Better Opamp in QUAD 606/909 Power amp

[arupg]

Hi All:

The Quad 606/909 comes with TLC271 opamp for its input duties and also feedback. Swapping this opamp for a OPA134 opens up the sound, tightens the bass BUT on shut-down of the amp, huge noises appear via the speakers which means either offset is high or some problem that I'm not sure of exists. Two number 0.1uF/100V Vishay 1837MKPs were attached directly on the PS pins for Rail-to-Gnd bypass.

Checking with a DMM on the speaker outputs show 18mV on switch-on while 19V on shut-down!!! with the OPA134. This really worries me.....

Quickly reverting to the Stock TLC271 opamp, the shut-down is absolutely noiseless. Usual turn-on 'thump' is present with Stock & the OPA134.

Does anyone recommend I attempt to swap the opamp with a more low & controlled offset one like the OPA602 or OPA606? Some recommended the OPA604 but my earlier experience with 2604 & its distortion kept me far away from this family.

Your suggestions will be most appreciated.

Thanks in advance.

--------------------

Arup

[AMV8]

Hi

An alternative solution would be to ground the op amp inputs and then use a pair of transistors/fets as the input pair.

This was done by Musical Fidelity on the mk 2 versions of the P270 and A370 power amps.

Don

[Ouroboros]

The op-amp in the 606 is only used for the dc servo. The rest of the circuitry is discrete. The TLC271 is a perfectly good op-amp for a dc servo. Certainly an OPA134 would be a foolish part to replace it with.

[arupg]

The op-amp in the 606 is only used for the dc servo. The rest of the circuitry is discrete. The TLC271 is a perfectly good op-amp for a dc servo. Certainly an OPA134 would be a foolish part to replace it with.

Many mod stores in the UK offer to change the opamp for better varieties like OPA227 or OPA134 and claim it makes a difference. There's been a debate that the Opamp in the 606/707/909 is "strictly for DC Servo" BUT

it does have its signature on the final output. I for instance, immediately noticed a complete change for the better. The Stock opamp is a slough in the bass dept that just slows it down. The bass definition is muddy. OPA134 just cleans up all this. So, with such an improvement, why wouldn't one use it. Only thing is that probably that Opamp is drawing more input current that upon shut-down is unable to offer the quiet decay.

[peterhickman]

I have had a lot of experiance with changing these op amps and can tell you the original or pehaps its 'B' version is the best to have. OPA134 cause the switch off bumps etc, the OPA227 can cause the power supply to lock up with a high DC offset and blown capacitors, particularly with low impeadence loads. Two customers have said the original sounds better then and 134 (better bass control) and I have to say I would agree. There are plenty of other componnets to get to work on in the 606, 707 and particulary later 909s.

[CopperTop]

Could I ask a silly question?

In the various Quad 606/707/909 schematics that are kicking around, pin 8 of the op amp appears to be connected to the output, pin 6. Now the TLC271 is slightly unusual in that it is "programmable", with pin 8 being used to select one of three "bias levels". Among other things, the GBW varies with bias setting.

From the data sheet:

The TLC271 offers a bias-select feature that allows the user to select any one of three bias levels depending on the level of performance desired. The tradeoffs between bias levels involve ac performance and power dissipation (see Table 1)...

...Bias selection is achieved by connecting the bias select pin to one of three voltage levels (see Figure 1). For medium-bias applications, it is recommended that the bias select pin be connected to the midpoint between the supply rails. This procedure is simple in split-supply applications, since this point is ground. In single-supply applications, the medium-bias mode necessitates using a voltage divider as indicated in Figure 1. The use of large-value resistors in the voltage divider reduces the current drain of the divider from the supply line. However, large-value resistors used in conjunction with a large-value capacitor require significant time to charge up to the supply midpoint after the supply is switched on. A voltage other than the midpoint can be used if it is within the voltages specified in Figure 1.

Is there something fiendishly clever going on here, or is the output merely a convenient point to tie pin 8 to? Why have Quad used such a relatively unusual op amp?

(Assuming pin 8 really is connected to pin 6. Can anyone with a 606 confirm the PCB layout..?)

[Ouroboros]

It is a CMOS opamp used as an integrator with a 2.2Meg input resistor and a 680nF feedback capacitor, giving a time constant of 1.5 seconds. The output of the op-amp is connected via a 120k resistor to the amp input. It can have no effect on the audio quality when the amplifier is operational, but a change of op-amp may well change the behaviour of the amplifier at switch-off, causing a nasty LF thump as the power rails collapse.
The Quad schematic certainly shows the two pins connected together by the way.

[CopperTop]

Originally Posted by Ouroboros

It is a CMOS opamp used as an integrator with a 2.2Meg input resistor and a 680nF feedback capacitor, giving a time constant of 1.5 seconds. The output of the op-amp is connected via a 120k resistor to the amp input. It can have no effect on the audio quality when the amplifier is operational, but a change of op-amp may well change the behaviour of the amplifier at switch-off, causing a nasty LF thump as the power rails collapse.
The Quad schematic certainly shows the two pins connected together by the way.

Thanks Ourboros.

I realise that the op amp has no effect on the sound, but any explanation as to why pin 8 is connected to pin 6 and not tied to a 'fixed' voltage? I don't suppose there's any chance at all that they are dynamically varying the op amp's characteristics at power up/down to prevent instability/'thump' etc..?

[arupg]

My work on the 606-II abt a year back led me to accept the fact that any deviation from the stock opamp, and you'll definitely get severe power up and down problems that can ruin your speakers too.

I have left the stock opamp in its place and have attacked at all its weaknesses in sound quality by performing other mods. All in all, the Quad amp sounds waaaaaayyy too good. No complaints whatsoever. Sad that some of the passive parts I used are no longer available (even if u throw your $$$).

Or else I would build up a 2nd 606-II with all the mods for the purpose of vertical bi-amping. (That ain't gonna happen anymore).

[Ouroboros]

Yes, that aspect of it is baffling! The TLC271 has very different characteristics at the different bias settings, from a slew-rate of 3.6V/usec at high bias, down to only 0.05V/usec at low bias. Even though the output pin will only change its level very slowly, I don't see what is gained by this. If you were modifying the board, I would connect the bias pin to medium bias (OV).
By the way, the standard quad 606 uses simple shunt 6.8V zener diodes to provide the op-amp supply. If you fit an op-amp taking any more than 4mA supply current, the rails will go out of regulation. I know people sometimes add an NPN and a PNP buffer to the regulators to beef up the supply current, though you would need to change to 7.5V Zeners to keep the rails at the specified voltages.