In general any modern design will be quite noiseless by itself and definitely reject supply ripple.
Like indianajo says, your success depends way more in "mechanicals": proper layout, grounding, I´d add proper shielding and of course starting with a good PSU to begin with, it´s not any more a "circuit" problem.
Meaning you have a lot of choices around, no need to search for "magic/exceptional" designs.
Thanks God we have advanced a lot, and modern "normal" amps are very very good.
Like indianajo says, your success depends way more in "mechanicals": proper layout, grounding, I´d add proper shielding and of course starting with a good PSU to begin with, it´s not any more a "circuit" problem.
Meaning you have a lot of choices around, no need to search for "magic/exceptional" designs.
Thanks God we have advanced a lot, and modern "normal" amps are very very good.
Power supply noise isn’t the only source of noise in an amplifier. It can be a major contributor, and you do need to take care of it or you're dead in the water.
Beyond that, there is the “hisssss” produced by thermal noise sources, mostly in the input stage. That can be almost non existent, or at least 100 dB down with modern components. A “low noise” input pair might get you -112 dB, and jellybean 2N3904’s might do -100. Neither is audible, but some older Ge type might sit there and hiss at -70dB.
High-order IM (due to crossover artifacts) is heard more as noise than distortion, because spectrally it’s far away from the signals which produce it. My guess is that’s what’s so “quiet” about the Circlophone and many class A amplifiers.
Beyond that, there is the “hisssss” produced by thermal noise sources, mostly in the input stage. That can be almost non existent, or at least 100 dB down with modern components. A “low noise” input pair might get you -112 dB, and jellybean 2N3904’s might do -100. Neither is audible, but some older Ge type might sit there and hiss at -70dB.
High-order IM (due to crossover artifacts) is heard more as noise than distortion, because spectrally it’s far away from the signals which produce it. My guess is that’s what’s so “quiet” about the Circlophone and many class A amplifiers.
Thanks, yes all good points. I agree the design itself is only one small aspect of noise, so it is just my starting point. There will no doubt be more questions on the mechanicals when the time comes, though I have learnt quite a lot about it over the years so hopefully not too many.
But it is still a factor; I have in the past built designs that were in themselves mediocre for noise (as I later realised with more experience), so become a limiting factor after getting everything else right. I guess old or average designs are popular and persist for a number of reasons; we all have different interests and objectives.
This latest application is somewhat more sensitive towards noise than might be typical (and I really, really dislike hearing it). So, I just want to be sure that decently low noise is considered from the start this time, by only choosing from designs where it is actually measured (or at least mentioned) as being quite good. Exceptional magic isn't a firm requirement 🙂
But it is still a factor; I have in the past built designs that were in themselves mediocre for noise (as I later realised with more experience), so become a limiting factor after getting everything else right. I guess old or average designs are popular and persist for a number of reasons; we all have different interests and objectives.
This latest application is somewhat more sensitive towards noise than might be typical (and I really, really dislike hearing it). So, I just want to be sure that decently low noise is considered from the start this time, by only choosing from designs where it is actually measured (or at least mentioned) as being quite good. Exceptional magic isn't a firm requirement 🙂
Mr Kev06,
I have built both BJT as well as MOSFET versions of Circlophone in the prototype setup and experimented with different output devices e.g 2N3055, 2N3773, IRF350, IRF540 and IRFP240.
MOSFET version is in operation for around a year now using IRF350 MOSFET. I am very much satisfied with the schematic designed by Mr Elvee. There is no noise and hiss produced by the amp. The best part is that after the assembly you donot require the cabinet to be opened frequently for adjusting the bias. There is no adjustment for bias in the present schematic and it is very very stable . Sound quality is quite similar to the vaccum tube designs.
You can have a try on Circlophone.
Katiyar
I have built both BJT as well as MOSFET versions of Circlophone in the prototype setup and experimented with different output devices e.g 2N3055, 2N3773, IRF350, IRF540 and IRFP240.
MOSFET version is in operation for around a year now using IRF350 MOSFET. I am very much satisfied with the schematic designed by Mr Elvee. There is no noise and hiss produced by the amp. The best part is that after the assembly you donot require the cabinet to be opened frequently for adjusting the bias. There is no adjustment for bias in the present schematic and it is very very stable . Sound quality is quite similar to the vaccum tube designs.
You can have a try on Circlophone.
Katiyar
Thank you, Katiyar. That is encouraging to hear.
The the tolerance for different output devices is very appealing. Both for obtaining them now, and for future repair and maintenance. Less need to spend lots on harder to find ones, or postage from distant countries. In fact it is likely I already have some power transistors that would work. I certainly have some exicon lateral mosfets, which I'll have to check for suitability.
Thanks, Kev
This design accepts a lot of BJT and MOSFET out put devices but I donot think lateral MOSFET would be suitable as the threshold GS voltage is very small compared to other MOSFETS. People have built Circlophone with Darlington transistors as well.
You may consult Mr Elvee regarding the same.
If you can provide type numbers, and ideally a spice model, I can probably make a virtual C based on them. The topology is highly versatile, and until now has gracefully accepted any device type thrown at it; laterals are probably no exception, and if the virtual amplifier appears to work, chances are that its physical counterpart also will
Ah that is interesting, thank you! I've never used spice unfortunately, but will have a look for the exicons after work and see what their model numbers are.
I just found my order for the exicons; they're 10N20R and 10P20R.
But I believe I've also got some MJL21193 and MJL21194 left over from older projects. They should be more similar to the original C design.
But I believe I've also got some MJL21193 and MJL21194 left over from older projects. They should be more similar to the original C design.
Thank you, that is encouraging, and very helpful. I see it worked well, but in post #474 you still seem to slightly prefer the BJT version. That would make sense to me since it is closer to the original conception, and perhaps (for a first attempt, at least) it would be better and easier for me to use output devices quite similar to those that are well proven to work without modification.
A little later on, I see you tried a lower voltage version, with success. That is also of interest to me, since some of my applications require only quite modest power and it would make sense to use an amp which isn't massively more powerful than needed. Good stuff.
Thanks again,
Kev
A little later on, I see you tried a lower voltage version, with success. That is also of interest to me, since some of my applications require only quite modest power and it would make sense to use an amp which isn't massively more powerful than needed. Good stuff.
Thanks again,
Kev
https://www.diyaudio.com/community/...ssories-beginner-friendly.209907/post-5712992
Please checkout the layout of pronk !
Please checkout the layout of pronk !
Thank you, for the link. It is interesting to see so many different approaches to making the boards for this amp. From dual layer PCBs to even copper-less boards.
I suppose a little caution might be worthwhile as whilst their builders have been subjectively happy with them, many aren't actually measured objectively. But it still supports the idea of the circuit not being very fussy or sensitive to such things.
I haven't yet decided what method I shall use but, at least to begin with, will likely make a single-layer PCB or even something more basic for trial purposes.
I suppose a little caution might be worthwhile as whilst their builders have been subjectively happy with them, many aren't actually measured objectively. But it still supports the idea of the circuit not being very fussy or sensitive to such things.
I haven't yet decided what method I shall use but, at least to begin with, will likely make a single-layer PCB or even something more basic for trial purposes.
Choose the heatsink, Choose how you will mount the PCB and outputs to it for good heat dissipation, and then design a PCB layout.
Thanks, yes that is roughly the intention. Though the first step is to look at parts availability, and what I already have etc. Depending on the outcome, there is some chance that I could start with the VHex instead; it might not be quite as simple to build but, I find the operation/schematic a little more conventional and familiar.