It does not really matter: the drivers will make up for the difference, and they have plenty of margin
With 32V and 8 ohm, that is the maximum.
You can increase the supply voltage: the current source will track automatically, and it will stay good for 8 ohm, but of course the transistors will dissipate more, in proportion to the square of the supply voltage.
You see the heatsink on my prototype? It already becomes pretty hot with the basic version. If you want more power, you have to be prepared for larger sinks (times N for the number of channels of your amp, of course)
If you want use use that kind of accessory, opting for class A seems a strange idea....The ClipNipper was designed for a 10 watt to 8 ohm amplifier, and if using it for 11 watts, can just exchange its 1N5819 for a standard diode. That circuit (in my signature, 1st post schematic) will make Tringlophone exchange clipping for compression instead, just like a tube amp.That's borderline enough for a recording of a piano to play as loud as the real thing. To get it the rest of the way to impressive, an 11 watt amp needs some efficient woofers. Its doable.
"I'll get you, my pretty! And, your little amp too!"
(obscured Wizard of Oz reference--I live in Kansas)For the best way to explain: I've sent you a pair of NSL7053 CDS LDR optocouplers and some of the exact same InGan LED's used in the ClipNipper detector circuit and 10 of Vishay 1n4007G's of the exact same model used in the detector. What I didn't send was the resistors or the 22u caps.
This particular accessory spends most of its time off, and it doesn't cut into your average power output margin at all. Its not a blocker. Dampening momentarily excess signal is transparent. The LED's also serve as clipping indicator. There is some de-clip capacity since the detector's reaction to studio compressed tracks with square/sharp edges is about 150% and thus there may be the occasional flash during normal replay if also using normal/inefficient speakers. While mitigating momentary overdo conditions, it can mimic a tube amp.
To make sense with a Class A amp, the accessory can be trimmed to compress a few errant notes that would have otherwise caused audible clipping. I've added trimming information to post 1 of the clipnipper thread in my signature. You might like this adaptive headroom boost better than increasing transformer voltage on a Class A solid state amplifier. The fun way to find out should be arriving in a week or two.
Last edited:
OK, I'll give it a tryTo make sense with a Class A amp, the accessory can be trimmed to compress a few errant notes that would have otherwise caused audible clipping. I've added trimming information to post 1 of the clipnipper thread in my signature. You might like this adaptive headroom boost better than increasing transformer voltage on a Class A solid state amplifier. The fun way to find out should be arriving in a week or two.
.Now, let's have a closer look at the ground/supply system of the Tringlophone.
A single supply with a positive ground is somewhat out of the ordinary.
There are valid reasons to opt for such an unusual scheme.
Although the output is floating, its natural, implicit "anchor" is referenced to the negative rail.
By contrast, due to its voltage translation effect, the input stage has its input referenced to the positive rail, which thus becomes a natural choice for the general ground.
Some conflicts remain though: the CCS feeding the tringle is necessarily referenced to the positive, but it has to receive some signal from the bottom for its modulation.
Fortunately, it is not the noble path, and no voltage translation is required provided the ripple voltage remains of a reasonable amplitude.
The DC operating point is referenced to the positive by C7 and C9, with the modulation simply injected R16/C10 (along with some ripple).
Other parts of the tringle are either AC-referenced to the negative by C8, or referenced internally by C1 for level shifting.
This game of yoyo between the rails results in some residual leaks in PSRR.
These leaks are compensated by the R27/C14, which also creates a rejection notch at 100Hz.
As a result, a simple good quality filtered supply is sufficient: no regulation or electronic filtering is required.
Please, i'm looking for a nice and simple idea to clip an imput signal, and to can tune-it just under the clipping point of any amp. I would be very pleased if you have some idea better than my adjustable zeners's ones in the input line.
Why not use the classical text-book diode clipper? It is easily adjustable with a single resistor.
But please, go and discuss it elsewhere, it has nothing to do with this thread
Attachments
I agree! There's no need to decrease the power output of Tringlophone.
Therefore, here is a new thread for soft clipper: Soft Clipper Central Station
Elvee, a link to your 4n25 soft clip is there, and a general purpose overload blocker (can use with Tringlophone), and maybe soon, an answer to Esperado's question (a highly "accessorized" overload blocker), which I will try to do and post at: Soft Clipper Central Station I also solicit your help in casing as much decorous clipping as possible at Soft Clipper Central Station
Meanwhile back to the regularly scheduled Tringlophone broadcast:
Is this a "find its own center" or "auto-center" amp?Elvee said:
Does it remove ground loop noise from grounded source to grounded amp?
I don't know exactly what distinction you make between them. Anyway, it is an open loop centering, defined by resistances ratios, there is no servoing.
It is not extremely accurate, and you may have noticed that the input bias resistor is 120K on the "official" schematic against 130K in the sims: there is small discrepancy between sim and reality, and it might be necessary to fine tune this value on different amplifiers.
It is not critical however, since the output is AC coupled.
Not particularly: it amplifies the voltage it sees between its input and ground terminals, and there is no differential input or "ground-scissor" there. Pretty standard stuff.
I've got an audio (and compensations) related question.
Did the amp compensations make most of the output device selections perform near-identically for no significant change to the audio when faster outputs were installed?
Or
Did you like any of the faster devices well enough to say that there was any audible benefit?
Since this amplifier has no feedback, it does not require compensation.
There are some degenerations, stoppers and stabilization networks to cope with the realities of the wiring, amongst other things, but that's quite different.
In general, the performances will be better with better devices, but I didn't test that aspect: I simply made sure the amplifier remained stable.
This also means that there is room for tuning and tweaking: I didn't want to run any risk, and this probably means the tringlotron effect has been excessively dumbed down.
If the degenerations are tailored to a specific build, higher performances could be attained without endangering the stability.
Stability here has a different meaning than in a feedback context: it means absence of local oscillations caused by an uncontrolled positive reaction from the tringlotron structure (negative resistances, etc)
There are some degenerations, stoppers and stabilization networks to cope with the realities of the wiring, amongst other things, but that's quite different.
In general, the performances will be better with better devices, but I didn't test that aspect: I simply made sure the amplifier remained stable.
This also means that there is room for tuning and tweaking: I didn't want to run any risk, and this probably means the tringlotron effect has been excessively dumbed down.
If the degenerations are tailored to a specific build, higher performances could be attained without endangering the stability.
Stability here has a different meaning than in a feedback context: it means absence of local oscillations caused by an uncontrolled positive reaction from the tringlotron structure (negative resistances, etc)
I apologize!
Well, I thought 11w might not be enough output power; however, I was probably mistaken.
The output cap in the range of 3300u or 4400u or 4700u, decreases the load at subharmonic, but right before it gets that done, it gives a "progressive current drive" effect. When I tried this with my own 11 watt amp and medium size sealed box speakers, the sudden appearance of plentiful extra low bass was surprising--a very nice surprise. I used a parallel pair of Nippon ChemiCon LXF 2200u 50v (makes Low ESR 4400u) when doing this observation. Don't worry, the palm mark faded from my forehead in about half an hour.
Well, I thought 11w might not be enough output power; however, I was probably mistaken.
The output cap in the range of 3300u or 4400u or 4700u, decreases the load at subharmonic, but right before it gets that done, it gives a "progressive current drive" effect. When I tried this with my own 11 watt amp and medium size sealed box speakers, the sudden appearance of plentiful extra low bass was surprising--a very nice surprise. I used a parallel pair of Nippon ChemiCon LXF 2200u 50v (makes Low ESR 4400u) when doing this observation. Don't worry, the palm mark faded from my forehead in about half an hour.
Perhaps there's a friendly discotheque where you can do some loud fun tests during the off hours?
I'd just love to see their bizarre facial expressions when you tell them it is 11 watts.
Here is one version. Maybe not exactly identical to final graphic schematic, but enough to get you going and have some fun
Elvee
thank you for the asc file.
Please look at the picture, this is a proposition for input cfp pair to make it constant power, the performance should be better after that
Attachments
???Elvee
thank you for the asc file.
Please look at the picture, this is a proposition for input cfp pair to make it constant power, the performance should be better after that
I don't see what you mean, nor where it fits into the schematic
ok, I should be clearer, please attached sim file
Elvee, could you set it up proper regarding dc conditions?
thank you in advance
I don't know what you mean by "constant power".
Regarding the "improvement", the THD has jumped from 0.0068% to more than 0.3%; OK, it's almost pure H2, some might find the sound more pleasant, but I prefer to leave it as it is.
In addition, the current in Q3/Q16 is now 75mA, meaning they will each dissipate 1.2W and require a heatsink.
What is the point? of this mod?
Regarding the "improvement", the THD has jumped from 0.0068% to more than 0.3%; OK, it's almost pure H2, some might find the sound more pleasant, but I prefer to leave it as it is.
In addition, the current in Q3/Q16 is now 75mA, meaning they will each dissipate 1.2W and require a heatsink.
What is the point? of this mod?
ups, you were faster
please read PS in upper post
the reason of proposing cfb constant power triple was minimizing memory distortion (if you believe that thay exist
)Do we need this big output transistors for 11W?
The NJW3281G are excellent devices, currently in production and big enough for this. Actually i'm thinking of something even smaller (but that i could find matched).
Among the monolithics, there is something suitable with power dissipation?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.