|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
67.72.98.45
In Reply to: Re: Need some clarification... posted by Janos on October 21, 2006 at 21:07:59:
The problem about the thinning out with higher volume control settings is real for a TVC or AVC. Neither of these should necessarily be better than the other except in your specific cases you tested. Another AVC can outperform your TVC. I received an AVC that I thought might have a bit too few windings on too small a core. So I designed my own with a larger core and more primary windings. The result is less of the problems you describe when you crank up the volume setting.Now the idea of using high level signals with another gain stage in front of it is a good one in this case, or in other words, it's now better to place the AVC after the gain stage rather than in front as in a traditional preamp just because of the nature of tranformers compared to resistors. With the transformer after the gain stage you have a high signal level from which to further attenuate. This means that the turns ratio of the transformer will be higher to drop the signal back down where you want it compared to not having the gain stage there. This reduces distortion in the core by having plenty of voltage swing up front with a high turns ratio - and this means the B field in the core is actually lowered than if the transformer has to "work" harder by using a smaller turns ratio. So for the same volume levels the extra gain stage can improve the situation, at the cost of the distortion and signal degradation of that gain stage.
It's a balancing act. If your AVC is not working well with a line level DAC output, usually an op amp but sometimes a tubed cathode follower, then that might mean the AVC is not designed as well as it could be if a TVC proves not to exhibit as much of this. The AVC might need to be bigger in core size and in number of primary turns. That was the solution I came up with. I would not use another gain stage for my system with my new AVC at this point because overall the results would be worse.
Yes it is true that inductance falls with increasing frequency with all magnetic-cored inductors and transformers, but that doesn't mean a low impedance op amp is not good for driving the AVC. The opposite is true. An op amp's low output impedance is especially good for driving the AVC for this reason. But it might not be swinging enough volts to get it operating in the AVC's sweet spot. It's not the impedance, it's usually the voltage swing that's a problem in these TVC/AVC devices. I believe that the physics for that is pretty solid.
A TVC will likely have less bandwidth than an AVC. That's just because of the close intimacy of the turns in an AVC compared to a TVC. This could be good or bad in a digital system, I don't know.
I like all the observations taken down here, and the solutions generated to overcome them in an appropriate way. I wanted to point out that it is possible to design these autotransformer volume controls with too small a core and too few primary turns. This will lead to such artifacts in the sound. Now it is true that you cannot make it completely go away. But a good AVC still sounds better to me than by using anything active added in the chain. So it is to me the ultimate "line stage", one AVC designed properly to operate well at line level signal levels over the range that best fits the gain structure of the entire system. Do it right and it's damn tough to beat.
Kurt
Follow Ups:
Hi Kurt,Sorry for the late reply. Thank you and Dave for the info on tranformers. It's good to learn.
With better quality ATC/TVC the distortion due to high volume setting would be probably lower. I think my AVC is not so bad in this aspect, has high turn numbers, and decent size mumetal M lams - which were drawn 20+ years ago, so most of the stress of drawing is since gone. :) Interestingly, the thinning of the sound is not so apparent with this unit as with the commercial ones I heard, however, I got the feeling of going from excellent sonics in the low volume settings to good sound in the high volume, if driven directly by the CD player.
Now I have an OPT after the CD player, and after that comes the preamp, and what comes out of the preamp is lower signal level compared to the signal directly from the CD. The output imoedance is also much higher, in the order of 1K instead of the 100ish of the digital. In addition, you get the tube distortion and noise of the preamp. Every reason indicates worse performance, yet, at any volume level the AVC performs excellent, no feeling of shrinkage in sonics in higher volumes. And at any volume the sound is vastly, vastly better in every aspect than directly from digital.
Hi Janos,I know your observations are well perceived, and I can understand where you're coming from. Although I have a degree in electrical engineering and I have designed an AVC myself, I very easily forget a lot of the formulae, basic physics, etc. that's necessary to try to figure out what makes these things good or bad. Dave is into the ATV business himself and is more up to speed on these basics than I am because I am still just an experimenter (a real amateur to be exact) and it's been a long time since I designed my own AVC. He is modest with his questioning of my statements thankfully, and he could have chosen to rip me to shreds. But he is a gentleman.
I have to go back to my books all the time to see where my recall has failed me. But I do understand your thoughts and observations. Some day there will be a line stage preamp that has a gain stage followed by a TVC or AVC, I am sure of it. It's a good idea for the case where more than 0 dB gain is necessary in a line stage.
Kurt
hey kurt,This reduces distortion in the core by having plenty of voltage swing up front with a high turns ratio - and this means the B field in the core is actually lowered than if the transformer has to "work" harder by using a smaller turns ratio.
could you please clarify this? if you have an inductive volume control and increase the signal level, the overall B field on the core goes up and since the B field is directly related to the distortion, i would expect core related distortion to go up too.
> > could you please clarify this? if you have an inductive volume control and increase the signal level, the overall B field on the core goes up and since the B field is directly related to the distortion, i would expect core related distortion to go up too. < <Well, yes, this does require some clarification because I did not say enough, and I am in error. I forget there is some actual load present, although a high resistance usually. When you just increase the input voltage to the transformer and do nothing to the volume control setting, the B field will increase. This is because the power applied by the energized core of the transformer has to go up to supply the added power that is delivered to the load since the load didn't change and the turns ratio didn't change.
But if you increase the voltage input and drop the turns ratio to maintain the same signal output then the power delivered to the same load is identical. This should maintain the same B field in the core, not decrease it.
The only way to decrease the B field in the core and get the same signal output is to do one or more of three things: increase the core volume, increase the number of primary turns, or decrease the permeability of the core material.
The limit to increasing core size and primary turns is that the B field can also become too small and begin a different form of distortion - the effect of stuck "domains" in the core that need enough energy to overcome it. Somewhere in this design there is some optimal size and turns ratio for a given signal level, load, and core material.
I have found I like a certain mu metal formula for the core laminations, 80% nickel, a little oversize for the core size, and lots of primary turns with small gauge wire. The mu metal type was selected for low distortion at low B fields. In other words, those "domains" unstick readily at small magnetic levels. These core types may all say 80% nickel, but they do have differences, and you can pick one that is lower in low level distortion than others. And those hysteresis patterns differ, too, for low level signals. Some are more rectangular, for example.
hey Kurt,This is because the power applied by the energized core of the transformer has to go up to supply the added power that is delivered to the load since the load didn't change and the turns ratio didn't change.
I still think you are in error. Lenz Law suggests that in an ideal lossless situation a transformer of any size could deliver infinite power (current). It is my contention that this means that power delivered to the load has zero practical effect on the B-field.
But if you increase the voltage input and drop the turns ratio to maintain the same signal output then the power delivered to the same load is identical. This should maintain the same B field in the core, not decrease it.Look at the formula for b-field, it simply states that if you increase the voltage, you increase the flux. Nowhere is there a provision for the power delivered to the load. In other word, if you believe Lenz, loading of a transformer has no effect on the B-field.
The only way to decrease the B field in the core and get the same signal output is to do one or more of three things:
i don't see where the "same signal out" comes in here, that should simply be a matter of turns ratio, and the associated impedances.
increase the core volume
yes, this is the Ac (AREAcore not Vac in the formula)
increase the number of primary turns
yes, but i would simply state increase the turns. Adding the word primary only holds for a stepdown. if it were a step up you would have to say decrease the secondary turns and then it just gets confusing.
or decrease the permeability of the core material.
the perm of the core has no effect on the flux in the core, so i'm not sure what you mean here.
that the B field can also become too small and begin a different form of distortion - the effect of stuck "domains" in the core that need enough energy to overcome it.
blasphemy, don't you know that the lower the signal, the lower the distortion :-) Seriously, i agree 100% with you on this one and have been trying to at really low level behavior of cores and lets just say it ain't easy. this is probably getting a bit too technical so lets just say i may be your only supporter in this claim.
To be complete, you can also decrease the B field by increasing the frequency. (what a can of worms that statement opens :-)
Somewhere in this design there is some optimal size and turns ratio for a given signal level, load, and core material.
absolutely, I would also include source impedance, and consider signal level and source impedance to be the two most important things to know when starting a design. Those two numbers tell you how much flux and inductance you need, then the rest needs to be fit into place.
You're right Dave. I have forgotten too much to comment on this. I was trying to do it from my old memory. Disregard my postings on this. I don't know what I was thinking anyway.
This post is made possible by the generous support of people like you and our sponsors: