I will have about 400000 uF of capacitance in my PSU and I 'am wondering that the inrush current will be too high when the amplifier is powering up. In the original schematics (Hiraga 20W amp) there are resistors (about 0.5 ohm per chanel) in series to the large capacitors which reduce this current. Is there are other way to reduce this inrush current because this resistors will reduce the voltage. My transformer is large 550W EI core.
You can put a GE Type CL Thermistor on the primary of your transformer, or ahead of the filter caps. (you've got to size the NTC thermistor for the application). I use CL-60 for amplifiers -- at room temperature it's 10 ohms, at idle (2.5A) about .44 ohms -- here's the PDF: http://www.gesensing.com/products/resources/datasheets/cl.pdf
you can add a bypass relay across the current limiting resistor and/or thermistor.
The secondary located limiter should have a bypass delay of 5 to 10seconds.
The secondary located limiter should have a bypass delay of 5 to 10seconds.
So, there are next solutions:
1. one thermistor in the primary windings of the transformer
2. resistor also in the primar but with bypas relay (with time delay of about few seconds)
3. resistor/thermistor at the secondary windings of transformer
4. combination of all of these
I will use microcontroler for remote controll therefore I can programm it to control this delay for the bypass relay. Which of theese solutions is the best for my capacity and how to choose proper resistor?
1. one thermistor in the primary windings of the transformer
2. resistor also in the primar but with bypas relay (with time delay of about few seconds)
3. resistor/thermistor at the secondary windings of transformer
4. combination of all of these
I will use microcontroler for remote controll therefore I can programm it to control this delay for the bypass relay. Which of theese solutions is the best for my capacity and how to choose proper resistor?
the resistor and/or thermistor in the primary is there primarily to reduce the start up current that flows in the primary of the transformer just as it passes the first few cycles of AC from off.
The delay needs be between 100ms and 500ms to ensure that the transformer starts without blowing the mains fuse and that the resistor does not overheat.
slow charging of the smoothing caps is better addressed by inserting a resistance in the secondary before or after the rectifier. This time the delay for the bypass should be of the order of 5 to 10seconds.
I think you might get less ripple current in the DC if you split the large filter caps into two and connect the two with a resister. Then you'd have a CRC filter rather then just a "C". You do not need to split in equal parts. If you decide to use a resistor, this is a better way to configure the same parts. CRC is better than RC. Size the first C to limit in rush to acceptable amount.
Also you can place a thermistor in series. These devices start out at 100 or 200 ohms then drop resistance as they heat up. If you put one of these on the AC mains side of the transformer it will limit current to the primary which will also limit current in the secondary. The advantage of placing on the primary is the resistance does NOT drop the DC voltage. I would suggest this even if you do the above also.
Another way to limit in-rush that was very common 50 or 60 years ago was to use a choke. Rectifiers back then could not handle much in-rush current so they designed power supply filters that did not have much in-rush. Chokes and CRC filters where common. You can get any desired level of ripple reduction this way without a large capacitor bank. But it is not used so much today because they require more skill to design and waste a bit of power.
A thermistor in the primary does reduce the secondary voltage and PSU supply voltage.
Worse, the thermistor variation with changing current demands, modulates the secondary voltage. You can easily detect this by monitoring the voltage of the PSU on a scope.
I decided for power resistor on the primary winding of transformer which will be bypassed with the relay. This is the simpliest solution for me and does not reduce output voltaga such in case of CRC filter. (voltage drop on the series resistor).
Yes if you lower the voltage on the primary you also lower the voltage on the secondary but the effect is determined by the voltage ratio of the transformer. So given a choise of which side to place the device, may as well put it on the higher voltage, lower current side. The voltage drop will be I*R. Assume primary current on a 20W amp is 0.2A and a 1ohm thermistor you get 0.2volt drop on primary and less than 0.2 V drop across secondary.
One could wire the SPST switch across the thermistor or use a time delay relay to close a switch automatically but really what is the effect is a 1 ohm resistor in series with a
The modulation effect is interesting that it is large enough to be measured. I guess it would be a kind crude voltage regulator, letting in more current when there is greater demand on the power supply.
But putting it in the primary side also reduces the effectiveness of what you are trying to do by the turns ratio. The correct place to put it is before the filter caps, on the secondary.
The ultimate, but it is a bit more expensive, is to use a device called TSR from a german compay.
It's operation is very clever, it sends pulses to the transformer to let the transformer complete an hysteresis cycle and switch on when not saturated in highest impedance state.
It controls also the initial loading of DC capacitor banks.
This device also acts as mains relay controlled by a low level signal. This is a saving in case you need a relay
The other methods have drawbacks even dangerous ones in case of failure of one component or reswitching on too often.
JPV
It's operation is very clever, it sends pulses to the transformer to let the transformer complete an hysteresis cycle and switch on when not saturated in highest impedance state.
It controls also the initial loading of DC capacitor banks.
This device also acts as mains relay controlled by a low level signal. This is a saving in case you need a relay
The other methods have drawbacks even dangerous ones in case of failure of one component or reswitching on too often.
JPV
I once heard that resistors are available in different resistance values. I don't know whether it's true or not though. 😉
If put on the primary, a slow start can not only prevent transformer turn-on surge but also provide slow charge of the capacitors. If put on the secondary side, it can only slow-charge the capacitors.
Last edited:
If you put a big resistor in the primary such that it's actually effective at limiting inrush current of the caps, it will not allow enough flux to build up, so when the resistor is bypassed another mini-surge takes place due to the establishment of full flux level.
That is why I suggested the thermistor. It starts out being about 200 ohms, then over the period of a few seconds the resistance falls to about 2 ohms. You need to select the correct size thermistor so the rate is "correct" but if you get this right, then when you close the switch or relay to short out the thermistor it is only shorting out a 2R resistance.
About that comment that a thermistor will change resistance with current demand. that should never happen, not unless the thermistor is to large. A properly sized device will heat up and drop to minimum resistance and will have a rather long thermal time constant so resistance can't change quickly, but more than that, the thermistor should be "pegged out" and unable to move below its minimum value even if the current through it doubled
No.
it starts out at it's cold resistance.
Over a period of one or two mains cycles it drops to a warm resistance.
Then the smoothing starts charging. This further warms the thermistor and it starts falling to even warmer resistance. Then the charging currents drops off and the resistance rises again to suit the cooler temperature over a period of a few seconds.
Finally the thermistor rises to near cold temperature resistance when the amplifier is drawing quiescent current.
That senario is a far cry from what most builders assume. It leads to two completely different views on how this slow charging should be implemented.
That, I believe is why there is so much argument, most do not consider what really happens during the first few milliseconds and then what happens in the following seconds.
The two requirements for soft start and slow charge are different and require different solutions.
Last edited:
I ask for series resistor (not NTC) in the primary or other solution. The goal is to avoid series resistor behind the first lagre capacitor (original Hiraga schematic PSU) and so on the voltage fall on it because I'am running on 15VAC transformer which produce about 18.5 VDC therefore this voltage falling can decrease this voltage to level which is not prefered.
How far would you let the capacitors charge in your proposed secondary side soft start before cutting the resistors out? If the capacitors are not fully charged there will be a surge anyway from the capacitor charging, regardless of which side of the transformer the soft start is put on.
Are you thinking of class A amplifiers perhaps? In a class AB amplifier there should be no problem getting the capacitors close enough to fully charged not to cause any significant surge when the resistor is cut out. You could cut bias of a class A amplifier during startup to ease the job of the soft start.
Last edited:
- Status
- Not open for further replies.