Hello apelizzo. Thank you for your comments, and for catching my error calling T2 power output transformer T1 instead.
Suppose one uses a dual voice coil woofer per my suggested application:
1. Each coil is 8 Ohms for example.
2. Each coil is naturally a power resistor.
3. At an idle current of 3/4 Amp passing through each coil, the resultant power dissipation is 4.5 W for each.
4. The net 9 W dissipation in this woofer is like playing music through it so as to dissipate the equivalent amount of power [9W]; but in the absence of this high idle current.
5. This Mosfet power output stage is described as Push-Pull in the Audiophonics website.
6. As the upper Mosfet is progressively turned on by the positive-going music signal, the increased current flowing through it also passes through the upper voice coil. It is like a Class A amp normally driving a single voice coil woofer.
7. Simultaneous to the state in point 6, the lower Mosfet progressively turns towards off.
8. The states in points 6 and 7 are next applied to the lower and upper Mosfets respectively.
Its doable.
Suppose one uses a dual voice coil woofer per my suggested application:
1. Each coil is 8 Ohms for example.
2. Each coil is naturally a power resistor.
3. At an idle current of 3/4 Amp passing through each coil, the resultant power dissipation is 4.5 W for each.
4. The net 9 W dissipation in this woofer is like playing music through it so as to dissipate the equivalent amount of power [9W]; but in the absence of this high idle current.
5. This Mosfet power output stage is described as Push-Pull in the Audiophonics website.
6. As the upper Mosfet is progressively turned on by the positive-going music signal, the increased current flowing through it also passes through the upper voice coil. It is like a Class A amp normally driving a single voice coil woofer.
7. Simultaneous to the state in point 6, the lower Mosfet progressively turns towards off.
8. The states in points 6 and 7 are next applied to the lower and upper Mosfets respectively.
Its doable.
So basically you want to apply 6 Volts on each 8 Ohm voice coil (assuming 8 Ohms is the DC resistance of the coil) so you have 750mA DC passing through each coil to make them together dissipate 9W. That 9W is not like playing 9W music. 9W DC is not 9W continuous audio power. You should calculate how much audio power you could push through a voice coil to make it dissipate 9W of average heat.
Also to cool down the coil as it becomes a heater in idle (no music) you need a forced fan cooling. Coil through the air or air through the coil.
Also to cool down the coil as it becomes a heater in idle (no music) you need a forced fan cooling. Coil through the air or air through the coil.
Hello apelizzo. Thank you for your helpful comments. I fully agree with you that my suggested approach is problematic; because of heat buildup and its entrapment in and around the coils. Add to this a highly probable degradation in sound quality because the DVC woofer is a poor transformer which is simultaneously burdened by its thermal state.
My approach is clearly a bad idea.
Best regards.
My approach is clearly a bad idea.
Best regards.
It's not a bad idea. It's just an idea. I like when people think out of the box.
I would experiment a layer of your idea not in the linear dominion but in PWM. You want to power a woofer right? Computing power it's not a problem today... so why don't you try to experiment a PWM driving the voice coil with cone position feedback using an infrared emitter on the cone and a photo resistor on the magnet's pole running the wire through the centre vent of the magnet? Sample both input and photo resistor with an ADC. In theory you could set via software virtually some of the Small parameters. That's something I started back in Italy in 1996 when micro controllers were not as powerful as today then I got lazy.
I would experiment a layer of your idea not in the linear dominion but in PWM. You want to power a woofer right? Computing power it's not a problem today... so why don't you try to experiment a PWM driving the voice coil with cone position feedback using an infrared emitter on the cone and a photo resistor on the magnet's pole running the wire through the centre vent of the magnet? Sample both input and photo resistor with an ADC. In theory you could set via software virtually some of the Small parameters. That's something I started back in Italy in 1996 when micro controllers were not as powerful as today then I got lazy.
Last edited:
Hi Guys
DC current through a speaker voice coils is ALWAYS a bad idea. The voice coil is in no way like a power resistor since the coil can physically move and is immersed in a magnetic field.
DC is what burns up speakers when an amp fails or if you are foolhardy and connect it prior to doing proper checks with a new amp assembly.
Cooling of the voice coil with a music signal relies on the fact that the coil is in motion the whole time.
The simple mosfet buffer riven by a transformer has been used as a line amp and as a full power amp. The OP's innovation of the cross-wired bias for the gates is good and keeps distortion down. But, you still have free-running mosfet source-followers and their typical THD of anywhere between 0.08% and 0.5%.
If the circuit had the output transformer in the drain, then it would be analogous to tube circuits from the 30s and 40s which actually sound quite neutral despite 2-3% THD. Vox did a guitar PA this way with mosfets but did not use a driver transformer. Car power boosters were built this way, with input and output TXs and collector-driven OTs. It was a way to overcome the limited power of using 12Vdc directly.
Have fun
DC current through a speaker voice coils is ALWAYS a bad idea. The voice coil is in no way like a power resistor since the coil can physically move and is immersed in a magnetic field.
DC is what burns up speakers when an amp fails or if you are foolhardy and connect it prior to doing proper checks with a new amp assembly.
Cooling of the voice coil with a music signal relies on the fact that the coil is in motion the whole time.
The simple mosfet buffer riven by a transformer has been used as a line amp and as a full power amp. The OP's innovation of the cross-wired bias for the gates is good and keeps distortion down. But, you still have free-running mosfet source-followers and their typical THD of anywhere between 0.08% and 0.5%.
If the circuit had the output transformer in the drain, then it would be analogous to tube circuits from the 30s and 40s which actually sound quite neutral despite 2-3% THD. Vox did a guitar PA this way with mosfets but did not use a driver transformer. Car power boosters were built this way, with input and output TXs and collector-driven OTs. It was a way to overcome the limited power of using 12Vdc directly.
Have fun
Hello Struth. Thank you for your added and valuable input. I author a thread entitled Class aP amplification in the Passlabs Forum. I use in this diy application a DVC woofer as load with 100 mA idle current flowing in each 8-Ohm voice coil. I'll give you and apelizzo a heads up when I post this week.
Best regards.
Best regards.
winding the interstage transformer
Hi apelizzo,
I'm not sure I understand well the winding scheme you posted.
1000 turns (secondary)
200 turns (primary)
1000 turns (secondary)
for each section of the bobbin
then you wire the primary in parallel and all the secondary in series
you have 200:4000 turns
Am I correct?
Thank you,
Vincenzo
Hi apelizzo,
I'm not sure I understand well the winding scheme you posted.
1000 turns (secondary)
200 turns (primary)
1000 turns (secondary)
for each section of the bobbin
then you wire the primary in parallel and all the secondary in series
you have 200:4000 turns
Am I correct?
Thank you,
Vincenzo
Yes It's correct.
Make sure the winding orientation is correct:
From Left section of the bobbin:
1) start with secondary 1000 CW and the end of it will be GND (it minimises the capacitance between Primary and secondary)
2) 200 turns primary CCW start GND end +
3) connect the start of 1) to start of the second section of the secondary and wind up 1000 turns CCW. The end will be G1
From Right section of the bobbin:
1) start with secondary 1000 CCW and the end of it will be GND (it minimises the capacitance between Primary and secondary)
2) 200 turns primary CCW start GND end +
3) connect the start of 1) to start of the second section of the secondary and wind up 1000 turns CW. The end will be G2
Connect the 2 primaries in parallel.
Remember that this winding works well SE to PP 1:10
Make sure the winding orientation is correct:
From Left section of the bobbin:
1) start with secondary 1000 CW and the end of it will be GND (it minimises the capacitance between Primary and secondary)
2) 200 turns primary CCW start GND end +
3) connect the start of 1) to start of the second section of the secondary and wind up 1000 turns CCW. The end will be G1
From Right section of the bobbin:
1) start with secondary 1000 CCW and the end of it will be GND (it minimises the capacitance between Primary and secondary)
2) 200 turns primary CCW start GND end +
3) connect the start of 1) to start of the second section of the secondary and wind up 1000 turns CW. The end will be G2
Connect the 2 primaries in parallel.
Remember that this winding works well SE to PP 1:10
the woofer approach may work electronically, but sending full spectrum audio to a heavy cone woofer will make for awful midrange response.
early AM/FM transistor radios used this topology and some of them sounded quite good.
I believe 2 NPN outputs with 680 ohm bias and diode tied to center tap of output transformer 3 - 6v battery supply
early AM/FM transistor radios used this topology and some of them sounded quite good.
I believe 2 NPN outputs with 680 ohm bias and diode tied to center tap of output transformer 3 - 6v battery supply
Last edited:
Having built the Pass M2, I see some similarities with the subject amp. With the M2, the big transformer on output is not used. Both are zero feedback and if you use a pre-amp on this Parker Zeus amp, might as well use the 2sk170/sj74 JFETs that the M2 has to get some buffering on the input to the signal gain trafo.
Parker Zeus:
Pass M2:
Parker Zeus:
Pass M2:
Interesting design, but the big transformer on the output that is not used on M2 to some may look unnecessary but IMO it's what that give zeus a firm hand on the speaker coil and the magic sound of inductive coupling and lower distortion. It also has single power supply and same polarity matched depletion mode mosfets that are self biased through the DC resistance of the output transformer (on the latest design). I got full peace of mind with my speakers behind the output transformer and three times the power of the M2.
How does the M2 sound anyway? I will try one actually. Thanks for sharing.
How does the M2 sound anyway? I will try one actually. Thanks for sharing.
Last edited:
Something is really wrong. I used M6 laminations for mine and I done the winding in the order and direction as per instructions.
I have tried to wind up the bobbins with the wire in perfect order next to each other. It took me 4 weeks to do it because the size of the wire and at the end I was disappointed with the bandwidth. Then I have done one in half hour with the wires crossing and it' was the best. With 120K without the gates connected is flat to 100KHz
I have tried to wind up the bobbins with the wire in perfect order next to each other. It took me 4 weeks to do it because the size of the wire and at the end I was disappointed with the bandwidth. Then I have done one in half hour with the wires crossing and it' was the best. With 120K without the gates connected is flat to 100KHz
