Author Topic: Simple amplifier soft start for inrush protection?  (Read 4522 times)

0 Members and 1 Guest are viewing this topic.

Offline Per HanssonTopic starter

  • Supporter
  • ****
  • Posts: 428
  • Country: se
Simple amplifier soft start for inrush protection?
« on: October 17, 2021, 08:56:58 pm »
I have been fixing up an ancient LEM amplifier for a friend of a friend.
It has been a learning experience: it has two 550VA toroidal transformers and originally one of the 4x 15000µF 80v bulk capacitors was shorted above 50v.
That is to say it tested fine on my ESR meters but when you hooked it up it blew the fuse.
Through that I also learned the hard way how much energy such a transformer requires with no load!
So I felt quite confident that with new capacitors it will work fine but it does not. (Probably because their ESR is much lower than the originals).
I bought these on Mouser rated for 100v instead since the unit runs nominally at around 80v.

Now with these to get the amplifier to start is a matter of luck:
With the toroidals on their own you could at least get lucky and time it with the AC mains cycle at its peak and then it worked.
But with the new capacitors it pretty much trips the 10A house circuit breaker 9.5/10 times...
I first researched to use inrush limiting NTC thermistors but soon found that they are not available to handle such large capacitance values.

So a soft start seems mandatory now, and I thought I had figured out the most elegant solution:
Just a 100VDC relay with its coil connected to the bulk capacitors.
They would charge through a high wattage resistor and as they reach 70v the relay should pull in (at 70% of its rated voltage).
But then I read that this is bad because the pull-in will be "soft" and create more arcing and wear on the contacts than normal.
So I made the circuit a bit more complex by using a MOSFET with built in zener protection diode:
I have tested this on a breadboard with a 24VDC relay and have made the following conclusion:
With the resistor values shown in the attached schematic I reach 70V before it pulls in, so that is very good.
However when it does pull in it does so at only 12VDC, i.e. half of the rated coil voltage, so the pull-in is very soft.
My question is is there any way around this, i.e. to give the relay a hard turn-on with 24VDC?
(I should clarify that the coil does see 24V+ when I reach 80v).
Or would I then have to admit defeat and go for one of the more complex designs like these: Project39 or SST-01 that use a dedicated power rail?
I have tried to understand if for example a SCR would have a less linear region than a MOSFET does.
Or if some relay driver could help but I'm not sure really when ramping up the voltage like this?
« Last Edit: October 17, 2021, 09:06:08 pm by Per Hansson »
 

Offline Benta

  • Super Contributor
  • ***
  • Posts: 5839
  • Country: de
Re: Simple amplifier soft start for inrush protection?
« Reply #1 on: October 17, 2021, 09:42:45 pm »
Your analysis is good.
Bringing two 550 VA transformers on line is a challenge in itself, and then charging the caps afterwards makes it even worse.

Your approach is good, and I hope the 100-ohm resistors are powerful enough.

What you're looking for is a "snap-on" response for your relay.
An idea would be to insert a Zener diode in series with R5. As a first try: 56 V Zener, R5 and R6 both 10 kohm. Additionally, I'd reduce R1 and R2 to 33 ohms just to get the voltage to increase faster.
 

Offline Per HanssonTopic starter

  • Supporter
  • ****
  • Posts: 428
  • Country: se
Re: Simple amplifier soft start for inrush protection?
« Reply #2 on: October 18, 2021, 08:37:11 pm »
This morning when I woke up I had the idea "schmitt trigger!" in my head.
So since I anyway needed some power resistors and other misc stuff I added a TCA345A that I found on the same website I order small orders from.
To be continued when the parts arrive :)

Your analysis is good.
Bringing two 550 VA transformers on line is a challenge in itself, and then charging the caps afterwards makes it even worse.
Thanks, yes as I alluded to this has been a learning experience, these two threads where really helpful: can a transformer alone blow fuses? and AC Active Soft Starter for Inductive Load.

Your approach is good, and I hope the 100-ohm resistors are powerful enough.
Additionally, I'd reduce R1 and R2 to 33 ohms just to get the voltage to increase faster.
It is a 50W power resistor I happened to have, when the design is finalized I might go for something lower both in wattage and resistance, but then again that will require some testing to see what works best.

What you're looking for is a "snap-on" response for your relay.
An idea would be to insert a Zener diode in series with R5. As a first try: 56 V Zener, R5 and R6 both 10 kohm.
I'm not at home now so I can't bredboard this but I did the next best thing and ran it through a simulator but the end result seems to be quite the same as I have now?
I.e. the MOSFET will see a linear increase in voltage, so at around 63v or so it will not be fully on anyway?
 

Offline Benta

  • Super Contributor
  • ***
  • Posts: 5839
  • Country: de
Re: Simple amplifier soft start for inrush protection?
« Reply #3 on: October 18, 2021, 09:12:12 pm »
You're right, I had a brain fart in my first response. I meant DIAC, not Zener.

Use 68 kohm for R5 plus an ST DB3 DIAC in series. Stay with 10 kohm for R6. That should trigger the MOSFET and after triggering keep the gate voltage at 10...12 V.

Unfortunately DIACs are rare today.
« Last Edit: October 18, 2021, 09:17:31 pm by Benta »
 

Offline Per HanssonTopic starter

  • Supporter
  • ****
  • Posts: 428
  • Country: se
Re: Simple amplifier soft start for inrush protection?
« Reply #4 on: October 24, 2021, 01:51:48 pm »
You're right, I had a brain fart in my first response. I meant DIAC, not Zener.
(TLDR: Skip to my last paragraph)
I had a brain fart moment too, I tried using the TCA345A I mentioned in my previous post to drive the MOSFET gate.
But it has a NPN transistor output, doh!
So I connected it to drive the relay directly, as that should work I thought, and well, it did once!
Those that already had a coffee before me might realize the mistake:
The open ciruit voltage is up to 70v before the relay closes.
But the TCA345A has a built in 13.6v Zener to protect its output, bye bye zener :D
Use 68 kohm for R5 plus an ST DB3 DIAC in series. Stay with 10 kohm for R6. That should trigger the MOSFET and after triggering keep the gate voltage at 10...12 V.

Unfortunately DIACs are rare today.
I tried the DIAC with the recommended circuit.
But it has the same problem with the 24v relay switching on at only 12v as in my original schematic.
The reason is that the voltage formed after the DIAC is not sharp either:
I.e. it switches on at 32v but then if we increase the voltage to 35v we have only 3v at the output.
So same problem...

So now I drew the attached schematic, compared to the TCA345A here I have a complexity advantage:
The LM393 comparator I'm using has the advantage that the non-inverting input is not tied to VCC.
So as I ramp up the voltage it becomes much easier to calculate the divider ratio for the input resistors.
Since both the TCA345A & LM393 will start working with a supply of only 2v, so we must make sure to not trigger early!
As the voltage rises beyond the D1 Zener voltage the divider ratio gets smaller.
And with the selected components it triggers at around 70v.
The MOSFET sees instant 18v from the Zener and the relay triggers with a hard snap-on.
Then the selected hysteresis resistor turns it off at around 67v.
The comparator even offers 16mA sink so in theory I could directly drive a sensitive relay with it if freewheeling protection diodes where added.
However I'm still not so convinced that the initial intention / subject of this thread has been maintained anymore ;)
At least I had some fun! Have I made any significant mistakes in my current design? This is my first ever op-amp circuit...
« Last Edit: October 24, 2021, 04:43:41 pm by Per Hansson »
 

Offline Terry Bites

  • Super Contributor
  • ***
  • Posts: 2353
  • Country: gb
  • Recovering Electrical Engineer
Re: Simple amplifier soft start for inrush protection?
« Reply #5 on: October 24, 2021, 04:56:34 pm »
The real solution is to just not use torroids of course.
 
A separate supply for the control circuit seems like a good idea. Dont get involed in solid state switching and Zeners that will give you more compexity and issues down the road when the go bad- and they will. I love the smell of scorching dust in the evening, better than a single malt.
See Mr Elliots idea(s)
https://sound-au.com/articles/soft-start.htm a fab website for audio- not for audiofools, just striaight up years of experience.

Mmmm NTCs. According to www.ametherm.com/blog/inrush-current/transformer-inrush-current-40va-transformer You would have 70A of inrush per transformer! Requirng a pair of 1.2KJ thermistors. That's what, a 1000 kr. Madness. So a 44089 18V  1.5VA or similar is more like 50kr. The 2N7000 is a bit of a fossil but dirt cheap and still readily available. It looks like your relay draws about 20mA? So a 200mA MOSFET will do. I dont see a freewheel diode acoss your coil. Its a must. Delay type fuses too. You might also fit a snubbers across your NO and COM contacts- 100R in series with 100n will do. Get some big fat bleeder resistors on those massve caps. 80V from a stored 50J will shorten your life to a brief look of surprise.
 

Offline Benta

  • Super Contributor
  • ***
  • Posts: 5839
  • Country: de
Re: Simple amplifier soft start for inrush protection?
« Reply #6 on: October 24, 2021, 07:24:36 pm »
A problem I see with yout latest circuit is, that your resistance values are far too high. This makes parasitics and bias currents play roles in the functionality.
As an example: an LM393 powered from +18 V is fine. But using 1M resistors to bias an input from there is utopic.
The same thing probably happended with the DIAC. The leakage current prevented it from working as a DIAC.
Don't be squeamish, you need to run several mA through this kind of circuit, not uA. And if a couple of resistors protest, use larger ones.
 

Offline Benta

  • Super Contributor
  • ***
  • Posts: 5839
  • Country: de
Re: Simple amplifier soft start for inrush protection?
« Reply #7 on: October 24, 2021, 07:28:04 pm »
The real solution is to just not use torroids of course.

Wrong.
The real solution is to use quality transformers that are not so starved on copper and iron that they saturate at the slightest overload. Whether toroids or EI makes no difference.

But cheapness is cheapness, and Alibaba products rule.
 

Offline james_s

  • Super Contributor
  • ***
  • Posts: 21611
  • Country: us
Re: Simple amplifier soft start for inrush protection?
« Reply #8 on: October 24, 2021, 07:45:00 pm »
I have used a very similar approach to what you proposed except it was to bring up a 4kVA distribution transformer wired backwards to get 4160V to power a series lighting regulator. I used a large power resistor in series with the primary and then a contactor with a 240V coil wired in parallel with the primary with the contacts bypassing the resistor when closed. I have not had any problems with contact wear, the contactor pulls in quickly, there is some inherent hysteresis due to the way magnetic fields behave. Try it and see what I mean, connect a relay coil to a variable bench supply and try to make the relay pull in slowly, it won't happen. Due to the inverse square law, the closer the relay armature pulls to the coil, the stronger the pull, you get inherent positive feedback.
 

Offline pqass

  • Frequent Contributor
  • **
  • Posts: 704
  • Country: ca
Re: Simple amplifier soft start for inrush protection?
« Reply #9 on: October 24, 2021, 08:10:17 pm »
...
I have tested this on a breadboard with a 24VDC relay and have made the following conclusion:
With the resistor values shown in the attached schematic I reach 70V before it pulls in, so that is very good.
However when it does pull in it does so at only 12VDC, i.e. half of the rated coil voltage, so the pull-in is very soft.
My question is is there any way around this, i.e. to give the relay a hard turn-on with 24VDC?
(I should clarify that the coil does see 24V+ when I reach 80v).
...

Break R4 into two separate series-connected resistors; R4a and R4b.
R4a is sized so that the coil receives 24V when the secondary is fully powered-up (assume R4b=0ohm for this calculation).
R4b is sized to say, half the coil resistance.  Then put a 470uF cap in parallel with R4b.

The 470uF cap in parallel with R4b will give you snap-on action but after it is charged R4b will limit the holding current to save power.  See Figure 4 at https://sound-au.com/project39.htm

In case I'm not understanding the last part, "...see 24V+ when I reach 80V", you can over compensate.  That is, at 70V, choose R4a such that the coil sees 24V, then add some more resistance to R4b so that it doesn't cook the coil at fully-on and save some power too.

Also, WRT the relay switching, why not just use two 24V zeners+resistor(s)?  See attached screenshot.
Notice the LED turns-on only when close to the top of the hump (includes a slight mechanical delay).  The capacitor appears as a short (1.2V) effectively removing one of the resistors giving a decisive snap-on action.

Play with it here:  http://www.falstad.com/circuit/circuitjs.html?ctz=CQAgjCAMB0l3BWEAWB0DsZnMwZgEwBsAHAJySkgKRVU0ICmAtGGAFC7Ij7IBu44DFlwJ8xMKTFh4+ZuijdY8ZSsjEuPcGwBe4dIXBh8eg4Q3c+OkyDPWwB4z15sw6YtzUWa+TwRoGIGGMYSDEEMARqXBIcSARiXChYJGkVJNCPdIj0uK0AJ2seGlcA-QVUyDYAc2s-WtEFSv4WMuR3JjaUd2LwGmJgxugENgATEBaDTpYjcYJuEBGGADMAQwBXABsAF1HxzqnWYyY540XVzZ2x6eMjdpKu+bP17d3r8DE9907T5eedmret3Gh3A2EabAA7sCZscbmVYVA2AV8AgDKQHKjZgMKpV7I5MejuJiio9fhcmBsGGNAoNIOwoSiDAj7gjKlCJMY6vc6pUAA7A0pM+xeAQ0NkgQlcso8yEShzIYplEmVMaSnxykU-c4vKGE+76sqVaJ3GFzN5MWxi2U+dwkxkiyoFA2CwwDWKyljymiE5Ue4X3PWGtgAYw1Aa9tPgEDAxGgIkI9nQ+iICFwMZy7Fw6EVTLm+oC4IA9oYbFwaArSO57GLoDSbhBEgA1ADCzbYxeFlpQIi4CqSNIgAAUAPIAdQAogAlJgAZQAgo3xwAdADOk-HM4AkjOACrDyft8CJLsV+RIEJGc-zMCJRIAGXHABFD-h5CeKPJq+kr455Ilm8Om53quzYAKqTuuAByO5sEAA
« Last Edit: October 24, 2021, 10:47:45 pm by pqass »
 

Offline nigelwright7557

  • Frequent Contributor
  • **
  • Posts: 689
  • Country: gb
    • Electronic controls
Re: Simple amplifier soft start for inrush protection?
« Reply #10 on: October 24, 2021, 10:10:26 pm »
1/ relay and resistor.
2/ relay and thermistor
3/ Triac controlled soft start

1 or 2 are simplest. 3 needs a microcontroller to change phase angle for triac.
 

Offline MarkL

  • Supporter
  • ****
  • Posts: 2120
  • Country: us
Re: Simple amplifier soft start for inrush protection?
« Reply #11 on: October 25, 2021, 04:42:49 pm »
I have used a very similar approach to what you proposed except it was to bring up a 4kVA distribution transformer wired backwards to get 4160V to power a series lighting regulator. I used a large power resistor in series with the primary and then a contactor with a 240V coil wired in parallel with the primary with the contacts bypassing the resistor when closed. I have not had any problems with contact wear, the contactor pulls in quickly, there is some inherent hysteresis due to the way magnetic fields behave. Try it and see what I mean, connect a relay coil to a variable bench supply and try to make the relay pull in slowly, it won't happen. Due to the inverse square law, the closer the relay armature pulls to the coil, the stronger the pull, you get inherent positive feedback.

This is exactly what Heathkit does (did) in their higher power amplifiers.  See snippet below from their AA-1640 2 ch, 200W/ch amplifier.  I've had one for 35 years, and the surge protection relay, and everything else, is still working fine.

Another option is a settable time delay relay.  Typical delay ranges are 0.1 to 30 seconds, and are adjustable via a built-in knob.
 
The following users thanked this post: Per Hansson

Offline james_s

  • Super Contributor
  • ***
  • Posts: 21611
  • Country: us
Re: Simple amplifier soft start for inrush protection?
« Reply #12 on: October 25, 2021, 07:26:32 pm »
Interesting that Heathkit wired the relay coil before the resistor though, I wired it after the resistor, the thought being that once the surge has passed and the transformer core is magnetized the primary voltage rises and the relay pulls in. The way Heathkit has done it is going to create a delay that is apparently enough.
 

Offline Per HanssonTopic starter

  • Supporter
  • ****
  • Posts: 428
  • Country: se
Re: Simple amplifier soft start for inrush protection?
« Reply #13 on: October 25, 2021, 08:14:43 pm »
A problem I see with yout latest circuit is, that your resistance values are far too high. This makes parasitics and bias currents play roles in the functionality.
As an example: an LM393 powered from +18 V is fine. But using 1M resistors to bias an input from there is utopic.
The same thing probably happended with the DIAC. The leakage current prevented it from working as a DIAC.
Don't be squeamish, you need to run several mA through this kind of circuit, not uA. And if a couple of resistors protest, use larger ones.
I did use the resistance values you suggested for the DIAC circuit, do you suggest I go much lower?
As for the comparator circuit I posted I did breadboard it and it works fine on the breadbord.
Well with the obvious side effects of adding the 10M of a multimeter in circuit can make it act up of course ;)
What would you suggest, around 100K instead?
The real solution is to just not use torroids of course.

Wrong.
The real solution is to use quality transformers that are not so starved on copper and iron that they saturate at the slightest overload. Whether toroids or EI makes no difference.

But cheapness is cheapness, and Alibaba products rule.
Agreed, and this is not a cheap product, it is made in Italy, transformers included.
It is also a rack mount unit and only 2U high (88mm) so good luck fitting a big EI core transformer in there ;)
I don't have a decent picture of it but I shamelessly stole one I found on an auction site and attached it to this post.
It also has a very manly solution to the problem it seems  :-DD
I have used a very similar approach to what you proposed except it was to bring up a 4kVA distribution transformer wired backwards to get 4160V to power a series lighting regulator. I used a large power resistor in series with the primary and then a contactor with a 240V coil wired in parallel with the primary with the contacts bypassing the resistor when closed. I have not had any problems with contact wear, the contactor pulls in quickly, there is some inherent hysteresis due to the way magnetic fields behave. Try it and see what I mean, connect a relay coil to a variable bench supply and try to make the relay pull in slowly, it won't happen. Due to the inverse square law, the closer the relay armature pulls to the coil, the stronger the pull, you get inherent positive feedback.
This is a good suggestion, my tests have been with a 24VDC relay and I can for sure make it pull in slowly, that is very obvious when increasing the voltage to it using a bench PSU...
But your idea of putting a 230VAC contactor directly on the mains does sound feasible, I'm just not sure if the caps will have time to charge enough so maybe I will blow the fuses (or power resistor) anyway?
Break R4 into two separate series-connected resistors; R4a and R4b.
R4a is sized so that the coil receives 24V when the secondary is fully powered-up (assume R4b=0ohm for this calculation).
R4b is sized to say, half the coil resistance.  Then put a 470uF cap in parallel with R4b.

The 470uF cap in parallel with R4b will give you snap-on action but after it is charged R4b will limit the holding current to save power.  See Figure 4 at https://sound-au.com/project39.htm
Thank you for explaining that circuit for me, I had actually read it before creating this topic but not fully realized the implications of it.
Also, WRT the relay switching, why not just use two 24V zeners+resistor(s)?  See attached screenshot.
Notice the LED turns-on only when close to the top of the hump (includes a slight mechanical delay).  The capacitor appears as a short (1.2V) effectively removing one of the resistors giving a decisive snap-on action.
Remember that a 24v Zener lets through any voltage lower than 24v, so this creates the problem of a slow pull in (around 12VDC using a 24VDC coil in my tests)
But maybe together with your other ideas above that may indeed be feasible.
 

Offline Benta

  • Super Contributor
  • ***
  • Posts: 5839
  • Country: de
Re: Simple amplifier soft start for inrush protection?
« Reply #14 on: October 25, 2021, 08:36:52 pm »
Wow, nice rack! I like it.

 

Offline Benta

  • Super Contributor
  • ***
  • Posts: 5839
  • Country: de
Re: Simple amplifier soft start for inrush protection?
« Reply #15 on: October 25, 2021, 08:42:15 pm »
What would you suggest, around 100K instead?
More like 10k for the LM393 biasing.
Look, you're trying to power a couple of 550 VA transformers. This kind of "green" power saving is too Thunberg, somehow.   ;)

 

Offline Per HanssonTopic starter

  • Supporter
  • ****
  • Posts: 428
  • Country: se
Re: Simple amplifier soft start for inrush protection?
« Reply #16 on: October 25, 2021, 08:55:06 pm »
What would you suggest, around 100K instead?
More like 10k for the LM393 biasing.
Look, you're trying to power a couple of 550 VA transformers. This kind of "green" power saving is too Thunberg, somehow.   ;)
Well, I'm not trying to power them with the comparator :D
And just to clarify again that picture was something I found online, the LEM amplifier sits one up from the bottom.
Do note the 3-phase power cable on the top! :D
 

Offline pqass

  • Frequent Contributor
  • **
  • Posts: 704
  • Country: ca
Re: Simple amplifier soft start for inrush protection?
« Reply #17 on: October 25, 2021, 09:16:00 pm »
...
Break R4 into two separate series-connected resistors; R4a and R4b.
R4a is sized so that the coil receives 24V when the secondary is fully powered-up (assume R4b=0ohm for this calculation).
R4b is sized to say, half the coil resistance.  Then put a 470uF cap in parallel with R4b.

The 470uF cap in parallel with R4b will give you snap-on action but after it is charged R4b will limit the holding current to save power.  See Figure 4 at https://sound-au.com/project39.htm
Thank you for explaining that circuit for me, I had actually read it before creating this topic but not fully realized the implications of it.
Also, WRT the relay switching, why not just use two 24V zeners+resistor(s)?  See attached screenshot.
Notice the LED turns-on only when close to the top of the hump (includes a slight mechanical delay).  The capacitor appears as a short (1.2V) effectively removing one of the resistors giving a decisive snap-on action.
Remember that a 24v Zener lets through any voltage lower than 24v, so this creates the problem of a slow pull in (around 12VDC using a 24VDC coil in my tests)
But maybe together with your other ideas above that may indeed be feasible.

No.  Look closer at the attached screenshot or at the falstad.com simulation.
The zeners are stacked and in-series with the coil.  They effectively block 48V from Vcc before they let any current flow.
They behave like a switch.  The LM431 is effectively used as such in 99% of all switching power supplies out there.
See first IV plot here: https://en.wikipedia.org/wiki/Zener_diode

Just like a forward-biased diode has a voltage drop, zeners can have a lot more drop when they are connected in reverse.
« Last Edit: October 25, 2021, 09:29:32 pm by pqass »
 

Offline Per HanssonTopic starter

  • Supporter
  • ****
  • Posts: 428
  • Country: se
Re: Simple amplifier soft start for inrush protection?
« Reply #18 on: October 26, 2021, 05:24:37 pm »
In reply #2 I show the reason for the slow turn on of the relay in my OP.
The circuit suggested by Benta has the same drawback as in my OP:
The voltage on my MOSFET increases linearly and as such the coil sees a gradual increase in voltage.

Then in Reply #4 I switched to a DIAC but it behaves much the same way.
And I think I understand why now, it really is just a diode that works in both directions.
So the behavior is the same and the MOSFET is still not out of the linear area as the voltage ramps up.

Then in the same message I posted the comparator based design, this works fine and fires the relay hard.

I'm trying to understand the circuit by pqass because I really like it's simplicity.
But I just run into thinking that it has the same issues as the first two above.
That is that the 24VDC relay will pull in at around 12VDC or around 11mA of its rated 22mA as in my first circuits.
Unfortunately I'm not at home so I can't breadboard it right now for a test...
I'm also thinking that maybe I'm overthinking it :D
« Last Edit: October 26, 2021, 07:20:35 pm by Per Hansson »
 

Offline MarkL

  • Supporter
  • ****
  • Posts: 2120
  • Country: us
Re: Simple amplifier soft start for inrush protection?
« Reply #19 on: October 26, 2021, 06:35:48 pm »
Interesting that Heathkit wired the relay coil before the resistor though, I wired it after the resistor, the thought being that once the surge has passed and the transformer core is magnetized the primary voltage rises and the relay pulls in. The way Heathkit has done it is going to create a delay that is apparently enough.
Ah, right.  They are clearly relying on the relay delay.

Reading a bit in the "Circuit Description" section, they say the relay delay is about 15ms.  So that's enough for about two 60Hz half-cycles of pre-charging for the main capacitors.
 

Offline Per HanssonTopic starter

  • Supporter
  • ****
  • Posts: 428
  • Country: se
Re: Simple amplifier soft start for inrush protection?
« Reply #20 on: October 27, 2021, 03:50:26 am »
Interesting that Heathkit wired the relay coil before the resistor though, I wired it after the resistor, the thought being that once the surge has passed and the transformer core is magnetized the primary voltage rises and the relay pulls in. The way Heathkit has done it is going to create a delay that is apparently enough.
Ah, right.  They are clearly relying on the relay delay.

Reading a bit in the "Circuit Description" section, they say the relay delay is about 15ms.  So that's enough for about two 60Hz half-cycles of pre-charging for the main capacitors.
Unless my math is off to achieve a full charge in 15ms would require a 0.25Ω resistor, its peak power would be around 25kW!
60000µF at 80v is 192J of energy
Or put another way the capacitors have an ESR of 12.5mΩ that should result in a initial circuit load of 512kW.
So while that 25kW number sure looks silly at first for sure it is a big improvement over no resistor!

Another idea I had this morning would be to use two relays: this way snap action of the main relay can be assured.
The resistance value R5 in this example is critical and might not be correct: it should be sized so it pulls in when the main relay sees 24VDC.
 

Offline pqass

  • Frequent Contributor
  • **
  • Posts: 704
  • Country: ca
Re: Simple amplifier soft start for inrush protection?
« Reply #21 on: October 27, 2021, 08:04:03 am »
...
I'm trying to understand the circuit by pqass because I really like it's simplicity.
But I just run into thinking that it has the same issues as the first two above.
That is that the 24VDC relay will pull in at around 12VDC or around 11mA of its rated 22mA as in my first circuits.
Unfortunately I'm not at home so I can't breadboard it right now for a test...
I'm also thinking that maybe I'm overthinking it :D


So I decided to play around some more with both my proposed approach and yours, then ended-up with a third.
See attached screenshot for a side-by-side comparison.

Since the last simulation, I've now correctly defined the relay (1.09K@22mA=24V) and lowered my zeners to a total of 36V (12V+24V) to block the low and slow Vcc initial ramp-up. I've also added a "power-save" resistor to both circuits (the one with the parallel cap) to give an initial punch of over-current before settling back to a lower coil holding current.  Wherever you see a 1R resistor, I've effectively removed it from the circuit.

As per the simulation, my [first] circuit would ramp-up the current faster (first to reach 22mA in the COIL CURRENT graph, 0.75 vs 1 division).  But I'm not sure this would make much of a real difference.  However, using the power-save resistor in your [second] circuit did improve it a bit.

Then I remembered how crow-bar circuits work in power-supplies.  The third circuit uses a triac (or scr) that is triggered when 72V is reached. You can use 3, 24V zeners or a Vbe multiplier circuit (the BJT+2 base resistors arrangement) to make the equivalent of a 72V diode drop. As you can see, this results in an instant sub-millisecond coil current ramp-up!

Watch out for the power rating of: the zeners (500mW), BJT (900mW), and coil resistor (>1W).  Also BJT Vce >80V.

Let it run for 20ms and then stop it.  See:
http://www.falstad.com/circuit/circuitjs.html?ctz=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-Zx1LrINKyMckWB6ZexOZNqUA4AY3Dkwr8YjVce-WU7HkXGI0CmliSbLqu1xgyUrFQAZiZtmsX1B0BtvTtozDFW63Te1yIEQ-S4wvSGHtm8Px4j+SP8oZ7CPttbN-tYj2PtPK8uyE+J4OIofb9fDraVQ41-B1E23ACDzXM13C2QsKwWY1VSzJleW2A5J1g5CoN5TR5yWStPVAp880IdZCHtJCmGtdIKw1OltV1AR2xKMY9HI4sqIDNDwGUOMeHtVNOJozgDmNdcsPYmVRJWOpbSCNdZI2S0ZM-TZsNtTZpKUuTzHEnIggOLFRB0jS4I0aNNWhY1jMTZs9NTetTJSExtNskSO3gLs93Ycp2CkdxyjyDBdyGVj1LmS49NiKhunlMToJcmyYK0hNZxiTQUtbGDQOSlgAAdG0rBMYLw40RBw3SZwylgGwYS8sPqowfkeJRyCPaB10QXB7mmcocFwDtWAZRqKVhAVWXZZ1jUa3NQNzA4CpbcUXITVM8MWzB-VdeTPVK7aPSdJawxS+SiyOttnQ207m02ksmKm-bhMKtznSSSKMietcPv0zgCk6Youh89Q1DULgsDAZAewKFpAI6Epuh4voWLHOARmA8DEB2tTMZE-L0nAwTypMvawLMpyXoW-HyRcomXpJ2nsqS+Y8dJysGZMPC8gKaUcZANljv6QolHuHiqEhUZQToSDCBMqiBa7K82M-Q6ZZyc80zdJWJM+Eq83GDGnWYDcnQvLTRplcL5rjVXc26C1LpAZtYmbNa825F2K2EJtcYKj3fo9II9ubXmvZMg43u4RTUyhky-qKbyuAcLBJ0sNREEgVZEdh5Q9x6GHBZR4YtHd5MNHS3Hxjtp6q4A7g8mk0OXprhyYUbkPpWjiy6RdNughr3npJr3MP2Qweo4rEeKYPSekMn+bRflUid2XE1l8+TuiQxBi9X2QW2E1oe-nL+7pINqgMYM-D+JM6+112PNljbw3pWIl8hItm20JZ-mIveT9A7dn-ZylxVa5WND-XmcsfrOkhFrRMN4dyFgQW0cM0BjDTF2GAUgwhcAiWgFgVIHAbzkXUHkFAu52hd1jCKVWYoQE63mDuVo8oaHrwQVAgeTDT4biCATbc3CnrgTAYg4BcDhEcN4Tw3GAB7cAEB4IgAyN1O4MAuzsGMF1VG5UQr+H1D0JAIAABqABhYxLBZGskdu5JRewbywHuGKLqicWhiCaB5fUEAABaABRAAct4gASgAZXMXXKxCgbGKL8LgnyHA1ioHwDUdxvgwm7gAAoAHkADqgSAA6ABnIJABBQx3iCkBO8UEgAkkEgAKhkgJoTNAKKUfIIYpR+gEAwMOCoEpMAoBQBIdyIRNC7mMRkqpAAZApxiACqASKm+NqaEiG4SqCtOoCxHsKTVm7kmd4gAIjM+ZizlmyOYGsxRLQ9jtTURo5hEpcg6PgM4OIBiTFmNkQQS5kT4aSEiGKDOGR05NCFi8gQIRvnpOyXkwpJSyn5IqdUupDSml7Badcig0AHA42tIid0xAHHJIgCMrcIAADiRTaneKaeQCs8glEDTHK8oWBBtkAFkMlBIAGLeNqQUzJOTGlfN3Bik4yT4blTEGDKYuQKjVHkCS5eO4QDjKmcchZfizk7jpdYlo8gUYOBGNYEgKr9lHPyXMzVSzQlpDFTcmAyBIjBWQH1GoA12lKpCGkXcHzbWir1co28MAlUQEMQAIzoAU9lszJm1KqWkyZVTAmhPvmKg1e5yAjLuGStV0zLUnK1am+Q6bNmGm2dmxoZrDkatOWwPY4qUFykyAEug+SACW+TITAgAHZ1ilkAA
 

Offline pqass

  • Frequent Contributor
  • **
  • Posts: 704
  • Country: ca
Re: Simple amplifier soft start for inrush protection?
« Reply #22 on: October 27, 2021, 05:26:00 pm »
CORRECTION: Forget my first proposed zener+resistor solution.  Go with the last (triac) circuit.

My last simulation was unrealistic. Its time-frame was on the order of a quarter 50Hz wave-length. The one below is 1000x slower!

I've since added in the 15MF cap and 100R slow-start resistor and that changes everything.  It takes at least 4 seconds (2 divisions) to full charge the rail to 80V!   In my first circuit, even if the zeners block passage until 36V appears, there is still a very gradual slope to 80V that doesn't have the ummpf to pull-in the relay. And the power-save resistor/cap is too short to do anything useful.

With your circuit, the slow rise also lets the MOSFET linger in the region between Vgs(threshold) and when it's fully-on.  Still 2 seconds spent there (1 division).

In the last circuit, the Vbe multiplier snaps the triac fully-on when the rail reaches 72V however long that takes. With the 2k2 coil resistor @ 72V gives enough current to pull-in the relay quickly.

Click the Reset button and stop 12 seconds into it:

http://www.falstad.com/circuit/circuitjs.html?ctz=CQAgjCAMB0mRBmBB2ayAcA2ZBOBZsEBWYokIyc8yogUwFowwAoBAFhACY2A3ccNGDbFO6MDlFg4kTg2RQusacpWR0HbuGYAvEPUyb6ySB32HkRTly69WHMJz4Qwg4ZbESx02UYWclqqrq4FYsumamFlYReuhw1g48zGAYekQSaciYaRn0bDjZ2RAwVjAyROlEWJhwolii4JA4lFIqULCcoU2UZWBkZZBkLABOegammGyUMUYmCiwA5nqcyNHuy6t6bJMKkMx89DhZelnosWxnRphnLdTgfbvtRMwAJnpH2VeXOA7nZ1YvWgAMwAhgBXAA2ABdXn8Ttd3r96Op-iBAaDITC3ockddvscvtZ0eDobCcdE8XDCQDgSSYUtyWN0N9cZseswAO6xeLIi4nTjTFFQZijehENiRAVpCVwlrKZIGaWS6bi0xEY40jHQ+gQ2hvYq7WAsLlimW8y7jWWcsZqLYFMaGfLZPYm4ymJ0O932l0nOZigmW-3Osluu2fUNBom0zHW2YTKaesYJvYIBGM83vXJC9km9JrAmqzLB0WBybTQPq50ga57ADGvvj0wjlcNKjY8hS0CmyHEPwFCBWWHacBYKHLhiFMzit3ZQL0SE+hfoKzWMoHIAQw84Iq2cSVu+mFx61enewZR4Pl7yOxzV6XC-3e1FF9N7r3yPicpUsZf06264-dlRRXBsNgpZMtFzM1JyIJFs1YYwwKpEDAP4dkGUDD0Zm2Dh2QIaICCbKVGBqJNKE1OkdT1NDWxGPRCP5aYGIzL84FjbhywTEjONvZdNgzbirRNTgiEXUjl1EtIGh9CTFwaPi1mk2MRDScSUNE9lsRUsVxO0ssoy1GETT0riUP0ut50sUD6BM48BmkdstzYEh0E4TBLAcZy3IeAY6IU6yUL-ciBRklCI200MZIjP8p0-WMIwjQNIuYAAHayYoSuYDRk4RohiwK4vrVCBJiykeiNMgxC7FI4gQeAwDEbZh3gZgGVK5luXLNlhTSuNrMShNstSk5gkjIxRo9IbXQm+0+ryb14tGls5pbPZeowRsRrVY4pq2sjrPM5JUmXLj9JOlVSKKdpSg6ZyEC8AoGru9VGmaRo2jKTpXvK+B+iUIYfwAjTLLWUi1v-cDDyB0jdryACztsnqIes3L9th1G5oiuZwbhvKeQx0MhoFIozWB2aPR+4olFc96QChYYwVoHcSNUpjPgpkB2zYhkCDDej2b3dkx35qleb6obRgAyakQ5hBmljWC8r5fyWNjMWhU6Ux4LSnZUN1-Tdv1wazUOnXNt14HDbVGHSbBo7-kk-QExEsTKGyG6bpkZyaiaTgB2QbYVkcqQ3taZQty6N7ej+uAAa5XW-01zrhVGJOneCrXPy4ELrSETa08O-VbZaE3nbRaNSS5PPWez63bzToUH1Vrk6s2pvDsl1NGI3LuWPetj8J78MpXbsviUxKj9UeXoEIzuEG-iXapftaH0K4DqdPIjecN2VrwDNSaD5vYVXWIydA2195wz9WbLd3k1ZojMXkof12Rc34UoXfy14c0W4jH8MyVYTRjBCFcpgZ00AIHGFTFkfAECKi4FpuPUkooxaBmftjDczQ4BQL8ugwwYtgaUDlj6VeG4KS2jIZDChoFqHWUwbhNWb9yFPkRErS4MVszvWYYKZW6s+Q+g4jXYRYpfg+gxugWafZYgLQfnBaRcFBHWhdrEd8XDBasARE3IUUw+E3BPsnZE6ieTwS5Ko9OXBHbMnrlxRWnMgbiOtBjexkiFoAHt7jVlwpzJoZxCLQBwPkEwGB4BqEmJA6QBpQiIBAAANQAMIJOYJ4-I3iFBTBwGcC4sBXISjlpwEOogA7NXgPMEAAAtAAogAOSqQAJQAMopOzukygmT-guHbMTZAqYGoFA+KU6JrTNwAAUADyAB1BpAAdAAzo0gAgnEqp8z6lVMaQASUaQAFXGfUlpmhj6ZPkPZKmkB8DoFEsyHA4oMDIB7EEvwNgNwgASeMzZAAZeZCSACq9T1k1J2S0sAm5jkgKoL0ESzzQWvM+VUgAIj8-5gLgWeKYG03xWTwA1ECcE9scQpBYCamc-gMTXmJOSZ4-AmKOlbgkOqCUXsLmQADuckcZKNyxImdM+p8ylkrLWRs7ZeyDmeIaOC7FUhoDcCyBIBqAofiTFOQEZw1gGibgAOKLJ2VUw5ZAdjyEyTgSF0A1VU3wNCgAsuMxpAAxKpOz5k8oaS0pAtKIWkrKBcSYolLASHcrgOW9hrDus3O8r5yKAW1LRRuA1PiTmmvgNCiAVhiBwsRVG1FLSfgeqlbi7AfQIEQJQHLBqQz1UgFzZuSlOarCSrOILM1CR4kACNaDzKtb8z5OzNmjM+Zs11nj0DyElfIGO0KrBWBHa8iN3y5l-OjUClpQ4x1JocGQKdIAhybnhUihdKKY2sDON6fgYd4j1NoHMgAlnMqEIIAB2tYmZAA
« Last Edit: October 27, 2021, 05:30:45 pm by pqass »
 

Offline mawyatt

  • Super Contributor
  • ***
  • Posts: 3192
  • Country: us
Re: Simple amplifier soft start for inrush protection?
« Reply #23 on: October 27, 2021, 06:50:03 pm »
Don't know if this has been mentioned, but a simple method would be to place a light bulb in series with the mains, then bypass the bulb with a proper sized relay that is driven by the supply output voltage. The inrush current won't be any higher than the bulb by itself, and the large amount of energy to charge the supply caps will get spread over the bulb since the bulb resistance increases as it heats up.

Best,
Curiosity killed the cat, also depleted my wallet!
~Wyatt Labs by Mike~
 

Offline JohanH

  • Frequent Contributor
  • **
  • Posts: 620
  • Country: fi
Re: Simple amplifier soft start for inrush protection?
« Reply #24 on: October 27, 2021, 08:23:30 pm »
I've used this circuit (from Elektor 7-8/97) for many years for a subwoofer amplifier with a 330 VA toroid transformer. For some reason the R3 resistor has burned at some point. This resistor limits inrush current at switch-on, whereafter C1 limits current (C1 should preferably be X2 rated or similar). I changed it to a 2W resistor now to be on the safe side. Otherwise the circuit has worked fine. The circuit is meant for 230V and the relay coil is 24V (1200 ohm).
« Last Edit: October 27, 2021, 08:35:03 pm by jukk »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf