[Luis MC]

Hi,

     I have tested with MOS125 and BUFAMP preamplifier the OPA132 and OPA627.
     Changes with OPA627 is a great upgrade. Highs are more natural and clear, and soundstage is better.
     Next week or so, I can try the AD825, AD8610 and AD8065.

     Regards,

     Luis.

[Raj]

Hi Luis,

Have you upgraded the resistors as David suggested (to supply more current to the ouput devices so the amps slew rate increases)? I have not done this yet, currently I am running with opa627's, but I am considering ad825's or 8610, let me know how they compare to the 627's please!

Thanks
Raja

[Luis MC]

Hi Luis,

Have you upgraded the resistors as David suggested (to supply more current to the ouput devices so the amps slew rate increases)? I have not done this yet, currently I am running with opa627's, but I am considering ad825's or 8610, let me know how they compare to the 627's please!

Thanks
Raja

Hi Raj!

With OPA627 I understand no changes are needed with quiescent current in driver transistors (for raising slew-rate).
When I put more fast opamps I must be change quiescent current.
Of course, I tell you what sound I get with these changes !!

Regards,
Luis.

[Raj]

Hi Luis,

I changed the resistors that David mentioned in my amp (mos 250),

R19 and R22 from 220R to 180R; changed R24 from 470R to 390R (3W at least ); and change R18 from 82R5 to 75R ( 1W )

At the moment I am still using the opa 627's, but the ad825's should arrive  soon, I was quite pleasently suprised at the difference the resistor changes alone made, worth doing whether you wish to change opamps or stay with the opa 627's. Will report back to let you know how the amp sounds with the ad825's. Any news from you yet Louis, on the other opamps you mentioned?

Thanks
Raja

[Luis MC]

Hi Luis,

I changed the resistors that David mentioned in my amp (mos 250),

R19 and R22 from 220R to 180R; changed R24 from 470R to 390R (3W at least ); and change R18 from 82R5 to 75R ( 1W )

At the moment I am still using the opa 627's, but the ad825's should arrive  soon, I was quite pleasently suprised at the difference the resistor changes alone made, worth doing whether you wish to change opamps or stay with the opa 627's. Will report back to let you know how the amp sounds with the ad825's. Any news from you yet Louis, on the other opamps you mentioned?

Thanks
Raja

Hi Raj,

     I have testing the AD8610, AD825 and AD8065 in the Buffer Amplifier.  I am having problems with the stability of the AD825 and AD8065, perhaps they are damaged or are oscillating by other causes (at 14 MHz region).  The AD8610 works well, and the sound is impressive, seems more natural to me and a richer and realistic tonality than the OPA627, that has a flatter sound.

     Regards,
     Luis.

[Raj]

Hi Raj,

     I have testing the AD8610, AD825 and AD8065 in the Buffer Amplifier.  I am having problems with the stability of the AD825 and AD8065, perhaps they are damaged or are oscillating by other causes (at 14 MHz region).  The AD8610 works well, and the sound is impressive, seems more natural to me and a richer and realistic tonality than the OPA627, that has a flatter sound.

     Regards,
     Luis.

I have the buffer amp also, I've never tried opamps other than opa627 in this position. With faster opamps you may be getting oscillation due to the lack of small decoupling caps at the power input pins of the opamp. David should be able to provide you with info as to what size and type of cap to use, I think the oscillation is curable.

Thanks
Raja

[Raj]

Hi,

Having made the above changes to increase the slew rate of the mos250. I changed over to the ad825's - they far outperform the opa627's and help demistify the obsession of percieved sound quality versus distortion argument. Looking at the spec's of opa627's and indeed the prices one would be lead to believe that they are the bona fide top dog's of opamps, replacing them with the ad825's just puts them to shame in every aspect of presentation. Keeping spirits up though, the opamp change takes the performance of the mos250 to a whole new level.

Thanks
Raja

[Luis MC]

Hi,

Having made the above changes to increase the slew rate of the mos250. I changed over to the ad825's - they far outperform the opa627's and help demistify the obsession of percieved sound quality versus distortion argument. Looking at the spec's of opa627's and indeed the prices one would be lead to believe that they are the bona fide top dog's of opamps, replacing them with the ad825's just puts them to shame in every aspect of presentation. Keeping spirits up though, the opamp change takes the performance of the mos250 to a whole new level.

Thanks
Raja

Raj,

I agree, today I put an AD825 in my MOS (modified SR) and sound is really impressive !! OPA627 sounds some dead to me, AD825 adds life and rich and full tonality.
If I can, I will try AD065 next days.

Regards.

[DVV]

You guys should also try AD826AN. Outstanding! Puts most Burr-Brown op amps to shame.

I may have missed it, but exactly what will those op amps be doing?

Cheers,
DVV

[Luis MC]

You guys should also try AD826AN. Outstanding! Puts most Burr-Brown op amps to shame.

I may have missed it, but exactly what will those op amps be doing?

Cheers,
DVV

Hi DVV,

     AD826 is a double opamp, it's no suitable for MOS125.
     
     Regards,
     Luis.

[DVV]

Hi DVV,

     AD826 is a double opamp, it's no suitable for MOS125.
     
     Regards,
     Luis.

Which is why I asked what do you need it for. Both application and circuit.

While it's an outstanding op amp, it has only average noise figures, which rules out its use in noise critical operations, such as, for example, phono RIAA eq amps, where you'd be better off with say LT1115.

But if you want a very high quality line stage, the fact that it's a dual op amp allows you to use one half as an audio amp and the other half as a DC servo amp, thus kicking out any and all capacitors from the direct signal path.

Cheers,
DVV

[Raj]

Hi Dvv,

no technical genius but I'll try to explain as I know it (may be wrong), the opamp is at the input of the power amps and provides some of the voltage gain, it is also in the feedback loop of the amplifier, in the case of the mos 250 I think we only have local feedback and perhaps for the mos 125 too. Perhaps David can explain a bit better.......

Thanks
Raja

[Raj]

Hi Dvv,

no technical genius but I'll try to explain as I know it (may be wrong), the opamp is at the input of the power amps and provides some of the voltage gain, it is also in the feedback loop of the amplifier, in the case of the mos 250 I think we only have local feedback and perhaps for the mos 125 too. Perhaps David can explain a bit better.......

Thanks
Raja

[Luis MC]

Hi Dvv,

no technical genius but I'll try to explain as I know it (may be wrong), the opamp is at the input of the power amps and provides some of the voltage gain, it is also in the feedback loop of the amplifier, in the case of the mos 250 I think we only have local feedback and perhaps for the mos 125 too. Perhaps David can explain a bit better.......

Thanks
Raja

Hi Raj,

     MOS125 have several feedback loops.
     IC1 have "local" (?) feedback using R5 resistor.
     There is a global FB loop using R21 / R3 network, carrying signal from output to inverting input in IC1.
     And there is some feedback from output to Tr4 and Tr5 emitters through R23/R24 network.
     I see three ways for negative feedback.

    Regards.
    Luis.

[davidw]

Sorry guys, been a bit remiss in keeping up with the forum - too many people distracting me by ordering stuff. Firstly let me say that Raja is right about the mods for increasing the standing current through the output stage drivers, its worth doing even if you don't experiment with the opamp because it ensures plenty of drive current for the mosfets at high frequencies. I'm going make this change permanent for all new kits once I get around to changing the manual.
Luis is also correct in his analysis of the feeback loops. There are three of them, one around the opamp, one around the output stage, and a global feedback loop.
The discussion about the AD825 is very interesting because I'm coming to the conclusion that slew rate is much more important than previously realised. I first came across this when I swapped the OPA551 for a LM6171 in the previous version of the headphone amplifier. The LM6171 blew the socks off the OPA551 and the only difference in the specifications of the two opamps is that the LM6171 ( 3200V/us ) is very much faster than the OPA551 ( 20V/uS ). The difference was most obvious on massed orchestral or choral works - with the LM6171 you could pick out every instrument and/or singer individually, but with the OPA551 it sounded like a muddled mess by comparison. All the textbooks tell you that 20V/us is more than enough for any audio system because transients that fast never appear in real music. The fact remains though, that amps with fast slew rates always have lower intermodulation distortion than a version of the  same amp with a lesser slew rate ( all other things being equal ). I think that this is at least part of what you may be seeing with the AD825 as its faster than the OPA627. The open loop bandwidth of the AD825 is also much greater than that of the OPA627 ( kHz as opposed to Hz ) but this doesn't appear to have the beneficial effect that you might expect on the hf thd. Incidentally the LM6171 won't work in the MOS125, even with the modified output stage. Its stable on the bench with no input but as soon as you bring a hand near to the pcb or provide an input signal the amp bursts into oscillation.

[Matthew Wilson]

Hi fella's, I read with great intrest your musing's over these i/c's. One that popped up recently was AD8066 and on the face of it looks like it should be better still than AD825 which I'm currently using in my cd player on conversion boards to dual op-amps. The funny thing is a year and half ago when I first tried these 825's, 627's were so expensive I gave them a miss and always wondered if I'd missed out. Apparantly not. If you've tried 8066 let me Know what you think....Realised I made a boob 8066 is dual 8065 is single, also really interesting is the new AD8099. Take a look on thier site at the spec's!!!

regards to Dr White and you all. I'm pleased to see things are going well.

[Raj]

HI David,

Although I'm really happy with the ad825's, I'd like to see how they measure up against even faster opamps, the  ad8065 looks interesting, as does the ad8099, the latter has a very fast slew rate in excess of 1000v/us. The ad8065 looks like an improved ad825, however it requres lower power supply, 13v being the maximum, I haven't checked the power supply requirements of the ad8099's....

1) Looking at the spec sheets of these opamps do you think either of them would suit the mos250?

2) should one desire to try these opamps which resistors would need to be changed to get the required voltage rating?

3) What is the safe maximum with regards to speed of the opamp, so that nasties like oscillation can be avoided?


Thanks
Raja

[davidw]

I've got some ad8065s so I'll give them a go sometime during the next few weeks and let you know how they pan out. The ad8099 is ( unfortunately ) a no no because the supply voltage is only +/-5V max. This might be ok for a low power class A variant of the mos125 but won't give enough voltage swing for normal use. I don't know how fast one can go with the opamp. Certainly 1-200V/us is perfectly ok but 3200V/us isn't. I'll kep you posted as I try more opamps

[Raj]

Hi David,

Any news on the ad8065's yet or others???

Thanks
Raja

[DVV]

As a general comment, an op amp's rail window is generally not a problem in technical terms. This is a limitation which can be overcome rather easily. I stay reserved and say "generally" because obviously, to increase the overall voltage swing from a +/- 15 V op amp to say +/- 50V or whatever, you will need additional electronics to "translate" the voltages.

This can be a problem, depending on the implementation. For example, if it's a phono RIAA eq/amp stage you are making, additional electronics mean additional noise, exactly what you do NOT want in that case. But if it's a power amp, the little additional noise will be quite insignificant even in absolute terms, let alone in terms of what you gain.

Regarding the slew rate, I'd like to make a comment. The actual slew rate on its own is NOT a problem; calculate what you need for a +/ 12V output swing (typical for +/-15V op amps) at TWICE the max frequency, i.e. not at 20 but at 40 kHz, and you come up with a number of +/- 4.25V/uS, or 8.5V/uS. Thus, say 80+% of modern op amps make the grade hands down.

The problem is not in the sheer value of the slew rate, but much more in the way HOW it has been obtained. It could have been done by biasing the input pair rather high and that's it, or it could have been done by making the whole op amp really fast. Compare say LF357 (first group, FET inputs) with say AD817 (second group, fast all around) and you will notice that the full power bandwidth varies by far more than the slew rate specs would lead you to believe.

In other words, you don't want slew rate as such if it's above 15V/uS, you want a wide bandwidth. OP27, for example, has a slew rate of 4V/uS, but a bandwidth of 60 MHz (sic!), which explains why it sounds great even if it appears to be a slowpoke.

But a high slew rate can be a problem, and a serious one. If it has been achieved by having a wide bandwidth, any such circuit is more and more prone to oscillation than a limited bandwidth one. Also, such cicruits will pick up RF far easier than slower ones. All this means is that to apply properly, they need special consideration and precautions, which includes outlay - exactly what you can't change when modding.

Lastly, what you definitely MUST look at, in my view much harder than the slew rate, is the settling time of the op amp. Yes, we want them to be fast and have good rise times, but by the same token, we want them to settle down as fast as possible once the signal has passes; if they tend to take their time, or overhang, your ambience and spatial information goes out of the window.

Typical op amps have settling times of 2,000 nanosecs, which is long. Better op amps (e.g. OP275) cut this to just 1/10, 200 nanosecs, and really cool cucumbers like AD817, AD825, AD826, etc decrease this further to just 45 nanosecs. 45 nanosecs is 44 TIMES faster than 2,000 nanosecs. Hardly surprising then that AD826AN has such wonderful voicing of ambient and spatial information, if applied properly.

Just a few thoughts.

Cheers,
DVV