Eldartford, of course no-one should be surprised that I don't like transformers :)
I did, however, outline some of the difficulties that one faces, bandwidth, ringing (distortion) and hysteresis losses (yet another kind of distortion).
You can dance through the parameters, trying to dodge the silver bullet and some good compromise is available. For example, by sacrificing a little LF bandwidth, you can reduce hysteresis losses, but they will still be there. Ultimately you are still dealing with a compromise.
This is not to say that a phono section with enough gain to start with does not represent some sort of compromise. For example, I prefer the sound of tubes, and tubes have a reputation of making more noise. What I found is that you have to treat the tubes the same way you treat transistors: fully-differential, with proper constant current sources, in order to make them be quiet. I found that making the proper constant current source was a huge part of that recipe.
So- we can get them to be quiet enough with the lowest output cartridges and at the same time the phono section is easily transparent enough to easily hear the insertion of the best SUTs we can lay our hands on. Now we all know that at the phono input, connections are very critical so for our evaluations we had to solder everything in. What seems to be coming out of this is that the more complexity, the less bandwidth and the less transparency.
So we endeavored to build a simple phono section. It has only two stages of gain and employs differential passive EQ. I don't know how much simpler it can get, in fact I've seen a lot of MM stages that have more complexity.
I guess the point is that my dislike of transformers is not **just** because I'm an OTL manufacturer, it really stems out of having worked with audio transformers a lot (and while they have always had some nice advantages, such as ground isolation): they always have an audible artifact. I really wish they did not, but there appears to be nothing that can be done to change that. You just have to avoid them if you can, that's all.
I did, however, outline some of the difficulties that one faces, bandwidth, ringing (distortion) and hysteresis losses (yet another kind of distortion).
You can dance through the parameters, trying to dodge the silver bullet and some good compromise is available. For example, by sacrificing a little LF bandwidth, you can reduce hysteresis losses, but they will still be there. Ultimately you are still dealing with a compromise.
This is not to say that a phono section with enough gain to start with does not represent some sort of compromise. For example, I prefer the sound of tubes, and tubes have a reputation of making more noise. What I found is that you have to treat the tubes the same way you treat transistors: fully-differential, with proper constant current sources, in order to make them be quiet. I found that making the proper constant current source was a huge part of that recipe.
So- we can get them to be quiet enough with the lowest output cartridges and at the same time the phono section is easily transparent enough to easily hear the insertion of the best SUTs we can lay our hands on. Now we all know that at the phono input, connections are very critical so for our evaluations we had to solder everything in. What seems to be coming out of this is that the more complexity, the less bandwidth and the less transparency.
So we endeavored to build a simple phono section. It has only two stages of gain and employs differential passive EQ. I don't know how much simpler it can get, in fact I've seen a lot of MM stages that have more complexity.
I guess the point is that my dislike of transformers is not **just** because I'm an OTL manufacturer, it really stems out of having worked with audio transformers a lot (and while they have always had some nice advantages, such as ground isolation): they always have an audible artifact. I really wish they did not, but there appears to be nothing that can be done to change that. You just have to avoid them if you can, that's all.
