It happens that I lately become heavily engaged a project that covered me from music-listening and the Morons-insulting person into a soldering and a constantly questionsasking Moron. Well, it is life.... :-(

I lately have some reasons to think about optimum requirements for SET’s out transformer if the amp handles juts a narrows bandwidth. For instance let pretend that we have 2 monoblock and each of them drive: LF <100Hz and HF >100Hz. So, would it be possible to optimize a specific transformer to push an envelop of an each reproduced range?

So, Chris, Thorsten, Dima, Joe or whoever “get” it, could you come up with some, preferably radical, ideas?

I wonder if an oversized trans with a lot and lot of iron in its core, with huge inductance of primary, with few turns and thick weirs will “do” on the LF amp. If for instance, the amp that does not handle anything under 1000Hz would be benefited if to make it with an air core?

Did anyone ever try it or hear about such an attempts? Are any seeds of an engineering wisdom that you might spread?

Thanks,
Romy the Cat

chaps,
I found the design details; it was posted by on the Joenet a few years ago.
Basically, it involved making a vertically sectioned bobbin; start with a 70mm diameter cylinder of whatever material (I used white nylon cos it was available and cheap; teflon or something like maple may be the way to go), 78mm length.

Then cut 17 grooves in, each 3mm wide, separated by about 1.6mm of material. Groove depth of a few mm.

Primary is 9 sections of 34 swg wire, about 200-250 turns each, secondary 8 sections, 22 gauge, wound to the desired turns ratio (about 18 for the original design). All sections wired in series, with alternate winding of Primary/Sec/Primary/Sec etc

Now this gave about 200mH of inductance, which is about right for a tweeter application with low Z tubes. DC resistance (DCR) of primary was about 200 ohms, which is probably on the high side for a 6C33, so may want to experiment with thicker windings and more/deeper sections, although more turns will degrade top end response. I think with the 6c33 you may get away with it though.

Now, given that the 6C33 has such a splendidly low internal impedance, it seems a shame to go and wind a high DCR transformer. Particularly for bass duty - which is why I think an enormous core with many turns (!) of thick wire is the way to go. You need the turns for the inductance and to keep the core operating at a low flux level at LF (less core and tube distortion).

The flux due to an ac signal component is proportional to (primary ac voltage) / (core cross sectional area * number of turns * frequency).

more later....

Romy the Cat wrote:

I lately have some reasons to think about optimum requirements for SET’s out transformer if the amp handles juts a narrows bandwidth. For instance let pretend that we have 2 monoblock and each of them drive: LF <100Hz and HF >100Hz. So, would it be possible to optimize a specific transformer to push an envelop of an each reproduced range?

Absolutely. Making a transformer for < 100Hz would mean to use a huge core and a fairly large airgap to get more inductance at a given standing current. This makes winding the transformer with low parasitic capacitance and low leakage inductance a problem that is increasingly hard to solve as size increases (though I believe S&B has a solution to that problem now).

Equally, if you make sure the Amplifier for > 100Hz does NOT RECEIVE any input below 100Hz (or at least an input that falls off progressively), meaning a line-level highpass or one build into the interstage coupling, then you can make the transformer relaively small with relatively few turns as it does not need to handle LF anywya, which makes getting low parasitic capacitance and low leakage inductance much easier.

Romy the Cat wrote:

If for instance, the amp that does not handle anything under 1000Hz would be benefited if to make it with an air core?

cv wrote:

Fairly sure it was a typo, but the bigger the airgap, the *lower* the inductance, all else being equal.We need a large airgrap for a 6C33 due to the high current. The advantage is that the airgap linearises the BH characteristic of the core, so that it's perm is more constant (generally a function of level and frequency, particularly at LF).

Allo T
Ah, I think we are agreed on all counts.
The suggestion of low bass via the deliberate introduction of its harmonics is something that has been done with guitar amps since way back, even if it wasn't the original intention...

cheers
cv

Chris,

Since I got that mind-boggling result with “Melquíades” (will finish the thread about the Melquíades sound) I went for an “ultimate solution” for my installation: a “Super Melquíades” amp. It will be dual chassis (amp+PS) 3-chennal amp to milliamp my horns by one amp per side.

When I asked Dima for an absolute craziest specification for LF channel he come up with something ridicules like .5A, full power down to 10Hz at –1dB and so on. In other words he proposed some kind of monster that will not run over 3Khz but would be an endless emperor at the bottom octave.

My transformer maker looked at this said “Hm…” and then decided to try making it. I spoke with him today and he said that he almost met already the Dima’s specifications, furthermore…. he so love the result that he is planning to convert it into a commercial product.

Considering that phenomenal bass that “Melquíades” do now with it’s 9H of primary and with 21Hz of cut off I wonder what “Super Melquíades” will be doing with a transformer the handle a full power down to 10Hz, having a lot of spare current in gap and with the inductance of primary over 15H? I really would like to here THAT sound. If everything turns out to be OK then I will do in a month.

Rgs,
The Cat

This will be very interesting indeed... that is an enormous airgap and will make for a very linear and unstressed hunk of iron. You may wish to apply similar tactics to the power supply (low DCR inductors) if you want to be really surprised...

Which reminds me, I must find time to measure that aircore.

Look forward to hearing all about it - I'm quite enjoying this vicarious little journey.

cheers
cv

cv wrote:

You may wish to apply similar tactics to the power supply (low DCR inductors) if you want to be really surprised...

I usually use my input chokes 4-5 times of critical inductance and twice of rated current, so it is very far form saturation. For instance the Melquíades’ output stage sucks 200mA and it uses input indictor of 10H and .5A. It has DCR of 53R. I might use 1A rated choke with lower DCR but what would be an effect. As I understand the chokes present inductance at a given current. So, when I “undercurrent” the chokes I’m loosing the effective inductance in them. Perhaps I might for my 200mA load try 20H choke at 1A but what sonic benefits I would get if I go there?

Also I would like to learn one more thing. I am planing to have a transformer for the B+ of the output stage rated at 1A. It will give me a nice spare current for 3 output stages even if I run all MF and HF hannels with dual anodes. My question is: would it be advisable to have the LF channel running it's own B+ supply? Theoretically it is class A operation and the current draw is constant but still something suggest me that I might consider this options. What would you your take on it? A separate supply with over-current transformer and a huge indictor in the filter?

Rgs,
Romy