I thought this was pretty simple, I'll have more later. The common mode loop needs a cap but I can show how little "work" it needs to do to make SE or Diff output.
I think I do understand how it works. Essentially the same as your phono pre but with less gain. For Vin you would need to use a pair of NJFET as LTP at the input stage ?
If you look at my XCEN 2018 linked earlier, it is essentially the same but fully complementary instead of single ended. I like these current conveyor circuits. You can do so many things with the same basic topology.
🙂
Patrick
If you look at my XCEN 2018 linked earlier, it is essentially the same but fully complementary instead of single ended. I like these current conveyor circuits. You can do so many things with the same basic topology.
🙂
Patrick
Hi Tournesol. My experience too with compensating driver inductive impedance rise. I never tried damping (RLC) mid/tweeter resonance.... do you do this ?. Did/do you apply RC and RLC compensation across each of the drivers ?. What are some ballpark values for each type of compensation network in your experience ?.
Yes, and effect of driver temperature dependencies is much reduced...more stable filter points.
Yes. IIRC somebody ages back (DPH ?) went to some length to 'prove' mathematically that such impedance compensations are unnecessary and will have no audible effect.
Yes, and this is what Joe is trying to say, albeit using some wrong/confusing terminologies. And further to this, as Joe says a fully compensated loudspeaker (appears as ohmic load) will not care a hoot about amplifier output impedance/damping, including amplifier frequency dependent output impedance variation ......IOW speaker performance is independent of amplifier performance, and the fact that the amplifier is driving a clean resistive load importantly moreover improves amplifier THD/IMD and dynamic noise floor performance.
Keeping reactive circulating currents contained within the loudspeaker effectively removes amplifier/cable reactive current dependencies. Removal of amp output/speaker line harmonic currents reduces a bunch of current noise mechanisms, and the benefits are clearly and pleasantly audible. Class AB Amplifiers run cooler, much cooler into clean R loading, this is significant. Resistance termination of speaker cable at each end also improves overall clarity and dynamic noise floor, cheap quick and easy.
Dan.
Yes 0 to 20db of gain, I tried just FET and FET/NPN CFP. The CFP gives a little less distortion for a little more noise kind of a small tradeoff.
I could try an LDR as a degeneration to set gain. 😀
Question - If you set the basic gain structure up right in the first place are the lowest and highest settings on a typical volume control really useful?
I also considered using this as volume control quite a while back. It depends on whether it is current or voltage input.
For current input, it is obviously an IV converter. So the only place to alter gain is at R_iv.
For voltage input, the gain is set by the input degen resistor(s) and the R_iv. In this case, I prefer to vary gain by varying the degen resistor. This will avoid swinging excessive current at the front end, at the expense of noise (?). But either way it will work.
How do you intend to use this in conjunction with the phono ? Is 20dB gain enough P
Patrick
For current input, it is obviously an IV converter. So the only place to alter gain is at R_iv.
For voltage input, the gain is set by the input degen resistor(s) and the R_iv. In this case, I prefer to vary gain by varying the degen resistor. This will avoid swinging excessive current at the front end, at the expense of noise (?). But either way it will work.
How do you intend to use this in conjunction with the phono ? Is 20dB gain enough P
Patrick
Right now with my MI 15dB is more than enough, I need you to help with all the fancy $$$ carts. 🙂
No no I meant the high gain circuits for all those MC lovers. I have a lifetime stash of FET's
Has this been simulated?
Seems clean. I'll take a better look at it tomorrow, when I can turn the schematic 90 degrees. Now, I have a crick in my neck....
I was thinking voltage drive, and using the difference between the 2 vc's in some fashion taking advantage of the direct measurement of the real component of dissipation.
I was thinking of integrating the output voltage in the normal servo manner for dc offset, but using the second coil for the fb.
jn
Oh, now I see.
You already have my Spice files.
How else can I help ? PCBs ?
I am using 2SK372s for all the NJFETs, which is essentially 2x 2SK170 in parallel.
Do have some if you need them, but not unlimited.
Of course one can also solder 2x 2SK170 // top and bottom of the PCB instead.
Patrick
You already have my Spice files.
How else can I help ? PCBs ?
I am using 2SK372s for all the NJFETs, which is essentially 2x 2SK170 in parallel.
Do have some if you need them, but not unlimited.
Of course one can also solder 2x 2SK170 // top and bottom of the PCB instead.
Patrick
A well designed amp can easily handle the q2 and q4 dissipations.
As far as I can see, your solution can only dull the transient response of the speaker. I'd rather just get an amplifier that has no issues with reactance.
I see far too many designers worry about the sine sweep response. It's tough enough worrying about the group delays of multi-order cabinet designs without adding more delays to satisfy terminal reactance presented to the amp.
jn
Your suggestions for changes are enough. I need to get back in my own lab to make any circuits and measure some results.
You realise that an amplifier can have two different sets of harmonic distortions, one for voltage and another for current when connected to a load that is not purely resistive? Maybe that sounds too 'heretical' for some and does not sound politically correct, because they never thought that possible?
You don't understand what you say is as much nonsense to them as what they say is to us.
No, I just realized what to do.
Remove the current feedback of the .47 R. Tie the coil 1 bottom to ground. Tie the coil 2 top to the coil 1 top. Use the bottom of coil 2 as the feedback.
Coil 2 is picking up all the reactance of the speaker and none of the IR drop. tied as I just said, the coil 2 bottom will present only the real component of the speaker voltage back to the amplifier. Any variation in coil inductance will affect both coils exactly the same, so the difference will not have any consequences due to system inductance or system reactive non linearities. In fact, if you drive this arrangement with a crossover inductor between the amp and the speaker terminal, this feedback system will ignore the inductance and compensate for it, rending the crossover inductor useless until the amp runs out of voltage compliance. Needless to say it completely compensates for vc inductance changes due to position within the gap, totally eliminating the need for a shorting ring. The only cost is a slight increased need for amplifier voltage..so what.
The transfer function will now be entirely dependent on the resistive component of the speaker, the summation of the coil resistance and the effective resistance of the real part of the energy transferred to the air. The coil resistance will be fairly stable, so in essence, this feedback looks at the delivered acoustic energy signal. If you wanted, you could add some DSP to look at the dc resistance of the driven coil and compensate for compression.
Attempting to measure the amplifier output node in this setup for distortion will be entirely useless. The output node will be compensating for any non linearities in the magnetics.
The only way to measure this system for distortion is to use a microphone...
To make the voice coil, do not use the same gauge wire for the drive coil and the feedback coil. Use a smaller gauge wire for the feedback coil. Use a sufficiently small gauge that the feedback coil will nest in the interstitial voids of the drive coil. (yes, this only works with round conductors). This will minimize the added mass of the pickup coil, provides an automatic nesting and exact same turns ratio, and does not increase the voice coil thickness. It will however, see the exact same magnetic system that the drive coil sees. If you use a dual vc speaker that is front to back coils, each coil will see different non linearities and the method I describe will be useless.
There will be some inter-coil capacitive pickup, but because most of the voltage derived will be inductive, the two wires will almost have the same voltage, especially if the conversion efficiency of the driver is in the 1 to 2 percent range. That reduces the effect of capacitive coupling.
This method is not bandwidth limited per se, it would be fun if someone would put it together for a spin.
jn
Sigh...now I can go to sleep. There was no way that was happening until I got this off my chest..
edit: Scott, the model for this is within the material I sent you back in 2012.
Last edited:
yes, I believe there are dual VC drivers which are also double layer and bifilar wound. Those can be tried first to see how well it works on a first cut.
THx-RNMarsh
Last edited:
This is a nonsense. Higher distortion in current, driven from voltage amplifier, is because of load nonlinearity, not because the load is not purely resistive. You may have purely resistive load like thermistor or varistor and current is distorted even if source voltage has zero distortion.
- Status
- Not open for further replies.