How Many Microfarads before Soft Start is Necessary?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Well the title says it all, when should one begin to worry that he/she has too many uFD's of capacitance to just power the amp on without accounting for a soft start? Assume a 500-800 VA toroid.

For example, I don't have a soft start on my Gainclone (has a 500 VA toroidal transformer and about 12,800 uF total between all channels). Seems to not be an issue, but I'm going to add more with some snubbing and was wondering when the inrush would just be too much so that I'd need to add some soft starting.

Heck, my BOSOZ preamp has 12,000 uF total but there I've got two power tranformers. Oh wait, I've got 3 and a few more power supplies in there. Forgot about that, need to add another 4,000 uF.
 
softstart

I have no objection to the discussion regarding cost, but how about answering the question as well? Soft start at ????uF?

I suppose a single thermister in-line with one primary winding would be a cheapie solution. I'm making the elliot circ, and the relay itself cost over $10.00, the mounting is another 3, and the xformer is at least 5, I think its going to be much higher than $40.
 
Digikey? My goodness, don't you people know how to scrounge a relay? You're making a one-off, not going into production.

In any case, the relay and an antikickback diode, a resistor and a zener to tap power (may not even be necessary if your rail voltage is somewhere south of 30 volts), a small resistor and cap to set a time constant, a 100 ohm 2 watt resistor (value may vary depending on your transformer and caps), and a small transistor to trigger the relay. And you're all done.

Looking in my solid-state woofer amp, the relay cost a buck (two bucks at Radio Shack), the transistors, resistors and cap came out of one of my coffee can storage bins (ok, ok, if you don't have a coffee can, figure another buck or two)...
 
With that old Elektor circuit you can do even better.
Metal casing Dale 25/50watt 40/50 Ohm resistors can be had on the web for a buck.
Same with top quality NOS relays, a dollar and cents.
Surplus electrolytic caps and 6"x4" pcb boards for next to nothing.

imho, a toroid at 300va or over needs soft starting. 300va's dont blow the fuse but does make the lights flicker.
I have a graph somewhere that shows toroidal inrush current, which is impressive to say the least.
Soft starters reduce inrush current to the capacitors, which lengthens the lifespan of the caps.
High inrush creates heat in the capacitors because of esr, heat reduces capacitor life.

SY is right (as usual, boring)
 
This is a bit of a "how long is a piece of string" type question !,
I have used 0.5 million uF before now without any soft start circuit. I would agree, however, that with transformers above 500va it becomes an issue, with blowing up rectifiers being the biggest hassle (tripping breakers and blowing fuses can be irritating as well!).
The method I have used in the the past consists of putting a 8r2 50W metal clad resistor in line with the mains supply(for european mains) , you then have a relay across this resistor, which is fed from the amps dc supply via a suitable dropper resistor which is sized to pull in the relay when the dc rails get to half/two thirds of their eventual value. This method avoids any extra transformers or active circuitry and is pretty foolproof, the only thing to watch out for is that the amp chassis and metalclad resistor MUST be earthed and the amp must have an appropriate size fuse, I have known these resistors become short, case to element:hot: rarely.
A useful side effect of soft start, of any type, is that it can allow a much more suitable fuse rating to be used with a toroidal TX, which normally need an overrated fuse to cope with switch on current.
 
Jez, precisely. And all it takes to fire the relay is a source of voltage (hey, there's a power supply rail already in there!), a transistor, and an RC fed by that selfsame voltage rail. Add a diode for protection and you're done.

I've seen all kinds of precision circuits adapted for this and ask myself, "Why?" Does it really matter whether the delay time is (say) 2.5 instead of 2 seconds?
 
SY. infact you do not even need the transistor or the RC!
The primary resistance (TX and wirewound resistor) and reservoir cap/s form a suitable RC time delay in themselves. The 8r2 resistor may need to be scaled for differant PSU's, this is the value I have used for banks of 22,000Uf caps (like 10 off).
 
SY said:
JI've seen all kinds of precision circuits adapted for this and ask myself, "Why?" Does it really matter whether the delay time is (say) 2.5 instead of 2 seconds?
According to my experience you need 0.3-0.6 seconds to catch the really huge peak from the transformer itself. Can be made like this.

If you have a really huge cap battery you may have longer time but thing is not to blow the fuses and save the power switch.
 
SY said:
Jez, precisely. And all it takes to fire the relay is a source of voltage (hey, there's a power supply rail already in there!), a transistor, and an RC fed by that selfsame voltage rail. Add a diode for protection and you're done.

I've seen all kinds of precision circuits adapted for this and ask myself, "Why?" Does it really matter whether the delay time is (say) 2.5 instead of 2 seconds?

If the soft start circuit is to use/rely on the amp's power supply rail isn't it defeat the purpose of the soft start circuit? Then why would RodE recommend to use a separate transfomer/ power source? Please help out a newbie here.

Fuse already blown by inrush current so power supply rail has no power so the soft start circuit has no power to make it work.. something is not logical here I think..
 
Because the 8r2 50W resistor on the mains jez talks about causes the inrush current to limit (I=U/R). If this resistor wasn't in place the resistance would be low and a virtual short circuit would be created at turn on. After a time the rails are high enough to activate the relay, which is connected parralell to the 8r2 50W resistor, effectivly shorting it out of the circuit.

Like lgreen proposed earlier, a thermistor would do the job.
(also recommended by Mr.Pass!)

/Niclas
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.