[LokiS]

Quote:

 

Thanks for the information, I currently have the amp at 80%-90% on the knob with a 18db setting and the volume to 50% on PC. So far it seems fine, I'll see how it goes after using it for a bit and report back if issues occur.

 

[K3cT]

Quote:

Suggestions?
I only have MKP4's and MKS2/MKS4's myself.
 
Also, would bypassing the input caps be good?

 
Best caps is no caps so if the original design doesn't have it then don't use it I suppose.
 
And I guess we're talking about the row of cyan-colored 'lytics behind the MOSFETs? I don't know what brand they are but I have good results using Panasonic FM, Elna Silmic II or Nichicon KZ so you can replace that with those if the size permits. Silmic II tends to be big though so be aware.
 
After hours and hours of trial and error with the analog output of my AMB γ2, my conclusion is that bypass is bad and should be skipped whenever possible. Just use a good quality cap and call it a day. 

 

[purrin]

Quote:

Suggestions?
I only have MKP4's and MKS2/MKS4's myself.
 
Also, would bypassing the input caps be good?

 
The input caps are already bypassed with the small caps right next to the big 1uF caps. MKP4s are probably similar to the existing bypass caps. but you can always experiment though.  The cyan colored caps behind the transistors are not the input caps - they are the power supply caps.  The input caps are the big blue / red pair just south of the inputs (assuming inputs are at the top).

 

[purrin]

Quote:

I really don't suggest bypassing with small films as it has the tendency to smear the treble.

This has not been the case here. I've always had good results bypassing 'lytics with smaller film (even tiny ones) in power supplies. I hear you shouldn't do this with boutique parts though (bypass Blackgates). Although you got me wondering though.

 

[K3cT]

Quote:

This has not been the case here. I've always had good results bypassing 'lytics with smaller film (even tiny ones) in power supplies. I hear you shouldn't do this with boutique parts though (bypass Blackgates). Although you got me wondering though.

Ah, power supplies might be a different case. My experience was mostly based on the analog section of things but I still wouldn't do it though.

 

[Mad Max]

Quote:

This has not been the case here. I've always had good results bypassing 'lytics with smaller film (even tiny ones) in power supplies. I hear you shouldn't do this with boutique parts though (bypass Blackgates). Although you got me wondering though.

 
Prickley Peete says he likes to bypass BG's with BG's.
 
Yes, I know the cyan caps are the power caps, lol.

 
Quote:

 
Best caps is no caps so if the original design doesn't have it then don't use it I suppose.
 
And I guess we're talking about the row of cyan-colored 'lytics behind the MOSFETs? I don't know what brand they are but I have good results using Panasonic FM, Elna Silmic II or Nichicon KZ so you can replace that with those if the size permits. Silmic II tends to be big though so be aware.
 
After hours and hours of trial and error with the analog output of my AMB γ2, my conclusion is that bypass is bad and should be skipped whenever possible. Just use a good quality cap and call it a day. 

Have you ever played with Rubycon MCZ or YXH?  How do they compare to the KZ's and FM's?  I can't figure out what brand and series those cyan caps are.  I don't remember if KZs come in 25V, I've only seen 50V and up.  I think I will go for FM's as far as the opamp power caps.

 

[K3cT]

Quote:

Have you ever played with Rubycon MCZ or YXH?  How do they compare to the KZ's and FM's?  I can't figure out what brand and series those cyan caps are.  I don't remember if KZs come in 25V, I've only seen 50V and up.  I think I will go for FM's as far as the opamp power caps.

I've never encountered them here unfortunately. Well, Panny FM is a perfectly fine cap for power supply so go for it.

 

[francisdemarte]

Quote:

This has not been the case here. I've always had good results bypassing 'lytics with smaller film (even tiny ones) in power supplies. I hear you shouldn't do this with boutique parts though (bypass Blackgates). Although you got me wondering though.

Can you post some pictures of this mod when you get a chance?

 

[purrin]

Yup - Will do when I get back home this Friday.

 

[dlaloum]

OPAMP Mods / Rolling
 
So  I ordered an M-Stage from JTAM
 
He also helpfully (at my request) added both a Class A Mod socket and a twin OPA627 socket to his pricelist....
 
While waiting for these to arrive... I have a few questions...
 
1) What opamps respond best to Class A modding?
2) How much Class A? (ie what resistance) - and is heatsinking therefore required?
3) are twin singles + Heatsink a better option for this sort of thing ? (does that allow biasing deeper into class A ?) - this assumes the singles are DIP8 on side by side adapter, not top bottom adapter.
ie: split heat load into 2 seperate dips with larger hear dissipation surface area, and potentially provide larger surface area for heatsinking.
 
I have also read some references in the forums to some opamps already being Class A biased ... is this right? or is this someone misinterpreting Class AB for Class A? (Pure Class A would probably run WAY hot!)
 

 

[jrosenth]

Quote:

V2 is said to sound tad better.

Who said that?  
 
I know they  extended the board, etc. and changed the transformer (cased, smaller) but never saw whether that transformer had different/better specs/output and must have missed (after reading 60 pages) where folks said they heard head to head that it actually sounded better as opposed to repeating supposition that lack of bare wire must shield it better.

 

[purrin]

 
Quote:

Originally Posted by dlaloum /img/forum/go_quote.gif
 
1) What opamps respond best to Class A modding?
2) How much Class A? (ie what resistance) - and is heatsinking therefore required?
3) are twin singles + Heatsink a better option for this sort of thing ? (does that allow biasing deeper into class A ?) - this assumes the singles are DIP8 on side by side adapter, not top bottom adapter.
ie: split heat load into 2 seperate dips with larger hear dissipation surface area, and potentially provide larger surface area for heatsinking.
 
I have also read some references in the forums to some opamps already being Class A biased ... is this right? or is this someone misinterpreting Class AB for Class A? (Pure Class A would probably run WAY hot!)
 

• OPA2107 and OPA627 respond very well to Class A. OPA2134 amd LM4562/LME49720 slightly less so, mainly because OPA2134, although pretty good, isn't all that great and can't be made great. The LM4562/LME49720 is already very good to start with. These are the only ones I have tried.
• Aim for between 2mA and 5mA - you can push to 7mA if you really crazy and curious.  Just 2mA (assuming the op-amp is feeding a suitable buffer/output circuit) gets you the sonic benefits (it's probably going to run class A 100% of the time). Adding more juice may change the sound balance, but you need to experiment yourself. Divide 15V by current (amps) to get resistor value. These op-amps will survive in hot environments up to 80C, but this is not ideal. With a 2.2k resistor (6.8mA), I did not need a heatsink, but that was with ambient temps at a cool 62F degrees at night. I've run 3.3k (4.5mA) in 95F ambient temps with no problem.  
• Top/bottom adapters may have more problems with heat. The SMT chips are smaller, and therefore have less surface area for heat dissipation. Also the bottom chip won't be able to breath.  I'd be conservative with top/bottom adapters and run 3.9K (less than 3.8mA) just to be safe, especially if you are running expensive chips. There are two advantages to running the top/bottom adapters compared to side-by-side DIP8 adapters: you are keeping the chips as close to the socket as possible; and you are limiting the number of socket contacts to only one set. If it means anything, soldering the resistors directly to the op-amp and plugging the chip straight into the socket sounds better than plugging the op-amps over a class A adapter and then into the socket. I can surmise better electrical contact (only one set vs two) and closer distance to the small PS decoupling caps (the small value 22nF probably takes into the account a good layout on the PCB). Some contact cleaner helps too.

 
Finally, yes there are some chips that can easily (or more easily) be configured to run class A.

 

[dlaloum]

Thanks Purrin
 
I have an M-Stage with 3k9 class A adapter and OPA627's on Brown dog on the way - I assume the OPA627's will be top/bottom 
That will be the starting point of my listening (opa2134 stock, then OPA627 class A)
 
Although I do own a Soldering iron it is not something I have used in the last 15+ years....
 
But I might pop down to the local electronics supermarket and get some addition sockets so I can make up a few with different resistors.... (risk level is low there...)
 
Depending on how I go we will see whether I feel brave enough to mod an adapter directly!
 
I lashed out and ordered a grab bag of different opamps to try out - Would love to have a table of opamps and recomended Class A resistors for best results.... But I think that is a pipe dream!
 
My grab bag will include (early next week when it all rocks up!)
 
OPA627
OPA637
AD797
AD744
LT1028
LME49860
OPA2107
OPA2111
OPA228
OPA49720
 
So this little exercise is likely to take some time...
 
Any experiences people have had with the above in the M-Stage and especially with the M-Stage and Class A mod very much appreciated
 
But as the saying goes " Nothing succeeds like excess, and nothing exceeds like success"

 

[dlaloum]

Quote:

So I just swapped an AD797BR into the M-Stage. The main difference that I notice is that gain seems to be lower now (subjectively need to turn up further to get the same volume out of it) - is that possible or is there something wrong?

Hi Buz
 
there was never any follow up to this post.... how did you go in the end with the AD797BR?
(I have some on the way to me.... )
 
thanks

 

[daisangen]

I just replaced LM4562NA with a LME49720NA and it's definitely bassier than the LM4562NA. I know they are supposed to be identical, but there's no way in hell this is placebo. I tried going back to LM4562NA and yep, my findings still stand. Other than the bass amount, I think they sound very similar.