Now that the 50W latfet amps that I designed for the speakers in my study have failed to meet expectations, I've been re-evaluating my options.
Firstly, I've redone the layout for the 50W Tilbrook amp - back to TO-126 VAS transistors, neatly arranged so that they can all share a common heatsink. That should (I hope!) allow me to get some reasonable performance. Actually I'm reasonably sure the performance will be good, as it's basically my 100W amps with one pair of FETs removed.
But while I was casting around, I revisited the TPA6120. I'd seen a rather cool design using some twenty odd of these bridged and paralleled to get reasonably high power levels into 8 Ohms, with vanishingly low distortion levels. Really very nice, but I want just a bit more power for my speakers.
Now as it turns out, my little bookshelf speakers (from Rabbitz design, described in the speakers forum) have a reasonably high crossover frequency, of 3KHz or thereabouts.
According to Rod Speed, with this high crossover frequency, the power requirements for the tweeter are really quite modest - only 15% or so of total speaker power.
So I got to thinking. Five TPA6120s chips in parallel (not bridged) driving the 6 Ohm tweeter each see 60 Ohms (the TPA 6120 is a stereo device, so there's a pair in each package). At 700mW per amp (+/-15V supplies), that's a total of 7W, which isn't too shabby.
But the nice bit is the THD, which rises (we'll use the 64 Ohm curves in the datasheet) from 0.0002% at 3KHz to 0.0005% at 20KHz. Yeah, there's three zeros after that decimal point. It's beyond the floor of my THD test set, which is a pretty nice one.
So given this 7W for the tweeter, and Rod's 85/15 split we ideally want 40W from the amp used to drive the mid-woofer. The 50W amp I've been mucking about with (with new layout to sort the VAS thermal matching issues) will do this nicely, and will (hopefully!) be able to manage something of order 0.001% THD at 40W, up to a few KHz...
The combination, I think, might be rather nice.
What are people's thoughts? Is it a totally naff idea? I'm encouraged at the thought of being able to use the same +/-15V supply to power both the tweeter and the input preamp/active crossover opamps, and I reckon it'd be a fun intro to active crossovers. Further, playing with the TPA6120 is something I've been itching to do.
Firstly, I've redone the layout for the 50W Tilbrook amp - back to TO-126 VAS transistors, neatly arranged so that they can all share a common heatsink. That should (I hope!) allow me to get some reasonable performance. Actually I'm reasonably sure the performance will be good, as it's basically my 100W amps with one pair of FETs removed.
But while I was casting around, I revisited the TPA6120. I'd seen a rather cool design using some twenty odd of these bridged and paralleled to get reasonably high power levels into 8 Ohms, with vanishingly low distortion levels. Really very nice, but I want just a bit more power for my speakers.
Now as it turns out, my little bookshelf speakers (from Rabbitz design, described in the speakers forum) have a reasonably high crossover frequency, of 3KHz or thereabouts.
According to Rod Speed, with this high crossover frequency, the power requirements for the tweeter are really quite modest - only 15% or so of total speaker power.
So I got to thinking. Five TPA6120s chips in parallel (not bridged) driving the 6 Ohm tweeter each see 60 Ohms (the TPA 6120 is a stereo device, so there's a pair in each package). At 700mW per amp (+/-15V supplies), that's a total of 7W, which isn't too shabby.
But the nice bit is the THD, which rises (we'll use the 64 Ohm curves in the datasheet) from 0.0002% at 3KHz to 0.0005% at 20KHz. Yeah, there's three zeros after that decimal point. It's beyond the floor of my THD test set, which is a pretty nice one.
So given this 7W for the tweeter, and Rod's 85/15 split we ideally want 40W from the amp used to drive the mid-woofer. The 50W amp I've been mucking about with (with new layout to sort the VAS thermal matching issues) will do this nicely, and will (hopefully!) be able to manage something of order 0.001% THD at 40W, up to a few KHz...
The combination, I think, might be rather nice.
What are people's thoughts? Is it a totally naff idea? I'm encouraged at the thought of being able to use the same +/-15V supply to power both the tweeter and the input preamp/active crossover opamps, and I reckon it'd be a fun intro to active crossovers. Further, playing with the TPA6120 is something I've been itching to do.
Hi,suzyj said:
that's a comparison of average power delivery.
It gives no clue to voltages sent to the respective drivers.
If we assume that your bass/mid driver and your treble driver have similar impedance and similar sensitivity, then both need the same voltage to reproduce the same SPL.
That means the voltage from the treble amp must match the voltage from the bass/mid amp.
The driver impedance will determine the current drawn from their respective amplifiers when the voltages are the same.
The big saving will be in the PSU feeding the treble amps and a slightly smaller saving in the heatsink due to less average power.
Do not skimp on the treble drive ability, or you end up with typical disco type sound where the treble is screaming at you because it is almost permanently clipping.
However,
if your drivers are of different sensitivity, then you need to calculate the voltages and currents into the drivers to get matching SPLs from them. This can save a lot of power in mid and treble amps when their much higher efficiency is compared to typical HiFi bass drivers.
Re: Re: Biamping silliness
Hi,
Quite true.
in practise I'd say a 3-way in terms of peak power (with similar
sensitivity drivers) splits around 44% bass, 33% mid, 22% treble.
In a two way the above numbers do make approximate sense.
Of course you can design mid and treble with bass voltage swing if you want,
but IMO the above numbers are already skewed towards the mid and treble.
The only real answer to this is real music statistics, bit thin on the ground.
🙂/sreten.
edit :
tweeter is a lot more efficient (~ 6db)than the BSC'd bass
unit, so overall I'd say the idea in principle will work fine.
AndrewT said:
Hi,
Quite true.
in practise I'd say a 3-way in terms of peak power (with similar
sensitivity drivers) splits around 44% bass, 33% mid, 22% treble.
In a two way the above numbers do make approximate sense.
Of course you can design mid and treble with bass voltage swing if you want,
but IMO the above numbers are already skewed towards the mid and treble.
The only real answer to this is real music statistics, bit thin on the ground.
🙂/sreten.
edit :
tweeter is a lot more efficient (~ 6db)than the BSC'd bass
unit, so overall I'd say the idea in principle will work fine.
- Status
- Not open for further replies.