AD797 not sounding good

[belgiangenius]

I dropped in some AD 797 into an Adcom GDA-700. Circuit diagram attached - see page 4.

I recapped the unit and added 100uF Silmic and 0.1uF MKP bypass to all V+ and V- pins.

In short, it doesn't sound that great. I put some OPA627 in a similar unit last year and Frankly the 627s smoke the 797.

I know the circuit implementation of the DAC is probably not ideal for the AD797. Any suggestions on simple changes to make this work?

 

[Ken Newton]

The AD797 is very high performance but, as a consequence, is somewhat notorious for exhibiting very high frequency oscillation unless carefully implemented. You would want to have low inductance bypass capacitors located right at the chip's supply pins. Also, try placing 100 ohm resistors in series with the 120pF feedback loop capacitors, C440 and C441.

 

[jcx]

this I/V circuit requires a unity gain stable op amp - the AD797 isn't without added parts that would also affect the response

I wouldn't be satisfied applying the mods with a AD797 without looking with a >50 MHz 'scope to verify

and the AD797 input current noise isn't suited to the >5k feedback R, the PCM 1702 low I out
some fet input op amps will give lower noise in this circuit than the AD797

 

[Mark Johnson]

The 3rd order Sallen-Key lowpass filters, operate their opamps at gain=+1.00 which often leads to stability issues with the AD797. I'd suggest you leave those as non-797s and only use 797s for the I-to-V stages on the left of page 4.

 

[anatech]

Hi belgiangenius,
Then remove the AD797's and put the OPA627 op amps back in. You have to accept that there is no one "best op amp". Sometimes other op amps will work better in some applications.

-Chris

 

[DPH]

Stole the words right out of my keyboard, Chris.

But there are a number of bad-for-the-application opamps. This certainly seems one case.

As has been flogged, the AD797 needs decent attention to what surrounds it. Incredibly linear when used right? You bet.

 

[belgiangenius]

I see the recommended 100 Ohm resistors for C440 and C441 based on the data sheet. What happens when these are left out as they currently are?

Sounds like, when you add up these responses, the short of it is that the 797s aren't suitable for either of these stages. And, I take it, no simple modifications will make these work acceptably?

The OPA627 were actually used in a GDA-600 and are SOIC mounted on dual adapters, so aren't suitable for this unit.

I do, however, have some LME49710 available soon. Anyone see a problem for those?

 

[DPH]

AD711's are FET-input. That, along with the feedback network's impedance suggests like-for-like is your best recourse.

I'd be looking at the OPA1641 instead. You'll have lower noise for sure than the 49710.

 

[belgiangenius]

Looks like OPA1641 not available in DIP, which makes it a giant PITA.

Anyone think 49710 will do well in this unit, or should I spring for more OPA627?

 

[DPH]

I'm honestly unsure what modern DIP FET-input opamps are available.

Audio Operational Amplifier | Products | Amplifiers and Linear | TI.com

opa134, opa604 look the best option if you don't want to use a pdip-soic8 adapter.

 

[anatech]

Hi Daniel,
I agree with your take on things also.

Hi belgiangenius,
Try the OPA134 J-Fet op amps in this one. I know you'll like the results. The LME49xxx series are BJT inputs, so not really suited if the circuits did need the J-Fet type op amps. The OPA134 is unity gain stable.

There isn't anything wrong with socket adapters. Sometimes they are the only choice.

You really do have to examine the spec sheets for the original device and your candidates. When you're done that , read the application notes. What you have doe here was merrily picked a "best op amp" and stuck it into the application without configuring the circuit for it. Everyone answering your question does read the data sheets and the application section. Some of us even read the general application notes that these companies generate at great expense, but they are free tools for you to use.

-Chris

 

[belgiangenius]

Thanks, everyone.

I'm going to spring for the OPA627. Some reviews characterize the OPA134 as being very similar to the OPA627, with the OPA627 presenting perhaps the "last 5%" in improvement. Unfortunately, it is that last 5% that we audiophiles are pursuing.

 

[DPH]

Hopefully the board's decoupling is rather tight, otherwise the 627 might get sideways on you. Be cognizant of that--it's fast and may oscillate if the layout isn't so good.

 

[Mooly]

The AD797 was the device I used in this example:
http://www.diyaudio.com/forums/anal...u-have-checked-see-its-stable-havent-you.html

 

[belgiangenius]

The OPA627 worked beautifully in a GDA-600 i worked on last year.

The unit has 100 uF Silicon and 0.1uF MKP added to every V+ and V- pin.

 

[DPH]

Sounds good: I hope it works well for you!

Edit: I'd rather a SMT x7r to the MKP for lower inductance, but it should work.

 

[belgiangenius]

On reflection, maybe I will get AD4627. Same opamps, much cheaper!

 

[DPH]

Just make sure it's the 4627--the 37 is compensated into G= +5

That said, the datasheet shows no DIP package. Maybe older stock?

 

[belgiangenius]

Just make sure it's the 4627--the 37 is compensated into G= +5

That said, the datasheet shows no DIP package. Maybe older stock?

Dang! No DIP. That's out.

My local supplier only has 2 OPA627 left and a bunch of OPA604.

Considering buying the 2 OPA627 for the gain stages and 4 OPA604 for the I/V stage and balanced inverter stage. I assume the gain stage would be more influenced by the opamp than the I/V stage?

 

[DPH]

Other way around, typically. I/V is pretty demanding and a good/fast/emi-resistant opamp is best there. In audio, I'd say I/V and phono input are the two most demanding spots for an opamp.

That said, OPA604's or OPA134's will also work well throughout.