OMNIFEX said:
Once you get to about 20,000 uf per rail you run into diminishing returns. Your pole frequency is about 7.5 Hz. Might be a little low for LFE, but you're playing dance music with a big peak centered in the mid 30's, right? That's enough cap for the purpose. Especially if you brick-wall at 20 or 25.
I don't know about you, but I clip my amps. I run in both low power and high power modes. In high power, I run minumum impedance (4 ohms bridged, typically) and just tap the clip lights. In low power, I run 4 ohms per side and clip the snot out of it - typically 6dB into clip pretty much continuously. It's the same average power either way. "High power" is cleaner and more dynamic, but "low power" actually sounds subjectively louder. Even so, the bass still sounds reletively clean if you have enough cap in the power supply.
This one isn't a gainclone type....it's got 2 devices per channel for the output, a couple of regulator discrete devices and discrete drivers. The inputs are all buffered with opamps and the headphone circuit is a tiny class-a discrete amp.
wg_ski said:
Cheers wg_ski!
I roll-off the bass @ 25 Hertz, 12 dB Octave. The heart of the bass notes range from 50 - 25 Hz.
I have this system (I own two) configured so I will have more than enough headroom for the bass. With 12 Peavey CS 800s I will have 14,400 watts RMS for the Subs. The Mid –Highs are only getting 4200 watts. The ratio amongst the two is 5 dB in favour of the bass. On those rare occasions when the amplifiers do clip the percentage is so low (I have a 90% chance they will not clip), it is not worth my concern.
tomman888 said:
You couldn't get 75W out of a gainclone anyway. You just can't run the Vcc high enough on those. Even with an IC front end, it's not a "gainclone". I've seen 225 w/ch amps with IC front ends (and discrete drivers/outputs).
Funny thing about all these DIY gainclones - people spend ridiculous money on all the pseudo-science to polish a turd. Those things were designed for *really* low cost amp solutions - like powered speakers and TVs.
Yeah I hear that, man. I have an LM3875 amp DIY on my list. Best case, with beefy PSU, it'll do ~55wpc at 8 ohm. Perfect for high-eff speakers and a nice passive pre.
Anyway, I spoke with Cambridge audio themselves about this and they told me: It was designed prior to Self working at the company. It's a Class-AB design (which I knew) utilizing an open-loop design which 'has little rate of change across the audio frequency band')
Guess he means they tried to make the thing as stable as possible before adding loads of NFB. Good thing, IMHO.
Anyway, I spoke with Cambridge audio themselves about this and they told me: It was designed prior to Self working at the company. It's a Class-AB design (which I knew) utilizing an open-loop design which 'has little rate of change across the audio frequency band')
Guess he means they tried to make the thing as stable as possible before adding loads of NFB. Good thing, IMHO.
Which means apply local feedback first, then a smaller amount of global. In theory you get the same total error signal (distortion) by either doing the feedback in stages or applying it all globally. No matter how you stack it it's the same amount of open loop gain, really. In reality, the result is design dependent. Some have poor dynamic behavior doing it all globally, some do not (like the Leach, which gets the same result with or without VAS loading).
Does applying the feedback in local loops decrease the THD only in that stage? If so, in an AB OP stage amp, with a good front end, should most of the OL gain be used to decrease crossover distortion? I ask because I am presently simming an amp and trying different amounts of local and global feedback loops, but not getting much difference in total THD. Wondering if I am on the right track? (and other factors are hiding the difference).
That's what I said - theory says you won't see a difference in THD driven by a single tone. Dynamic behavior (TIM, risetime and overshoot, stability margin, gain peaking at HF) can and will change. Local feedback in a stage reduces the distortion in that stage, so there is less error for the global loop to correct. Lest anyone go worrying about not having enough loop gain to correct crossover distortion - all is not lost. In the usual topologies feedback in the VAS makes the VAS better able to drive the outputs, which in itself reduces crossover distortion.
Interesting info, thanks wg_ski!
Today I used a 1kHz test tone to measure the amp's headphone circuitry unloaded. I do not have a dummy load available, but my cans are 32 ohms nominally.
Question: If the headphone amp can output 600mVolts per side unloaded, what, approximately, am I looking at with real music waveforms and a 32 ohm load?
Roughly? Maybe 50mWatts per channel? 100? It's a tough call without the proper test gear, I realize....Just curious, yet again. 😀
Today I used a 1kHz test tone to measure the amp's headphone circuitry unloaded. I do not have a dummy load available, but my cans are 32 ohms nominally.
Question: If the headphone amp can output 600mVolts per side unloaded, what, approximately, am I looking at with real music waveforms and a 32 ohm load?
Roughly? Maybe 50mWatts per channel? 100? It's a tough call without the proper test gear, I realize....Just curious, yet again. 😀
About a milliwatt. Max output would be 11 mW. That's loud enough for a lot of people - maybe not for Omni or myself, but we could add some bass shakers 🙂
Well I measured the jack's output imedance and its at roughly 50 ohms. Time for a new seperate amp I guess, considering I'm probably in negative damping factor territory with my Grados. Might not matter, but this doesn't sit well with me.
Also after reading the other thread about the SAP15 devices, I am about ready to unload this amp. I'm sure it'll sell since it's practically brand new and sounds good.
Also after reading the other thread about the SAP15 devices, I am about ready to unload this amp. I'm sure it'll sell since it's practically brand new and sounds good.
Low DF isn't near as much an issue with cans. With loudspeakers, the high Z will detune the box if it's high enough, but with phones the whole bottom end is way below resonance so you don't rely on the amp at all for "tuning". A lot of amps just use 220-470 ohms straight off the output for the phone jack. Sounds better with a small class A amp, but that's because it's a small class A amp not because of DF.
- Status
- Not open for further replies.