making a zobel output RL, options

[Sebastiaan]

Dear,

One quick question.

To make a RL in series with an amplifier output. Is there any disadvantage to make 14 turns around a 10ohm 5 watt resistor? Or is it better to make a air inductor separate from the resistor?

If the last option, If I make a inductor by 14 turns of magnet-wire, does it matter how thick the cylinder is I turn the manget-wire around to form an inductor? (after the turns are done I remove the cylinder of of course)

With kind regards,
Bas

 

[jmaf]

Is there any disadvantage to make 14 turns around a 10ohm 5 watt resistor?

The coil, as small as it may be, will induce some current on the conductors in that resistor. That current will be out of phase with the current through the resistance, meaning you're inducing distortion and noise in there.

Why would you wind the coil around the resistor?

 

[Sebastiaan]

The coil, as small as it may be, will induce some current on the conductors in that resistor. That current will be out of phase with the current through the resistance, meaning you're inducing distortion and noise in there.

Why would you wind the coil around the resistor?

Dear jmaf,

First because I saw it at the Leach site. Second it will save space on the PCB.

With kind regards,
Bas

 

[woody]

If you change the length,diameter,or number of turns the inductance changes.

 

[Sebastiaan]

If you change the length,diameter,or number of turns the inductance changes.

Thank you. I will seek for an online diameter/length ratio calculator online, since I forgot most of my math (Admitted being spoiled by software who calculates everything for you)

With kind regards,
Bas

 

[jaycee]

I would say it is negligible. I have done both and not seen any difference in performance. The amount of inductance is not really that critical, it is just enough to isolate the capacitive effect of the cable, and the resistor acts as a stopper against RF ingress.

 

[jmaf]

I would say it is negligible. I have done both and not seen any difference in performance. The amount of inductance is not really that critical, it is just enough to isolate the capacitive effect of the cable, and the resistor acts as a stopper against RF ingress.

Pretty interesting. I've seen similar tricks in guitar amps, like taking hookup wire soldered to +B on one of the power tubes and winding the other end of it around the input cable(no connection), to stabilize the amp, but never around a resistor.

 

[jaycee]

well, winding it around the resistor isnt important, its just a good way to save space. The inductor is what usually passes the signal, but at higher frequency the resistor comes into play. It forms a parallel RL circuit.

 

[Bonsai]

I would say it is negligible. I have done both and not seen any difference in performance. The amount of inductance is not really that critical, it is just enough to isolate the capacitive effect of the cable, and the resistor acts as a stopper against RF ingress.

The resistor is there to reduce the Q of the coil - i.e. it acts as a damper and ensures that the total coil+resistor impedance can never be higher than the resistor value - it really does not help with RF ingress. If you are worried about RF getting into the feedback summing node via the feedback resistor, you really have to try some other tricks . . . but most people here would frown at that. How about a Murata RF ferrite for example?

 

[AndrewT]

air core is theoretically better.
Larger diameter core gives more inductance and more copper length.

I like the full Thiele Network of R+C & R//L
But I have combined that with the other version of R//L & R+C across the speaker terminals.
This results in a Pi filter of sorts.
47nF to 100nF + 4r to 10r fitted very close to the output devices to minimise inductance.
Then in the cable route from amp to speaker terminals air core L//R. 1uH to 2uH and 4r to 10r.
Finally across the speaker terminals 68nF to 150nF + 4r to 10r.

It seems to sound alright and seems to perform OK.

 

[Sebastiaan]

Thank you guys for those useful responses.

I read the big old debate about weather to use a RL on the output or not. Me too can hear a difference in SQ with or without those RL in series with the output. Many commercial amplifiers leave them out as well. However in my experience the RL does make the amplifiers more stable with unexpected inductive or capacitive loads. Better be save then sorry (worry)

With kind regards,
Bas

 

[Bonsai]

If you are using fast output devices, you will not need more than 1-2uH for the output L. This will not have any effect on the sound unless you are driving speakers and cable that are very capacitive. If you are building active speakers, there is a case for no output inductor. If however you cannot pre-define the load, an inductor is always a sensible route to take. On a well compensated amplifier (so phase margin of 60 degrees and 12-15dB of gain margin) you end up with an unconditionally stable amp.

 

[Sebastiaan]

If you are using fast output devices, you will not need more than 1-2uH for the output L. This will not have any effect on the sound unless you are driving speakers and cable that are very capacitive. If you are building active speakers, there is a case for no output inductor. If however you cannot pre-define the load, an inductor is always a sensible route to take. On a well compensated amplifier (so phase margin of 60 degrees and 12-15dB of gain margin) you end up with an unconditionally stable amp.

Dear Bonsai,

Agree. It is always a trade off between loosing some slew-rate and not output inductor, or keep the slewrate and use the output inductor. Even with active speakers, the voice-coil load can be unpredictable by different temperatures.

The best sound ever obtained was no output-inductor and a minimum miller compensation. The sound was phenomenal, but the stability no reliable. With some wrong speaker cable routing it get wrong..

As designer it is all about decisions, and sometimes they are hard to make. You want your amplifier 100% reliable in all circumstances. The same time you want to obtain the best sound quality possible. Both seem to be natural enemies.

With kind regards,
Bas