They seem bit overkill for this project. With 16-18V rails, I'm only expecting less than 30W @ 8ohm. Those devices can handle 200W if I'm not mistaken. Got any suggestions for slightly lower power? My transistor selection knowledge is a bit dated...
Hi njswede,
That's cool. Now, later build it using good parts, 2 EF output stage. Look at the Nakamichi 620 amplifier or the Marantz 300DC for inspiration. There are other amplifiers out there using a similar output configuration. They all sound great and have good reputations.
Those types are selected for their linearity and consistency between devices. To get performance there are times you need to use parts intended for higher power applications. Also try NJW0281 and NJW0302. Don't use a TL072 or 071, try NE5534 or NE5532, an LM4562 is probably going to be the pinnacle of performance but I'm not sure how much better the entire circuit would be over the NE553x series. An NE5534A will perform better than the NE5532A on paper, but make sure you buy the TI version, not something unknown from a non-authorized distributor.
That's cool. Now, later build it using good parts, 2 EF output stage. Look at the Nakamichi 620 amplifier or the Marantz 300DC for inspiration. There are other amplifiers out there using a similar output configuration. They all sound great and have good reputations.
Those types are selected for their linearity and consistency between devices. To get performance there are times you need to use parts intended for higher power applications. Also try NJW0281 and NJW0302. Don't use a TL072 or 071, try NE5534 or NE5532, an LM4562 is probably going to be the pinnacle of performance but I'm not sure how much better the entire circuit would be over the NE553x series. An NE5534A will perform better than the NE5532A on paper, but make sure you buy the TI version, not something unknown from a non-authorized distributor.
Peak voltage has to be divided by root 2 to yield rms voltage. Formula above does this more efficiently.
Yup, I just ballparked it in my head. 0.7071 ~ 0.7. Right now the question is more distortion performance. Power, that's just voltage. I've built many up to around 100 wpc, so performance should be quite good. Commercial examples are held in high regard from what I have seen. I have a couple examples of models that are diamond buffer designs for reference. Nothing better than looking at successful examples.
The real plus is the extremely linear output stage. The negative is supplying high peak currents to the load as it may current limit. Depends on how you drive it.
The real plus is the extremely linear output stage. The negative is supplying high peak currents to the load as it may current limit. Depends on how you drive it.
Yeah, I think there’s some improvements I can make on the build, including (but not limited to) better parts. 🤣
But it was a very successful proof of concept.
It's a bipolar supply. I'm running the prototype on +-18V, so I'm getting about maybe 14V peak, which translates to about 25W peak across 8ohm. The swing is still a bit anemic, but I don't think I can fix that without adding transistors, and the whole point was low component count.
I'm sorry, but that's wrong.
If the ground to peak voltage is 16V, the rms voltage of a sine wave is 16/sqrt(2) = 11.32Vrms. Power is 11.32^2/8 = 16W
If the ground to peak voltage is 16V, the rms voltage of a sine wave is 16/sqrt(2) = 11.32Vrms. Power is 11.32^2/8 = 16W
Not for peak power. It's defined as the highest power across the cycle, So if my voltage peaks at 16V, I'm putting 16V across 8ohm at the very instance the wave reaches its highest value. Average power is the same as RMS power and is always half the peak power, since the squaring of the voltage makes the square root vanish.
https://en.wikipedia.org/wiki/Peak_power#:~:text=Peak power refers to the,average power ratio (PAPR).
https://en.wikipedia.org/wiki/Peak_power#:~:text=Peak power refers to the,average power ratio (PAPR).
If we're arguing the distinction between peak power and the power delivered when a sine wave is driven to threshold of clipping, I agree completely.
But If I'm not mistaken, common practice is to specify power (and distortion) when driven by a sine wave. I remember when "peak music power" was castigated as advertising hype. 😉
But If I'm not mistaken, common practice is to specify power (and distortion) when driven by a sine wave. I remember when "peak music power" was castigated as advertising hype. 😉
Look Ma! A PCB!
Kind of cute, if I may say so.
Kind of cute, if I may say so.
Hi No ideas,
What do you mean? It would work with TO3, TO-220 or any other case style.
Normally you would put the drivers on another smaller heat sink and the bias transistor or diode connected to the drivers.
What do you mean? It would work with TO3, TO-220 or any other case style.
Normally you would put the drivers on another smaller heat sink and the bias transistor or diode connected to the drivers.
by the same flavour i meant for example two bd911/912 pairs instead of one bd911/912 pair and one bd139/140 pair.
and by to 220 i meant the pinouts and leg sizes, also i could just use wires for dffrent transistors.
and by to 220 i meant the pinouts and leg sizes, also i could just use wires for dffrent transistors.
Hi No ideas,
I am missing what you are getting at. Those two are vastly different transistors, nowhere even close. One is a driver or Vas, the other an output or driver for a high current output stage.
Can you rephrase your question?
You would typically use the same drivers in each channel, and the same outputs. But since they operate independently of each other you could use whatever you want in each channel as long as they were appropriate for the job. The channels would probably differ in performance, but they would work.
I am missing what you are getting at. Those two are vastly different transistors, nowhere even close. One is a driver or Vas, the other an output or driver for a high current output stage.
Can you rephrase your question?
You would typically use the same drivers in each channel, and the same outputs. But since they operate independently of each other you could use whatever you want in each channel as long as they were appropriate for the job. The channels would probably differ in performance, but they would work.
As soon as I verified that the PCB design isn’t completely bonkers, I’ll post all the CAD and simulation files on GitHub (like I do with all my projects). You should be able to tweak the layout very easily to allow any component footprint.