Maybe it's time you did consider this. Maybe you have already, since you bypass for SS ClassAB.
Some long while ago you jumped in to a thread stating that the voltage sag at the power pins of power amplifiers must be enormous and all your results proved that to be the case. I tried to argue the case that this is not necessarily so. But you did not want to hear the message.
Now that you are revealing that you leave the Thermistor in circuit on the primary side, it begins to fall into place why YOU get high voltage sag when delivering power. The source resistance/impedance at the transformer is modulating the supply voltage whenever current demand changes.
Leaving the thermistor in circuit WILL INCREASE voltage modulation at the transformer input.
If you remember my recommendations over the years, I always say that the added resistance be bypassed after start up.
My resistors/thermistors will be switched out after 100 ms so I am certain that the resistance of the NTC devices is only going one way. Decreasing.
No.
NTC means the resistance goes down as the temperature goes up.
After you take them out of the current carrying route they cool down and that means the resistance is increasing. Ready for the next time they are needed for their current limiting duty.
You don't want them hot when the power fails and comes straight back on again after a short 2 second delay.
NTC means the resistance goes down as the temperature goes up.
After you take them out of the current carrying route they cool down and that means the resistance is increasing. Ready for the next time they are needed for their current limiting duty.
You don't want them hot when the power fails and comes straight back on again after a short 2 second delay.
Thanks PRR & Andrew. Very thorough.
I started a spreadsheet to calculate the self-heating and current flow through an NTC. My only guiding principles are that I should see if I can fit 3x 10R NTC thermistors in series (20% tolerance!) and I'll be looking for a specific heat capacity that gives a substantial reduction in resistance after 2-3 cycles, before the 100ms interval expires. Seamless transition would be good.
I started a spreadsheet to calculate the self-heating and current flow through an NTC. My only guiding principles are that I should see if I can fit 3x 10R NTC thermistors in series (20% tolerance!) and I'll be looking for a specific heat capacity that gives a substantial reduction in resistance after 2-3 cycles, before the 100ms interval expires. Seamless transition would be good.
Noted... but in my current configuration my soft-start board is on the singular mains power supply prior to division of power for each transformer.
So I planned to use 30R NTC instead of 2x 60R in parallel (left-right). Assuming my left+right channels are identical (fingers crossed), does this arrangement seem ok?
you worry about a lot of things.....me, i just build them, and my amps are all over this board....
yes, but first, current needs to flow, and then the ntc heats up and then the resistance goes down...
An NTC is preferably selected such that it won't be damaged by the amount of energy passed through it at turn-on. The NTC datasheets usually provide a chart of capacitance and supply voltage for the typical mains rectified filter scenario. The energy transferred to that filter capacitance, to raise it to the peak of the mains AC rectified voltage, is the benchmark energy level that needs to be related to any application.
For your application, if the NTC is passing energy during turn-on for transformer primary in-rush, and for secondary side heater in-rush, and for charging secondary side capacitors, then the cumulative energy passed to those items is well worth assessing, and compared to the NTC rating.
There is an interesting chart at the end of the linked NTC datasheet. It probably indicates steady-state conditions. It indicates that voltage drop is somewhat constant over a wide range of operating current, although transient changes in current would take time to shift from a constant resistance locus back to the steady-state locus.
I guess it depends what transient change in AC mains current occurs in a particular amp as to whether an NTC does affect the audio signal at all.
http://dscelec.co.kr/board/main.cgi/5D20.pdf?down_num=1151457156&board=data&command=down_load&d=&filename=5D20.pdf
I guess it depends what transient change in AC mains current occurs in a particular amp as to whether an NTC does affect the audio signal at all.
http://dscelec.co.kr/board/main.cgi/5D20.pdf?down_num=1151457156&board=data&command=down_load&d=&filename=5D20.pdf
^yes, once the ntc has heated up, it stays that way....hot....
i even monitor the voltage at one time when the amp was playing,
it stays below a volt.....who am i to complain?
ntc's are not linear resistors and so simple ohm's law does not apply....
i even monitor the voltage at one time when the amp was playing,
it stays below a volt.....who am i to complain?
ntc's are not linear resistors and so simple ohm's law does not apply....
Thanks trobbins. Definitely will not be damaged by current at switch on. Even if the NTC 30R is the only resistance in circuit, RMS current will be reduced initially to 8A. Voltage could be 253 and resistance could be -20%, so 10.5A is worst case. I'd like to use 15A NTC parts as they are about 3x cheaper than 18A parts... (10R thermistors).
Of course, the resistance will quickly fall and the caps are still 'short-circuit' uncharged.
Of course, the resistance will quickly fall and the caps are still 'short-circuit' uncharged.
another thing to consider is the ampacity of your switch, if you use one that
can handle say 16A, then you have one less to worry about...
can handle say 16A, then you have one less to worry about...
look the datasheet and/or app notes from the NTC manufacturers.
They guide you through the design procedure to select an NTC with sufficient heat capacity (Joules) to survive the high charging current. That duty is specifically what these NTCs are made for.
Look at the NTC inside a mains SMPS. It charges up a 400Vdc cap direct off the mains through a small rectifier. And survives.
You are doing it again, not listening.
As the current drops the temperature drops. It does not stay hot when the capacitors have charged up.
Constant voltage across the NTC does not imply constant temperature.
If you see 1Vrms across the NTC when the current is 100mArms, that equates to 0.1W of dissipation and will result in a particular temperature. If the current decreases to 50mA and the voltage drop stays at 1Vrms, then the dissipation is now reduced to 0.05W. The NTC will drop to a lower temperature and that lower temperature brings with it a higher resistance.
And the opposite happens when the current steps up to a higher level. Increased current gives a lower resistance since it operates at a higher dissipation resulting in a higher temperature.
Last edited:
millions upon millions of ATX psus out there.....the ntc's are never shorted out by a relay....why is that....?
- Status
- Not open for further replies.