|
Posted by Romy the Cat on
08-05-2005
|
|
If you are an adherent to single-ended amplifiers and high-efficiency loudspeakers then undoubtedly you came across to the thought what to do with LF sections and how to drive them. The high sensitivity at LF comes only along with the non-resolvable cost of delays (if it done properly) and therefore it is highly likely that your LF would require much more power then your MF-HF channels would requires.
It is very common that people use SS PP to drive the LF sections. People attracted by many in theory objective benefits of SS amplification and amp-speaker interface. However, the truth of the subject is that the SS amplification (ANY SS amplification) can’t do bass. The bass of SS amp, although it is instantaneously impressive (and particularly for the people who have their audio taste at a typically barbarian level) is wrong bass, purely synthetic and NEVER real. I am not mentioning the fact that in biamping the different amps must not be used at all…. So, no matter what you do but do not biamp with SS PP at LF.
From another point of view the SE amplifier can’t do REAL bass by the verdure of the topological limitation. So, what to do? The answer is ridiculously simple, so simple and reasonable that I have no idea why the proposed solution is not a dominating solution in a given situations: use two SE amplifiers when one of them will be specifically optimized for LF operation only.
The dedicated LF SE amp is so powerful tool on context of your SE high-efficiency installation that it will take your “result” way out there. Perhaps it will even teach you REPRODUCED bass might sound….
Rgs,
Romy the Cat
|
|
|
|
Posted by Teflon on
06-04-2012
|
|
fiogf49gjkf0d I'm interested in bi-amping so you have my interest - if you say a SET has a 'topological limitation' what is the ridiculously simple change needed to make it good for LF operation ?
I was previously thinking about using triodes in PP, like Sakuma-san bi-amping.
|
|
|
|
Posted by Romy the Cat on
06-05-2012
|
|
fiogf49gjkf0d  Teflon wrote: | | I'm interested in bi-amping so you have my interest - if you say a SET has a 'topological limitation' what is the ridiculously simple change needed to make it good for LF operation ?I was previously thinking about using triodes in PP, like Sakuma-san bi-amping. |
|
So, Teflon, I do not think you got what I meant. SET has no 'topological limitation' because exists DSET. DSET is free from any, even theoretical limitation, of cause if you prepared to deal with price and power demands. The DSET optimization for LF operation is well covered in multiple threads if this site. It might depend mildly from the specifics of your amps that you would like to use for LF but the underlining principles are the same across the board. I do not remember that Sakuma-san use the DSET concept. The bi-amping that he demonstrated to me did not use DSET, used tow identical amps and in fact was implemented in what I call the “ridicules fashion”.
|
|
|
|
Posted by Teflon on
06-06-2012
|
|
fiogf49gjkf0d
| Romy the Cat wrote: | | So, Teflon, I do not think you got what I meant. SET has no 'topological limitation' because exists DSET. DSET is free from any, even theoretical limitation, of cause if you prepared to deal with price and power demands. The DSET optimization for LF operation is well covered in multiple threads if this site. It might depend mildly from the specifics of your amps that you would like to use for LF but the underlining principles are the same across the board. |
|
Ah, I understand what you meant now when you said 'topology', you refer to having a dedicated SET for each frequency band. I did read some of the threads on this site about your amplifier and what I see is dedicated amplifier channels for each frequency band, each one uses same basic topology but with differences at the input for frequency band-pass and uses different output depending on the demand from the load - meaning more power for lower frequency.
| Romy the Cat wrote: | | I do not remember that Sakuma-san use the DSET concept. The bi-amping that he demonstrated to me did not use DSET, used tow identical amps and in fact was implemented in what I call the “ridicules fashion”. |
|
this link explains what I was referencing. There are three pages, the schematic is on the 3rd page.
http://www10.big.or.jp/~dh/msm/msm1.html
What is the weakness you see in this implementation ? - is it because of the use of Push Pull for LF ?
|
|
|
|
Posted by Romy the Cat on
06-07-2012
|
|
fiogf49gjkf0d The Sakuma schematic the was presented in there:
http://www10.big.or.jp/~dh/msm/msm3.html
…was not the one that I referred in my comment above. The bi-amp that seen/heard in Sakuma place was different. The one at the picture above is legitimate way to do the thing, well, legitimate if you subscribe the way how Sakuma builds amps. Still, even if to accept the Sakumas’ idea of amp than I do recognise a lot of things that I personally not like in Sakuma’s amp from poorly bi-amping stand point.
The Cat
|
|
|
|
Posted by Teflon on
06-07-2012
|
|
fiogf49gjkf0d Romy, can you say what you do not like in the Sakuma amp that is poor from a bi-amping stand point ?
|
|
|
|
Posted by Romy the Cat on
06-08-2012
|
|
fiogf49gjkf0d This Sakuma’s is kind of BS. What I mean is that it has the appeal to be “good thinking” but only in the realm of marketing hype, something that the editor of “Direct Heating” had created around Sakuma. The reality is that the Sakuma’s TWO channel Amplifier is not truly interesting multi-amp solution, in fact it is VERY bad solution. For sure the Sakuma’s TWO channel Amplifier is a wonderfully vehicle for the “Direct Heating” to sell the Tamura iron, that is why the Sakuma’s syndrome was even created.
To understand the bogusness of the Sakuma’s TWO channel Amplifier you need to understand Sakuma’s personal objectives. He drivers a single speaker - stock Altec 7 – that is all that he cares. Sakuma never intend to use any DSET concepts, in fact he did not even understand them or have a need for them. He did clearly expressed it in his TWO channel Amplifier design.
Take a look, he has many bandwidth limiting elements sitting in chain. By doing it he sharpens the roll offs. The amp itself has no line-level filtration - the transformer after the second stage juts splits signal and then he uses the Altec’s speaker-level filters to shape the channel response. Very freaking intelligent! The performance of the former stages impact the performance/bandwidth of the LF amplifier – this is the horrible valuation of DSET concept and the bandwidth-specific optimization become impossible. Take a look at the transformer A4714 juts after EL34 tube. It has to be wide-band transformer as if it will be HF optimized than it will have inductance to give good LF for LF amp. If you however make this A4714 transformer with a lot of inductance, large with low section then you get a lot of capacitance hat will compromise you HF channel. The transformer TN6 and SYU-001 have to be bandwidth optimized but they are not. The whole amp is just a driver with transformer splitter with two separate power amps and speaker-level filtration – this is as stupid as it could be of one promotes the amp as “TWO channel Amplifier”. The idiocy of the design is even in the fact that Sakuma use two attenuators – one for HF and one for LF. This is foolishness as you need ONLY ONE attenuator.
I can go much longer but frankly whatever Sakuma’s does has very little interest in my view.
|
|
|
|
Posted by Teflon on
06-08-2012
|
|
fiogf49gjkf0d Hi Romy,
I see why you don't like the design. And in particular, the choice of where he chose to limit the frequency responses for the two channels. I see the benefit of the DSET approach in this context as you explained it. If I understand you, your amplifier uses line level frequency selection at the input to each channel, and the output of each channel is optimized for the frequency range of interest -you worry not about the performance of a channel outside of the pass-band providing it doesn't create problems. Am I understanding this properly ?
So, say I take an off-the-shelf stereo SET amplifier of good quality - i.e. designed to cover the full audio spectrum. Then say I take this amplifier and use it to bi-amp a single 2-way speaker. I would then install passive line level frequency filters at the input to each channel, one for LF and the other for HF. The output transformers would be wide band, so they would not be optimized for the frequency range they were used at but it would still work as a DSET topology ?
If I were to replace the output transformers such that the HF channel used a transformer that is 'smaller' with less parasitic inter-winding capacitance and the OPT in the LF channel replaced with a 'larger' one with higher inductance - then I am creating an optimized DSET topology ?
I don't claim to be familiar with the works of Sakuma-san, I did read the web site with interest though. It seems to me that he had developed a high level of familiarity with the performance, the limitations and the sound of the components he used - especially the Tamura transformers. This allowed him to create the sound he was looking for. And I perceive that he knew how to listen to the sound and knew what he needed to change to achieve his goals. I think this aspect of his work is something you can appreciate. And the benefit he appears to have brought forth was to create in others, the interest to build and listen to tube amplifiers and to encourage others to listen for themselves and learn what they should do to achieve the sound they like - as far as I can tell he did not ask others to copy his work, but rather to develop their own. Unfortunately the website dedicated to his work is not his words and was not created by him and so reflects the interests of it's creators instead.
|
|
|
|
Posted by Romy the Cat on
06-09-2012
|
|
fiogf49gjkf0d | Teflon wrote: | | If I understand you, your amplifier uses line level frequency selection at the input to each channel, and the output of each channel is optimized for the frequency range of interest -you worry not about the performance of a channel outside of the pass-band providing it doesn't create problems. Am I understanding this properly ? |
|
Yes, it is correct. Each individual amp of DSET configuration is optimized (and not only output) for very specific pass-band, still using the very same basic topology and the optimization methods that do not violate the given topology.
| Teflon wrote: | | So, say I take an off-the-shelf stereo SET amplifier of good quality - i.e. designed to cover the full audio spectrum. Then say I take this amplifier and use it to bi-amp a single 2-way speaker. I would then install passive line level frequency filters at the input to each channel, one for LF and the other for HF. The output transformers would be wide band, so they would not be optimized for the frequency range they were used at but it would still work as a DSET topology ? |
|
Yes, this is viable approach. The filters might be installed NOT only at input but wherever the topology allows. The main key in such conversion would be replacing the wide band output transformer (that is compromised by default) with two DSET band output transformers – one made for LF and one for HF.
| Teflon wrote: | | If I were to replace the output transformers such that the HF channel used a transformer that is 'smaller' with less parasitic inter-winding capacitance and the OPT in the LF channel replaced with a 'larger' one with higher inductance - then I am creating an optimized DSET topology ? |
|
Sure but it might be taken further than this. For instance for HF you might use some faster core or even air core transformer is you cross very high. For LF you use no sectioning, huge core, ability to care larger current, very fat wire with ultra low impedance as parasitic capacitance is irrelevant at LF. Then you can load your HF and LF out tubes differently and slightly differently power them. So, the possibilities are truly endless.
| Teflon wrote: | | I don't claim to be familiar with the works of Sakuma-san, I did read the web site with interest though. It seems to me that he had developed a high level of familiarity with the performance, the limitations and the sound of the components he used - especially the Tamura transformers. This allowed him to create the sound he was looking for. And I perceive that he knew how to listen to the sound and knew what he needed to change to achieve his goals. I think this aspect of his work is something you can appreciate. And the benefit he appears to have brought forth was to create in others, the interest to build and listen to tube amplifiers and to encourage others to listen for themselves and learn what they should do to achieve the sound they like - as far as I can tell he did not ask others to copy his work, but rather to develop their own. Unfortunately the website dedicated to his work is not his words and was not created by him and so reflects the interests of it's creators instead. |
|
I had interest in Sakuma unit I did hear the results, did see his amps and was able to observe what references he personally had in Sound. I did/do not care about what I have witnessed. I do not particularly criticize his crazy use of transformers. I have my reason do not like this direction but it is not the thread about Sakuma. What however need to be understood is that whatever Sakuma was pitching was in context of Tamura transformers from 80s. At that time the top of the line Tamura were good but they do not exist anymore. Tamura was bought out long time ago and nowadays Tamura is made by Chinese corporation that makes transformers for kid’s toys and Home Depot table’s lamps. Sure, they bare the same name but they have absolutely no resemblance sonically with what Tamura did in 80s.
The Cat
|
|
|
|
Posted by Teflon on
06-09-2012
|
|
fiogf49gjkf0d This makes it clear. I will give some further thought to my next projects based on these ideas.
There is a trade-off when optimizing the output transformers in this way - when listening to each channel during construction and testing etc. it will be much more difficult for me to judge the sound quality as I construct each part since I can not listen wide-band to each channel. Only the composite DSET can be listened to.
|
|
|
|
Posted by Romy the Cat on
06-10-2012
|
|
fiogf49gjkf0d Yes, to “voice” the DSET is tricky and requires a lot of listening experience, or more accurately saying – it requires to make a lot of mistakes and understand own mistakes. It is not difficult –juts require patience. Still, if one “get” how DSET sounds and how loading sounds then the rest might be “tuned” by hearing. I for sure would advised to use with DSET the speaker drivers of the very familiar sonic quality.
|
|