Two things here do you see the filament glowing bright for a short duration once you turn on the tube amp?
If so its called as filament flash what you need to do is to add a series resistance for the filament for about 15secs just to warm up and then trigger a relay after that.
Considering Primary V is 230VAC Coming to the total VA is 150VA which is quite low and you still need an inrush even for 150Va nothing wrong to have one.
You need to have a NTC in primary with about take 40ohm Joules will be min 10Joules capable if you have more cap take 20 or even 40J of NTC in primary.
Now why 40ohm is that as per primary voltage DC/ 40ohm will give 8.1Amp as max inrush which happens in primary. The main challenge for you is to identify the Energy in Joules that you need the NTC for.
Now if you want to increase NTC cold resistance value you can still do to 60ohm but usually 40ohms falls under the CE standards of max permitted inrush so they should be fine.
Now check your ckt with NTC in series and check if any of the classic bulbs dims badly after putting such NTC in series if its blinking without much dimming then you have got the right value.
As far as I know 40ohm NTC with 50J should be fine for your application.
If so its called as filament flash what you need to do is to add a series resistance for the filament for about 15secs just to warm up and then trigger a relay after that.
Considering Primary V is 230VAC Coming to the total VA is 150VA which is quite low and you still need an inrush even for 150Va nothing wrong to have one.
You need to have a NTC in primary with about take 40ohm Joules will be min 10Joules capable if you have more cap take 20 or even 40J of NTC in primary.
Now why 40ohm is that as per primary voltage DC/ 40ohm will give 8.1Amp as max inrush which happens in primary. The main challenge for you is to identify the Energy in Joules that you need the NTC for.
Now if you want to increase NTC cold resistance value you can still do to 60ohm but usually 40ohms falls under the CE standards of max permitted inrush so they should be fine.
Now check your ckt with NTC in series and check if any of the classic bulbs dims badly after putting such NTC in series if its blinking without much dimming then you have got the right value.
As far as I know 40ohm NTC with 50J should be fine for your application.
Replying to your first quesite, no I don't appreciate an alteration in glowing for the filaments....It looks like there is a sort of "blocked" voltage at pass, maybe I have to check if I assembled the circuit well.
The B57236S0800M EPCOS from datasheet has a Ipk=1.6A, and based upon your indications seems to be a too much low value. Anyway I chose it because the CL60 or CL80 (frequently used in this kind of power supplies) are not easily available here.
Then I'd like to bypass that with a relay (like in the Elliot's circuits), it's OK for me also using a power resistor bank or maybe I'll try dimensioning my psu.
The B57236S0800M EPCOS from datasheet has a Ipk=1.6A, and based upon your indications seems to be a too much low value. Anyway I chose it because the CL60 or CL80 (frequently used in this kind of power supplies) are not easily available here.
Then I'd like to bypass that with a relay (like in the Elliot's circuits), it's OK for me also using a power resistor bank or maybe I'll try dimensioning my psu.
So, people are thinking that a bigger KVA transformer needs bigger NTCs. Remember V=I*R, or I = V/R, its just Ohm's Law. What matters is the Mains voltage 120 or 240, you set the NTC resistance cold to limit the current rushing TO the transformer primary. I limit to 12A which is 120 volts divided by 10 ohms. As the NTC warms up the resistance will go down and the current will increase but by that time the main in-rush event is over, 3-10 seconds. Then you remove the NTCs from the circuit with a bypass relay so fully current can flow when your Amp is working hard. So the circuit works for any KVA rating, big or small. Smaller transformers may not need the soft start but will still work, like chicken soup for a cold, it can't hurt. The bigger KVA transformers, or bigger capacitor banks, with just take longer to charge up so the time is set to the full 10 seconds. You can control the relay with a PIC or a simple switch, I chose to charge a capacitor and use a voltage comparator to trigger the relay bypass event.
However you want to do it, build it, test it, validate the design, have FUN.
However you want to do it, build it, test it, validate the design, have FUN.
Perhaps a little more attention to NTC datasheets is worthwhile. There is a joule limit that the NTC can pass safely - so you need to estimate that initial surge joule level. One contributor to the joule level is charging the secondary side caps. Another contributor is heating up heaters - both for 5Vac and 6.3Vac - as they conduct significantly more current initially. Another contributor is transformer primary in-rush, although that is suppressed somewhat by including the NTC for starters.
The choice of NTC may well need to cope with a long delay for any relay bypass, or even a failed relay, so the amp's operating continuous current should be somewhat below the NTC max current rating, and also somewhat above about 30% of that rating.
If you can use a calculator then those design aspects are discussed in a section in https://www.dalmura.com.au/static/Valve amp fusing.pdf.
The choice of NTC may well need to cope with a long delay for any relay bypass, or even a failed relay, so the amp's operating continuous current should be somewhat below the NTC max current rating, and also somewhat above about 30% of that rating.
If you can use a calculator then those design aspects are discussed in a section in https://www.dalmura.com.au/static/Valve amp fusing.pdf.
Actually, I know I don't really need an inrush current limiter for my amp's VA, but I'd prefer to have one since I like the idea to have a kind of soft start. I have already made the choice of the fuse based on my parameters (1A), that of the thermistor I think is to be calculated as follows:
E = 1 / 2CVp ^ 2 = 0.5 * 324 ^ 2 * 0.00022 = 12J (thus a 30J NTC for safety could be ok)
Rmin = Vp / IL = 324/4 = 81R
Imax = Pout / efficiency * Vin = 150 / (0.95 * 230) = 0.7A
So, if my calculation is correct, the choice of a 80-ohm NTC isn't far fetched. Yesterday I attached the B57236S0800M alone in the hot line: from an initial voltage of 20Vrms falling rapidly to to 7-8Vrms final. The NTC seems to be lukewarm.
PS: for johnhenryharris, if you like to post the relay circuit with comparator it would be appreciated (it would be doable with NE555 too?). Thanks
E = 1 / 2CVp ^ 2 = 0.5 * 324 ^ 2 * 0.00022 = 12J (thus a 30J NTC for safety could be ok)
Rmin = Vp / IL = 324/4 = 81R
Imax = Pout / efficiency * Vin = 150 / (0.95 * 230) = 0.7A
So, if my calculation is correct, the choice of a 80-ohm NTC isn't far fetched. Yesterday I attached the B57236S0800M alone in the hot line: from an initial voltage of 20Vrms falling rapidly to to 7-8Vrms final. The NTC seems to be lukewarm.
PS: for johnhenryharris, if you like to post the relay circuit with comparator it would be appreciated (it would be doable with NE555 too?). Thanks
trobbins - you make good points. I limit the in-rush current to 12 amps, the NTCs are rated for 16 amps continuous, more for surge, once the initial in-rush is over the transformer and rest of the power supply draws 5 amps for the mains so even if the by-pass relay were to fail to pull in the NTCs would live fine, very warm but fine.
indaco - 80 ohms for the NTC is excessive, you want to limit the in-rush not strangle it.
The schematic, BOM and gerbers are in the first post.
indaco - 80 ohms for the NTC is excessive, you want to limit the in-rush not strangle it.
Here's the manufacturer's datasheet for the Inrush Current Limiter device mentioned in post #168.
If the amplifier is class-D (what else has 95% power efficiency?) then you need to operate the amp at full output power into an 8 ohm, 200 watt load resistor, when measuring the ICL's temperature and its voltage drop.
_
If the amplifier is class-D (what else has 95% power efficiency?) then you need to operate the amp at full output power into an 8 ohm, 200 watt load resistor, when measuring the ICL's temperature and its voltage drop.
_
Attachments
This morning I stumbled across yet one more Soft Start Circuit (link) . The unique (in my opinion) features of this one are
Disclaimer: I have no data and no opinion about whether this is a "good" or "bad" or "dangerous" or "safe" or "economically reasonable" design. You'll have to make those decisions yourself. I'm just pointing out that this project exists, and might be of interest to some readers.
- It's a completely free, open-source project with freely downloadable schematics, Gerbers, BOM, and software. Hosted by PCBWay.
- microcontroller based
- includes two relays with 30 amp contacts (!), one for on/off and the other for soft-start bypass
- inrush current limiter device is a 20 watt, 27 ohm power resistor; no mention of different resistor values for 115VAC versus 230VAC
- the temperature of the resistor is continuously measured by firmware, using a bead thermistor in contact with the power resistor. If the temperature of the power resistor crosses above a "too hot" threshold, the microcontroller declares "FAULT!!" and opens both relays, thus cutting off all AC power to the load
- quite a few voltage isolator slots, cut in the PCB
Disclaimer: I have no data and no opinion about whether this is a "good" or "bad" or "dangerous" or "safe" or "economically reasonable" design. You'll have to make those decisions yourself. I'm just pointing out that this project exists, and might be of interest to some readers.
Interesting ideas. Kind of hitting a nail with a sledge hammer using a Arduino for that but to each there own.
And I AM a fan of using Arduino for many projects.
I am not a fan of power resistors for this rather than thermistors, after having experimented with both.
And the projects on this forum are, for the most part, free.![Smile :) :)](data:image/gif;base64,R0lGODlhAQABAIAAAAAAAP///yH5BAEAAAAALAAAAAABAAEAAAIBRAA7)
And I AM a fan of using Arduino for many projects.
I am not a fan of power resistors for this rather than thermistors, after having experimented with both.
And the projects on this forum are, for the most part, free.
For a transformer up to a few hundred watts a thermistor is a good, simple choice. But at higher powers you run out of suitable choices and are forced to use resistors.
In that case, your best choice is a wire or, even better, band wound GLAZED type with a short, thick and heavy body because those offer the high heat capacity needed to cope with a short time pulse of high power. If you use one of those metal cased they need to be rated many times higher than a glazed ceramic one because they have almost no heat capacity. A thick heavy body ceramic can take up to 10 times the short time power compared to a metal cased of the same power rating.
In that case, your best choice is a wire or, even better, band wound GLAZED type with a short, thick and heavy body because those offer the high heat capacity needed to cope with a short time pulse of high power. If you use one of those metal cased they need to be rated many times higher than a glazed ceramic one because they have almost no heat capacity. A thick heavy body ceramic can take up to 10 times the short time power compared to a metal cased of the same power rating.
Last edited:
So, from the safety side of things, if the relay that bypasses the resistors doesn't then the resistor wattage has to be so over sized to handle the heat they can be expensive. Also is the resistors stay in the input line path they restrict the current to the transformer. The Thermistors decrease in resistance as they warm up so if the relay doesn't bypass them there is no harm to the circuit and sound, other than a bit more heat inside the amp. The thermsistors must be sized for continuous duty as a safety factor. I took all that into account in my design.
Or, some kind of fault detection can sense failure-to-bypass and take corrective action. The PCBWAY open hardware design linked in #171 uses a temperature sensor to detect excessive resistor heating and shuts everything off. Other circuits simply place a thermal fuse in contact with the resistor, which again shuts everything off if a fault occurs. A cute implementation uses two power resistors in series, placed side by side with the thermal fuse between them and in contact with both.... if the relay that bypasses the resistors doesn't then the resistor wattage has to be so over sized to handle the heat
Because of the simplicity of thermistors i looked some years ago for a suitable candidat.We are using large thermistors for 2400 to 4 kva toroids without issue.
I could not find anything that could have handled the inrush current of my 1kWdc powersupply.
To be fair, the toroid has been wound to have a regulation of 2.5% at 1kWdc load, so 2KVA core and much more copper than usual but still should not even come close to yours.
Just out of curiousity, where did you find thermistors that can handle the peak currents of a decent 4kVA on a 120V line ?
Ametherm's AS35 series of inrush current limiter devices, include models rated for 50 amperes of steady state current and 800 Joules of inrush energy. Mouser has them in stock.
They are BIG. Diameter = 36 mm , thickness = 6.4 mm , lead to lead spacing = 19 mm (plus stagger)
They are BIG. Diameter = 36 mm , thickness = 6.4 mm , lead to lead spacing = 19 mm (plus stagger)
Also, the resistance of the thermistor controls the amount of current, therefore wattage of the in-rush current. Read back through this thread and you will see why the values were chosen. The BOM has the part number of some large thermistors that I used in the softstart for large power amplifiers.
- Home
- Amplifiers
- Power Supplies
- Yet Another Soft Start Circuit