On soft start circuits and safety

[Pwent]

Hello all
Im currently planning my first "real" amplifier project, only previous experience was a head phone amplifier. The amplifier will use a 500 or 625VA toroid, and as I understand it, it is in need of some kind of soft start circuitry to control the inrush current.

I plan to implement this through a relay controlled by a micro-controller. The same mcu will also operate the speaker muting relays and the two relays needed for stand by operation (so it can be turned on from a future pre amplifier). The relays and the mcu will be fed from a second transformer of suitable size, ~30VA 0-15.

The question is what to use to limit the current. A resistor will have to withstand some serious power, albeit for a very short time. A thermistor perhaps? I understand those are used sometimes for soft starting. While in "soft start mode" the main fuse are not likely to blow since the current is limited anyways, so it is a good idea to keep the soft start delay as short as possible? I have seen suggested (at ESP for instance) that 100ms is suitable. Commercial soft start modules seem to use ~3 seconds however. What are the ideas behind these different choices of delay length?

Any input is most welcome.


Cheers!

[Pano]

Thermistors work for me. 3 seconds in not a bad delay for everything to stabilize.

[Pwent]

Thanks for your prompt reply panomaniac!

Yes, three seconds should be great for stabilization. However, what can happen during those three seconds when the amplifier is essentially "unfused"? Perhaps I'm worrying to much over a non issue?

[danny_66]

In my F5 I've put the following soft start with thermistors and relays:
softstart circuit

The DC protection is based on upc1237: basic circuit like in the datasheet, DC protection, time delay at power on and AC level detection.

Regards,

Danny

[Pwent]

Thanks a lot danny_66, I'll get to brooding over that right away. =)

[jrenkin]

Somewhere I bought some ready made softstart modules for around 15-20 bucks. I will try to find out where tonight and post it. Gotta go to work...

[Pano]

Originally Posted by Pwent

what can happen during those three seconds when the amplifier is essentially "unfused"? Perhaps I'm worrying to much over a non issue?

Yes, you are worrying over a non-issue. The amp is just as protected by a fuse during soft start as any other time.

[star882]

Calculate the RC time constant of your circuit, then let 5 of those constants pass before finishing the precharge.

[Pwent]

panomaniac:
Yes, true enough. But what I was thinking is that if there is a fault in the amp and something short circuits, then the fuse will not blow while the current will be limited by the soft start circuitry. Hence the amplifier will be "unfused" while the soft start operates?

star882:
The RC time constant for the filter caps in the power supply, right?


Thanks again for input!

[5th element]

The inrush you want to reduce is produced by two things.

1) The formation of the magnetic field within the transformer.

2) Charging the reservoir capacitors.

The huge inrush current created by these is only present for a very short amount of time and hence only needs limiting for fractions of a second.

Rod Elliot has two sections, one dedicated to soft start circuits and one dedicated to protection circuits that control a relay on the amplifiers output.

Here is the first. I have built something similar to the last design on that page. The opamp acts as a comparator and changes state when the voltage between R1 and C2 exceeds that present on the non inverting input.

My version is set to around 1/2 a second and it works great. To be on the safe side I used a 50 watt resistor.

If you're constructing a class A/B amplifier where the quiescent conditions require a typically small amount of standing current, then you could quite easily increase the delay to a longer time period. However if it's a class A amp, where the quiescent current is very high, then increasing the time delay will really cause the inrush limiting resistor to get hot quickly.

The other page is here. I have also built this minus the, Loss of AC Detector. It also works very well.