diyAudio (
-   Power Supplies (
-   -   Transformer soft-start with relays (

len_scanlan 22nd July 2010 02:57 AM

Transformer soft-start with relays
2 Attachment(s)
Here's a circuit I've been playing around with in LT Spice.

I'm building several channels of Pass F5 amplifier, and wanted to avoid the HUGE surge current when these things are powered on - I know dimming the house lights when you switch your system on is kinda impressive, but it just can't be good for the amp?? Also I will have 2 or 3 of these in an active setup, so triple the power-on surge and you know.. circuit breakers tripping..

So this circuit is intended to sequence two power relays for power supply inrush current limiting.
There is an auxilliary 24VDC supply which is always on and powers this circuit. When the amplifier is off, the power drain should be negligible.

Relay 1 switches 230VAC to the transformer primary. I wanted to have a remote 12V trigger input, so several amplifiers can be powered on/off in sequence. In the schematic, V2 is the control voltage, and Q5 will be replaced with an opto-coupler.

The power transformer has a nominally 120 ohm wirewound resistor in series with its primary winding. This limits the power-on surge current while the large capacitor bank on the secondary side charges up.

After 2 seconds, Relay 2 switches on, shorting across the wirewound resistor thereby directly connecting the power transformer to the mains.

Here's how it works:

Power On:
Initially the remote trigger is OFF, and transistors Q1, Q4 are OFF. Both relays are unpowered.
When the remote trigger switches ON, it turns on Q5 which pull the bases of Q1 and Q2 low.
Q2 switches on immediately, and starts charging the RC time constant on the base of Q3.
Q1 switches on within 100mS, as delay capacitor C1 discharges.

Q3 has a couple of diodes connected between its emitter and ground. This raises the Q3 turn-on voltage by two diode drops - from about 0.6V to about 2V. This helps during the switch off phase.

After about 2 seconds, Q3 switches on, pulling the base of Q4 low, which switches Q4 on and powers relay RLY2

Power Off:
The remote trigger goes OFF, switching off Q5.
No more current is pulled from the base of Q2 so it switches OFF.
The RC time constant on Q3 discharges via D3 and R8 within 100mS, and relay RLY2 drops out. The amplifier's power transformer is now connected to the mains via the 120 ohm wirewound resistor.

Q1 does not switch off immediately - C1 is still discharged, so it holds the base of Q1 low via R2.
After about 2 seconds, C1 has charged up enough that Q1 switches off and relay RLY1 drops out.

To do:

There may be a power-on glitch when you first connect the system to mains power. When power is first applied, C1 is discharged. I suspect this might cause relay RLY1 to 'tick' briefly. Need to sort this out before breadboarding..

Then do a PCB.. :-)

ChrisA 23rd July 2010 04:40 AM

So much easier if you simply use a in-rush limiting thermister. It does the job and cost less than two bucks.

These little things are (about) 100 or 200 ohm resisters when the circuit is off and then slowly over several seconds lower to about 2 ohms. they are designed exactly for soft starts in high current devices. I got about 2 dozen of them as free samples. They came in all power ratings and time constants.

You can place one in series with the fuse on the primary side of the PT and/or on the secondary side.

Think of them as automatically controlled variable resistors.

cliffforrest 23rd July 2010 08:25 AM

The major problem with the thermistor solution, which is otherwise very elegant, is what happens on multiple rapid power cycles?

The thermistor cool-down time is too long.

A relay solution should be designed for an instant disconnect and very rapid re-arm.

AndrewT 23rd July 2010 09:59 AM


Originally Posted by cliffforrest (Post 2251461)
The major problem with the thermistor solution, which is otherwise very elegant, is what happens on multiple rapid power cycles?

The thermistor cool-down time is too long.

A relay solution should be designed for an instant disconnect and very rapid re-arm.

Yes, the Power Thermistor is a very good solution, but it must be bypassed before it gets too hot, ready for the next restart.

The auto fault clearing switches on the UK mains can try to re-establish the mains voltage every one to three seconds for two to four attempts before finally cutting off the power. I wish we could get the real info from a "Power" engineer/technician in the UK.

tangmonster 23rd July 2010 10:03 AM

I have asked this in another forum but got no answer. To me personally a lm555 monostable circuit is the most simplest relay trigger. with far less components.

Is there a reason why everyone chooses the complication of a much higher component count above an lm555 timer

cliffforrest 23rd July 2010 10:12 AM

Could be - but then you need an auxiliary supply, on all the time.

tangmonster 23rd July 2010 10:29 AM


You get a real el cheapo 200 to 500mA 9V AC transformer. This feeds the power to to the lm555 which after x time (I used 100ms) switches on relay 2 which bridges the resistor.

In my current amp i have 2 relays and 2 power inputs.

power input 1 = MAIN high wattage power input for 500va toroidal.
power input 2 = low wattage 220V input feeding small 9V transformer. This is fed from my surround sound processor aux output.

AndrewT 23rd July 2010 10:35 AM

An extra, non audio, transformer used to power all the other circuits that can be optionally used in a Power Amplifier is a good addition.
It is cheap and it is small and it is light and generally it keeps some extra interference out of the audio PSU.

But it is not zero cost. Some will do anything "on the cheap" irrespective of what it does or doesn't do for the quality of the audio.

tangmonster 23rd July 2010 10:45 AM

What i am getting at is:

In the above circuit 3 x mje350 comes VERY close to the cost of a small 9V transformer.

other than that you need
lm555 dip ic
small bridge rectifier
220uF capacitor (should be enough you might have to experiment)
a few 1n4007 for the relay back emf
and around 1uF and 22K (i think) for the 100ms timer.

AndrewT 23rd July 2010 11:19 AM

If I saw a second transformer inside an amplifier, I would think "Ahh, this designer/manufacturer has thought about the performance of the equipment".
I may not agree entirely with his philosophy, but we are thinking along the same lines.

All times are GMT. The time now is 07:01 AM.

vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2015 DragonByte Technologies Ltd.
Copyright 1999-2015 diyAudio

Content Relevant URLs by vBSEO 3.3.2