May I add another LM3886 based commercial amplifier?
Cambridge Audio Azur 351A
http://www.cambridgeaudio.com/media/azur351a_servicemanual-1356952111.pdf
If you want to keep it simple for a DIY, plenty of things can be suppressed, for instance the microcontroller section to select inputs, that can be done simply with the traditional rotary switch button. Other stuff for protection... tone control...
Drawing a well routed PCB and using good quality components, may even end in better performance than the original, because you don't have to aim the amp for a specific price range market, and so you can be more "esoteric" with caps and op-amps, for instance.
Thanks to bring this to my knowledge
Cambridge Audio Azur 351A
http://www.cambridgeaudio.com/media/azur351a_servicemanual-1356952111.pdf
If you want to keep it simple for a DIY, plenty of things can be suppressed, for instance the microcontroller section to select inputs, that can be done simply with the traditional rotary switch button. Other stuff for protection... tone control...
Drawing a well routed PCB and using good quality components, may even end in better performance than the original, because you don't have to aim the amp for a specific price range market, and so you can be more "esoteric" with caps and op-amps, for instance.
Thanks to bring this to my knowledge
Last edited:
The gainclone as it is popularized has three very significant drawbacks that make its performance system dependent and hard to market.
Low capacitance in the PSU which makes it poor in bass, and lowish input impedance which again makes it even poorer in bass if anything less than 10uF is used for dc coupling... Both can be dealt with, but still we keep seeing the old shigaclone implementation.
Add to that the inability of LM3875 to drive hard loads at decent watts and you see that it is not something that can be easily marketed.
Tubes, SET, Class-A and big Watts are sexier and easier to market...
We just need to accept that we need easy speakers and a moderate power/voltage implementation to keep the amplifier in a comfortable output level at all times.
It's not its fault if it clips or if the Spike kicks in when we power it for max power output and use it with difficult loads at high volumes expecting it to knock us down with its bass... It' s one chip for crying out-loud... And the max power rating is just what it is on everything else... an overhyped and overrated nominal value.
The same inability in transients is true for SET class-A super duper rare unobtainium discrete amps with crappy damping factors, but I have never heard anyone using such harsh words like crippled for them... Because they are sexier and we take as a given that they need an easy load.
It s a bit a matter of perspective and what label you want to attach to it. Whatever you feel like doing towards it... punish it for what it is or pet its ears...
Currently I am loving mine
Does it show? the market on the other hand...
Wrong.
There was a link to a set of lab tests that showed just how much current was drawn by a variety of speakers to actual music excerpts.
Surprisingly the selected mild reactance speaker and the moderate reactance speaker and the severe reactance speaker, each drew maximum currents that (very roughly) exceeded the current of a nominal resistive load by a factor of 5 !
This is equivalent to 50W into 8ohms needing a peak transient output current of >17Apk into the speaker instead of 3.5Apk into the resistor load.
If the Transformer on the output of the SET can get to somewhere approaching 3times the nominal maximum power current then I would not call it "crippled".
BUT,
Chipamps are current crippled. They cannot approach 3 times, never min 5times the current demand of typical speakers.
But if you keep a 3886 cool, it can manage 2times (7Apk) into an 8ohms speaker. Forget about decent SPLs into 4ohms, unless you do a Peter Daniel and develop the chipamp amplifier into 95dB/W @ 1m speaker..
Last edited:
Apples, oranges. I'm talking drive units, you cite speakers.
So your claim of 'wrong' appears to be based on a misreading of my post.
I was speaking specifically of drive unit reactance - in the electrical model of a driver the reactive elements appear in series with the DCR (DC voice coil resistance) and hence the modulus of the drive unit's impedance cannot drop beneath this value.
Do please explain what's wrong about my above statement. Also a link to the claim of 5X current peaks would be handy.
Do please explain what's wrong about my above statement. Also a link to the claim of 5X current peaks would be handy.
i recall Doug Self (and perhaps Bob Cordell too) referring to a specifically crafted waveform which produced surprisingly large (5X springs to mind, but my memory is hazy) current peaks - however this waveform is acknowledged to be vanishingly unlikely to occur in any real music played through a speaker. So the claim of 5X the peak current on normal music does seem rather extreme. We're to take your word for this figure of 5X then?
Last edited:
I have read Self and Cordell and many others.
Yes they do say that contrived test/artificial signals give these high peak transient currents.
They generally agree that any conclusions reached using these tests signals are unlikely to apply to audio systems when used as audio reproduction systems.
It was in the light of that, that I took particular interest in and carefully read the lab report (reported in an Audio magazine) to ensure I did not get "conned" into believing some conclusion that could not be true.
As a result I changed my amplifier design procedure to adopt 3times as the peak current for transient demands instead of only 2times.
This is current into a load, not current into a short circuit or 0r1, which is rather different in that virtually no voltage is required on the "near shorted test".
You can call me gullible if you want, just don't post that opinion here.
But certainly you are not allowed to abuse me, or my reputation, by calling me a liar on this Forum.
Yes they do say that contrived test/artificial signals give these high peak transient currents.
They generally agree that any conclusions reached using these tests signals are unlikely to apply to audio systems when used as audio reproduction systems.
It was in the light of that, that I took particular interest in and carefully read the lab report (reported in an Audio magazine) to ensure I did not get "conned" into believing some conclusion that could not be true.
As a result I changed my amplifier design procedure to adopt 3times as the peak current for transient demands instead of only 2times.
This is current into a load, not current into a short circuit or 0r1, which is rather different in that virtually no voltage is required on the "near shorted test".
You can call me gullible if you want, just don't post that opinion here.
But certainly you are not allowed to abuse me, or my reputation, by calling me a liar on this Forum.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.