Hi,
ClassA topolgies draw constant current when idling.
Almost all ClassA output stage topologies draw varying current from the supply rails when a signal is present. The current variation is proportional to the signal level.
At the onset of clipping the current in one rail is almost zero and in the other rail is almost double the quiescent current. One half cycle later the rail currents have swapped over and the rail that was double current is now almost zero current. These rail current variations are generally ignored by many ClassA afficionados who unwittingly and erroneously quote constant current for ClassA.
There is one single ended ClassA topology that does genuinely draw constant current when signal is present but it is very rarely used. I believe that bridged single ended ClassA output stages can also draw a combined constant current from shared supply rails. That makes two topologies that I am aware of.
Your RCLC supply has been designed to specifically suit the ClassA stage it is supplying. The low level of ripple on the PSU output will be attenuated by the amplifier PSRR and appears on the output. When idling this attenuated ripple will be heard as hum and buzz. With careful design and build this can be reduced to acceptable levels.
With the varying currents described earlier, any and all resistance in the PSU will lead to voltage modulation in the supply rails when signal is being amplified. Any extra resistance will add to these voltage modulations and in turn affect the sound quality. Will your sound assessor accept any hint of designed in quality reduction, but, will it be audible.
The Cs before and after the L will make a good job of reducing ripple and voltage modulation, so if any resistance is be to be added then the best place to add it is before the first C and this is exactly where your cap protection resistor needs to be. The first thing in your favour.
I cannot assess whether any added R here will affect the sound quality of the final output signal, the only way is by experiment or some other builder jumping in with their experiences.
Finally, the constant current draw, while idle, will allow you to calculate the required resistance to ensure the caps do not feel excess voltage when mains are at nominal voltage and by using a larger resistance some some level of mains over voltage. Lots of room for try it and listen.
It should be very simple to undo the added R, but then you need another solution.
ClassA topolgies draw constant current when idling.
Almost all ClassA output stage topologies draw varying current from the supply rails when a signal is present. The current variation is proportional to the signal level.
At the onset of clipping the current in one rail is almost zero and in the other rail is almost double the quiescent current. One half cycle later the rail currents have swapped over and the rail that was double current is now almost zero current. These rail current variations are generally ignored by many ClassA afficionados who unwittingly and erroneously quote constant current for ClassA.
There is one single ended ClassA topology that does genuinely draw constant current when signal is present but it is very rarely used. I believe that bridged single ended ClassA output stages can also draw a combined constant current from shared supply rails. That makes two topologies that I am aware of.
Your RCLC supply has been designed to specifically suit the ClassA stage it is supplying. The low level of ripple on the PSU output will be attenuated by the amplifier PSRR and appears on the output. When idling this attenuated ripple will be heard as hum and buzz. With careful design and build this can be reduced to acceptable levels.
With the varying currents described earlier, any and all resistance in the PSU will lead to voltage modulation in the supply rails when signal is being amplified. Any extra resistance will add to these voltage modulations and in turn affect the sound quality. Will your sound assessor accept any hint of designed in quality reduction, but, will it be audible.
The Cs before and after the L will make a good job of reducing ripple and voltage modulation, so if any resistance is be to be added then the best place to add it is before the first C and this is exactly where your cap protection resistor needs to be. The first thing in your favour.
I cannot assess whether any added R here will affect the sound quality of the final output signal, the only way is by experiment or some other builder jumping in with their experiences.
Finally, the constant current draw, while idle, will allow you to calculate the required resistance to ensure the caps do not feel excess voltage when mains are at nominal voltage and by using a larger resistance some some level of mains over voltage. Lots of room for try it and listen.
It should be very simple to undo the added R, but then you need another solution.
Since cap voltage ratings come in discrete steps precise calculations are not really needed. A simple ruke of thumb from the "belt and suspenders" crew is take the nomina rail to ground voltage and select the next higher cap rating that is at leat 20% greater.
Example it the rails are +35/0/-35 use two 50V caps. For +42/0/-42, 50V is just a tiny bit low so use 75V or 80V. Etc.
Example it the rails are +35/0/-35 use two 50V caps. For +42/0/-42, 50V is just a tiny bit low so use 75V or 80V. Etc.
Yes, I have measured the voltage with 230 on the primaries from a variac. This is the same as the nominal in the house mains. The secondaries measured 46v on the recifier with no capasitor. With a capasitor I stopped the primary at 215 with a 0.2ohm power resistor in series with the + and one 66000uF capacitor. The voltage was then 58v. I was hoping that a 1,5 ohm resistor would bring down the voltage with about 6v and just below 60v.
The spikes in the net should be kept below 75v and therefore I should manage these.
There is obviously something that I have not thought of. Please help me get this right! My part supplyer ment that by using resistors on the secondaries would help me bring the voltage below 60. He has built Alephs, but I don't know if he has used 60v capacitors.
I need to get this to work....soon....!
Tnx for all the help!
The spikes in the net should be kept below 75v and therefore I should manage these.
There is obviously something that I have not thought of. Please help me get this right! My part supplyer ment that by using resistors on the secondaries would help me bring the voltage below 60. He has built Alephs, but I don't know if he has used 60v capacitors.
I need to get this to work....soon....!
Tnx for all the help!
For what it's worth...
In 1992 I built some Synder based Krell mono blocks. The loaded rails were 53 volts and the unloaded circuit was 58 volts with a CL 30 (5ohms) on the line.
I used 59,000 uf caps at 50 volt 63 volt surge ratings. In no way am I reccommending this is a good idea, but since I had them on hand at the time and couldn,t find anything with that much mfd at a higher voltage... well a few beers later and a gun to my head.. just do it!..--- They lasted 12 years before they needed replacement--- (This was indicated by hearing a soft crackling distortion in one amp) This is when I pulled them out and replaced them. Their ratings are conservative enough (based on my experience only) that a little gambling was in order since tring to find big capacitance at higher voltages is always tough at DIY prices. Regards David
In 1992 I built some Synder based Krell mono blocks. The loaded rails were 53 volts and the unloaded circuit was 58 volts with a CL 30 (5ohms) on the line.
I used 59,000 uf caps at 50 volt 63 volt surge ratings. In no way am I reccommending this is a good idea, but since I had them on hand at the time and couldn,t find anything with that much mfd at a higher voltage... well a few beers later and a gun to my head.. just do it!..--- They lasted 12 years before they needed replacement--- (This was indicated by hearing a soft crackling distortion in one amp) This is when I pulled them out and replaced them. Their ratings are conservative enough (based on my experience only) that a little gambling was in order since tring to find big capacitance at higher voltages is always tough at DIY prices. Regards David
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.