Can you explain a little more. I have no feeling how it will sound. I can only say I can make it work electrically with confidence, but it might sounds like a dog!!!!
Theory and engineering can only go so far, experience and "the touch" is what make the sound. That's the part I don't have
Last edited:
i doubt your "theory and engineering"........
but carry on and see how it goes...😉
experience is still the best teacher....
but carry on and see how it goes...😉
experience is still the best teacher....
The more complex any system, the more unfamiliar features the designer tries to incorporate, the more problematic the build, the more difficult the debug. These problems can compound each other so that a few small added complexities can add up to very big problems.
A less ambitious approach would be more incremental. This allows you to gain experience and insight into the problems involved while they are still bite-sized and not interacting.
You are the best judge of your own capabilities, however. Good luck.
A less ambitious approach would be more incremental. This allows you to gain experience and insight into the problems involved while they are still bite-sized and not interacting.
You are the best judge of your own capabilities, however. Good luck.
Thanks for you advice. I don't want to say I will not run into problems, but I did work in various fields for a long time. I think I have no issue in getting it working, fixing oscillation, crosstalk and the common engineering issue. I just don't know how it will sound and any experience that tells me whether it's going to sound good even if it works perfectly. I really need advice from you guys from experience what make the amp sounds good.
What I drawn is nothing new. CCS tail for the input LTP is strongly adviced by Morgan Jones. Everyone here said using follower to drive the power tubes. I follow the differential power tubes from Vacuum State. Then I use UL. To me, this is really simple and straight forward. I don't even see any aggressiveness in this design.
Thanks
Last edited:
What's the amp's output impedance? This may not be suitable for many speakers.
Last edited:
That's a good question. Without NFB, the output impedance is going to be the plate output impedance of two tubes in parallel. Assuming OPT is 4.5K:4 ohm. Then the output impedance to the speaker is the output impedance of the two UL power tubes in parallel times 4/4500.
That's one concern I have that the DF is low, but people keep saying don't do global NFB.
Still, a low output impedance amplifier is best for most speakers, which typically have a quite variable input impedance.
Check out some tube amplifier tests on Stereophile's web site, which use a simulated speaker load.
The frequency response at the speaker terminals as a result of the voltage divider action with the amp's output impedance
can result in significant frequency response errors, which tend to follow the speaker's impedance curve.
This is just a rough design to get opinion. If the output impedance is too high, I can change the 6SN7 to 12AX7. Then remove the emitter followers and put cap across the two plates of the LTP to create a dominant pole. Then create a NFB from output of the OPT to the LTP. This will lower the output impedance.
The reason I draw the schematic is because people have been telling me that NFB is not good. A lot of the SE amp has no NFB and their output impedance is just as bad as my drawing.
Thanks
Hi Mrcurwin.
I am glad you join in. I just read from Jones that the important thing is to have CCS tail to balance the symmetry of the LTP. The 2nd harmonics created by the plate load resistor is going to be cancelled in the push pull action of the OTP.
Is there any advantage of doing CCS top or even mu-follower for the LTP? I am open to advice and changes. This is just a first draft to let people comment.
Thanks
Yes, but many using no GNF to lower output impedance use atypical loudspeakers. What speakers do you plan to use with it?
What do you mean by using no global NFB to lower output impedance?
I have a pair of JM Lab 913 like this oneFS : Jmlab Focal Spectral 913.1 floor standing with 2X8" woofer, 5" tweeter and a reverse dome tweeter. That's the only pair of speaker I have. Whatever I design will have to work with this pair.
Also to NFB, I have no idea whether it is good or bad. I just read a lot of post here. I read a lot of comments suggesting not to use global NFB. Maybe this is wrong, I don't know. I just did this draft design based on my understanding of what people said here. A lot of the SE design has no GNFB, don't they see the same problem as you described? Then how come people keep doing that?
Thanks
Last edited:
😀 you see, this is the point, if you are who you say you are, shouldn't you be telling us instead of you asking us? 😉
Some speakers don't need a low source impedance and it constrains some designs to give it undue concern. Horns for one example. A damping factor even an order of magnitude below unity won't bother some horns, provided the impedance has been conjugated or used wisely.
Last edited:
I am stuck with this pair of speaker FS : Jmlab Focal Spectral 913.1 So it has to work with this pair.
Last edited:
One series RLC (shunting the speaker) per crossover works in many cases.
The woofer resonance might be OK, as the amplifier Zout compensates for the rolloff but another RLC can be used for tuning. Try the speaker on different output transformer taps to see whether this would help.
Top end tuning/compensation can be done with a series RC.
The woofer resonance might be OK, as the amplifier Zout compensates for the rolloff but another RLC can be used for tuning. Try the speaker on different output transformer taps to see whether this would help.
Top end tuning/compensation can be done with a series RC.
Can you elaborate a little. I am not familiar with what you suggest at all. Do you mean you actually put a series or parallel conjugate impedance with the speaker to neutralize the reactance? Like if the speaker impedance is R+jX, you put -jX in series so the resultant impedance of the two in series is R?
I thought that would change the sound!!! Do you have any article talking about speaker reactance compensation that you can give me the link? This is totally new to me.
Thanks
- Status
- Not open for further replies.