Hello All,
I am finishing my version of the ZenV4 amplifier. I started this project a long time ago when I was a student and didn't have much money to spend. "My version" mostly implies a fancy chassis, the circuit itself remains the same. Feel free to look at the pictures at http://pages.infinit.net/deafboy/Pass/Pass.html. The power supply is not regulated as in the original article by Pass. Instead, it is a capacitor input, pi filtered with a large choke. I built both the choke and the toroidal power transformer. For the choke I dismantle a regular IE core transformer and reorganized the lams to build an air gapped choke. I had to wind a coil for it. I used an old Variac as the basis for the power transformer. I simply wound a secondary on the core it to make a nice toroidal transformer. Variacs are very cheap at Hamfests. The components values for the pi filter are 6800 µF, 14 mH and 66 000 µF. The first capacitor value is low by "audiophiles" standards because I wanted to keep the peak charging currents at reasonable values. The inductor and the 66 000 µF capacitance that forms the filter attenuates 120 Hz voltage ripple from the input capacitor by a factor of 350. When the supply is feeding a 4A load the ripple is about 3.5 Vp-p before the inductor and 10 mVp-p after. The supply feeds both channels. My first intention was to build a choke input supply but the inductor was buzzing loudly during my tests so I decided instead to build a pi filter supply.
I have a question regarding noises at the output of the amplifier. I noticed there is a 120 Hz noise at the output when the input is shorted or connected to a signal generator. This noise does not appear when the input is left open. The voltage of the noise can be substantial (several hundred mV). Playing with the amplifier components, I discovered that I can practically eliminate the noise by physically orienting the wire connecting the star ground to the channel circuit board ground. The audio band noise can be reduced to 200 µV by carefully orienting this ground wire and the orienting the toroidal power transformer.
My reasoning is that shorting the input forms a loop of wire (input wire and above mentioned ground wire) that can pick-up the magnetic field of the transformer. The picked-up signal is amplified by the amplifier and the amplified signal appears at the output. Since I would like to make the wiring a little less critical for low noise I twisted the wires together, thinking that the loop would have a reduced area to pick up the field. Unfortunately, the noise is not reduced by twisting the wires. I also encased the transformer in a steel case to reduce the stray magnetic field. The case is in fact material used for house ducts (two caps). This metal is hardly µmetal, but it does have some magnetic permeability (magnet sticks to it). Again, this procedure did not reduce the noise pick-up. I'm quite certain that the noise is generated by the magnetic field of the transformer. I devised a "magnetic field sensor"(simply a 0.13 mH air-core inductor connected to an oscilloscope - see photo on my website) to determine the shape of the magnetic field around the transformer. The signal picked up by the sensor is very similar in shape to the one at the amplifier output. It turns out the transformer has a stray field shaped like that of a bar magnet. I get the lowest noise on the amplifier output if I orient the ground wire at 90 degrees of the field lines that I measured with my homemade detector. Furthermore, the sensor confirms that the transformer casing does not do much to reduce the magnetic field. I'm surprised that the toroidal transformer causes much stronger stray magnetic field than the air gapped choke. I'm guess the cause of this is that the peak current is higher in the transformer than the choke.
My question is why can't I reduce the noise by encasing the transformer or twisting the wires?
Michel
I am finishing my version of the ZenV4 amplifier. I started this project a long time ago when I was a student and didn't have much money to spend. "My version" mostly implies a fancy chassis, the circuit itself remains the same. Feel free to look at the pictures at http://pages.infinit.net/deafboy/Pass/Pass.html. The power supply is not regulated as in the original article by Pass. Instead, it is a capacitor input, pi filtered with a large choke. I built both the choke and the toroidal power transformer. For the choke I dismantle a regular IE core transformer and reorganized the lams to build an air gapped choke. I had to wind a coil for it. I used an old Variac as the basis for the power transformer. I simply wound a secondary on the core it to make a nice toroidal transformer. Variacs are very cheap at Hamfests. The components values for the pi filter are 6800 µF, 14 mH and 66 000 µF. The first capacitor value is low by "audiophiles" standards because I wanted to keep the peak charging currents at reasonable values. The inductor and the 66 000 µF capacitance that forms the filter attenuates 120 Hz voltage ripple from the input capacitor by a factor of 350. When the supply is feeding a 4A load the ripple is about 3.5 Vp-p before the inductor and 10 mVp-p after. The supply feeds both channels. My first intention was to build a choke input supply but the inductor was buzzing loudly during my tests so I decided instead to build a pi filter supply.
I have a question regarding noises at the output of the amplifier. I noticed there is a 120 Hz noise at the output when the input is shorted or connected to a signal generator. This noise does not appear when the input is left open. The voltage of the noise can be substantial (several hundred mV). Playing with the amplifier components, I discovered that I can practically eliminate the noise by physically orienting the wire connecting the star ground to the channel circuit board ground. The audio band noise can be reduced to 200 µV by carefully orienting this ground wire and the orienting the toroidal power transformer.
My reasoning is that shorting the input forms a loop of wire (input wire and above mentioned ground wire) that can pick-up the magnetic field of the transformer. The picked-up signal is amplified by the amplifier and the amplified signal appears at the output. Since I would like to make the wiring a little less critical for low noise I twisted the wires together, thinking that the loop would have a reduced area to pick up the field. Unfortunately, the noise is not reduced by twisting the wires. I also encased the transformer in a steel case to reduce the stray magnetic field. The case is in fact material used for house ducts (two caps). This metal is hardly µmetal, but it does have some magnetic permeability (magnet sticks to it). Again, this procedure did not reduce the noise pick-up. I'm quite certain that the noise is generated by the magnetic field of the transformer. I devised a "magnetic field sensor"(simply a 0.13 mH air-core inductor connected to an oscilloscope - see photo on my website) to determine the shape of the magnetic field around the transformer. The signal picked up by the sensor is very similar in shape to the one at the amplifier output. It turns out the transformer has a stray field shaped like that of a bar magnet. I get the lowest noise on the amplifier output if I orient the ground wire at 90 degrees of the field lines that I measured with my homemade detector. Furthermore, the sensor confirms that the transformer casing does not do much to reduce the magnetic field. I'm surprised that the toroidal transformer causes much stronger stray magnetic field than the air gapped choke. I'm guess the cause of this is that the peak current is higher in the transformer than the choke.
My question is why can't I reduce the noise by encasing the transformer or twisting the wires?
Michel
The address for my link
http://pages.infinit.net/deafboy/Pass/Pass.html
in the original text is OK but has to remove the "." at the end of the hyperlink to make it work.
Michel
http://pages.infinit.net/deafboy/Pass/Pass.html
in the original text is OK but has to remove the "." at the end of the hyperlink to make it work.
Michel
Here's my progress on the ZenV4 amplifier noise pick-up.
I managed to make the noise less sensitive to component placement following Nelson's suggestions regarding the magnetic field from the toroidal transformer. I first put the transformer a few feet away from the amplifier. It was then easier for me to find the source of noise. It turns out that I had two places noise would be generated in the amplifier. I mentioned in my first post I could eliminate the noise by carefully orienting a ground wire. It seems that this compensated exactly for canceling both noise pick-up. This is why twisting some cables didn't eliminate all the noise. It eliminated only one source.
The first source came from the transformer's magnetic field inducing noise through the loop formed by the ground wire and other wires in the amplifier.
The second source came from having my 110Vac wiring parallel to my input signal cables. This wiring harness is about 2m in length.
Anyway, this is what I did:
- Replaced the signal cable in the wiring harness from a coaxial to a twisted pair coaxial.
- Added several layers of duct material around the toroidal transformer. One layer didn't do much. 8 layers reduced noise pickup by about a factor of 10.
- Orienting the toroidal transformer for lowest pick-up.
I can now place my transformer in the original position, next to the wires, without any noticeable pick-up.
Michel
I managed to make the noise less sensitive to component placement following Nelson's suggestions regarding the magnetic field from the toroidal transformer. I first put the transformer a few feet away from the amplifier. It was then easier for me to find the source of noise. It turns out that I had two places noise would be generated in the amplifier. I mentioned in my first post I could eliminate the noise by carefully orienting a ground wire. It seems that this compensated exactly for canceling both noise pick-up. This is why twisting some cables didn't eliminate all the noise. It eliminated only one source.
The first source came from the transformer's magnetic field inducing noise through the loop formed by the ground wire and other wires in the amplifier.
The second source came from having my 110Vac wiring parallel to my input signal cables. This wiring harness is about 2m in length.
Anyway, this is what I did:
- Replaced the signal cable in the wiring harness from a coaxial to a twisted pair coaxial.
- Added several layers of duct material around the toroidal transformer. One layer didn't do much. 8 layers reduced noise pickup by about a factor of 10.
- Orienting the toroidal transformer for lowest pick-up.
I can now place my transformer in the original position, next to the wires, without any noticeable pick-up.
Michel
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.