Bumping the thread as it was posted over the weekend hoping there is some discussion to be had regarding the topic.
Active FETs vs. Step-up phono stage designs
I've noticed quite a few threads on A-gon discussing step-up transformers recently, and it got me thinking of the different applications in the phono stage design. I'm curious of the advantages and disadvantages of designs that employ step-up transformers vs. active designs with FETs, especially in tube preamps. Examples would be the new Manley Chinook vs. the Rogue Audio Ares with its custom Cinemag transformers, or the highly-regarded Andros, which also went the step-up route. I'm thinking of upgrading my solid stage phono preamp in the near future, and would like to hear your opinions (and I know we have a few designers posting regularly) regarding this subject.
22 responses Add your response
I guess the following document has some relevance. http://gramophone.net/Issue/Page/December%201978/197/834601/Movingcoil+stepup+devices |
This link might be useful. http://www.gramophone.net/Issue/Page/December%201978/197/834601/Movingcoil+stepup+devices |
Dan_ed got that right... Its been my experience that you don't want semiconductors at the input of a phono stage- they just don't have the same low level detail as an all-tube phono section. I don't like transformers for the same reason- they limit bandwidth and loose detail. But noise is also an issue. If I had to compromise (meaning that I can't use a phono section that is all-tube, is quiet enough yet has the needed gain), I think I would go with an SUT before using semiconductors. I guess I don't like the semiconductor sound (brighter due to emphasis of odd ordered harmonic distortion, lacking in tubes and SUTs). |
Aesthetix is all tube gain stages I believe. I believe they are all tube gain stages. I do think the mm with SUT is perhaps the "standard". However, there can be more work involved with getting SUT loading optimized. Not a big thing, just a bit more work involved. In my case, the phono stage I use has a FET mc gain stage that is configured conservatively. I.e., most of the gain is realized through the MM stage. I can run my XV-1s through my MM stage. Obviously, it sounds much better through the mc inputs, but I think that gives an idea of how much gain is applied at each stage. I leave the discussion now since Actusreus is really looking for designer input. |
There is no general rule for a specific Design. You will find good ones, you will find bad ones, you will find right ones, you will find wrong ones and every designer will tell you why he went that way and not the other one....:-) To design a good Phonostage which sounds like the real thing is quite a task, specially when you need higher gain. The main problem today is profit and good parts are simply expensive, see resistors, caps, wire and so on and when everything will be tested in a lab, the costs can ruin the mark up. Hyping is more cost efficient and it works also pretty good today. I used for a lot of years high gain transistor stages, now I am with SUT's based units. Here it is the comparable problem, the normal SUT's are in general cheap, lousy and horrible from the datas. But the price counts. Top SUT's which have excellent datas, superior channel separation and uncolored sound, they exist, but no one buys them. Too expensive for "High End" units. No matter how much money you are willing to spend, you don't get it. Professional Users, yes, they do, but they tell the manufacturer what datas they want to have and they pay the price for it.. |
Dan_ed, Everyone's opinion is more than welcomed :) Your comment about optimization is one of the issues that I was thinking of as well. It seemed to me it would pose a bit of a challenge in designing a phono stage that utilized step-up transformers rather than FETs. I have it from a good source that Fremer's review of the Chinook is coming out, I believe, in the next issue of Stereophile. It should be an interesting read. |
One challenge facing designers who wish to use solid state these days is the rapidly dissappearing discrete transistor. Good low noise parts (both bipolar and FET) are becoming harder and harder to find, because the mass market electronics industry is moving to integrated circuits soley. That means OP amps at the minimum, with other mass market parts incorporating analog audio into LSI and VLSI parts. I-Phones, I-pads, MP3 players have no need of discrete semis. If there is no volume or demand, there is no product. High end audio demand is not even on the map. I have trying to procure a discrete transistor hybrid class A operational amplifier that is used in the professional microphone amp industry. The manufacturer tells me the parts he has used as low noise input transistors have been discontinued by the manufacturers so he is scrambling to find replacements. Ironicly, a technology long considered obsolete (tubes) seems to have better support from a parts perspective than discrete semis. I am a solid state biased designer, as I don't care for coupling caps and transformers in the audio chain if they can be avoided. |
Dear DHL, The late Allen Wright told me much the same thing about the scarcity of high quality discrete transistors that he experienced in building his two phono stages, FPV and RTP. I was implementing a hybrid dual differential cascode input stage in my own phono stage, and AW advised me to try the MAT02 bipolar for the bottom active element in the cascode. With some effort I finally found some, and I bought a bunch, just in case. But when I asked AW why he was not using that part in the RTP3C (or whatever), he said he could not afford to risk it because of the difficulty in obtaining a stockpile for production and future repairs. The MAT02 had already been discontinued at that time. As I understand it opamps and other ICs are also constantly being discontinued in favor of newer types that offer some real or imagined advantage. |
Lewm: I think that part MAT02 is the same part I was refering to. As the older analog device producers get aquired by larger companies, many of these parts get lost in the shuffle, unless the customer starts screaming. Couple that with the fact that fabs for discretes are becoming more expensive to maintain...not even the Chinese want to make discretes. Large power transistors are still available, but small signal are getting tougher to find. Suprisingly, many of the early op amps are still available, probably due to replacement requirements in military and industrial equipment. All the late 70's TL070 and 080 series are still with us, as are the the very first 741s. But special parts from Burr Brown and PMI have dissappeared. |
The MAT02 certainly lives up to AW's opinion of it in the context of my preamp. I have two extra pairs, just in case. (As you know, the MAT02 is a dual section device, so one of them suffices for each balanced channel, and the section matching in terms of transconductance, etc, is way better than one could hope for in a dual section tube, plus it stays that way over time.) Everything else downstream is tubes. |
Interesting addendum. I received an issue of TAS yesterday that contains a roundtable discussion of amplifier and preamplifier design by a collection of well known designers, e.g., Rowland, Pass, Curl, Carver, Lamm, etc. It's mostly self-serving drivel, a la TAS, but at least one or more of the panel notes that "audio grade" solid state devices, meaning devices specifically made with a view to no compromise audio component design and build, are becoming ever more scarce. He fears that his work will eventually be hampered by sheer lack of the optimal parts. Other panelists were more optimistic, however. |