Suggestions for delay circuit for dc power

[Salvador]

Hi, I'm sure there are such devices both factory made or diy versions , basically what I need is a relay that switches on a few secs after the power supply itself has turned on, because there is or can be a voltage spike occurring during the first few moments when the power supply is switched on.

the output from my power supply is about 180v DC so I would want to feed that through a relay which then with a delay connects the power to my amplifier boards.
maybe there are some simple IC's specifically for this purpose together with a relay.

thank you.

 

[donpacino]

Hi, I'm sure there are such devices both factory made or diy versions , basically what I need is a relay that switches on a few secs after the power supply itself has turned on, because there is or can be a voltage spike occurring during the first few moments when the power supply is switched on.

the output from my power supply is about 180v DC so I would want to feed that through a relay which then with a delay connects the power to my amplifier boards.
maybe there are some simple IC's specifically for this purpose together with a relay.

thank you.

there are a few ways you can do this.

1. use a micro-controller. detect when the supply is turned on and up and running (voltage sensor or time based).
then trigger enable circuitry to drive the relay.

2. create an RC circuit tied to the output of the supply. the output of the RC network will go through a voltage divider to a comparator. when the voltage is over a certain level, enable the relay.

i'm not sure of any off the shelf parts that will do what you want to do.

note, depending on your expected current and capacitance you may have to precharge the relay.

 

[jim hardy]

search for "time delay relay"

EDIT

555 timer IC is a natural-born time delay element and can drive most any relay with a low voltage coil

http://www.ti.com/lit/ds/symlink/ne555.pdf

 

[Nidum]

Can't you use a Zenner diode to ground or other circuitry to clip the spike ?

Or use some RLC to smooth it out ?

 

[meBigGuy]

You can do a search on "time delay relay" and "time delay solid state relay" and look over your options.

You may find the power supply also spikes when you finally connect the load, or something about the load is causing the spiking at turn on.

Have you tried a manual switch just to test that what you are planning to do will solve the issue?

You can add a power fet and delay circuit, but 180V means you need to be careful about part selection

 

[Salvador]

well i would rather not use a semiconductor here , first of all it outputs 180v DC and it can output those 180 v into an almoust 1KW load as far as I have tested.

Also its not the addition of a load that causes the voltage to fluctuate , since i have tried plugging into mains both with a heavy load attached and with nothing but just 22k idle resistors across the outputs of my secondaries.Rather weirdly but in both cases the supply sometimes manages to push as high as 250v through the load for a very brief moment at first turn on , then the voltage settles and is stable.
well after all it's a homemade high power smps , using leftover things i had on my desk so given that i'd say apart from the turn on problem it works rather well.
I think a simple time delay relay would probably be the best option , ofcourse a more sophisticated IC together with a relay that could also disconnect the relay on overvoltage conditions would be nice but it depends on how complicated i want to go for this simple experiment.

 

[donpacino]

well i would rather not use a semiconductor here , first of all it outputs 180v DC and it can output those 180 v into an almoust 1KW load as far as I have tested.

Also its not the addition of a load that causes the voltage to fluctuate , since i have tried plugging into mains both with a heavy load attached and with nothing but just 22k idle resistors across the outputs of my secondaries.Rather weirdly but in both cases the supply sometimes manages to push as high as 250v through the load for a very brief moment at first turn on , then the voltage settles and is stable.
well after all it's a homemade high power smps , using leftover things i had on my desk so given that i'd say apart from the turn on problem it works rather well.
I think a simple time delay relay would probably be the best option , ofcourse a more sophisticated IC together with a relay that could also disconnect the relay on overvoltage conditions would be nice but it depends on how complicated i want to go for this simple experiment.

if it is a home made power supply, you might consider adjusting the PI controller (assuming you have one) in your power supply.

 

[jim hardy]

I'm still suspicious of your meter and that's why i suggested an analog one.

 

[donpacino]

well i would rather not use a semiconductor here , first of all it outputs 180v DC and it can output those 180 v into an almoust 1KW load as far as I have tested.

Also its not the addition of a load that causes the voltage to fluctuate , since i have tried plugging into mains both with a heavy load attached and with nothing but just 22k idle resistors across the outputs of my secondaries.Rather weirdly but in both cases the supply sometimes manages to push as high as 250v through the load for a very brief moment at first turn on , then the voltage settles and is stable.
well after all it's a homemade high power smps , using leftover things i had on my desk so given that i'd say apart from the turn on problem it works rather well.
I think a simple time delay relay would probably be the best option , ofcourse a more sophisticated IC together with a relay that could also disconnect the relay on overvoltage conditions would be nice but it depends on how complicated i want to go for this simple experiment.

do you know that the 22k resistors are enough?
a system typically has more of an underdamped response with a light load (the heavy load provides more damping).

can you post an image of the voltage spike you are seeing?

 

[Salvador]

my scope is not that advanced I doubt I could catch that spike but quite frankly i haven't scoped the secondary outputs after rectifier.

As for what Jim said, he is right given its a very cheap multimeter it might actually lie on situations were high frequency spikes or interactions are concerned.
I will call my friend he has a much more advanced multimeter, just to see what that shows before going out and buying myself an analog one, even though i need an analog one once in a while so I just might get one anyway.

I will post what I find out.

 

[Salvador]

So it seems I won't be needing the delay relay anymore , what I did was put less resistance across the secondary of my smps , the resistance I use as idle load for the smps.
lowering the resistance means more mA now goes through my idle load and this solved the voltage spike.Now there is not a single volt going over the desired level.
I didn't get the analog meter but I got another digital one , more advanced than my cheap meter and it didint showed any voltage spikes even when they still were there now after the resistor correction both multimeters show a steady climb in voltage and no overvoltage spikes.

 

[jim hardy]

Glad to hear of your success.

Would you post again that schematic drawing link ? I'd like to understand this phenomenon a little better.
Some kind of startup transient ? Asymmetry?

 

[Salvador]

Quite frankly Jim I think it's magic :D More on the point it could be because my smps was more of an experiment named "can I build a high power smps" much of the parts were salvaged and reused from other equipment rather than specially purchased for this application.I think much of the fault lies there.But in the end I got what I needed anyways and I leaned many things, one of them is why buy a more expensive MOSFET for high power switching when these days one can get really cheap IGBT with higher power ratings for half the price.even the frequency is not a problem anymore.

What schematic drawing link were you referring to ?I will have more time tomorrow if you say what you mean I will try to find that.

 

[jim hardy]

What schematic drawing link were you referring to ?I will have more time tomorrow if you say what you mean I will try to find that.

Hey don't go to any trouble.

You had a schematic of your SMPS

the few I've built were sensitive to layout -
with fast switching the inductance of a loop of wire can give you surprising voltage spikes
so how one routes his current carrying wires becomes important -
trace out your current path, if it makes any loops with area as much as a few square inches look out...
That's why "twisted pair" is so common.

old jim

 

[tech99]

A traditional method is to use a thermistor, maybe with a switch in parallel, so you manually give a slow start. But I don't know if the steady current is too great for a thermistor.

 

[jim hardy]

A traditional method is to use a thermistor, maybe with a switch in parallel, so you manually give a slow start. But I don't know if the steady current is too great for a thermistor.

There's a thought... A thermistor load would pull current until it heats up.
Picking one with enough thermal mass might absorb his "spikes" but waste hardly any energy thereafter..

 

[tech99]

There's a thought... A thermistor load would pull current until it heats up.
Picking one with enough thermal mass might absorb his "spikes" but waste hardly any energy thereafter..

Just for clarification, the thermistor is placed in series with the load. This arrangement was often used for slide projectors to protect the lamp at switch on.

 

[Salvador]

Well if one connects it in series with the load I don't know how good that is because this amp and psu can supply and both draw quite a bit of load, with the resistors I am having now I'm wasting some 7 watts at idle at least upon my calculations , isn't that bad for a psu that's only used when loud music is played.

As for the smps circuit.

 

[jim hardy]

Just for clarification, the thermistor is placed in series with the load. This arrangement was often used for slide projectors to protect the lamp at switch on.

i didnt complete my thought

indeed a NTC thermistor would go in series with load to limit inrush current
a PTC would go in parallel, to absorb substantial energy at startup and less thereafter.

7 watts isn't a lot - one Christmas tree lamp.

Thanks Salvador

i'd try slower startup, bigger softstart capacitor (pin 8?)
though it looks like you have sizeable part of a second as is .
Hmmm is that 33 uf softstart cap known to be good?

 

[jim hardy]

re Tom.G's analysis in earlier thread...

how hot are your switching transistors ?

 

[Salvador]

I actually increased the softstart capacitor to 47uF.When observing with a scope it first kind of establishes the peak of the square wave and then more slowly expands the duty cycle.
the cap itself should be good.The IGBT's actually get only mildly warm after some 5 mins of extreme tests that I did with boiling water in a kettle and having a big bulb and another heating element all hooked up the psu, all in all I measured the voltage under that load and then did some simple ohms law and the psu can basically go up to some 1.5KW of power output total. i made two separate transformers to kind of balance the load as I will be driving 4 channels so each stereo pair uses one transformer.
Actually to my surprise the switching transistors are among the coldest parts.even the mains rectifier bridge got hotter, and especially the secondary side high frequency rectifying diodes get quite hot,even the transformer cores got like really hot so what I did was I made heatsinks for all of those parts.my box is made out of aluminum so I use the very box as a heatsink since it has large surface areas.I made a heatsink even for the transformer core so that in case it will be driven under some hard load it could stay cool to keep the winding isolation safe.

the problem now is that one of the channels distorts while playing music and the other one doesn't it sounds ok.their both essentially the same layout.
so i will have to dive into that for now, also if i won't find the problem maybe you guys could help me out.