DC Coupled amplifier capacitor vlaues
- Thread starter mart.s
- Start date
Hi, it depends on several factors !
How low are the speakers "supposed" to go down to in Frequency ?
How low do you want the speakers to go down to in Frequency ?
The impedance of the speakers.
With "nominal" 4 Ohm speakers & 2200uF the -3dB point is around 18Hz. I say around, as the capacitors tolerance "could" be + or - 20%. With "nominal" 8 Ohm speakers & 2200uF the -3dB point is around 36Hz. That's ONLY in theory though, as i don't know the speakers Specs. If they naturally roll off anywhere near those Frequencies, then the response will not reach as low as you "might" wish, due to both things.
The Amp is not DC coupled due to the capacitor
Why is it there ?
How low are the speakers "supposed" to go down to in Frequency ?
How low do you want the speakers to go down to in Frequency ?
The impedance of the speakers.
With "nominal" 4 Ohm speakers & 2200uF the -3dB point is around 18Hz. I say around, as the capacitors tolerance "could" be + or - 20%. With "nominal" 8 Ohm speakers & 2200uF the -3dB point is around 36Hz. That's ONLY in theory though, as i don't know the speakers Specs. If they naturally roll off anywhere near those Frequencies, then the response will not reach as low as you "might" wish, due to both things.
The Amp is not DC coupled due to the capacitor
Why is it there ?
You can use any value cap. within reason of cost, the increased thump likely at at power up, power capability (Much higher power is required to extend the lower -3dB point by an octave, for example) or speaker type as Zero D discusses. The added advantage is to reduce the bass distortion which rises at frequencies below the roll off.
Bear in mind that single power rail amplifiers designed with only 1,000uF output caps were powered with appropriately small power supplies to suit that duty. Extending the lower roll off point to great depths may work if other filters in the amplifier, such as the feedback electrolytic and input cap. permit it but at some penalty to overall SPL.
Doubling the value of the output cap is hardly going to be a problem, even if particularly noticeable but quadrupling to say 4,700uF might be a value at which to consider the power rating of the amp, actual speaker impedance and the SPL you like to listen at.
Otherwise, it would seem a reasonable max. value IMO. We omit to say much about sound character because increasing the bass level subjectively makes a radical change. Add a subwoofer and prove the point but otherwise there is no change to higher frequencies than the original roll-off. There will always be caims that it can if using brand X or Y boutique components but that could apply to any value.
Bear in mind that single power rail amplifiers designed with only 1,000uF output caps were powered with appropriately small power supplies to suit that duty. Extending the lower roll off point to great depths may work if other filters in the amplifier, such as the feedback electrolytic and input cap. permit it but at some penalty to overall SPL.
Doubling the value of the output cap is hardly going to be a problem, even if particularly noticeable but quadrupling to say 4,700uF might be a value at which to consider the power rating of the amp, actual speaker impedance and the SPL you like to listen at.
Otherwise, it would seem a reasonable max. value IMO. We omit to say much about sound character because increasing the bass level subjectively makes a radical change. Add a subwoofer and prove the point but otherwise there is no change to higher frequencies than the original roll-off. There will always be caims that it can if using brand X or Y boutique components but that could apply to any value.
Last edited:
It's just that I have seen others with similar amp specs increasing the cap value (for reason I dont really understand) and I have a couple of new 10'000uf 63v caps spare.
Would I be better off using these in the psu section? instead, it currently has ONE 10'000uf.and would it be better to fit some upgraded faster rectifier diodes?
I had also read that a higher value DC blocking cap can reduce distortion in the lower registers.
I have some oldish Monitor audio speaker @ 6 ohm which have revised X-overs and are very easy to drive at high levels.
Would I be better off using these in the psu section? instead, it currently has ONE 10'000uf.and would it be better to fit some upgraded faster rectifier diodes?
I had also read that a higher value DC blocking cap can reduce distortion in the lower registers.
I have some oldish Monitor audio speaker @ 6 ohm which have revised X-overs and are very easy to drive at high levels.
@ Redshift187
Yes you're quite right ! I did mean to say that though
@ mart.s
To get a lower f3. Plus some say the increasing the cap value lowers distortion, as you mention.
It won't do any harm, & will reduce the ripple You could do both things.
Again, it won't do any harm, & will help rectify HF garbage instead of passing it through as noise.
Yes you're quite right ! I did mean to say that though @ mart.s
To get a lower f3. Plus some say the increasing the cap value lowers distortion, as you mention.
It won't do any harm, & will reduce the ripple
You could do both things.Again, it won't do any harm, & will help rectify HF garbage instead of passing it through as noise.
It will allow very high bass current to flow which will do harm if the volume is not reduced from previous levels that may have approached clipping. Take it easy on the amount of output capacitance. PS capacitance is fine as long as big thumps are.
I'm pretty sure the amplifier that the OP is referring to is a 15W 70's style model as per his previous thread. It's not a leviathan with 500VA transformers and DC-to-daylight bandwidth.
I'm pretty sure the amplifier that the OP is referring to is a 15W 70's style model as per his previous thread. It's not a leviathan with 500VA transformers and DC-to-daylight bandwidth.
Ok Thanks chaps I get the idea. If we are only talking negligible differences on distortion at low frequencies (which I probably wont notice @ below 35Hz) and I may have to go easy with the volume I may as well leave them as they are as I recently replaced them with Nichicons anyway. I think I'll stick an extra one on the psu as is is already space for one.
Attached is what a typical speaker impedance plot looks like. Because of resonant effects the lower end has a much higher impedance than the nominal impedance through the midrange. In practice full bass extension can be achieved with smaller capacitance than what is calculated for nominal impedance.
Attachments
