manufelices
New Member
Hi! What does it mean (Direct Coupled amp/receiver)? How can I know if my receiver (Onkyo TX-SV454) is DC or not?
DC/AC coupled amp
- Thread starter manufelices
- Start date
Register to hide this ad
Adda
Super Member
Direct coupled means that there are no capacitors or transformers on the output or input.
An amp can have a cap coupled input but a direct coupled output though.
Direct coupled amps can usually produce lower bass frequencies then other designs, without phase shift, and mates with both high and low impedance gear equally well.
Nothing is free though, without input or output caps, a direct coupled amp will amplify DC as well as the music signal, and can output DC on it's own, unless carefully calibrated.
An amp can have a cap coupled input but a direct coupled output though.
Direct coupled amps can usually produce lower bass frequencies then other designs, without phase shift, and mates with both high and low impedance gear equally well.
Nothing is free though, without input or output caps, a direct coupled amp will amplify DC as well as the music signal, and can output DC on it's own, unless carefully calibrated.
Last edited:
bowtie427ss
Lunatic Member
A direct coupled amp will pass current at zero frequency.
Direct coupled in an amplifier means there are no capacitors between any amplifier stages, not just I/O.
Direct coupled in an amplifier means there are no capacitors between any amplifier stages, not just I/O.
bowtie427ss
Lunatic Member
Agree.
Some amplifiers have selectable cap/direct coupling on the input. I've seen a handful of manufacturers state that their protection circuitry may be triggered by very small amounts of DC on the input, and therefore offer this option.
There's an urban legend that you can drive a Crown DC-300 with a 1.5 volt battery on the input, and jumpstart your car with it. Not sure id want to put my own DC-300 to that test, but if i were stranded....................
Last edited:
JURB
Super Member
The OP raises an interesting point, though it was probably inadvertent. I've seen those true DC amps and have had at least one. A Marantz 1180DC has that switch.
Most of the totally DC coupled amps at least have a coupling cap right at the input. this is one hting, its Xc is so low at 20 Hz it really doesn't matter. Using a single ended supply and a cap on the outout is another matter. It has to be pretty massive.
Here's a nify little thing someone on Usenet wrote up. It is saved in MHT so it might only open in IE and you'll have to allow active content but it is a pretty cool thing. Called an impedance nomograph :
https://dl.dropboxusercontent.com/u/29948706/Impedance Nomograph.mht
Let's assume your speaker though rated 8 ohm is actually 4 ohms at say 25 Hz. Plausible. To get 1 ohm at 25 Hz you'll need about 6,000 uF or so. Now at 1 ohm, remember the speaker is 4 ohms, it is not only losing over a third of the power, when Xc is that high there is considerable phase shift. I am not sure right now and don't feel loike math at the moment but I would guesstimate 10 degrees. I fergit right now but it is sometnhing like take the arcsine of the vector quotient, something like that. I don't design alot of tone controls and crossovers these days so I am rusty.
With that phase shift you have to be careful with the negative feedback. What's more, 6,000 uF is almost impractical. With the cost you could have gone DC coupled so alot of them only use like a 3,300 uF. That will have about double the Xc at 25 Hz so now it is losing over half the power.
In a DC amp with a split supply, even if it is AC coupled at the input, the power bandwidth in the low frequencies is pretty much down to DC.
There is another component in amps that are considered DC coupled that can sap your low end power bandwidth. The bootstrap cap. Well at least that is working into a few K ohms. If the value goes low it can hurt the low end, and if it gets high ESR it can make (aminly second order) distortion in the whole audio range.
If it uses a bootstrap - is it DC ? I realize that bootstrapping is actually positive feedback, so wuld that rightly disqualify an amp from being called DC ?
One of the things people don't realize these day also is that there uised to be solid state amps with not only output transformers, but interstage transformers. This was back when transistors were germanium and NPNs were hard to come by.
Later, when silicon came out, Sony came up with a "DC" amp. It was not DC coupled to the speakers but there was only an output cap, not a transformer, and no interstage transformer. However it did not use any PNP transistors. It was a nightmare, you needed the transistors to come from them because the were matched, the gain of the inverter stage had to be just right and all that. For the want of ONE PNP transistor !
Going DC, and going DC coupled output was a good thing for audio, but because of that now when the outputs blow they frequently take out a few other things with them. It didn't used to be that way. but then everything sounded like a teansistor radio. (well not everything)
Most of the totally DC coupled amps at least have a coupling cap right at the input. this is one hting, its Xc is so low at 20 Hz it really doesn't matter. Using a single ended supply and a cap on the outout is another matter. It has to be pretty massive.
Here's a nify little thing someone on Usenet wrote up. It is saved in MHT so it might only open in IE and you'll have to allow active content but it is a pretty cool thing. Called an impedance nomograph :
https://dl.dropboxusercontent.com/u/29948706/Impedance Nomograph.mht
Let's assume your speaker though rated 8 ohm is actually 4 ohms at say 25 Hz. Plausible. To get 1 ohm at 25 Hz you'll need about 6,000 uF or so. Now at 1 ohm, remember the speaker is 4 ohms, it is not only losing over a third of the power, when Xc is that high there is considerable phase shift. I am not sure right now and don't feel loike math at the moment but I would guesstimate 10 degrees. I fergit right now but it is sometnhing like take the arcsine of the vector quotient, something like that. I don't design alot of tone controls and crossovers these days so I am rusty.
With that phase shift you have to be careful with the negative feedback. What's more, 6,000 uF is almost impractical. With the cost you could have gone DC coupled so alot of them only use like a 3,300 uF. That will have about double the Xc at 25 Hz so now it is losing over half the power.
In a DC amp with a split supply, even if it is AC coupled at the input, the power bandwidth in the low frequencies is pretty much down to DC.
There is another component in amps that are considered DC coupled that can sap your low end power bandwidth. The bootstrap cap. Well at least that is working into a few K ohms. If the value goes low it can hurt the low end, and if it gets high ESR it can make (aminly second order) distortion in the whole audio range.
If it uses a bootstrap - is it DC ? I realize that bootstrapping is actually positive feedback, so wuld that rightly disqualify an amp from being called DC ?
One of the things people don't realize these day also is that there uised to be solid state amps with not only output transformers, but interstage transformers. This was back when transistors were germanium and NPNs were hard to come by.
Later, when silicon came out, Sony came up with a "DC" amp. It was not DC coupled to the speakers but there was only an output cap, not a transformer, and no interstage transformer. However it did not use any PNP transistors. It was a nightmare, you needed the transistors to come from them because the were matched, the gain of the inverter stage had to be just right and all that. For the want of ONE PNP transistor !
Going DC, and going DC coupled output was a good thing for audio, but because of that now when the outputs blow they frequently take out a few other things with them. It didn't used to be that way. but then everything sounded like a teansistor radio. (well not everything)
manufelices
New Member
Thanks for the answers!
I don't know about a jump start, but I do know that when the outputs rail they will toast speakers pretty well.
Stinks too.
fred soop
Super Member
Jurb, right on.
My Heathkit AR-15 receiver uses a 4000 µF output coupling capacitor which translates to 3 dB down at 20 Hz into 4 ohms. These designs are simply obsolete in this day and age and I can easily demonstrate that, in addition to the power loss and phase shift you mention, there is also a very low frequency DC shift when applying or removing a large signal. There is also a 5 volt DC offset at power up as the charging path for this capacitor is through the speaker. The worst part is that, with no center tap on the power transformer, there is no way to modify the circuit.
The bootstrap capacitor is more bad news, causes a few more artifacts, and should really be replaced with a current source.
The feedback capacitor is not as bad. It rolls off the gain at low frequency to be unity at DC, and, strictly speaking, is not in the signal path. Still better if it is eliminated, but there is a greater cost. My Harman Kardon Citation 24 (same as the Citation 22) is, in fact, DC coupled, except the input coupling capacitor. In this type of design, there is full gain down to DC so there must be an adjustment to minimize DC offset due to component tolerances.
Douglas Self has tested almost everything that he has written about and one of his simple tests is capacitor distortion. This test is simply a capacitor in the signal path driving a resistor to ground. He demonstrates that an electrolytic capacitor must be about 10 times the size required for desired low frequency response to bring distortion generated by that capacitor down to the noise floor. That means that my AR-15 example, above, would need a 40,000 µF coupling capacitor to the speaker.
My Heathkit AR-15 receiver uses a 4000 µF output coupling capacitor which translates to 3 dB down at 20 Hz into 4 ohms. These designs are simply obsolete in this day and age and I can easily demonstrate that, in addition to the power loss and phase shift you mention, there is also a very low frequency DC shift when applying or removing a large signal. There is also a 5 volt DC offset at power up as the charging path for this capacitor is through the speaker. The worst part is that, with no center tap on the power transformer, there is no way to modify the circuit.
The bootstrap capacitor is more bad news, causes a few more artifacts, and should really be replaced with a current source.
The feedback capacitor is not as bad. It rolls off the gain at low frequency to be unity at DC, and, strictly speaking, is not in the signal path. Still better if it is eliminated, but there is a greater cost. My Harman Kardon Citation 24 (same as the Citation 22) is, in fact, DC coupled, except the input coupling capacitor. In this type of design, there is full gain down to DC so there must be an adjustment to minimize DC offset due to component tolerances.
Douglas Self has tested almost everything that he has written about and one of his simple tests is capacitor distortion. This test is simply a capacitor in the signal path driving a resistor to ground. He demonstrates that an electrolytic capacitor must be about 10 times the size required for desired low frequency response to bring distortion generated by that capacitor down to the noise floor. That means that my AR-15 example, above, would need a 40,000 µF coupling capacitor to the speaker.
