best opamp for equaliser

[MurrayP]

I've built a 4 stage parametric bass equaliser based around Rod Elliotts design (http://sound.westhost.com/project28.htm) - it works well. It replaces a Behringer DSP8024 digital eq that worked well, but in my system levels ran too few bits so hashed the sound up a bit I felt. The analogue eq seems pretty transparent, though it is in the main signal path (well tape loop actually). Anyway, being completely obsessive I'd like opinions of what op amps might be better than the 5532's I've used so far. The OPA627 is well recommended in the op amp discussion on this board, but I just discovered what they cost! Any more sensibly priced recommendations please?

My circuit is effectively the 35-150Hz stage (trimmed here and there) repeated 4 times, though I only link in stages as required 1 on one channel, two on the other for now. The object is to judiciously cut room resonances only - works wonderfully when WAF prevents other solutions! The circuit is well laid out with good decoupling.

Any suggestions appreciated.

rgds

MurrayP

[Chucko]

People carp about the "lowly" 5532. But even some 20 years after its introduction, there aren't many op amps available that are significantly better... and those that are cost a LOT more money!

Having said that, I have had good results with the TI (Burr-Brown) OPA2134 as a general purpose device in musical instrument gear.

In an equalizer application, you can pick and choose your devices for each frequency band. It would be a waste to use an AD797 for the low frequency filters, but it might be appropriate on the high end.

The input buffer amp should be capable of driving low impedance loads with low distortion, and for this the 5532 is hard to beat. The main summing amp needs to have very low noise voltage at low source impedances as well as good drive capabilities, and because the loop gain is >> 1 you can often use a decompensated device like the Analog Devices OP-37 or Linear Technology LT1115 or LT1126. Again, the 5532 is not a terrible choice either.

[Chucko]

Just had a look at the schematic, it wasn't quite the topology I expected but is not too different. I was thinking of the one that has active filters in the feedback loop, instead of the series-resonant "LC" shunt to ground in this version.

Gee, I might have to build this one for myself!

The input buffer can use almost any device as long as the slew rate and noise aren't abysmal. The author suggests a TL072; I might suggest an LM833 (dual) or LM837 (quad), a bipolar input device which will be quieter than the '072 with most line-level sources, and is about as cheap. It's also not as power hungry as the 5532.

The output stage appears to need unity gain stability, so forget what I said about decompensated chips.

[MurrayP]

Sorry should have said 5534 (single version I think) - anyway - same thing. Anyway just thinking maybe there could be better. Should also have said rest of system is a bit flash - good valve preamp, Plinius SA250/4 and N801's so the eq is in pretty strong company, but I really don't think I'm loosing much going thru it, and gaining heaps by eliminating a really nasty room mode that masks other detail like crazy,

thanks for the ideas,

Murray

[JoeBob]

I quite like them LM7171 (or LM6172 for a dual). I don't know if they're worth using for just the low end though...

[paulb]

Careful - the NE5532 is a dual, compensated version of the single, uncompensated NE5534. The 5534 needs a compensation cap for unity gain, the 5532 doesn't (otherwise it wouldn't fit into an 8-pin DIP).

[Chucko]

Originally posted by MurrayP
Sorry should have said 5534 (single version I think) - anyway - same thing. Anyway just thinking maybe there could be better.

At the risk of repeating myself, I think the 5532 or 5534 is hard to beat in this application, and better than many people want to believe it is.

But don't let me stop you from experimenting with more exotic devices!

[djk]

You only need one op-amp here, the rest are just buffers.Get a McIntosh MQ104 service manual, it gives the formulas for calculating frequency and Q for this circuit.The early MQ104s use a single transistor in class A for each buffer.Later ones use cheezy op-amps.Two guesses as to which version sounds better.

[Samuel Jayaraj]

The output/summing opamp can be an OP27 instead of NE5534 and I can almost bet that there will be a significant improvement to the overall sound quality. The same principle of gyrators with frequency sweep is used by Rod Elliot in his Mixer Project. I recently used the tone control section in conjunction with a different input section. I faced problems with the value of the feedback resistor (2k7) in the output/summing opamp, in that a large negative dc voltage appeared at the output with hardly any signal at all; I had to increase the value above 18k (actually I used 22k in the final circuit board). Comparisons between NE5543, LM301, OPA604 and OP27 in this position, showed that OP27 performed significantly better. I had a TLE2141 at hand but forgot to use it; I assume this would have been a better choice.

[Bernhard]

Hi,

is the OP27 a good choice for a 24dB avtive crossover ?

Input / output / filter stages.


Bernhard