AD8620 or LME49710

Status
Not open for further replies.
Jump to Latest
#2
I have experience of the LM4562 which I understand is the dual version of the LME49710. I took it out of my active speaker crossover and replaced it with the generic NE5532 and it sounded better. It seems its highly sensitive to RF - I don't doubt that with intensive care focussed on input and power supply cleanliness its capable of superb results though.

No experience with AD8620 but a quick flick through the datasheet suggests that it'll be more RF resistant (JFET input, dielectrically isolated process).
 
#3
If it is for a composite amp for speakers, for which the main problem is usually stability, I'd be more concerned about the fact that these two opamps are very different (25mhz vs 55mhz gain bandwidth to start with) than about any "real world" subjective listening.

Without knowing the full circuit, it seems difficult to offer any meaningful advice.
 
#5
The whole point of a composite amp is to get as high loop gain as possible, thereby, maximizing the error correction on the circuit. The LME49710 wins here as it has the highest bandwidth, hence, the highest amount of gain available within the audio band.

Which op-amp sounds better is largely a matter of confirmation bias, expectation bias, and personal preferences.

~Tom
 
#7
abraxalito is the PSSR of an op-amp that critical, considering light loads, good local decoupling, and fed by well regulated supplies with nary a SMPS in confines? compared to the final PA , I doubt that. sure the baby is thrown out with the bathwater when designers throw some bad practices into a circuit.
 
Last edited:
#8
Depends what 'light loads' and 'good local decoupling' actually turn out to mean in practice. Ditto the word 'well' in 'well regulated'. But I agree that the biggest challenge of all is the final PA - I'm now permanently wedded to OPTs to lighten the load and hence ease the decoupling burden. I was very surprised how much improvement there was in clarity with a 1.5:1 step down trafo between my chipamps and the bass/mid drive unit in my active setup.
 
#11
No, no ferrites at LF, they suck really bad. Laminated steel works fine, I rewind EI power trafos for bass/mid duty. Ferrites are well suited to 3kHz+.

I've not found that those traditional ways of arranging opamp PSUs deliver sonic nirvana so I've adopted more extreme measures - loads >22k, TL431 shunts, LC filtering, towers of X7R 10uF ceramics on top of chips, multiple paralleled electrolytics 2,200uF and up. But I would agree that context is highly important here - the issues tend to arise with more complex music, not what Lynn Olson calls 'sparse spectrum' stuff, which is what most audiophiles go for. Since my diet is almost all classical orchestral, solo piano and choral, I recognise I'm in the minority here.
 
Last edited:
#12
abraxalito said:
No, no ferrites at LF, they suck really bad. Laminated steel works fine, I rewind EI power trafos for bass/mid duty. Ferrites are well suited to 3kHz+.

I've not found that those traditional ways of arranging opamp PSUs deliver sonic nirvana so I've adopted more extreme measures - loads >22k, TL431 shunts, LC filtering, towers of X7R 10uF ceramics on top of chips, multiple paralleled electrolytics 2,200uF and up. But I would agree that context is highly important here - the issues tend to arise with more complex music, not what Lynn Olson calls 'sparse spectrum' stuff, which is what most audiophiles go for. Since my diet is almost all classical orchestral, solo piano and choral, I recognise I'm in the minority here.
Click to expand...


what does winding a transformer buy you for driving VC on mid bass duties. I can see advantages for an electrostatic design. maybe class A would be more towards your nirvana eg non switching tech. but off loads more duties onto the power supply.
 
Last edited:
#13
Yes classA would be another way to skin the cat but I'm not so comfortable doing a fully active system with 4 channels of classA, given the inefficiency. Since my amps are around 20W per channel, I'd be idling at over 300W.
The trafo transforms the impedance seen by the amp - currently I have 2.5:1 in voltage ratio which is 6.25:1 in impedance. Thus a 4R bass/mid looks like 25R. Caps at higher voltages store more energy per unit volume so decoupling becomes more compact and compactness translates to a lower impedance supply.
 
#16
oh how do you prevent accidental core saturation > series cap on primary?
EDIT > ok just checked yer blog, I see split bobbins = lots of leakage reactance!
check over shoot on various waveforms I would then try interleaved winding > 1st using double primaries sandwiched around the secondary. IMO its easiest to use series over parallel connections on primary depending on wire gauge you have there.
 
Last edited:
#17
Yes split bobbins, also series caps. Leakage reactance is not so much of an issue when the top frequency's only 3kHz (as in an active setup) rather than running all the way to 20kHz. I've not done any measurements with squarewaves.....
 
#20
abraxalito said:
I'm now permanently wedded to OPTs to lighten the load and hence ease the decoupling burden. I was very surprised how much improvement there was in clarity with a 1.5:1 step down trafo between my chipamps and the bass/mid drive unit in my active setup.
Click to expand...

If you like the distortion of magnetics, that's certainly one way to go.

As far as getting good performance from a chip amplifier, I'd argue that the composite amplifier approach is the way to go. It is certainly challenging to design a good composite amp. Stability needs to be addressed. As does clipping/overload recovery. The resulting performance is well worth it, though.

~Tom
 
Status
Not open for further replies.
Top Bottom