A high-voltage, but low-current amplifier for midranges and tweeters?
Hi Phloodpants
What you want is an amplifier with a high dynamic headroom. For example, a 100W continuous 8 ohm that can do a clean, non-clipped 200W in short burst.
Regards
Mike
Hi Phloodpants
What you want is an amplifier with a high dynamic headroom. For example, a 100W continuous 8 ohm that can do a clean, non-clipped 200W in short burst.
Regards
Mike
IOW, exactly what you would build for a 200W high current subwoofer amp, but with a smaller heat sink and transformer.
I'm still unclear on one point: My understanding is that the reason that big amplifiers have many output devices in parallel has more to do with SOA than actual current delivery. Is there any reason to think that having only two pairs of outputs would insufficient in some way? Will they experience beta droop when the voltage/current spikes on these transients?
You're asking a somewhat oversimplified question. And, it does have to do with the current. Take a look at a SOA graph, it's all a matter of operating the transistor inside the limits, which are a function of voltage and current.
If you have an amp where you have a good safe margin, but go from an 8 ohm to 4 ohm load (doubling current) you've moved up closer to the edge of the SOA. If you raise the supply voltage, and leave the current the same, you also move towards the edge. Conversely, doubling the number of transistors will half the amount of current under the same conditions, and give you more margin away from the edge. As for beta droop, those brain cells are a bit dusty... I'd have to do some research...
So, to answer your question properly, you'd have to determine just where in the SOA that particular amp's transistors are operating with that particular load you're putting on it. Then, you could see if it would be dangerous to go from two parallel transistors to one. Or, if going to three might be a good idea...
The load/current aspect of this seems to not be considered in this discussion so far, since you've been looking at peak voltage. It occurs to me that the drivers will likely be different, as they usually are. A sub is nearly always a 4 ohm driver, midranges could be either but tend to be 8 ohm, and many tweeters are 4 ohm. Not counting the complex impedance over frequency that real drivers have. Are you looking at your power calculations in light of the load impedances?
Aha... just went and looked up your tweeter, (nice set of drivers, BTW!) the OW1 just happens to have a min impedance of just under 7 ohms. So, you do have a worse case in voltage requirement than the average 4 ohm tweeter.
I'm still unclear on one point: My understanding is that the reason that big amplifiers have many output devices in parallel has more to do with SOA than actual current delivery. Is there any reason to think that having only two pairs of outputs would insufficient in some way? Will they experience beta droop when the voltage/current spikes on these transients?
For a same current linearity is not the same, with only two pairs the distorsion at say 5A will be higher than with four pairs, so that s not a matter of SOA but of equal linearity between channels.
Yet it s the tweeter that require the most linear amp, so at least the same number of devices as the bass channel is required, the only difference, as pointed above, is that you ll need a smaller heatsink given the lower average power.
The SOA comes especially into play when the load is reactive, like with a passive xover. Meaning that with say pos Vout, it is the LOWER transistor that needs to pull current. In that case, the Vce of that device is high while still sourcing current, and SOA-wise that is worst case.
With just a simple, non-passive-xover tweeter, the load is pretty much ohmic, so SOA issues are absent.
Meaning, again, that you can save something on the output devices, one pair instead of two.
Jan
With just a simple, non-passive-xover tweeter, the load is pretty much ohmic, so SOA issues are absent.
Meaning, again, that you can save something on the output devices, one pair instead of two.
Jan
You cant...
At 5A a single pair will have higher distorsion than two pairs, so there will be more distorsion, the lower average power only allow smaller heatsinks if quality is to be left equal to a passive speaker..
You might want to limit the HF transient anyway. Filtering some music samples with Audacity shows occasional digital full scale transients close to crossover frequency. Faithfully driving a tweeter with this risks exceeding excursion limits and breaking the wires
Looks like I have some calculating to do... later.
But so far, I'm leaning towards this... I gather that there is no sound quality reason to use fewer output transistors.
The boards of a GFA-535 have empty spots to install a third pair of output transistors. (The GFA-545 uses the same board as the 535, but with three pairs.)
The GFA-555 uses four pairs of outputs, and handles abuse and low-impedances easily. For another example, the Phase Linear 400, also a 200WPC amp, uses three pairs, but is notorious for blowing out.
Since a midrange or tweeter is a far easier load than a woofer or sub, then I suppose three pairs would be perfect.
I'll do my homework on the SOAs first, but that's my guess.
But so far, I'm leaning towards this... I gather that there is no sound quality reason to use fewer output transistors.
The boards of a GFA-535 have empty spots to install a third pair of output transistors. (The GFA-545 uses the same board as the 535, but with three pairs.)
The GFA-555 uses four pairs of outputs, and handles abuse and low-impedances easily. For another example, the Phase Linear 400, also a 200WPC amp, uses three pairs, but is notorious for blowing out.
Since a midrange or tweeter is a far easier load than a woofer or sub, then I suppose three pairs would be perfect.
I'll do my homework on the SOAs first, but that's my guess.
Last edited:
- Status
- Not open for further replies.