I have been playing a lot with this design and i have to say that i was wrong suggesting to use the TIP-142. The MJL4281A sounds better and can be direct-driven by many opamps as long as you check for a beta over 100 and set a low idle current.
The direct-driven version is the best-sounding but the power is so constrained that i have developed a version that uses two parallel output devices and a driver.
The distortion is low and the circuit looks stable so far.
I post the schematic with some dissipation numbers to give a guideline to those who might want to build one.
I'm working on a board but it's my first serious project so i don't expect to have it finished anytime soon.
EDIT: I can't suggest them as i havn't already tryed it, but i plan to use the FJL4315 as output transistor because it's low output capacitance will make everything better for sure.
EDIT2: My current prototype is stable without zobel network and is dead quiet on 90dB/6Ohms loudspeakers even if powered from an smps.
The direct-driven version is the best-sounding but the power is so constrained that i have developed a version that uses two parallel output devices and a driver.
The distortion is low and the circuit looks stable so far.
I post the schematic with some dissipation numbers to give a guideline to those who might want to build one.
I'm working on a board but it's my first serious project so i don't expect to have it finished anytime soon.
EDIT: I can't suggest them as i havn't already tryed it, but i plan to use the FJL4315 as output transistor because it's low output capacitance will make everything better for sure.
EDIT2: My current prototype is stable without zobel network and is dead quiet on 90dB/6Ohms loudspeakers even if powered from an smps.
Hi there Mr Ion,
Now the parts are exclusive/rare here, n seems this a notfornoob ss
peace
Is your smps just OK?
i got my switching trafo ( +/- 18VDC 10/A ) droping to +/-11vdc in use. i put some more caps ( 4 x 3300uF ) then got the fuses ruptured.n always blow the fuses even after back remove the caps.
Is it silly to change the opam with pedja's buffer for the EF?
Now the parts are exclusive/rare here, n seems this a notfornoob ss
peace Is your smps just OK?
i got my switching trafo ( +/- 18VDC 10/A ) droping to +/-11vdc in use. i put some more caps ( 4 x 3300uF ) then got the fuses ruptured.n always blow the fuses even after back remove the caps.
Is it silly to change the opam with pedja's buffer for the EF?
eketehe said:Hi there Mr Ion,
Now the parts are exclusive/rare here, n seems this a notfornoob ss
Is your smps just OK?
i got my switching trafo ( +/- 18VDC 10/A ) droping to +/-11vdc in use. i put some more caps ( 4 x 3300uF ) then got the fuses ruptured.n always blow the fuses even after back remove the caps.
Is it silly to change the opam with pedja's buffer for the EF?
The project is relatively simple, i suggest the parts that i have tried or that i belive that would do the best, but this doesn't mean that a BD139 won't do as a driver and even a 2N3055 could be used at the output.
You can't replace the opamp with a buffer for three reasons:
( 1 ) the opamp has gain, while the buffer has not. This means that you will have to drive the amp with an already full-amplitude signal. If you want to do that then you can use a tube as gain stage.
( 2 ) The opamp keeps the offset low . If you remove it you will have at least 1.3 V dc at the output, probably more. Then you will have to use a single supply and a big capacitor at the output.
( 3 ) The opamp keeps the output impedance low. I belive this is the least important problem as the output impedance of the EF is probably on the 0.1 ohms range already (giving a damping factor of 80 or more), but it might change the sound in some loudspeakers [perhaps to better?].
Changing from 120 dB of feedback to no feedback will change the sound for sure.
My smps work fine. I use a couple of 18V 4A ones and they do just fine. If you get your fuses blown then there is either a short, a broken transistor or the fuse is underrrated. Smps usually have its current limited by the switching logic. This means that they give the current they are rated for without failing when overloaded. Of course there are also poor-designed smps that explode when overloaded.
There aren't "not for noobs" projects. The only thing you should never forget is to do things safely and to double check everything before connecting it to expensive loudspeakers. Class A projects can fail (and burn your loudspeakers) due to the big dissipation involved, that's the only thing that makes them more difficult to class B designs).
If you have suceeded with the original tip's then you might also want to try the JLH amp, which can be made with few readily-available parts and has a nice sound.
I have just rebuilt one of these with some modifications:
I have changed the LM4562 for a LM49860 and cranked up the voltage to 21V.
The free-running amp of the LME49860 has been wired as a 10x gain amp with its input tied to ground to be sure it does not interfere with the working one.
The power transistor has been changed from a MJL4281A to a FJL4315 which is cheaper and faster, with lower gate capacitance. The amp uses three output transistors biased at a total current of 3.5A, choosen so to give 30W on 6Ohms monitor audio loudspeakers, these have 90 dB/W sensitivity and go much louder than i listen to so i've made one of the current sinks swichable.
The driver has been changed from a faster MJE15034 to 2SC3503. I know it's having a hard time.
The opamp is now powered from a different power supply.
Parts are of the highest quality one can buy in a conventional online store (Digikey), no rare parts from audio-specific boutique vendors. The increase in parts cost is completely masked by the outrageous price of transformers, capacitors and heatsinks.
The amp uses a zobel network. An output inductor also helps getting cleaner scope images (I belive my scope injects garbage into the circuits because they suddenly become noisy when i touch them with the probes, even at 10x magnification.
The amp is plain beautiful. Noiseless, oscillationless. It's not a tube amp but it's not as cruel to bad records as other opamp-based amps are. For these occasions it's never bad to have a tube preamp, but i prefer to switch it off when i listen to good records.
I attach a picture of it.
I have changed the LM4562 for a LM49860 and cranked up the voltage to 21V.
The free-running amp of the LME49860 has been wired as a 10x gain amp with its input tied to ground to be sure it does not interfere with the working one.
The power transistor has been changed from a MJL4281A to a FJL4315 which is cheaper and faster, with lower gate capacitance. The amp uses three output transistors biased at a total current of 3.5A, choosen so to give 30W on 6Ohms monitor audio loudspeakers, these have 90 dB/W sensitivity and go much louder than i listen to so i've made one of the current sinks swichable.
The driver has been changed from a faster MJE15034 to 2SC3503. I know it's having a hard time.
The opamp is now powered from a different power supply.
Parts are of the highest quality one can buy in a conventional online store (Digikey), no rare parts from audio-specific boutique vendors. The increase in parts cost is completely masked by the outrageous price of transformers, capacitors and heatsinks.
The amp uses a zobel network. An output inductor also helps getting cleaner scope images (I belive my scope injects garbage into the circuits because they suddenly become noisy when i touch them with the probes, even at 10x magnification.
The amp is plain beautiful. Noiseless, oscillationless. It's not a tube amp but it's not as cruel to bad records as other opamp-based amps are. For these occasions it's never bad to have a tube preamp, but i prefer to switch it off when i listen to good records.
I attach a picture of it.
I have just rebuilt one of these with some modifications:
I have changed the LM4562 for a LM49860 and cranked up the voltage to 21V.
The free-running amp of the LME49860 has been wired as a 10x gain amp with its input tied to ground to be sure it does not interfere with the working one.
The power transistor has been changed from a MJL4281A to a FJL4315 which is cheaper and faster, with lower gate capacitance. The amp uses three output transistors biased at a total current of 3.5A, choosen so to give 30W on 6Ohms monitor audio loudspeakers, these have 90 dB/W sensitivity and go much louder than i listen to so i've made one of the current sinks swichable.
The driver has been changed from a faster MJE15034 to 2SC3503. I know it's having a hard time.
The opamp is now powered from a different power supply.
Parts are of the highest quality one can buy in a conventional online store (Digikey), no rare parts from audio-specific boutique vendors. The increase in parts cost is completely masked by the outrageous price of transformers, capacitors and heatsinks.
The amp uses a zobel network. An output inductor also helps getting cleaner scope images (I belive my scope injects garbage into the circuits because they suddenly become noisy when i touch them with the probes, even at 10x magnification.
The amp is plain beautiful. Noiseless, oscillationless. It's not a tube amp but it's not as cruel to bad records as other opamp-based amps are. For these occasions it's never bad to have a tube preamp, but i prefer to switch it off when i listen to good records.
I attach a picture of it.
I have changed the LM4562 for a LM49860 and cranked up the voltage to 21V.
The free-running amp of the LME49860 has been wired as a 10x gain amp with its input tied to ground to be sure it does not interfere with the working one.
The power transistor has been changed from a MJL4281A to a FJL4315 which is cheaper and faster, with lower gate capacitance. The amp uses three output transistors biased at a total current of 3.5A, choosen so to give 30W on 6Ohms monitor audio loudspeakers, these have 90 dB/W sensitivity and go much louder than i listen to so i've made one of the current sinks swichable.
The driver has been changed from a faster MJE15034 to 2SC3503. I know it's having a hard time.
The opamp is now powered from a different power supply.
Parts are of the highest quality one can buy in a conventional online store (Digikey), no rare parts from audio-specific boutique vendors. The increase in parts cost is completely masked by the outrageous price of transformers, capacitors and heatsinks.
The amp uses a zobel network. An output inductor also helps getting cleaner scope images (I belive my scope injects garbage into the circuits because they suddenly become noisy when i touch them with the probes, even at 10x magnification.
The amp is plain beautiful. Noiseless, oscillationless. It's not a tube amp but it's not as cruel to bad records as other opamp-based amps are. For these occasions it's never bad to have a tube preamp, but i prefer to switch it off when i listen to good records.
I attach a picture of it.
Attachments
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.