I have a question about driving Magnepan 3.7 speakers with the kypton v2. So what are the things to be considered before building the amp.
Im using about 3 of ALF16N20W and 3 ALF16P20W with bias of about 350ma for each mosfet at 45V.
The load impedance is about 4ohm per speaker.
psu capacitors about 30000uf Mundorf Mlytic AG per rail.
anything to be considered before I connect the amp to the speakers? especially magnepan.
A good speaker protection relay is always a good idea before connection to any speakers. You don't need to do anything special otherwise.
A balanced version of one of the amps would be cool. Would allow lower rail voltages, increased bias with no overall power loss.
I discussed higher bias with vzaichenko and was informed that these designs did not really benefit from higher bias. Something about high OLG.
Check out this: post 408
http://www.diyaudio.com/forums/soli...me-old-ideas-1970s-ips-ops-9.html#post4578177
Check out this: post 408
http://www.diyaudio.com/forums/soli...me-old-ideas-1970s-ips-ops-9.html#post4578177
Or if you use a non-switching / error-correction mechanism. Excessive bias can make things worth. In my case (NS-OPS), optimal quiescent current is 40-50mA per output pair. Thanks to the non-switching mechanism - THD is reduced for about an order, comparing to conventional EF3 OPS.
Going to balanced (or bridged) invokes increased power loss, not "no power loss".
Why do you said this ? The efficiency remain the same, and, if each side of the bridged amp is able to face to the current needed (each amp now have to face half of the impedance load) the power is increased by a factor of 4 (V²/R, double voltage) . The only trade-of a bridged amp concerns the damping factor.
balanced/bridged requires each half to drive HALF the load impedance.
Using lower rail voltages to drive half the impedance, so that the total output remains the same as the unbridged, will dissipate more power inside the amplifier.
Using lower rail voltages to drive half the impedance, so that the total output remains the same as the unbridged, will dissipate more power inside the amplifier.
Andrew is absolutely right and caught me in a statement that was not true or at best, mis stated. By default, amplifiers that are biased further into AB or even A are much less efficient, wasting more power...but they do sing.
One could bias the two amplifiers to exactly the same as the one unbridged amplifier. The losses in each bridged will be higher than the losses in the unbridged.
Now add the two bridged losses together and compare to the unbridged.
If you don't believe me, then use National's design spreadsheet for their chipamps, to investigate the losses.
It allows for bridging and it does not allow you to change the output bias of either chipamp.
Now add the two bridged losses together and compare to the unbridged.
If you don't believe me, then use National's design spreadsheet for their chipamps, to investigate the losses.
It allows for bridging and it does not allow you to change the output bias of either chipamp.
Last edited:
I will admit to being stuck in a Class A mindset when I think of power amps. For me bridging is always an opportunity to get voltage in the cheap and as a trade off, increase bias. I will be trying the krypton ND as well as Vzaichenko's TubSumo.
True. But sending the same power in both situations with musical signals at normal listening levels will not make a major difference in temperature. Currents in the speakers remain the same. And bridge brings an obvious advantage with (pseudo) symetrical amps : Both rails are symetricaly sucked on each half of signal periods, and you double the reservoirs capacitances during those transients.
Also, look at this one - VERTICAL front-end:
http://www.diyaudio.com/forums/soli...e-old-ideas-1970s-ips-ops-41.html#post4578177
Came out nicely, as well as the non-switching OPS 😉
More info on different options - see the link in my signature.
Cheers,
Valery
P.S. Sorry for a little bit of advertising in my brother Ostripper's thread - just coincidence 😛
Last edited: