Hi all!
I have built a nice soung amp (KT88 PP from Triodedick).
The only diffrens in my bulit is that i have a separate PSU.
My question is like this:
Where is the best place to but the large caps? In the PSU or in the AMP close to the OPT? And where to ground them? In the PSU or the AMP?
The reason for asking is that the sound is all greate but there is some buzz in the treble...
Thanks //Daniel
I have built a nice soung amp (KT88 PP from Triodedick).
The only diffrens in my bulit is that i have a separate PSU.
My question is like this:
Where is the best place to but the large caps? In the PSU or in the AMP close to the OPT? And where to ground them? In the PSU or the AMP?
An externally hosted image should be here but it was not working when we last tested it.
The reason for asking is that the sound is all greate but there is some buzz in the treble...
Thanks //Daniel
What you call buzz in the treble may be oscillations.
Try having the large capacitors like in option A and add a film capacitor of 0.1uF to 1 uF close to the OPT.
Try having the large capacitors like in option A and add a film capacitor of 0.1uF to 1 uF close to the OPT.
I'll give you a 90% probability that you have a grounding error. Make sure that the circuits connect to the high ripple current loop at a single point. That's the classic cause of buzz (as opposed to hum).
Unless you have a pile of oddly sized chokes, the chances that you have oscillation are... remote.
Unless you have a pile of oddly sized chokes, the chances that you have oscillation are... remote.
I would suggest that the reservoir cap needs to be near the rectifier, to reduce the size of the circuit loop containing the charging pulses. There may be some advantage to having the smoothing cap near the output stage, to reduce the size of the loop carrying the output current. However, grounding (i.e. electrical position) is far more important than location (spatial position).
As SY said, it's probably a grounding problem that's causing your buzz. Read a little about star grounding and it should become clear, even if you don't actually use classical star grounding.
But for your original question: If the reservoir capacitors are far from the current-consuming devices, then the inductance in the conductors from the caps to the devices might limit the ability to satisfy the demands for fast current changes. So it might be a good idea to put some more caps at the point of load. As a bonus, they will also help to filter out ripple and noise on the DC lines (as well as not let some of it get created in the first place).
Note, too, that it is important to form the absolute minimum "loop area" with conductor pairs, such as the input signal and ground return pair. Time-varying AC (and other) fields will induce currents in any loop (See Faraday's Law, or Maxwell's Equations.). And those currents will then induce voltages across any impedances in the loop, which is "a bad thing". So, for example, your input signal and ground conductor should be twisted tightly together, all the way from the input jack to whatever they connect to, such as across a resistor from grid input to ground. And the input ground should not connect to anything else, so the input jack must be isolated from the metal case, etc. (Better yet, you could use a cable that has two internal conductors and a shield, putting your input signal and ground on the two inner conductors and connecting (only) the shield to the chassis/ground, at the input end only.)
And that input-ground-reference resistor's ground return conductor should not share any length of conductor with other ground returns, all the way back to the power supply's central ground point. Otherwise, other time-varying currents will induce time-varying voltages back at the NON-ground ends of all of the ground-return conductors that share with them, making your input ground-reference resistor's "ground" into a time-varying voltage, which will arithmetically sum with your input signal, which is "a bad thing".
It turns out that the "loop area" minimization thing also prevents transmitting as well as receiving of over-the-air interference. So all other pairs should also stay as close together as possible. So, if you're using wires, twist together any pairs that have enough length to them, such as AC primary and secondary pairs, DC power and ground return pairs, AC or DC heater pair, output pair, etc.
But for your original question: If the reservoir capacitors are far from the current-consuming devices, then the inductance in the conductors from the caps to the devices might limit the ability to satisfy the demands for fast current changes. So it might be a good idea to put some more caps at the point of load. As a bonus, they will also help to filter out ripple and noise on the DC lines (as well as not let some of it get created in the first place).
Note, too, that it is important to form the absolute minimum "loop area" with conductor pairs, such as the input signal and ground return pair. Time-varying AC (and other) fields will induce currents in any loop (See Faraday's Law, or Maxwell's Equations.). And those currents will then induce voltages across any impedances in the loop, which is "a bad thing". So, for example, your input signal and ground conductor should be twisted tightly together, all the way from the input jack to whatever they connect to, such as across a resistor from grid input to ground. And the input ground should not connect to anything else, so the input jack must be isolated from the metal case, etc. (Better yet, you could use a cable that has two internal conductors and a shield, putting your input signal and ground on the two inner conductors and connecting (only) the shield to the chassis/ground, at the input end only.)
And that input-ground-reference resistor's ground return conductor should not share any length of conductor with other ground returns, all the way back to the power supply's central ground point. Otherwise, other time-varying currents will induce time-varying voltages back at the NON-ground ends of all of the ground-return conductors that share with them, making your input ground-reference resistor's "ground" into a time-varying voltage, which will arithmetically sum with your input signal, which is "a bad thing".
It turns out that the "loop area" minimization thing also prevents transmitting as well as receiving of over-the-air interference. So all other pairs should also stay as close together as possible. So, if you're using wires, twist together any pairs that have enough length to them, such as AC primary and secondary pairs, DC power and ground return pairs, AC or DC heater pair, output pair, etc.
- Status
- Not open for further replies.