O Timbreflex

4

Projeto dos transformadores de saída

Documento que iniciou as pesquisas de Eduardo de Lima

para os transformadores de saída.

http://www.alinden.mynetcologne.de/mc/amplifiers.pdf

Using Windings Of Output Transformer To Provide Negative Feedback To Cathode Of A Single-ended Output Stage

http://livinginthepast-audioweb.co.uk/index.php?p=feedbackwindings

Purpose: to improve output speaker damping in order to reduce loudspeaker bass resonance ('boominess') that is normally suppressed by closed loop negative feedback coupled with high open loop gain. This does not exist in 'minimalist' single-ended valve amplifiers of few valves, low gain and having little or no negative feedback at all.

Test Circuit
Output valve: KT66 as triode
Driver valve: EF36 pentode (gain = 40dB [x100])
Transformer: DB1087 (A, B); modified DB1090 (C)

Using Windings Of Output Transformer To Provide Negative Feedback To Cathode Of A Single-ended Output Stage

Purpose: to improve output speaker damping in order to reduce loudspeaker bass resonance ('boominess') that is normally suppressed by closed loop negative feedback coupled with high open loop gain. This does not exist in 'minimalist' single-ended valve amplifiers of few valves, low gain and having little or no negative feedback at all.

Test Circuit
Output valve: KT66 as triode
Driver valve: EF36 pentode (gain = 40dB [x100])
Transformer: DB1087 (A, B); modified DB1090 (C)

Electrical Model To Emulate Loudspeaker Resonance
DC resistance: 7.6Ω 

Performance Of Test Circuit Into The Electrical Model

'A': no feedback, Ra = 5,000Ω 
'B': using 16Ω speaker winding, Ra = 5000Ω 
'C': using 1300Ω section of primary, Ra = 3700Ω 

Results
It's interesting that the resonance with no feedback is still quite low at only +3dBV, so some damping is evident, mostly due to the low resistance of the speaker winding, but also because the Ra increases in proportion and so the power goes down. This is not a complete picture of course, since for a real speaker the effect is exacerbated by mechanical resonance of the speaker cone, resulting in exagerated sound output at the resonant frequency.

Best results were obtained using the full (16W) speaker winding in series with the cathode ('B'). This reduced the resonance peak to just +1dBV. Surprisingly, using a 1300Ω section of the primary in its place was worse, producing +2dBV ('C'), not very different from 'A'!

Total Voltage Gain Of Output Stage
I.e. input grid to 8Ω speaker tap:

Note the 1300Ω winding added an extra 37Ω to the cathode circuit, which of course could not be bypassed, and so this mostly accounted for the reduced gain for 'C', in addition to the feedback. Of course the speaker winding option ('B') has a negligible impedance by comparison.

My personal conclusion is that option 'B' offers the best compromise (and doesn't involve hacking the transformer primary!). I tried the amp with a small, very simple, closed-box 'music-centre' type speaker and couldn't discern very much in the way of resonance.

To hear is to forget
To see is to remember
To do is to learn
To learn is to know
To know is to grow
To grow is to give
To give is to live

Anon

Knowing is not enough; we must apply.
Willing is not enough; we must do.                 

Goethe

LM3 (Low Mu Triode with Higher Raw Efficiency Emulator – LMTHREE)

Esquemático dos quatro tipos de enrolamentos dos transformadores Maggiore 100

Cinza =ferragem tipo EI-50 com empilhamento de 50mm.

Enrolamentos rosa fio #26 = primário 1 (placa) e 26 (fonte). Enrolamentos vide tabela.

Derivação do primário = 23 (para grade auxiliar).

Enrolamento verde fio #31 = realimentação do catodo 11 e 12.

Enrolamentos azuis fio #24 = secundário três bobinas em paralelo 7+15+19 e 8+16+20

38 voltas fio cada uma.

Bolinhas dentro dos enrolamentos= início de espiralagem.

De acordo com o que foi explicado em http://www.novacon.com.br/audioctrasnf2.htm termos as seguintes RELAÇÕES ENTRE ESPIRAS nos quatro tipos de transformadores usados no Maggiore 100:

Tipo 1 =  3.500 x 2.5  =  1400   ∴ √1400 = 37.4

Tipo 2 =  4.900 x 2.5  =  1960   ∴ √1960 = 44.27

Tipo 3 =  2.500 x 2.5  =  1000   ∴ √1000 = 31.6

Tipo 4 =  5.900 x 2.5  =  2360   ∴ √2360 = 48.57

Desta forma teremos como ESPIRAS primário, secundário e terciário:

Todos os Secundários fixados em 3 x 38 espiras  fio# 24

Primários:  fio #26

Tipo 1 = 37.4  x 38 = 1421 (9 camadas de 158 espiras)

Tipo 2 = 44.27  x 38 = 1682 (9 camadas de 186 espiras)

Tipo 3 = 31.6  x 38 = 1200 (9 camadas de 134 espiras)

Tipo 4 = 48.57  x 38 = 1846 (9 camadas de 205 espiras)

Derivação dos primários a contar da linha de alimentação:

Ponto 23.

Tipo 1 = 2ª camada de 158 espiras  (316ª espira)

Tipo 2 = 2ª camada de 186 espiras  ( 372ª espira)

Tipo 3 = 2ª camada de 134 espiras  ( 268ª espira)

Tipo 4 = 2ª camada de 205 espiras  ( 410ª espira)

Terciários:  fio #26

Ligação ao catodo: Ponto 12

Tipo 1 = 158 espiras

Tipo 2 = 186 espiras

Tipo 3 = 134 espiras

Tipo 4 = 205 espiras

Placa da válvula: Ponto 1.

Alimentação +B: Ponto 26 .

Distribuição dos enrolamentos

XXXXXXXXXXXXXXX