Hi Guys
8 years ago I started to build my Audio system... Bought most of the parts and untill now didnt have the time to complete the task.
Now I am finaly going to finish a pair of Lynn Olsons Ariels - the cabinets are ready for tuneup as they are waiting ready for 8 years. Imagine such complex cabinets, amplifier parts, woofers and twiters sitting together in a dark place for years, just waiting for someone to give them life.
I also have all the following ingredients for a mullard 5-20
4 EL34 and sockets,
2 EZ34 (not sure) rectifier tubes
2 Hammond 1650P output transformers
2 Hammond BX300 transformers
2 boxes
2 Chockes
etc...
I dont have the resistors and capacitors
Now for my 40th birthday I ordered the Bottlehead Foreplay III and I am going to complete the task.
I would like to know how to choose the right Mullard 5-20 design to fit this system:
Foreplay III -> Mullard 5-20 -> Ariels
Which design and which modifications are needed.
Please provide as much details as you can - I promiss to finish it this time!
Thanks!
8 years ago I started to build my Audio system... Bought most of the parts and untill now didnt have the time to complete the task.
Now I am finaly going to finish a pair of Lynn Olsons Ariels - the cabinets are ready for tuneup as they are waiting ready for 8 years. Imagine such complex cabinets, amplifier parts, woofers and twiters sitting together in a dark place for years, just waiting for someone to give them life.
I also have all the following ingredients for a mullard 5-20
4 EL34 and sockets,
2 EZ34 (not sure) rectifier tubes
2 Hammond 1650P output transformers
2 Hammond BX300 transformers
2 boxes
2 Chockes
etc...
I dont have the resistors and capacitors
Now for my 40th birthday I ordered the Bottlehead Foreplay III and I am going to complete the task.
I would like to know how to choose the right Mullard 5-20 design to fit this system:
Foreplay III -> Mullard 5-20 -> Ariels
Which design and which modifications are needed.
Please provide as much details as you can - I promiss to finish it this time!
Thanks!
The rectifier is GZ34/5AR4.
A FP has a fair amount of gain. You end up with excessive gain using Mullard's small signal complement of an EF86 and a 12AX7.
A 6AB4 voltage amplifier and an ECC99 LTP should do nicely. Both are high gm types that provide resistance against slew limiting. Mullard topology has a global NFB loop and slew limiting is a matter of considerable concern.
The triode in the 6AB4 is the same as that in the 12AT7/ECC81. That triode is very well suited to service in PP amps. The NET harmonic distortion distribution is 2nd > 3rd > 4th > 5th. That "waterfall" is PLEASING to the ear.
You could entertain the notion of replacing the LTP cathode to ground resistor with a constant current sink. CCS loading LTP tails forces symmetry and ensures good balance between the 2 out of phase signals.
A FP has a fair amount of gain. You end up with excessive gain using Mullard's small signal complement of an EF86 and a 12AX7.
A 6AB4 voltage amplifier and an ECC99 LTP should do nicely. Both are high gm types that provide resistance against slew limiting. Mullard topology has a global NFB loop and slew limiting is a matter of considerable concern.
The triode in the 6AB4 is the same as that in the 12AT7/ECC81. That triode is very well suited to service in PP amps. The NET harmonic distortion distribution is 2nd > 3rd > 4th > 5th. That "waterfall" is PLEASING to the ear.
You could entertain the notion of replacing the LTP cathode to ground resistor with a constant current sink. CCS loading LTP tails forces symmetry and ensures good balance between the 2 out of phase signals.
Should I stick to the original design?
Hi Eli,
I understand that the 5-20 had many clones that were developed to overcome the 5-20 flaws.
I know that I dont have enough information at the moment and I want to focus on the right path.
Should I stick with the original design (from the mullard book) and make the needed changes?
Or choose anothe flavour of the 5-20 design?
Should I choose triod mode with KT88?
Regarding the 5-20:
A. So there is going to be excessive gain and to solve it:
1. replace the voltage amplifier ef86 with 6AB4
2. replace the pahse splitter with ECC99
- correct?
B. Do you think that my transformers are suitable for the design?
BTW, What does LTP stand for?
Thanks
Hi Eli,
I understand that the 5-20 had many clones that were developed to overcome the 5-20 flaws.
I know that I dont have enough information at the moment and I want to focus on the right path.
Should I stick with the original design (from the mullard book) and make the needed changes?
Or choose anothe flavour of the 5-20 design?
Should I choose triod mode with KT88?
Regarding the 5-20:
A. So there is going to be excessive gain and to solve it:
1. replace the voltage amplifier ef86 with 6AB4
2. replace the pahse splitter with ECC99
- correct?
B. Do you think that my transformers are suitable for the design?
BTW, What does LTP stand for?
Thanks
Re: Should I stick to the original design?
LTP = long tailed pair.
Both the EF86 and 12AX7/ECC81 Mullard called for yield considerable gain. That was OK, when less than 1 VRMS of drive from a ceramic phono cartridge or a FM tuner was the signal source. You will be dealing with the 2 VRMS max. O/P of a CDP that's been amplified by the FP line stage. In addition, both small signal types Mullard called for exhibit low gm and expose the builder to slew limiting. Two important demons in amps with global NFB loops are O/P trafo core saturation and slew limiting. You attack core saturation by rolling infrasonic noise off at the amp's I/P. Slew limiting is attacked by employing high gm small signal types that can swing the HF error correction signal properly.
The Hammond 1650P O/P trafo is quite suitable for Mullard's cathode biased EL34 O/P stage. You could switch to "fixed" bias 6L6GCs for higher O/P power. The low RP of the ECC99 allows it to work into the much lower valued O/P tube grid leak resistor "fixed" bias requires, without incident. Remember, the 'X7 is a WIMP.
The 300BX power trafo is quite hefty. Start with only 2.2 μF. in the 1st position of the CLC filter. Load the PSU down and increase the amount of capacitance in the 1st position until the B+ rail voltage is where it needs to be. Under no circumstances should you use filter caps. rated for less than 500 WVDC and 550 WVDC is probably a GOOD idea. 400 VRMS is approx. 565 V. peak.
drors01 said:
LTP = long tailed pair.
Both the EF86 and 12AX7/ECC81 Mullard called for yield considerable gain. That was OK, when less than 1 VRMS of drive from a ceramic phono cartridge or a FM tuner was the signal source. You will be dealing with the 2 VRMS max. O/P of a CDP that's been amplified by the FP line stage. In addition, both small signal types Mullard called for exhibit low gm and expose the builder to slew limiting. Two important demons in amps with global NFB loops are O/P trafo core saturation and slew limiting. You attack core saturation by rolling infrasonic noise off at the amp's I/P. Slew limiting is attacked by employing high gm small signal types that can swing the HF error correction signal properly.
The Hammond 1650P O/P trafo is quite suitable for Mullard's cathode biased EL34 O/P stage. You could switch to "fixed" bias 6L6GCs for higher O/P power. The low RP of the ECC99 allows it to work into the much lower valued O/P tube grid leak resistor "fixed" bias requires, without incident. Remember, the 'X7 is a WIMP.
The 300BX power trafo is quite hefty. Start with only 2.2 μF. in the 1st position of the CLC filter. Load the PSU down and increase the amount of capacitance in the 1st position until the B+ rail voltage is where it needs to be. Under no circumstances should you use filter caps. rated for less than 500 WVDC and 550 WVDC is probably a GOOD idea. 400 VRMS is approx. 565 V. peak.
I am not sure where I read about reducing the negative feedback.
Is that recommended?
What is the goal for reducing it? Power? Dynamics?
Wehre do you recommend I order the small parts?
- Any reommendation of brands of resistors? capacitors?
What does the following stand for:
gm = ?
slew limiting = amp's speed ?
'X7 = ?
Dror
Is that recommended?
What is the goal for reducing it? Power? Dynamics?
Wehre do you recommend I order the small parts?
- Any reommendation of brands of resistors? capacitors?
What does the following stand for:
gm = ?
slew limiting = amp's speed ?
'X7 = ?
Dror
hey-Hey!!!,
With the ~20:1 voltage gain of te FPIII, you can use a two stage amp. Go right to Eli's El-Cheap-O and scale it up a bit.
An ECC99 rigged as a LTP will provide plenty of gain given the FPIII. Rig for fixed bias and run a CCS as the LTP's 'tail' for good balance. Use lower than usual B+ so you can increase idle current to improve the loading.
cheers,
Douglas
http://www.aikenamps.com/LongTailPairDesign.htm
With the ~20:1 voltage gain of te FPIII, you can use a two stage amp. Go right to Eli's El-Cheap-O and scale it up a bit.
An ECC99 rigged as a LTP will provide plenty of gain given the FPIII. Rig for fixed bias and run a CCS as the LTP's 'tail' for good balance. Use lower than usual B+ so you can increase idle current to improve the loading.
cheers,
Douglas
http://www.aikenamps.com/LongTailPairDesign.htm
Dror.
'X7 = 12AX7. It's common to shorten 12A_7 tube types to '_7. You will also see things like 'SN7, where somebody is talking about either a 6SN7 or a 12SN7. Except for the heater voltage, the 2 'SN7s are identical.
The mutual transconductance of tubes is what gm stands for.
The slewing capacity of circuitry is given in V./μsec. Yes, it is an indication of "speed". High gm is conducive to slew limiting resistance.
The remark Doug (Bandersnatch) made about FP3 having a gain of 20X indicates the use of CCS loaded 'U7 voltage amplifiers. That comparatively large gain may be good for use with Bottlehead's DHT based power amps, but it's a PITA in combination with indirectly heated PP "finals". I know of no modest mu/high gm small signal type whose HD spectrum is skewed towards 2nd order, like the high mu 'T7. So, obtaining a net "waterfall" HD spectrum does not seem feasible.
Doug's suggestion of "El Cheapo" topology with an ECC99 LTP makes sense, from a gain structure perspective. A NOS 5687 (especially TungSol black plate) might work out even better. A Soviet surplus 68 nF. PIO cap. in combination with a 150 KOhm grid leak resistor at the I/P grid will yield both nice tone and a near perfect high pass "corner". The negative rail "El Cheapo" topology requires can be derived from the low current 5 VAC winding of the 300BX by using a multi-stage multiplier, along the lines of the 2nd schematic here . Use Schottky diodes in all the multiplier stages, except the last. Use a damper diode in the last stage for a SLOW start. B- should rise after B+ and the 5AR4 B+ rectifier is pretty darned slow.
Should you choose to use "fixed" (instead of cathode) bias, tap into the voltage multiplier before the last stage. An immediate start in a bias supply is desired.
'X7 = 12AX7. It's common to shorten 12A_7 tube types to '_7. You will also see things like 'SN7, where somebody is talking about either a 6SN7 or a 12SN7. Except for the heater voltage, the 2 'SN7s are identical.
The mutual transconductance of tubes is what gm stands for.
The slewing capacity of circuitry is given in V./μsec. Yes, it is an indication of "speed". High gm is conducive to slew limiting resistance.
The remark Doug (Bandersnatch) made about FP3 having a gain of 20X indicates the use of CCS loaded 'U7 voltage amplifiers. That comparatively large gain may be good for use with Bottlehead's DHT based power amps, but it's a PITA in combination with indirectly heated PP "finals". I know of no modest mu/high gm small signal type whose HD spectrum is skewed towards 2nd order, like the high mu 'T7. So, obtaining a net "waterfall" HD spectrum does not seem feasible.
Doug's suggestion of "El Cheapo" topology with an ECC99 LTP makes sense, from a gain structure perspective. A NOS 5687 (especially TungSol black plate) might work out even better. A Soviet surplus 68 nF. PIO cap. in combination with a 150 KOhm grid leak resistor at the I/P grid will yield both nice tone and a near perfect high pass "corner". The negative rail "El Cheapo" topology requires can be derived from the low current 5 VAC winding of the 300BX by using a multi-stage multiplier, along the lines of the 2nd schematic here . Use Schottky diodes in all the multiplier stages, except the last. Use a damper diode in the last stage for a SLOW start. B- should rise after B+ and the 5AR4 B+ rectifier is pretty darned slow.
Should you choose to use "fixed" (instead of cathode) bias, tap into the voltage multiplier before the last stage. An immediate start in a bias supply is desired.
Well, hhhm... well
I think I will stick with the original amplifier while making the changes Eli suggested to reduce the gain.
It is almost imposible for me to track what you guys are proposing and it is sure out of my current capabilities to go into desgin process of connecting the proposals to a safe
and working anplifier.
Is there a good source for the resistrs and capacitors?
thanks for the support!
I think I will stick with the original amplifier while making the changes Eli suggested to reduce the gain.
It is almost imposible for me to track what you guys are proposing and it is sure out of my current capabilities to go into desgin process of connecting the proposals to a safe
and working anplifier.Is there a good source for the resistrs and capacitors?
thanks for the support!
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.