Also, how is the Torroidal mounted?
I had an experience where the center mount screw and disc brace plate can was too close to the top of the chasis, causing a buzzer/relay effect. Also (from a foggy memory) even may have even induced a hot spot there.
As an alternative test, can you mount the torroid in a small cookie tin, surrounded by foam or other sonically absorbant material? Silicone caulk (never tried, but an idea none the less).
Lastly, what tune was it humming?
I had an experience where the center mount screw and disc brace plate can was too close to the top of the chasis, causing a buzzer/relay effect. Also (from a foggy memory) even may have even induced a hot spot there.
As an alternative test, can you mount the torroid in a small cookie tin, surrounded by foam or other sonically absorbant material? Silicone caulk (never tried, but an idea none the less).
Lastly, what tune was it humming?
I want to try this, but don't understand everything in this drawing:
http://www.diyaudio.com/forums/attachment.php?postid=328821
All i have to do is a 0,22 uF with 221 k.ohm accross L & N?
What about that rectifier bridge? Or does Bryston use this for something else.
http://www.diyaudio.com/forums/attachment.php?postid=328821
All i have to do is a 0,22 uF with 221 k.ohm accross L & N?
What about that rectifier bridge? Or does Bryston use this for something else.
"Surely you have that the wrong way round? The caps are essential as they block the DC, the bridge merely protects them from reverse bias and overvoltage."
Wrong, the caps are not needed.
LC Audio used to sell one similar to this with only a small snubber cap.
No different than using a triac as a switch.
Wrong, the caps are not needed.
LC Audio used to sell one similar to this with only a small snubber cap.
No different than using a triac as a switch.
Hi Richie & Djk,
are you still referring to the Bryston schematic?
The caps are essential to the DC blocker. No caps = no DC block = potential hum in the toroid.
The diodes are essential to the safety of the blocking system.
The peak voltage across the caps is determined by the cap reactance to the current and frequency.
During normal working currents and 50Hz/60Hz the voltage across the caps is chosen to be in the range 500mVpk to 1200mVpk. Cap values are selected accordingly.
During start up or overload the peak current flowing through the caps is much larger and one cap sees this as polarised one cycle voltage, the other cap sees this as a reverse voltage. The diodes provide a voltage clamp to prevent either cap seeing more than 700mVpk or 1400mVpk across them. The diodes are essential to ensure both caps are never overloaded during rare overvoltage events. Although start up is rare, it is normal and must be designed for. So the diodes conduct whenever the excessive current and consequent voltage try to exceed the diode clamping limit.
Bryston have chosen a quad to give 1400mV clamping and 6.3Vdc caps. This cap voltage is a bit lower than the range I have seen specified elsewhere so in light of their very long warranty I revise my recommendation down from 10Vdc to 6.3Vdc.
Note also the use of 33mF to minimise the reactance to suit Bryston's very big ClassAB amps
are you still referring to the Bryston schematic?
The caps are essential to the DC blocker. No caps = no DC block = potential hum in the toroid.
The diodes are essential to the safety of the blocking system.
The peak voltage across the caps is determined by the cap reactance to the current and frequency.
During normal working currents and 50Hz/60Hz the voltage across the caps is chosen to be in the range 500mVpk to 1200mVpk. Cap values are selected accordingly.
During start up or overload the peak current flowing through the caps is much larger and one cap sees this as polarised one cycle voltage, the other cap sees this as a reverse voltage. The diodes provide a voltage clamp to prevent either cap seeing more than 700mVpk or 1400mVpk across them. The diodes are essential to ensure both caps are never overloaded during rare overvoltage events. Although start up is rare, it is normal and must be designed for. So the diodes conduct whenever the excessive current and consequent voltage try to exceed the diode clamping limit.
Bryston have chosen a quad to give 1400mV clamping and 6.3Vdc caps. This cap voltage is a bit lower than the range I have seen specified elsewhere so in light of their very long warranty I revise my recommendation down from 10Vdc to 6.3Vdc.
Note also the use of 33mF to minimise the reactance to suit Bryston's very big ClassAB amps
Hi,
I just noticed that Bryston fit the DC blocker in the neutral line.
If you search back you will find a safety argument for either fitting the DC blocker in the live line or the neutral line.
I am still undecided which is safer.
Go look at the pros and cons and decide for yourself.
One final point, do not copy the version that has half the DC blocker in each line. This gives the worst case safety senario of both styles mentioned above.
I just noticed that Bryston fit the DC blocker in the neutral line.
If you search back you will find a safety argument for either fitting the DC blocker in the live line or the neutral line.
I am still undecided which is safer.
Go look at the pros and cons and decide for yourself.
One final point, do not copy the version that has half the DC blocker in each line. This gives the worst case safety senario of both styles mentioned above.
Hi Mzzj,
you say 1000uF (1mF) as a DC blocker.
The current capacity as a DCblocker is even lower than the ripple capacity you say corresponds.
The DCblocker value needs to be much higher for a power amp
I plan on using 3m3F//3m3F//3m3F 10Vdc in each direction for a 150W amp.
Bryston use 33mF 6.3Vdc, but the model number (power) is unknown.
With the values we are using ripple is not a critical factor. But you are correct, ripple capacity cannot be ignored.
I think if the reactance is low enough at normal working current to be at or below 500mVpk then ripple capacity will be adequate.
you say 1000uF (1mF) as a DC blocker.
The current capacity as a DCblocker is even lower than the ripple capacity you say corresponds.
The DCblocker value needs to be much higher for a power amp
I plan on using 3m3F//3m3F//3m3F 10Vdc in each direction for a 150W amp.
Bryston use 33mF 6.3Vdc, but the model number (power) is unknown.
With the values we are using ripple is not a critical factor. But you are correct, ripple capacity cannot be ignored.
I think if the reactance is low enough at normal working current to be at or below 500mVpk then ripple capacity will be adequate.
Hi Eva,
How does this eliminate a few mV of DC
Diodes on their own will not eliminate DC they simply add a voltage threshold above which current flows.
The unbalanced waveform still exists above the diode threshold and as a result the DC still flows. The DC is the net difference between the +ve half of the cycle and the -ve half of the cycle. Chopping out the bottom part (nearest the 0Volt level) of both cycles does not remove the net imbalance, although it may reduce it by a tiny amount. I wonder if one could measure the improvement, if any, that would arise from using side by side diodes alone?
How does this eliminate a few mV of DC
Diodes on their own will not eliminate DC they simply add a voltage threshold above which current flows.
The unbalanced waveform still exists above the diode threshold and as a result the DC still flows. The DC is the net difference between the +ve half of the cycle and the -ve half of the cycle. Chopping out the bottom part (nearest the 0Volt level) of both cycles does not remove the net imbalance, although it may reduce it by a tiny amount. I wonder if one could measure the improvement, if any, that would arise from using side by side diodes alone?
AndrewT said:
It allows the transformer to freewheel during the time it's not providing current to the filter capacitors (approx. 70% of the 50/60Hz period), so it can apply to itself either V(mains+1.5V) or V(mains-1.5V) depending on the sign of the magnetizing current.
That provides only a very small amount of correction, particularly because the magnetizing current should cross zero during the top of the wave, when the transformer is charging the capacitors and the diodes are forced to be biased in a certain direction.
Don't hesitate to try plugging your favorite electric heaters with a big diode in series to unbalance mains line, and then try to correct the transformer disaster
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.