Hi there!
swapped out that old cap for a cheap polyprop motor-run type, & it powered up fine with no pyrotechnics - then oscillated! Reversed feedback connection, & bingo, music issued forth!
Ran it for 5 mins with a CD walkman throu headphone outlet to a test speaker, then connected it to a proper speaker. Sounds pretty good! (America - Simon & Garfunkel!) Open channel with no signal is near silent, ear to tweeter there's a soft hiss. No hum at all.
Not taken many measurements yet - too chuffed to hear it working. Voltages are all a little high as I have 249v coming out of the wall, but none so high I'm concerned. I have built a little signal generator while waiting for transformers to arrive, & I'm aiming to borrow a 'scope at the weekend (just need to learn how to use it!)
Internal wiring needs tidying as the fault finding has wreaked havoc with my initial layout. Thinking of taking B+ from the 5v CT for neatness. Am I right in thinking this will also drop a few volts?
Oh, and 'course there's the other channel still to build...
Transformers stayed cool, but a bit surprised how hot the power tubes get. How hot is normal? I could hear the glass pinging as they warmed up!
Thanks to all who've given advice - this forum is excellent. I'll add a photo of my handiwork next week.
Regs
Eduardo
swapped out that old cap for a cheap polyprop motor-run type, & it powered up fine with no pyrotechnics - then oscillated! Reversed feedback connection, & bingo, music issued forth!
Ran it for 5 mins with a CD walkman throu headphone outlet to a test speaker, then connected it to a proper speaker. Sounds pretty good! (America - Simon & Garfunkel!) Open channel with no signal is near silent, ear to tweeter there's a soft hiss. No hum at all.
Not taken many measurements yet - too chuffed to hear it working. Voltages are all a little high as I have 249v coming out of the wall, but none so high I'm concerned. I have built a little signal generator while waiting for transformers to arrive, & I'm aiming to borrow a 'scope at the weekend (just need to learn how to use it!)
Internal wiring needs tidying as the fault finding has wreaked havoc with my initial layout. Thinking of taking B+ from the 5v CT for neatness. Am I right in thinking this will also drop a few volts?
Oh, and 'course there's the other channel still to build...
Transformers stayed cool, but a bit surprised how hot the power tubes get. How hot is normal? I could hear the glass pinging as they warmed up!
Thanks to all who've given advice - this forum is excellent. I'll add a photo of my handiwork next week.
Regs
Eduardo
Hi Ed,
Glad you have got your amplifier sorted out!
Does your mains transformer not have a 250v input? Many transformers are spec'ed at 0-220-250V. As long as your HT+ reads correctly, not too serious. You should have wire-wound resistors inserted between the rectifier plates (pins 4&6, IIRC) and you can always increase their values a bit to get the correct HT conditions.
The EL34 O/P valves do indeed get hot, you won't be able to touch them after a few minutes operation, so don't try!
Your B+ (HT) take-off point should be on pin 8 of the rectifier socket, best is to leave it there. Insulate the 5v centre-tap (strange, this is not the norm for a 5v rectifier winding) so it doesn't touch anywhere it shoudn't.
Sit back, relax, and enjoy your amplifier! I look forward to a photo of all views, chassis as well as topviews.
-Eric
Glad you have got your amplifier sorted out!
Does your mains transformer not have a 250v input? Many transformers are spec'ed at 0-220-250V. As long as your HT+ reads correctly, not too serious. You should have wire-wound resistors inserted between the rectifier plates (pins 4&6, IIRC) and you can always increase their values a bit to get the correct HT conditions.
The EL34 O/P valves do indeed get hot, you won't be able to touch them after a few minutes operation, so don't try!
Your B+ (HT) take-off point should be on pin 8 of the rectifier socket, best is to leave it there. Insulate the 5v centre-tap (strange, this is not the norm for a 5v rectifier winding) so it doesn't touch anywhere it shoudn't.
Sit back, relax, and enjoy your amplifier! I look forward to a photo of all views, chassis as well as topviews.
-Eric
Congrtulations on getting the first amp working. Doesn't it feel good when you finally get the problems licked! I don't think I'd worry too much about the voltage readings in the amp unless they are far off. Much of the time the voltages specified are design values or are measured with the ideal mains voltage. You never said what the schematic you have specifies but in their book, Mullard shows 465V at pin 8 of the rectifier tube and 433 on the plates of the output tubes. These are "nominal" voltages. Unless yours a much higher, I wouldn't worry, like Eric said. According to one of my RCA Receiving Tube manuals, the 6CA7/EL34 can take 800V max on the plate and has a max cathode current of 150 mA with a plate dissipation of 25 Watts max. As long as you are not too close to these values, you'll be OK. The one thing to check is heater voltage. This should be pretty close to 6.3V and 5V for long tube life. There have been several good discussions on this forum on heater voltage and it's affects on sound, tube life, and other things. It's worth searching for them as they were all good reading with input from many knowledgable folks.
Now go get the other one done so's you can have Steereeoo!
Now go get the other one done so's you can have Steereeoo!
I've constructed a DIY power amp based on the Leak TL25+ (EL34PP)
Although very similar to the Mullard 5-20, I find the Leak circuit a much nicer amplifier.
schematic here : http://werple.net.au/~kiewavly/Leak/Circuits/tl25Plus.gif
What refinements I've done and is finally staying this way.
Its so far been bullet proof!.
1. Replaced from EL34 to KT66
2. Replaced cathode bias resistors to 560ohm 6Watt resistor
3. Instead of EF86, go for 12AX7A, lovely mids and less hard & harsh.
4. Replaced feedback resistor to 12k to increase NFB. (fine tune it to your actual speaker load)
5. Used Hammond power(375-0-375V) & output tranny.
6. Used 1% metal film & Viterous enamel wirewound resistors
7. PIO coupling and EvoxRifa PEG long life caps.
8. Solid 16gauge mildsteel chassis A4 dimensions & laser cut holes. Spraypainted at autoworkshop. (autopaint would do fine!)
Shuguang KT66, JJ tubes for the rest are an excellent choice & my favorites..considering the price against NOS's.
Bottomline, it's fabulous!.
Although very similar to the Mullard 5-20, I find the Leak circuit a much nicer amplifier.
schematic here : http://werple.net.au/~kiewavly/Leak/Circuits/tl25Plus.gif
What refinements I've done and is finally staying this way.
Its so far been bullet proof!.
1. Replaced from EL34 to KT66
2. Replaced cathode bias resistors to 560ohm 6Watt resistor
3. Instead of EF86, go for 12AX7A, lovely mids and less hard & harsh.
4. Replaced feedback resistor to 12k to increase NFB. (fine tune it to your actual speaker load)
5. Used Hammond power(375-0-375V) & output tranny.
6. Used 1% metal film & Viterous enamel wirewound resistors
7. PIO coupling and EvoxRifa PEG long life caps.
8. Solid 16gauge mildsteel chassis A4 dimensions & laser cut holes. Spraypainted at autoworkshop. (autopaint would do fine!)
Shuguang KT66, JJ tubes for the rest are an excellent choice & my favorites..considering the price against NOS's.
Bottomline, it's fabulous!.
I am tuning in to this post after all the good advice has been given, especially by my friend Eric who has constructed a number of 520s.
So perhaps just a few general comments:
(1) It can be disasterous to try and sort out output transformer feedback phase with full feedabck on! If one is not exactly certain, first omit the feedabck resistor and replace by a high value (100K, depending), keeping one end floating so that it can be momentary touched down. Without feedback there will be some hiss in the loudspeaker, and it is quite perceptable whether touching said feedback resistor momentary, will increase the hiss or decrease it. Switch the transformer connections to get a decrease in hiss, which will obviously be negative feedback. For this experiment all compensating caps. in the feedback line should also not be connected initially. Trying this with a high degree of feedback from the outset can fry your expensive output transformer because of flash-through as a result of high transients, should the phase be wrong.
(2) ALWAYS first ground the input with a 1K resistor or so - open loop gain with an open input is also looking for trouble.
(3) Respectfully, I have a problem with an ECC83 as an input valve instead of an EF86. The Miller effect can be severe depending on the input impedance (affecting high frequency response/stability), which might vary from pre-amp to pre-amp or if the amp. input is fed from a volume control - check with a scope and square wave if available and you will see. If the rest of the circuit is in order, pentodes are to be preferred as they have less distortion at the low signal levels of input valves. The increase in noise is academic.
(4) Perhaps unnecessary advice, but remember that one cannot directly swap a KT66 in place of an EL34 - the latter needs a connection from pin 1 to pin 8 (separate G3). In the case of KT66/6L6GC this pin is a no-connection and may have been used for a tie-point for something else.
But I am glad that you are OK, Ed.
Johan
So perhaps just a few general comments:
(1) It can be disasterous to try and sort out output transformer feedback phase with full feedabck on! If one is not exactly certain, first omit the feedabck resistor and replace by a high value (100K, depending), keeping one end floating so that it can be momentary touched down. Without feedback there will be some hiss in the loudspeaker, and it is quite perceptable whether touching said feedback resistor momentary, will increase the hiss or decrease it. Switch the transformer connections to get a decrease in hiss, which will obviously be negative feedback. For this experiment all compensating caps. in the feedback line should also not be connected initially. Trying this with a high degree of feedback from the outset can fry your expensive output transformer because of flash-through as a result of high transients, should the phase be wrong.
(2) ALWAYS first ground the input with a 1K resistor or so - open loop gain with an open input is also looking for trouble.
(3) Respectfully, I have a problem with an ECC83 as an input valve instead of an EF86. The Miller effect can be severe depending on the input impedance (affecting high frequency response/stability), which might vary from pre-amp to pre-amp or if the amp. input is fed from a volume control - check with a scope and square wave if available and you will see. If the rest of the circuit is in order, pentodes are to be preferred as they have less distortion at the low signal levels of input valves. The increase in noise is academic.
(4) Perhaps unnecessary advice, but remember that one cannot directly swap a KT66 in place of an EL34 - the latter needs a connection from pin 1 to pin 8 (separate G3). In the case of KT66/6L6GC this pin is a no-connection and may have been used for a tie-point for something else.
But I am glad that you are OK, Ed.
Johan
Hi all
sorry for the long pause - hardly had a chance to touch my project while the kids have been on their summer holidays!
1 amp is now complete, the other is still in need of some metalworking! Hope this weekend to assemble & will post the promised photos next week.
2 new issues with the working one:
1 - because the amp was way too sensitive I have replaced the 1meg resistor at input with a 1meg pot. While this works as an attenuator, I'm sure the highs are now slightly rolled off. The pot is installed about 2 inches from the EF86, & is nothing special. Would a better quality part likely help things? Where can I get a good quality 1meg log pot?
2 - while my output t/x is a good quality one from Sowter I skimped on the power t/x & choke & used cheap ones from Hammond. They work fine, but there is a bit of mechanical hum. I suspect it is due to my mains (high, above 250v most of the time), but I was wondering about other people's experiences with Hammond transformers, and chokes in particular.
Thanks
ED.
sorry for the long pause - hardly had a chance to touch my project while the kids have been on their summer holidays!
1 amp is now complete, the other is still in need of some metalworking! Hope this weekend to assemble & will post the promised photos next week.
2 new issues with the working one:
1 - because the amp was way too sensitive I have replaced the 1meg resistor at input with a 1meg pot. While this works as an attenuator, I'm sure the highs are now slightly rolled off. The pot is installed about 2 inches from the EF86, & is nothing special. Would a better quality part likely help things? Where can I get a good quality 1meg log pot?
2 - while my output t/x is a good quality one from Sowter I skimped on the power t/x & choke & used cheap ones from Hammond. They work fine, but there is a bit of mechanical hum. I suspect it is due to my mains (high, above 250v most of the time), but I was wondering about other people's experiences with Hammond transformers, and chokes in particular.
Thanks
ED.
Hi Ed,
I apologise if I am repeating myself here, I have not re-read the entire thread. Here are some suggestions:
Because of the sensitivity problem, I wired the EF86 in triode mode, which reduced the gain somewhat, in fact, exactly to the degree I want. Details of this conversion are available in the Claus Byrinth notes. Howver, you can still use a 1M pot, ensure that the metal body of the pot is at chassis potential or you will indeed pick up hum. Your highs should not be affected by this modification.
I suspect your hum problem to be more from a possible earth-loop - are you using chassis mount aluminium caps for your filter caps? Have you earthed the centre spigot of the EF86 valve base as well as pins 2 & 7?
For the record, you should not hear any hum or hiss from the 5-20 (at least, not after modifying the EF86 to triode mode - I can't remember how it used to be before I lowered the over-all gain) when placing your ear next to the speakers with input shorted.
Also ensure that the wiring to the valve filaments is twisted and routed away from sensitive parts of the amplifier. I routed the wiring in mine along the folds of the metal chassis. The 6.3v filament windings should also be earthed at a centre-tap if there is one, otherwise two balancing resistors of around 100e each in series across the 6.3v winding with their centre-taps earthed will do the trick.
I am using Hammond O/P transformers (the 40w type) for three of my amplifiers with good results, the chokes are 10H chokes which were locally manufactured. If your voltages are constantly high, then add a series WW resistor of suitable rating between the plates of the GZ34 and transformer tappings.
I apologise if I am repeating myself here, I have not re-read the entire thread. Here are some suggestions:
Because of the sensitivity problem, I wired the EF86 in triode mode, which reduced the gain somewhat, in fact, exactly to the degree I want. Details of this conversion are available in the Claus Byrinth notes. Howver, you can still use a 1M pot, ensure that the metal body of the pot is at chassis potential or you will indeed pick up hum. Your highs should not be affected by this modification.
I suspect your hum problem to be more from a possible earth-loop - are you using chassis mount aluminium caps for your filter caps? Have you earthed the centre spigot of the EF86 valve base as well as pins 2 & 7?
For the record, you should not hear any hum or hiss from the 5-20 (at least, not after modifying the EF86 to triode mode - I can't remember how it used to be before I lowered the over-all gain) when placing your ear next to the speakers with input shorted.
Also ensure that the wiring to the valve filaments is twisted and routed away from sensitive parts of the amplifier. I routed the wiring in mine along the folds of the metal chassis. The 6.3v filament windings should also be earthed at a centre-tap if there is one, otherwise two balancing resistors of around 100e each in series across the 6.3v winding with their centre-taps earthed will do the trick.
I am using Hammond O/P transformers (the 40w type) for three of my amplifiers with good results, the chokes are 10H chokes which were locally manufactured. If your voltages are constantly high, then add a series WW resistor of suitable rating between the plates of the GZ34 and transformer tappings.
Hi Ed,
More advice, hopefully helpful. The 1 meg pot would be OK with an EF86 input tube. Thinking again of total input capacitance, Miller contribution would be 5 pF for this circuit + another about 5 pF for other input capacitances, which would give a -3 dB high frequency point at 20 KHz with a series resistance of about 800K. The highest series resistance (when the 1 meg pot is in the centre, resistance-wise) would be 250K, so that is OK.
With the EF86 in triode mode as Family Dog is using it (and adjusting the cathode bias resistor to give the same voltage at the ECC83 cathodes), the total input capacitance is about 30 pF - I do not have exact figures for EF86 in triode mode - maximum input resistance for 20 KHz cut-off will be 265K; still just in order.
As a comparison the total equivalent input capacitance with an ECC83 borders on the 100 pF, which reduces allowable input resistance to only 80K - this is why I do not like that tube as an input tube, unless a low input impedance can be guaranteed. Incidentally, the variation of phase angle with changing input impedance would also compromise feedback stability. (One can use both triodes of an ECC83 in parallel, still. . . .)
But lastly, why does one want a pot as high as 1 meg? This is a general value to prevent loading any pre-amplifier. I presume you know what you will be driving your amplifier with - to my knowledge most pre-amplifiers will be happy with a main amplifier input impedance of 100K or even lower. A pot of that value would thus serve you far better.
A more practical execution of this kind of circuit to my mind is to use a lower mu triode as phase inverter (ECC82) with the necessary changes in component values and stick to a pentode input because of its relative independence of input conditions - but not to confuse you with too many alterations; that is a circuit for another time. The bottom line of this comment is: Lower the value of your input pot. to something 5 - 10 times the output impedance of your pre-amplifier - it can only be an advantage.
Regards.
More advice, hopefully helpful. The 1 meg pot would be OK with an EF86 input tube. Thinking again of total input capacitance, Miller contribution would be 5 pF for this circuit + another about 5 pF for other input capacitances, which would give a -3 dB high frequency point at 20 KHz with a series resistance of about 800K. The highest series resistance (when the 1 meg pot is in the centre, resistance-wise) would be 250K, so that is OK.
With the EF86 in triode mode as Family Dog is using it (and adjusting the cathode bias resistor to give the same voltage at the ECC83 cathodes), the total input capacitance is about 30 pF - I do not have exact figures for EF86 in triode mode - maximum input resistance for 20 KHz cut-off will be 265K; still just in order.
As a comparison the total equivalent input capacitance with an ECC83 borders on the 100 pF, which reduces allowable input resistance to only 80K - this is why I do not like that tube as an input tube, unless a low input impedance can be guaranteed. Incidentally, the variation of phase angle with changing input impedance would also compromise feedback stability. (One can use both triodes of an ECC83 in parallel, still. . . .)
But lastly, why does one want a pot as high as 1 meg? This is a general value to prevent loading any pre-amplifier. I presume you know what you will be driving your amplifier with - to my knowledge most pre-amplifiers will be happy with a main amplifier input impedance of 100K or even lower. A pot of that value would thus serve you far better.
A more practical execution of this kind of circuit to my mind is to use a lower mu triode as phase inverter (ECC82) with the necessary changes in component values and stick to a pentode input because of its relative independence of input conditions - but not to confuse you with too many alterations; that is a circuit for another time. The bottom line of this comment is: Lower the value of your input pot. to something 5 - 10 times the output impedance of your pre-amplifier - it can only be an advantage.
Regards.
Thanks for those comments. This weekend I've managed to get both channels up & running, some internal tidying still to do on one side.
I've read Claus Byrinth's notes & I'm going to rewire the EF86 as a triode, when time allows! Not heard anyone mention replacing the phase splitter with a tube of lower gain before, might investigate that as well. Also I'll try a nice Panasonic 250k pot as I have one in my parts bin.
About the hum - sorry I wasn't very clear. There is no hum or noise from the speakers at all. The hum is vibration from the mains tranny & choke. I can live with it to be honest, it's not loud enough to be heard from the listening position. It's just THERE every time I go to the record rack, & I wonder if it would still be there had I bought a more costly unit. I suspect it would to be honest. Never had any trouble with small Hammond t/x's in preamps.
All in all I'm really pleased with these amps - sounding very sweet now they've been playing for a few hours.
Thanks again for all the help & interest shown.
Ed
I've read Claus Byrinth's notes & I'm going to rewire the EF86 as a triode, when time allows! Not heard anyone mention replacing the phase splitter with a tube of lower gain before, might investigate that as well. Also I'll try a nice Panasonic 250k pot as I have one in my parts bin.
About the hum - sorry I wasn't very clear. There is no hum or noise from the speakers at all. The hum is vibration from the mains tranny & choke. I can live with it to be honest, it's not loud enough to be heard from the listening position. It's just THERE every time I go to the record rack, & I wonder if it would still be there had I bought a more costly unit. I suspect it would to be honest. Never had any trouble with small Hammond t/x's in preamps.
All in all I'm really pleased with these amps - sounding very sweet now they've been playing for a few hours.
Thanks again for all the help & interest shown.
Ed
Hi Ed,
Hammond (or any other) Transformers shouldn't really hum - This might be caused by either loose laminations (unlikely, but check them anyway) or the transformer is incorrectly rated for the task in hand and is being "overdriven". The transformer will get hot after an hour or so in use, but can you place your hand on it without burning your hand?
Is the biasing of your two EL34 output valves in order? Should be in the region of 30-32V measured between earth and cathode.
Another thought - how good is the rectifier you're using?
Johan will no doubt come up with some more meaningful suggestions, you should definitely not be plagued with transformer hum.
-Eric
Hammond (or any other) Transformers shouldn't really hum - This might be caused by either loose laminations (unlikely, but check them anyway) or the transformer is incorrectly rated for the task in hand and is being "overdriven". The transformer will get hot after an hour or so in use, but can you place your hand on it without burning your hand?
Is the biasing of your two EL34 output valves in order? Should be in the region of 30-32V measured between earth and cathode.
Another thought - how good is the rectifier you're using?
Johan will no doubt come up with some more meaningful suggestions, you should definitely not be plagued with transformer hum.
-Eric
Hi Eric,
The transformer is rated generously at 200mA on the HT, 3a on the rectifier heater & 6a on the tube heaters. Voltage on the cathodes is spot on.
After 3 hours the tranny is hot but I can keep my hand on it without difficulty.
The rectifiers are chinese 5AR4's, but better ones (golden dragons) than the chinese GZ34's that I tried first. One died during testing, the other started making noise within a few hours.
I'll check those laminations are tight. Maybe I can squeeze them up a bit. Or mount them on rubber washers!
By the way the sound out of the amps has improved dramatically with use, now I have about 15 hours' listening from them. They sound effortless, not what I was expecting from 30w/channel! Also they don't sound old fashioned, for a 1950's design!
Ed
The transformer is rated generously at 200mA on the HT, 3a on the rectifier heater & 6a on the tube heaters. Voltage on the cathodes is spot on.
After 3 hours the tranny is hot but I can keep my hand on it without difficulty.
The rectifiers are chinese 5AR4's, but better ones (golden dragons) than the chinese GZ34's that I tried first. One died during testing, the other started making noise within a few hours.
I'll check those laminations are tight. Maybe I can squeeze them up a bit. Or mount them on rubber washers!
By the way the sound out of the amps has improved dramatically with use, now I have about 15 hours' listening from them. They sound effortless, not what I was expecting from 30w/channel! Also they don't sound old fashioned, for a 1950's design!
Ed
Finally, two channels installed! Here's some pics:
An externally hosted image should be here but it was not working when we last tested it.
Ed,
They look great! Here is a pic of mine:
http://www.imagestation.com/album/pictures.html?id=2120900008&idx=1
The photo shows a Denon AVR2803 which is used to control the amplifiers, with a Marantz DVD Player. I have since replaced the Marantz with a Denon 2200 DVD player.
Will have to find a pic of the under-chassis view; there is one on this forum somewhere.
-Eric
They look great! Here is a pic of mine:
http://www.imagestation.com/album/pictures.html?id=2120900008&idx=1
The photo shows a Denon AVR2803 which is used to control the amplifiers, with a Marantz DVD Player. I have since replaced the Marantz with a Denon 2200 DVD player.
Will have to find a pic of the under-chassis view; there is one on this forum somewhere.
-Eric
Hi Eric, thanks. Our amps look remarkably similar! Building 5 is no small feat - I've had enough hole-cutting in steel to last me a lifetime!
Mine were intended to be Beam Echo DL7-35 replicas (as the boards were made by late 1990's re-incarnation of the Beam Echo company). There was going to be an aluminium cover over the transformers, unlikely it'll ever get done now they're installed & running.
My system is analogue only - the Mullards are fed by Hagerman Clarinet linestage, Cornet phono pre & Linn LP12. I've been playing records for nearly 30 years now & with these 1950s designed amps they sound better than ever.
The mechanical hum problem is fixed (well, err.. hidden at least!). Fitted rubber grommits to the transformer mounting holes & placed a damper under the power t/x made from a mouse mat! As long as it doesn't melt or anything, it's great.
Thanks for all the interest & help you've given - I may give you a shout here when I alter the input stage to triode!
Regards
Ed
Mine were intended to be Beam Echo DL7-35 replicas (as the boards were made by late 1990's re-incarnation of the Beam Echo company). There was going to be an aluminium cover over the transformers, unlikely it'll ever get done now they're installed & running.
My system is analogue only - the Mullards are fed by Hagerman Clarinet linestage, Cornet phono pre & Linn LP12. I've been playing records for nearly 30 years now & with these 1950s designed amps they sound better than ever.
The mechanical hum problem is fixed (well, err.. hidden at least!). Fitted rubber grommits to the transformer mounting holes & placed a damper under the power t/x made from a mouse mat! As long as it doesn't melt or anything, it's great.
Thanks for all the interest & help you've given - I may give you a shout here when I alter the input stage to triode!
Regards
Ed
- Status
- Not open for further replies.