Behringer A500 as a DIY project

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Many amps are made this way. In this case, by using the RCA, you would make the signal pass one OP-amp stage less. Probably you will not hear this, if you do and it is not due to a different level, it might just as well be the coupling 50uF capacitor. These are a cheap solution and any modder would change them at first sight. But adding some 4700uF caps to the power supply should have a more audible impact, IMO, if the amp is used full range or for bass reproduction. 3300uF seem a bit on the lean side for an amp with more than 100 Watt.
 
Many amps are made this way. In this case, by using the RCA, you would make the signal pass one OP-amp stage less. Probably you will not hear this, if you do and it is not due to a different level, it might just as well be the coupling 50uF capacitor. These are a cheap solution and any modder would change them at first sight

Thanks for the very valuable advice. Do you think that some electrolytics can be tolerated also in series with the signal ? i read somewhere that at Millennia they use to put small plastic caps in parallel with bigger electrolytics in their mic preamps ... and they preamps are top notch from what i understand.
Even a 10uF plastic is big ... and the better the quality the bigger it is.

But adding some 4700uF caps to the power supply should have a more audible impact, IMO, if the amp is used full range or for bass reproduction.
3300uF seem a bit on the lean side for an amp with more than 100 Watt

thanks and yes ... that i have noticed also myself and i have to check the space but i could try even to place 10mF parts if space allows
I am a fan of very big caps ... they are like the lungs for a human
I have done some silly recapping in the past on old amps ... i am in the camp that the more uF the better Of course i needed to replace the diodes in the bridge with some monster bridge. And i used big cables as well.
Often especially with cheap amps i have the feeling that both uF and quality of the ps caps is very suspicious and inadequate. And the transformer as well.
Another thing i noticed is that they leave very little space to increase uF ... like they put constraints for modding
 
Last edited:
The required capacitance has little to do with the amplifier wattage. It has to do with the load impedance. 10,000 uf caps on each channel would be about right for an amplifier capable of 4 ohms. To go down to 2 ohms you need to double it, for 8 ohms only you might get away with half (which would be 10,000 uf for both channels combined - about the minimum you see for any “real” amplifier). To really belt out the bass you would run enough capacitance for 2 ohms, but run your load higher to keep that voltage stiff. But you really don’t want to go much higher. Going a lot larger (say 100,000 uf) is bad too, for other reasons.
 
The required capacitance has little to do with the amplifier wattage. It has to do with the load impedance. 10,000 uf caps on each channel would be about right for an amplifier capable of 4 ohms. To go down to 2 ohms you need to double it, for 8 ohms only you might get away with half (which would be 10,000 uf for both channels combined - about the minimum you see for any “real” amplifier). To really belt out the bass you would run enough capacitance for 2 ohms, but run your load higher to keep that voltage stiff. But you really don’t want to go much higher.

Thank you for the very useful advice. Looking at amps i see 4700uF/rail*channel a very common value. Personally i like 10,000uF but this usually fights with space available on board to place bigger caps. It seems that manufacturers do no like the idea ... i wonder why they space parts so closely ... weird It is not a personal computer ... it is a power amp
In power amps i would love to see high section cables for the high current paths Even in power transistors i would like to see bigger pins for power supply and output than input signal :eek: In the end Ohm law is valid also for power amp from what i understand

Going a lot larger (say 100,000 uf) is bad too, for other reasons.
So indeed there can be issues with too large capacitance in the power supply ? this i do not understand. :confused: Could you elaborate please ? :eek:
Better ... is there an optimal value for capacitance in a power supply ? a formula to calculate that ?
 
The bigger the capacitor, the higher the RMS secondary current, fir a given DC load. Power factor due to harmonics gets progressively worse (and so do turn-on surges). The only question is where you cry ‘uncle’ and stop. And after it gets big enough there is little benefit anyway.

As to an optimum value, the power supply cap and load impedance form a low frequency pole, just as if you had an output coupling cap. It is inside the feedback, so it is of little consequence closed-loop. But you lose power capability if the pole frequency is too high. And you get poor clipping behavior at low frequencies to boot. You want that pole frequency to be lower than the one created by your input coupling cap. How low do you want? Single digit Hz for sure. And remember the load impedance is both channels in parallel, and the caps are in series across the supply. 10,000 uf and 4 ohms is effectively 8 Hz when you look at the total series parallel circuit.
 
The bigger the capacitor, the higher the RMS secondary current, fir a given DC load. Power factor due to harmonics gets progressively worse (and so do turn-on surges). The only question is where you cry ‘uncle’ and stop. And after it gets big enough there is little benefit anyway.

Hi thanks and that is a good news because big and high grade caps are insanely expensive Probably some specific applications need them

As to an optimum value, the power supply cap and load impedance form a low frequency pole, just as if you had an output coupling cap. It is inside the feedback, so it is of little consequence closed-loop. But you lose power capability if the pole frequency is too high. And you get poor clipping behavior at low frequencies to boot. You want that pole frequency to be lower than the one created by your input coupling cap. How low do you want? Single digit Hz for sure. And remember the load impedance is both channels in parallel, and the caps are in series across the supply.
10,000 uf and 4 ohms is effectively 8 Hz when you look at the total series parallel circuit

this is very technical and i am not sure to get it right. However i will keep the value of 10mF/channel rail in mind. That would be 20mF per two when feeding both channels ... good ! thank you very much again.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.