that before and after the diodes. I really doubt the voltage can boost in any way especially with the diodes and capacitors, but I'll look into AC boosting and see if I accidentally made a booster and DC converter.
     
  17. Nutball

    Thread Starter New Member

    Jan 29, 2016
    23
    1
    How I have it wired with 8 diodes it should be possible to get double the voltage with the center tap disconnected, I don't know how full the wave rectification would be. Half of the diodes would not be used then, but the center tap is connected, so I would think it's voltage difference would cancel out the difference between the outer ends, so I should still get the lower voltage. Otherwise I might still have the same problem if I further isolated the two rectifiers which parallel immediately and directly share the center tap.
     
  18. tcmtech

    Distinguished Member

    Nov 4, 2013
    2,808
    2,705
    I wouldn't worry about it. The amount of capacitance you have there and its charge up time is insignificant compared to many other devices you use every day.

    You're trying to design around theoretically perfect ideal and not real world applications where there are numerous other effects in play such as the winding resistance of the transformers coils which in themselves keep the peak amperages well under control in anything but a dead short condition.

    My vacuum cleaner has a DC resistance of ~2 ohms yet runs off of 120 VAC that has a ~ 170 volt peak so in theory every time I turn it on at a peak in the AC sine wave it should draw around 85 amps and both blow its 15 amp rated switch bits plus trips the 20 amp rated circuit breaker every time I turn it on yet it never does. Any guesses why theoretically ideal and real life aren't matching and blowing switches and circuit breakers like theory says it should? o_O
     
  19. Nutball

    Thread Starter New Member

    Jan 29, 2016
    23
    1
    The theory doesn't take into account all that is possible due to ignorance. In my case I expect a surge, but I want to be on the safe side to avoid blowing fuses because I don't know the actual draw and duration. I expect some ripple on the output end, I expect voltage drop due to the diodes and other restrictions. I expect DC voltage very close to the AC voltage coming out of the transformer based on how rectifiers work and how they seem so simple and I have not found any obvious issues with them. I must be ignorant of something because I'm getting more than double the intended voltage, and I just can't see yet how I could have made a boosting circuit, maybe a flawed one that produces 50v, but not 64v
     
  20. KeepItSimpleStupid

    AAC Fanatic!

    Mar 4, 2014
    3,540
    667
    Generally, you don't need the slow turn-on. I have essentially 40,000 uF at 50 V in my amp and I NEED the slow turn-on,

    The way I did mine was to first turn off the audio input with a series optocoupler and turn off the speakers with a relay. the slow turn-on circuit has it's own power supply (Around 12 VDC). Then I insert a flame proof resistor in the AC line and wait until all supplies are about 2/3 their max value of 50 V, Then I turn on the speakers and ramp up the audio logarithmically by controlling the opto-coupler.

    The good part, is if a power rail fuse blows, there are 4, not much gets damaged, but that resistor has to be replaced.

    The circuit I built needs some sort of fail-safe where if the voltages are not 2/3 of Vcc , within a certain amount of time, the slow charging should cease and a fault light should illuminate.

    Most slow turn-on circuits are just based on time, For a certain amount of time after power up, the current is limited. At a minimum, you have to remove the speaker load generally until the timeout is over.
     
Loading...