don't do it, not without doing some checking of specs first.
1) basic capacitance voltage/current relationship is I=C*dv/dt. Simplistically, you're going from 1x to 3x the capacitance, so the peak current through all the soft-start circuit and transformer will triple. Assuming that your soft start circuit is a negative temp coefficient thermistor bypassed by a relay, you need to check the spec on the thermistor to see if it can handle the additional current. You also need to check the transformer spec. Chances are you have maybe a 30 - 50% margin, it may work the first time and the first N times but at the very least it would be a long term reliability problem.
2) The second problem is that the voltage rails will take longer to come up to their steady state value, so the delay for engaging the relay would have to be lengthened. Ideally the relay doesn't switch until the caps are at full charge so the current is very low. Also any other delays like output relays would have to be changed.
3) Once you have a sufficient amount of capacitance to handle the expected load, adding more isn't going to do you any good. I have been exploring this area lately, in my simulation test bench with 50V rails and 20,000uF, full power sine waves will make the rail droop about 10V. Real music, with lots of heavy bass will make it droop 2-3V. More capacitance won't change the end result, only the slope of the droop. It is much more effective to raise the supply voltage than throw more capacitance at it. Cheaper too.
1) basic capacitance voltage/current relationship is I=C*dv/dt. Simplistically, you're going from 1x to 3x the capacitance, so the peak current through all the soft-start circuit and transformer will triple. Assuming that your soft start circuit is a negative temp coefficient thermistor bypassed by a relay, you need to check the spec on the thermistor to see if it can handle the additional current. You also need to check the transformer spec. Chances are you have maybe a 30 - 50% margin, it may work the first time and the first N times but at the very least it would be a long term reliability problem.
2) The second problem is that the voltage rails will take longer to come up to their steady state value, so the delay for engaging the relay would have to be lengthened. Ideally the relay doesn't switch until the caps are at full charge so the current is very low. Also any other delays like output relays would have to be changed.
3) Once you have a sufficient amount of capacitance to handle the expected load, adding more isn't going to do you any good. I have been exploring this area lately, in my simulation test bench with 50V rails and 20,000uF, full power sine waves will make the rail droop about 10V. Real music, with lots of heavy bass will make it droop 2-3V. More capacitance won't change the end result, only the slope of the droop. It is much more effective to raise the supply voltage than throw more capacitance at it. Cheaper too.
Thanks for the response guys,
Here is more info on what it has .
1. The A/C input board has a power delay relay , there are jumpers to set the voltage supply range and an amp bias delay circuit.
2. The turn on delay relay is rated @ 125v/30A/24vdc
3. PS rectifier 400v/35amp
4. Trans is rated @`1,800VA max /ch. ( mono bloc)
5. Main resistor 4.7ohm/ 20W ( there is a mention not to use a variac to slowly bring up the voltage without first shunting this resistor , failure to do so will burn it out )
Based on what Michael suggested , it seems i should update the 20 W resistor to a 50 W.. I would also think the 30 amp rectifier /ch would be enuff .. if not suggestions ? as to lengthening the delay , how would i approach that ?
Bob where would i obtain that software ? PSU designer .............
Here is more info on what it has .
1. The A/C input board has a power delay relay , there are jumpers to set the voltage supply range and an amp bias delay circuit.
2. The turn on delay relay is rated @ 125v/30A/24vdc
3. PS rectifier 400v/35amp
4. Trans is rated @`1,800VA max /ch. ( mono bloc)
5. Main resistor 4.7ohm/ 20W ( there is a mention not to use a variac to slowly bring up the voltage without first shunting this resistor , failure to do so will burn it out )
Based on what Michael suggested , it seems i should update the 20 W resistor to a 50 W.. I would also think the 30 amp rectifier /ch would be enuff .. if not suggestions ? as to lengthening the delay , how would i approach that ?
Bob where would i obtain that software ? PSU designer .............
Last edited:
Here's a really simple simulation result, showing a basic power supply. The transformer is 120VAC input, 35VAC output (50V p-p), a 4.7ohm resistor shunted by a relay on the line side, 4 diode bridge and 35000uF capacitor.
The relay is modeled with a spice switch primitive and turns on at t=2s. At t=7s a 30Hz sine wave, peak 8 amps into 4 ohms, loads the rail. Current through the 4.7 ohm resistor (R1) and the rail voltage is shown.
You see how the rail voltage quickly droops from about 48.7V to about 40V when the load is applied. More capacitance will make the decay take a little longer is all.
The relay is modeled with a spice switch primitive and turns on at t=2s. At t=7s a 30Hz sine wave, peak 8 amps into 4 ohms, loads the rail. Current through the 4.7 ohm resistor (R1) and the rail voltage is shown.
You see how the rail voltage quickly droops from about 48.7V to about 40V when the load is applied. More capacitance will make the decay take a little longer is all.
Attachments
Just curious why? Do you have a design which is sensitive against ripple or is it a powerful amp? Are you aware of that you will create lot's of unwanted harmonics on the mains?
4.7 ohm is very low. I recommend 33-47 ohms. This should be enough to tame the inrush current surge. If the surge is still too high, increase the current limiting resistor. Seldom does it exceed 100 ohms.
Do not lengthen the delay for the relay. It does not reduce the inrush current. It is the Main resistor that does the current limiting. Keep the delay short, less than 1 sec. This will avoid stressing the current limiting resistor.
A variac should never be used to test this circuit. The Main resistor will burn out due to the prolong delay before the relay is activated. Also, some amplifiers, a small number, will blow if they are powered up with a variac.
A practical approach to determining the value of the current limiting resistor is to adjust it for about 2/3 of the rails voltage. Say the amp has working voltages of +/-80Vdc. During power up, adjust the Main resistor for 55-60V. Once the relay kicks in, the rails will shoot up to 80V. You will be able to see this with a scope.
A more scientific approach is a use a Clamp Meter that measures Peak current. That will tell you the actual current surge. Adjust the Main resistor accordingly.
Mike
Your simply sim. predicts there will be an improvement in transient response..
Yes it is a powerful amp and being used at low imp... Could you explain more on the unwanted harmonics.
Thanks Bob .
Thanks for the response Andrew , wouldn't 5 sec be 2 long ?
no, 5s to 10s for the slow charge thermistor bypass is OK.
The soft start resistor must be bypassed a lot sooner, 100ms to 300ms. I think 1s is too long for a mains resistor. 47r across 110Vac is about 257W. How hot will it get in 1s?
Hello Mike,
Thanks for the input .
You suggest increasing the resistor value to as much as 33 ohm from 4.7?
I had thought about upping it to a 30 watt , so i will now consider a higher imp by using a 33 ohm do you think the rest of the circuit will be sufficient ?
I do not have a clamp meter , so I'm not able to adjust as you suggest and was going to increase the resistor to 12ohm/30 watt .. do you still suggest 33 ohm ?
Hello andrew ,
so you agree to increase the 4.7 main resistor to 33 ohm ? of course any changes will include higher wattage to at least a 30 watt resistor , would you still add the thermistors to the circuit for the additional caps ? or will my original suffice with the resistor upgrade .....
Mr. Wayne, how low impedance do you have really?
The waveform of the current is short spikes which creates over tones the mains and they may interfere with other apparatus connected to the mains. Your rectifier bridge will have a harder life and also the transformer. The VA rating becomes half the value if you have a huge capacitor battery. A 500VA transformer can only take 250 VA if you have a PSU like this.
The waveform of the current is short spikes which creates over tones the mains and they may interfere with other apparatus connected to the mains. Your rectifier bridge will have a harder life and also the transformer. The VA rating becomes half the value if you have a huge capacitor battery. A 500VA transformer can only take 250 VA if you have a PSU like this.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.