Okay, I just got two really old output transformers (1950's). I was planning some kind of amplifier around these. Trafos has two primarywindings for push-pull outputstage and two secondary windings. I used 12VAC psu to measure impedance ratio.
input: 12VAC in single sec.winding caused 2x115V in "primary windings".
230/12=19.1
19.1*19.1=365:1 impedance ratio.
So if I use 8ohm speaker, there is 2900ohm load impedance for tubes. Is that correct?
Physical sizes of trafo:
10x12cm, cross section area is 12cm2, enough for 50W output?
Anyway, that is good starting point. Now I just need some tubes which can handle 2.9khom load. Any ideas? 5881, EL34 or similar would be nice.
input: 12VAC in single sec.winding caused 2x115V in "primary windings".
230/12=19.1
19.1*19.1=365:1 impedance ratio.
So if I use 8ohm speaker, there is 2900ohm load impedance for tubes. Is that correct?
Physical sizes of trafo:
10x12cm, cross section area is 12cm2, enough for 50W output?
Anyway, that is good starting point. Now I just need some tubes which can handle 2.9khom load. Any ideas? 5881, EL34 or similar would be nice.
These transformers lies in old tube amps. Sad thing is that there is no tubes(except preamp stage, EF40) and another one has no power transformer. Only chassis and components and both output transformers are still there.
I'm not 100% sure but I think this amp has used pair of EL51 pentodes. No other tubes has matching pinouts.
There is also phase shifting transformer for output tubes. preamp stage has EF40 and EL41. As much as possible, I want to use original circuit with transformer coupling. Power supply will be external.
Output transformers DC resistance: pri - 2x154ohm(308 p-p), sec: 2x1.3ohm.
Picture:
amps
I'm not 100% sure but I think this amp has used pair of EL51 pentodes. No other tubes has matching pinouts.
There is also phase shifting transformer for output tubes. preamp stage has EF40 and EL41. As much as possible, I want to use original circuit with transformer coupling. Power supply will be external.
Output transformers DC resistance: pri - 2x154ohm(308 p-p), sec: 2x1.3ohm.
Picture:
amps
Hi tonskulus ,
If you see the EL51’s datasheet , you will conclude that your
output transformer , was built regarding the old 12 ohms spea-
kers from old-time cinemas , so the primary impedance was 4300 ohms plate to plate , enough for a 65 Watts push-pull class AB cathode bias amplifier , BUT if you will use an 8 ohms speaker (no less ), the primary impedance will be 2900 ohms plate to plate as you had calculated .
In being so , IN MY OPINION , you have some options , but the
“best” ones are :
1st ( so/so good ) – KT88 , push-pull , triode connected , class AB1,
cathode bias , +B = 450 V , Pout ~ 30 W
2nd ( good ) – EL34 , single-ended , pentode connected , class A ,
cathode bias , +B = 270 / 300 V , Pout ~ 10 W
3rd ( best ) – EL34 , push-pull , pentode connected , class AB1 ,
fixed bias ( - 36V ) , +B = 380 / 400 V , Pout ~ 45 W
As I said above , these options are MY OPINION , other people
may has different one , BUT , I believe that , this is a good starting point for your job .
Regards ,
Carlos
If you see the EL51’s datasheet , you will conclude that your
output transformer , was built regarding the old 12 ohms spea-
kers from old-time cinemas , so the primary impedance was 4300 ohms plate to plate , enough for a 65 Watts push-pull class AB cathode bias amplifier , BUT if you will use an 8 ohms speaker (no less ), the primary impedance will be 2900 ohms plate to plate as you had calculated .
In being so , IN MY OPINION , you have some options , but the
“best” ones are :
1st ( so/so good ) – KT88 , push-pull , triode connected , class AB1,
cathode bias , +B = 450 V , Pout ~ 30 W
2nd ( good ) – EL34 , single-ended , pentode connected , class A ,
cathode bias , +B = 270 / 300 V , Pout ~ 10 W
3rd ( best ) – EL34 , push-pull , pentode connected , class AB1 ,
fixed bias ( - 36V ) , +B = 380 / 400 V , Pout ~ 45 W
As I said above , these options are MY OPINION , other people
may has different one , BUT , I believe that , this is a good starting point for your job .
Regards ,
Carlos
If you want to use the interstage phase splitter, back in the old days they seemed to have fairly limited frequency response(12Khz response was considered fairly good). You might need to redesign the front end to lower driving impedence. This should extend their useful response somewhat. Think about rapping some loop feedback around the interstage. The pentode driver doesn't seem like a good idea.
Shoog
Shoog
refference said:
That sounds good but I'm not sure about that low load impedance for two EL34 tubes. Under 3000ohm impedance gives 45W out? I think this will need four tubes, two in parallel to match 3000ohm load impedance. Well, I have test bench ready so I can test this configuration. Is that 380V for G2 and 400V for plate? Well, there should be at least 300V for G2 to get as low plate resistance as possible. However, let's see how it works.
Glowing plates and grids..Not so unusual for my tubeamp prototypes.
I still can't figure out everything necessary what comes to tube theory.. but practice makes makes champion.
Hi tonskulus ,
Sorry for the delaied reply .
The optimum plate to plate impedance for 2 x EL34 in the
push-pull way is 3500 ohms , BUT with 2900 ohms they
still work properly .
Don’t try to use 4 x EL 34 , being paralleled 2 by 2 , becau-
se the total current will rise A LOT , and the output transfor-
mer will “suffer” A LOT , perhaps blow out .
But if you have any doubt about my previous suggestion , START FROM A BIT LOWER POINT , then I recommend you to do the following experiment :
Class AB1 , EL 34 push-pull pentode operation , CATHODE BIAS
+ B = 350 Volts able to supply at least 190 mA
Supressor grids (G3) – Connect each one with the respecti-
ve cathode
Screen grids (G2) : Connect each one to the + B , using a
470 ohms x 2 watts resistors , obviously ,
one resistor for each grid .
FORGET ABOUT FIXED BIAS : Use cathode bias , strap
the two cathodes togheter , then connect a 135 ohms x 10 W
( two paralleled 270 ohms x 5 W ) wirewound resistor , paralle-
led with a 220 uf x 50 V cap , between the two cathodes and
Ground . The bias voltage at the cathode , must to stay around
22.5 ~ 25 Volts ( optimum 23.5 Volts )
Control grid (G1) : Connect each one to the earth , by a 220 K
x ½ Watt resistor .
If you build a convenient phase inverter , able to produce a maxi-
mum 15 Vrms of audio signal , at each output , you can do your
1st listening test . Be careful to choose the speaker , the impe-
dance value have to be 8 ohms , NO LESS . This assembly can
produce a maximum 35 Watts of power output , with only 1.5%
THD without feedback . With global feedback loop , this value
will lower to aprox. 0.15% or less .
Any doubt , feel free to ask .
Carlos
Sorry for the delaied reply .
The optimum plate to plate impedance for 2 x EL34 in the
push-pull way is 3500 ohms , BUT with 2900 ohms they
still work properly .
Don’t try to use 4 x EL 34 , being paralleled 2 by 2 , becau-
se the total current will rise A LOT , and the output transfor-
mer will “suffer” A LOT , perhaps blow out .
But if you have any doubt about my previous suggestion , START FROM A BIT LOWER POINT , then I recommend you to do the following experiment :
Class AB1 , EL 34 push-pull pentode operation , CATHODE BIAS
+ B = 350 Volts able to supply at least 190 mA
Supressor grids (G3) – Connect each one with the respecti-
ve cathode
Screen grids (G2) : Connect each one to the + B , using a
470 ohms x 2 watts resistors , obviously ,
one resistor for each grid .
FORGET ABOUT FIXED BIAS : Use cathode bias , strap
the two cathodes togheter , then connect a 135 ohms x 10 W
( two paralleled 270 ohms x 5 W ) wirewound resistor , paralle-
led with a 220 uf x 50 V cap , between the two cathodes and
Ground . The bias voltage at the cathode , must to stay around
22.5 ~ 25 Volts ( optimum 23.5 Volts )
Control grid (G1) : Connect each one to the earth , by a 220 K
x ½ Watt resistor .
If you build a convenient phase inverter , able to produce a maxi-
mum 15 Vrms of audio signal , at each output , you can do your
1st listening test . Be careful to choose the speaker , the impe-
dance value have to be 8 ohms , NO LESS . This assembly can
produce a maximum 35 Watts of power output , with only 1.5%
THD without feedback . With global feedback loop , this value
will lower to aprox. 0.15% or less .
Any doubt , feel free to ask .
Carlos
Seems to be pretty good now. There is pair of EL34, cathode biased. Not too much output power yet but this is good starting point as I'm very pleased for this.
My speakers has two 8ohm woofers so I can connect them in series to get that 16ohm..It might be better, 5600k load impedance for tubes.
picture
My speakers has two 8ohm woofers so I can connect them in series to get that 16ohm..It might be better, 5600k load impedance for tubes.
picture
I have now 180ohm cathode resistor. G1 voltage seems to be
-23.8Volts, G2 - 320V, B+ 330V.
There is still some slight red glow at screen grids..It might be just reflection from orange cathode..not sure about that.
And, at idle, there is strong blue glow which dims as signal is applied.
I havent measured output power yet but I think it is somewhere beetween 10 and 20watts, 8ohm load is used. Let's see how it works with 16ohm load.
-23.8Volts, G2 - 320V, B+ 330V.
There is still some slight red glow at screen grids..It might be just reflection from orange cathode..not sure about that.
And, at idle, there is strong blue glow which dims as signal is applied.
I havent measured output power yet but I think it is somewhere beetween 10 and 20watts, 8ohm load is used. Let's see how it works with 16ohm load.
Hi tonskulus ,
The voltage , seems to be OK , BUT if I was you , I would
try to rise the + B voltage to 350 V , and then lower a bit
the cathode resistor , may be 150 ohms or 165 ohms .
Don’t forget , to measure the ACTUAL output power , you
need to substitute the speaker , by a wirewound resistor of
8.2 ohms x 50 watts ( or any kind of resistor’s association
that gives this value ) , and then , measure the voltage across
this resistor , for a 400 or 1000 Hz , sinusoidal signal input .
The input signal and the output voltage , must be monitored
by a double trace scope .
You have to rise the input signal slowly , until you can see
in the scope a distorted sine wave at the output , and at this
point measure the voltage across the 8.2 ohms resistor and
then ( using the Ohm’s law ) calculate the ACTUAL RMS
maximum power .
You can use the 8 ohms ( NO LESS ) speaker , without fear ;
a 16 ohms speaker , will cause a loss of high frequencies , due
to impedance’s mismatch .
If you see any glow on the screen grid , you can rise the resis-
tor’s value , from 470 to 680 or 1000 ohms , but I don’t believe
that it will occur .
About the input stage and the driver stage , seems to be good ,
the idea to use the Mullard’s 520 design ( EF86 or 6AU6 and
an ECC83 , configured as a Schmidt phase splitter )
Good luck ,
Carlos
The voltage , seems to be OK , BUT if I was you , I would
try to rise the + B voltage to 350 V , and then lower a bit
the cathode resistor , may be 150 ohms or 165 ohms .
Don’t forget , to measure the ACTUAL output power , you
need to substitute the speaker , by a wirewound resistor of
8.2 ohms x 50 watts ( or any kind of resistor’s association
that gives this value ) , and then , measure the voltage across
this resistor , for a 400 or 1000 Hz , sinusoidal signal input .
The input signal and the output voltage , must be monitored
by a double trace scope .
You have to rise the input signal slowly , until you can see
in the scope a distorted sine wave at the output , and at this
point measure the voltage across the 8.2 ohms resistor and
then ( using the Ohm’s law ) calculate the ACTUAL RMS
maximum power .
You can use the 8 ohms ( NO LESS ) speaker , without fear ;
a 16 ohms speaker , will cause a loss of high frequencies , due
to impedance’s mismatch .
If you see any glow on the screen grid , you can rise the resis-
tor’s value , from 470 to 680 or 1000 ohms , but I don’t believe
that it will occur .
About the input stage and the driver stage , seems to be good ,
the idea to use the Mullard’s 520 design ( EF86 or 6AU6 and
an ECC83 , configured as a Schmidt phase splitter )
Good luck ,
Carlos
refference said:
There is one problem, I have no transformer for that voltage. Next step would be over 400volts..(can't remember for sure but it was something like 2x350V output). Is this too much? How about using fixed bias and 16ohm speakers (it makes 5600ohm load impedance for tubes)?
I was also wondering that there is 2x1.3ohm DC resistance in output trfmr secondary, it is a bit high for 8ohm speaker. Wire diameter is about 0.5-0.7mm. 16ohm needs less current so it might be better. I have used to that output DC resistance is in range of 0.2-1.0ohms in typical output transformers.
Hi tonskulus ,
There is NO problem . If you have a 400 V dc power supply ,
you can connect an additional RC filter into the + B circuit ,
where the R ( resistor value ) can be calculated by Ohm’s
law , to give the desired voltage ( 350 V or even 375 V ) at
the output . In doing so , as a bonus , you will get an ULTRA
LOW ripple and a fictitious “soft-start” .
Forget about fixed bias , this tube’s working point is ONLY for
CATHODE BIAS .
As I had said , use an 8 ohms load ( NO LESS ) , or even an
8 ohms speaker series connected with a 1.2 or 1.5 ohms x
20 Watts wirewound resistor , if you don't want to worry about .
The wire diameter , has relationship with the power allowable ,
there is nothing to do with the impedance . The MOST important
is the turns ratio between primary and secondary , and you had
said on your first post that it was 19.1 : 1
I think that you are geting 21 Watts over 8 ohms load , may be ,
because the input signal ( at control grids ) , is lower than the
necessary value , to drive the tubes to maximum power . If you
are using the scope , perhaps the distortion that you could see
is on the first stage , or even on the output transformer and not
on the final stage . That is a good reason because the amp could
not reach the maximum power .
Below I am reproducing the datasheet so you can analyze the
things , better :
EL34 – 6CA7
Class AB1 , Pentode operation , Push-Pull
Plate Voltage ............................................. 375 Volts
Grid # 2 voltage ......................................... 365 Volts
Grid # 2 resistor ......................................... 470 ohms
Cathode resistor ....................................... 135 ohms
Signal input voltage ................................... 20 Volts RMS
( Aprox. 28 V peak )
Zero signal plate current ............................ 150 mA
Maximum signal plate current .................... 188 mA
Zero signal grid # 2 current ........................ 25 mA
Maximum signal grid # 2 current ................ 39 mA
Load resistance ( plate to plate ) ............ 3000 to 3500 ohms
Power output ............................................... 35 Watts
THD ............................................................. ~ 1.5 %
Obs. Cathode resistor common to both tubes , current values
for both tubes ( togheter )
Bibliography : SYLVANIA - Technical Manual
As you can see , the grid bias value in the steady state will be
aprox. 175 mA x 135 ohms = 23.5 Volts , as I had said .
Try again . Good luck ,
Carlos
There is NO problem . If you have a 400 V dc power supply ,
you can connect an additional RC filter into the + B circuit ,
where the R ( resistor value ) can be calculated by Ohm’s
law , to give the desired voltage ( 350 V or even 375 V ) at
the output . In doing so , as a bonus , you will get an ULTRA
LOW ripple and a fictitious “soft-start” .
Forget about fixed bias , this tube’s working point is ONLY for
CATHODE BIAS .
As I had said , use an 8 ohms load ( NO LESS ) , or even an
8 ohms speaker series connected with a 1.2 or 1.5 ohms x
20 Watts wirewound resistor , if you don't want to worry about .
The wire diameter , has relationship with the power allowable ,
there is nothing to do with the impedance . The MOST important
is the turns ratio between primary and secondary , and you had
said on your first post that it was 19.1 : 1
I think that you are geting 21 Watts over 8 ohms load , may be ,
because the input signal ( at control grids ) , is lower than the
necessary value , to drive the tubes to maximum power . If you
are using the scope , perhaps the distortion that you could see
is on the first stage , or even on the output transformer and not
on the final stage . That is a good reason because the amp could
not reach the maximum power .
Below I am reproducing the datasheet so you can analyze the
things , better :
EL34 – 6CA7
Class AB1 , Pentode operation , Push-Pull
Plate Voltage ............................................. 375 Volts
Grid # 2 voltage ......................................... 365 Volts
Grid # 2 resistor ......................................... 470 ohms
Cathode resistor ....................................... 135 ohms
Signal input voltage ................................... 20 Volts RMS
( Aprox. 28 V peak )
Zero signal plate current ............................ 150 mA
Maximum signal plate current .................... 188 mA
Zero signal grid # 2 current ........................ 25 mA
Maximum signal grid # 2 current ................ 39 mA
Load resistance ( plate to plate ) ............ 3000 to 3500 ohms
Power output ............................................... 35 Watts
THD ............................................................. ~ 1.5 %
Obs. Cathode resistor common to both tubes , current values
for both tubes ( togheter )
Bibliography : SYLVANIA - Technical Manual
As you can see , the grid bias value in the steady state will be
aprox. 175 mA x 135 ohms = 23.5 Volts , as I had said .
Try again . Good luck ,
Carlos
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.