Rolling (replacing) Op-amps in Topping D10 DAC

[gorg]

Looks like collectively we may have misunderstood Topping's answer:



By "other" they may mean the soldered ones and hence the one we replaced is indeed the output buffer!

I will have to open it again to test to be sure as the answer without the question asked is still a bit ambiguous @Krunok, what did you ask him exactly?

Maybe it will be easy to check resistance (connection) between the socketed opa2134 outputs (legs 1&7) and appropriate "inputs" of relay next to rca sockets?

 

[amirm]

Maybe it will be easy to check resistance (connection) between the socketed opa2134 outputs (legs 1&7) and appropriate "inputs" of relay next to rca sockets?

That is my plan. Have to take the thing apart again though....

 

[RayDunzl]

The results were so close...

Does it work without the removable device?

 

[sonci]

The only time I definitely heard an improvement, was when replacing the opamps on my XFi soundcard, with some lme4562, but the original opamps were really crap.
then I got an Emu1820M, it was like diy heaven, 4 pairs of channels, you can route them to outputs through patchmix.
I replaced opamps on one channel and compared to the stock ones.
Not once I found smth that sounded better than the default one, some of them sounded really bad even expensive ones.
There's a lot of bullshit in the net about replacing opamps on this card, but you have to hear for yourself.
with that said, I believe you can a hear a difference when replacing opamps, so maybe the Topping is not the best choice for this test..

 

[restorer-john]

It's obvious from the PCB shot that the buffer is indeed the socketed opamp. There are even LPF test points for each channel I can see which line up via the SMD resistors to pin 1 (out) and pin 7(out) with an offset trace under the pcb most likely. There is zero doubt it is the buffer stage.

 

[amirm]

OK, I confirmed it. The socketed opamp outputs go to 1K ohm resistors that then go to the relay. And the relay controls the output. So for sure we are dealing with the output buffers, not I/V conversion. I will re-write the review to reflect this.

Edit: the series resistor is 100 ohm not 1K.

 

[Sythrix]

I wonder if there would be more of a change in a system where you can also swap out the I/V, like the ASUS STX II sound card.

As is, it just doesn't seem worth it, but I still wouldn't call it conclusively pointless without more testing on different devices.

 

[amirm]

Are they socketed?

 

[Addicted to music]

I’m starting to get the general consensus is that becuase the THD measures very close to each other for each opamp that it’s a waste of time rolling them....
Before you continue with this consensus I have to say I don’t have the equipment to measure but I sure use a CRO to tell me whether oscillation occurs. I can’t speak for the other opamps in the list but I have experimented with the OPA 2134, OPA 1612 and the LME Devices.
It’s no surprised to me that the THD results measure so closely. The OPA and the LME are all unity gain, the designer would have optimised the circuit for 02134 hence the slightly better measurement compared to the rest with data sheet guidelines. Most audio compliant Opamps uses a Feedback resistors to control gain and are heavily compensated.
But between the OPA devices here there is a difference in the input used where the 2134 uses FET and the 1612 uses BJT. To me in my subjective test; there is a difference in the SQ between the these 2 opamps.

 

[derp1n]

But between the OPA devices here there is a difference in the input used where the 2134 uses FET and the 1612 uses BJT. To me in my subjective test; there is a difference in the SQ between the these 2 opamps.

How would you descibe the audible differences?

 

[Addicted to music]

How would you descibe the audible differences?

I do not like the OPA2134. And you can see this with the latest products coming out from companies like Topping and SMSL. Notice they use the 1612

 

[Wombat]

I do not like the OPA2134. And you can see this with the latest products coming out from companies like Topping and SMSL. Notice they use the 1612

Possibly for a different reason.

 

[Addicted to music]

Possibly for a different reason.

There are many:

The 2134 is available in DIP the 1612 is only SOIC. What are the chances that Topping saw a benefit where the 2134 DIP version was heavily discounted and decided to purchased them just for this reason; opamp rolling! Just saying!
Notice in there D50 and Dx7s do not have the 2134 at all. The other reasons could be because with opamp rolling there is a chance that the product can be damaged by the client if it was inserted the wrong way, as there is no physical prevention to eliminate that possibility with 8pin dip devices.

 

[Sythrix]

Are they socketed?

The STX II? Yes.

 

[Deleted member 2944]

For info: I'm measuring a voltage gain of 3.81 (+11.6db) on the socketed OPA2134 in this unit.
It appears to be a standard non-inverting configuration. With a 100 ohm output resistance as already determined by Amir.

Dave.

 

[Addicted to music]

For info: I'm measuring a voltage gain of 3.81 (+11.6db) on the socketed OPA2134 in this unit.
It appears to be a standard non-inverting configuration. With a 100 ohm output resistance as already determined by Amir.

Dave.

Amirm said 1K output in post 26

 

[Deleted member 2944]

Amirm said 1K output in post 26

Yes, I misquoted Amir. Sorry about that.
But, what he did say was that the outputs "go to 1k ohm resistors and then the relays." The output resistance of the D10 is indeed 100 ohms. It's noted on the D10 specification sheet and I measured it to confirm. The 1k resistors are in some other usage (part of the relay muting scheme I think), but they're not load decoupling resistors.

 

[amirm]

Yes, I misquoted Amir. Sorry about that.
But, what he did say was that the outputs "go to 1k ohm resistors and then the relays." The output resistance of the D10 is indeed 100 ohms. It's noted on the D10 specification sheet and I measured it to confirm. The 1k resistors are in some other usage (part of the relay muting scheme I think), but they're not load decoupling resistors.

You are right. I misread the value in my quick glance. It is definitely 100 ohm. I corrected my post.

 

[restorer-john]

From the zoomed in figures I can make out 1000 on the SMD resistor, meaning it is a 100 ohm precision resistor. First three numbers are significant digits, last number is multiplier (power of ten).

If it were 1000 ohms on a 3 digit marking, it would be 102. Otherwise a 101 marking on a SMD 3 digit for 100 ohms.

 

[Deleted member 65]

Thx Amir, glad to see the Sonic Imagery opamp not failing. Think there's some solid engineering behind it.