Transformer soft-start with relays

[Redshift187]

Used one in anger? I've been using it in my amp for a long time.

I hardwired the low voltage switch on, so I just turn on the mains switch (which supplies both the soft start and then the main transformer through the soft start). My mains switch is built into my receptacle and I didn't want to add another switch.

[CLS]

How about a timer (like Omron H3Y, or the likes) bypassing a resistor? When turn on, the resistor limits the inrush current, and then the timer would close and bypass the resistor. Such timer is actually a relay with built-in control circuits.

Setting it at 1 or 2 sec., then both the flux in trafo and cap charging issues are taken care of. If you choose timers for mains voltage, then there're only 2 components.

[Rudi_Ratlos]

This is how I did it:



I use a no-longer used Laptop's PSU to switch on my AMP.
I do not like to switch on the main's 220V.

As soon as 12V is applied to connector X2 relay K1 switches and feeds the 220V main's voltage via NTC resistors R4 and R5
(both rated at 22 Ohm) to the output X1. The inrush current is thus limited to approx. 5A.

When the timer LM555 triggers (its delay being determined by R1 and C1 giving a delay of about 750ms) relay K2 switches
and shortens the NTCs.

This is the PCB for the Soft-Power-On circuit:



Its dimension is 50 x 70mm and it uses relays which I had on stock (1-pole relays would have been better).

I will etch it in a few days.

Best regards - Rudi_Ratlos

[smithy666]

Why not use the soft start that Elektor published many years ago. I've been using it for over 10 years without a problem

[AndrewT]

Originally Posted by Rudi_Ratlos

Its dimension is 50 x 70mm and it uses relays which I had on stock (1-pole relays would have been better).

nothing wrong with using 2pole relays. You have already paralleled one.

Borbely tells us to use 2pole instead of 1pole but when paralleling the poles to swap one pair, to be fed from the opposite side.

[Redshift187]

Rudi, that circuit seems almost identical to the one I bought. It has a ~12V connection (from it's own on board transformer) for switching on and off, so you don't have to switch the mains. Good idea, I just didn't feel like using it.

I really like everything about that functionality. Using thermistors, bypassing them after a short time, low voltage switch on.

[brickj]

Thanks Tyler - think I'll give one a crack.

Justin.

[ChrisA]

Originally Posted by CLS

How about a timer (like Omron H3Y, or the likes) bypassing a resistor? When turn on, the resistor limits the inrush current, and then the timer would close and bypass the resistor. Such timer is actually a relay with built-in control circuits.

Setting it at 1 or 2 sec., then both the flux in trafo and cap charging issues are taken care of. If you choose timers for mains voltage, then there're only 2 components.

Rather use a timer why not use the power supply itself?

In-rush current is inversly proportional to the voltage across the power supply's filter caps. When first powered up the caps are at zero volts and "look" like a short circuit and hence the problem. But quickly they are charge and then we no longer have the in-rush. We all know this already. The problem is knowing when to close the relay.

If you place a zenier diode in series with a high value resistor from the power suppy to ground there will be zero volts at the resistor right after power up but then as the filter caps charge above some limit there will be the zenier voltage on the resistor. This voltage come on nearly instantly when ever the caps reach a given state of charge.

So you can use the zenier diode to switch the relay, or to switch a transistor which then switches the relay. The advantage over a timer is that this is self adjusting and only introduces the _exact_ required amount of delay. You can choose the zenier for (say) 90% of the power supply voltage. If you need two relays, one to switch the irush limitor and another to conect the audio to the speakers then ue two diodes with different voltage thresholds. So first the in-rush limiters is removed thenthe audio is un-muted so as to avoid a "thuink" sound.

Yes this work for all cases including quick on/off cycles because the diode in effect is measuring in-rush current and keeping the resistor in place only when there is such current.

In effect we are still using a timming but an RC timmer but with a C and and R aalready built into the amp.


All that said a simple device to use is a time delay relay. These are designed to have the contacts close so many seconds after power is applied to the control pins.

[AndrewT]

Chris,
No.
For the same reason that you should not use a Variac to slowly bring up an amplifier fitted with a soft start system.

If the voltage is highest, all the components need to be designed to survive and the time delay is suitable for this worst case slow charge system.
When mains is lowest the delay becomes very much longer and in extreme cases may not time out. The slow charge system stays on line.

If this slow charge system has been used to double up as a soft start system with a resistor in the primary feed then that resistor is going to get mighty hot and may burn out.

BTW,
I think you are confusing the slow charge problem with the topic of this thread - Soft Start.

[metalsculptor]

Rather use a timer why not use the power supply itself?

It works fine I posted examples in a prior thread, I had the problem that Andrew T mentioned once because the series resistor was too high, aim for 5x the full load current at start for a soft start. Usually a quality transformer has an impedance of around 5% which means the initial charging current can be 20x the normal running current, reducing that to 5x is a major improvement while maintaining reliable starting under load.

No need for a zener in series with the relay that just makes the system more sensitive to mains voltage variations. most relays start to pull in around 70% voltage and by the time they make contact a few more cycles have passed so if the relay closes at 80% voltage the 2nd surge will only be 4x normal load current at this point hardly a problem, if it is the rectifier or fuses are too small.