[noviygera]

Forgive me if this is a stupid question. Since it makes sense from a sound quality perspective to replace resistors with transformers in an attenuator as Bent audio does, does it make any sense to follow the signal path and further replace resistors with tiny transformers?

For example, in an opamp based phono stage there are resistors in the passive RIAA eq circuit. Probably 1/4w resistors right in the sensitive signal path. Why not replace them with small possibly surface mount transformers?
Does this make any sense?

Thanks, herman

[FullRangeMan]

It make sense to audio quality, but transformers are expensive and may need made to order.
Few audio builders use transformer as volume control, an example:
http://spbsound.narod.ru/SPbSound/schematic/schematic2.html

[Speedskater]

Not that anyone could afford to build that circuit, but if one were to do so.  They should connect the input circuit as a balanced interconnect.
That is, no input connection to circuit common (ground).
For balanced:
XLR pin 2&3 to transformer
XLR pin 1 to chassis
For RCA unbalanced
RCA pin to transformer
RCA shell to transformer & to chassis

[John Chapman]

Hello!

Transformers are used where you have AC signals in the circuit - so they can't be used in place of resistors in all that many places. The TVC's are used to lower the voltage of the AC (music) signal - an alternate to using a standard resistor pot which at any given setting is a voltage divider with 2 resistors from the circuit point of view. You example of the RIAA is a spot where they would not fit application wise - although there are some really nice phono stage circuits out there that use inductors and caps to set the RIAA curve you could not just take an existing RC RIAA design and replace the resistors with transformers and have it work out. Comes down to the application and for most of them the use of a transformer is an integral part of the circuit design - TVC's and autoformer volume controls are an exception in that they can almost always be 'inserted' into the circuit at the location of a resistor control without other circuit changes needed. Like the circuit Full RangeMan posted the level control is most often located at the input to the circuit so the resistor control can be interchanged  with the AVC or TVC. If it is 'burried' in the middle of the circuit (rarely the case) then you would indeed need to understand the circuit in detail to be able to safely swap one type of level control with the other.


Thansk!

John

[noviygera]

Hello John,

thank you for the detailed reply. So yes, I brought up the topic because of my phono stage. It's a simple one, and I was curious to see if at all possible to improve it. To remove/replace those 2 electrolytic caps in the signal path and, looking at the active EQ section, to replace the resistors with transformers. Is it possible with this type of opamp phono stage?
I think it would be interesting to see how good a single opamp based phono can sound... this is why I brought up the topic.




-Herman

 

[John Chapman]

Hello!

I have had a few oem customers making very nice high end gear that have mentioned op amp based phono stages over the years that they really like - and these are guys who are great designers making tube gear. That says that they can be pretty decent I think. As for transformers you would not be able to drop one into that circuit - the RC based EQ wouyld be hard to replicate with a transformer in a design like that......

Have fun playing!

John

[Pete Schumacher]

Hello John,

thank you for the detailed reply. So yes, I brought up the topic because of my phono stage. It's a simple one, and I was curious to see if at all possible to improve it. To remove/replace those 2 electrolytic caps in the signal path and, looking at the active EQ section, to replace the resistors with transformers. Is it possible with this type of opamp phono stage?
I think it would be interesting to see how good a single opamp based phono can sound... this is why I brought up the topic.




-Herman

There are actually 3 electrolytic caps in the signal path, all of them performing DC blocking functions.  The 1uF caps would be easy to replace with some high quality film types.  The big ones, 47uF and 470uF would need to be HUGE if they are to be film types.  So replacing them would likely require an additional op amp to do DC servo control to make sure that any DC drift would be compensated.

Otherwise, choosing a very low-offset op amp, or one that is easily adjustable, and that is also quite temperature stable, could be substituted and the caps eliminated resulting in a DC coupled circuit. 

The circuit could also be reworked somewhat and those bigger caps replaced with 1uF caps.  That would take a small change to the feedback network around the op amp, but the cost is only a single additional resistor.

[noviygera]

Pete,

>>The circuit could also be reworked somewhat and those bigger caps replaced with 1uF caps.  That would take a small change to the feedback network around the op amp, but the cost is only a single additional resistor.

How would I do that with a single additional resistor? Then the 470uf and 47uf can be substituted with 1uf film caps? Maybe it's worth it... need to try.

Thanks. Herman

[Pete Schumacher]

Pete,

>>The circuit could also be reworked somewhat and those bigger caps replaced with 1uF caps.  That would take a small change to the feedback network around the op amp, but the cost is only a single additional resistor.

How would I do that with a single additional resistor? Then the 470uf and 47uf can be substituted with 1uf film caps? Maybe it's worth it... need to try.

Thanks. Herman

Looking at the (-) input to the op amp, you'd hang a large resistor from there to the output, a few MegOhm.  Then you'd add a 1uF cap between the (-) input and the feedback node that used to be connected to the (-) input.  The op amp input would ideally consist of something like low noise J-fets so the input impedance is extremely high.  Then you can drop the 470uF cap completely since the additional components are performing the DC blocking function.

That large feedback resistor will keep offset voltage low by closing the loop with a unity gain feedback at DC.  As frequency goes up, the original feedback network dominates the voltage at the node.

[noviygera]

Hello Pete,

I am glad I found another phono schematic. Before I post my sketched diagram of your described modifications (I'm working on that) I would like to point out that this new schematic is of a famous (and reliable) mixer Xone 62 -- industry standard for sound quality for what it is. It is also a Single opamp but I believe they are doing what you propose, already. I see no coupling caps in the signal path, compared to the previous schematic. It seems this second one is a better phono, no? Better opamp, no caps, is this the best one can do with a single opamp? Can I then eliminate all coupling caps in the previous phono schematic in a similar way? In other words can I use this for ideas to incorporate into my phono?




regards,
Herman

 

[Pete Schumacher]

Hello Pete,

I am glad I found another phono schematic. Before I post my sketched diagram of your described modifications (I'm working on that) I would like to point out that this new schematic is of a famous (and reliable) mixer Xone 62 -- industry standard for sound quality for what it is. It is also a Single opamp but I believe they are doing what you propose, already. I see no coupling caps in the signal path, compared to the previous schematic. It seems this second one is a better phono, no? Better opamp, no caps, is this the best one can do with a single opamp? Can I then eliminate all coupling caps in the previous phono schematic in a similar way? In other words can I use this for ideas to incorporate into my phono?




regards,
Herman

This new schematic is very similar to the other except the absence of two coupling caps.  However there is still the large electrolytic C51 in the signal path.  It may not look like it's in the path but that cap is in the feedback circuit which determines the shape of the RIAA curve.

C51 would be the big offender then and would benefit by being replaced with either a high quality film cap or modifying the feedback circuit in a manner similar to what I described before so that one of smaller value can be used.

This circuit however is only calling for a 33uF cap.  Lower voltage film types aren't overly huge and should probably be substituted.  If you're building this yourself, just lay it out with a high quality 33uF cap instead of the lower quality electrolytic.

This would be a nice cap for that job.

http://www.mouser.com/ProductDetail/WIMA/MKS4C053306F00KYSD/?qs=sGAEpiMZZMv1cc3ydrPrFwgzP98AO9idlpSl64Eu2%2fg%3d

Pete Schumacher
Vapor Audio

[noviygera]

Dear Pete,

Going back 2 posts, I drew up the modified (per your instructions) schematic. Please take a look and verify that I understood correctly. If so I will proceed with both channels, I have the needed parts to do this. Only showing one channel mods for clarification.



Also, can I simply omit the (1uf) output cap, as done with no issues in the other, posted above, schematic? Or do you still recommend to keep it?

best regards,
Herman

[Pete Schumacher]

Dear Pete,

Going back 2 posts, I drew up the modified (per your instructions) schematic. Please take a look and verify that I understood correctly. If so I will proceed with both channels, I have the needed parts to do this. Only showing one channel mods for clarification.



Also, can I simply omit the (1uf) output cap, as done with no issues in the other, posted above, schematic? Or do you still recommend to keep it?

best regards,
Herman

You almost got it.  The input and output caps can be omitted but you need to move one component.  R121 needs to connect to the node that connects R129, C117 and R125.  So disconnect R121 from the (-) input and move it to the other side of the new 1uF cap.

The only components connected to the (-) input will be the 2M resistor and the 1uF cap.

That particular circuit does have what seems to me an unusually high low frequency gain.  As it stands, the gain is 60dB which seems a bit high to me.  The other circuit, with the 390 Ohm feedback resistor had a more reasonable gain.  Depends on the cartridge of course, but if its putting out 2.5mV, then the circuit with the 100 Ohm resistor in the feedback will be putting out more than 2V which may or may not be an issue depending on how the downstream volume control is implemented.

Were it me, I'd increase R121 to at least 150 Ohm.  That will lower the gain from 1001 to 668, more in line with the circuit using 390 Ohm in the feedback.  And consider increasing the 2M resistor to perhaps 5M.

[noviygera]

Dear Pete,

Like this?


I updated R121 to 150ohms, but can do any value now that I am changing it. As far as the stock opamp -- it's mc4580.
http://www.unisonic.com.tw/datasheet/MC4580.pdf

Is there any point of replacing it with NE5532? Just going to be a different sound, not necessarily better? Should I bother with this, in you opinion?

Thank you for all the help! I am looking forward to comparing stock with your mods, and plan to clearly post my findings.

Best regards,
Herman

[Pete Schumacher]

Dear Pete,

Like this?


I updated R121 to 150ohms, but can do any value now that I am changing it. As far as the stock opamp -- it's mc4580.
http://www.unisonic.com.tw/datasheet/MC4580.pdf

Is there any point of replacing it with NE5532? Just going to be a different sound, not necessarily better? Should I bother with this, in you opinion?

Thank you for all the help! I am looking forward to comparing stock with your mods, and plan to clearly post my findings.

Best regards,
Herman

That schematic is correct.  And rather than the NE5532, go with something like the LT1113 which has J-Fet input devices and a much higher input impedance.

[*Scotty*]

The OPA1611 or OPA1612 with 1.1-nV/√Hz noise and a THD of  0.000015% might also be a good choice. This same opamps are used to good effect in the Auralic Vega.
Scotty

[Pete Schumacher]

The OPA1611 or OPA1612 with 1.1-nV/√Hz noise and a THD of  0.000015% might also be a good choice. This same opamps are used to good effect in the Auralic Vega.
Scotty

My only concern with those particular Op-amps is that they do not employ J-fets for input, and with the large feedback resistor (5M) a higher input impedance will reduce output offset voltages.

That said, if the builder isn't averse to adjusting the offset once the circuit is built then the OPA1611 would be an excellent choice.

[noviygera]

Hello again, just wanted to thank you for the help provided regarding the above project. I am on a little break and will return with results of your advice in a few days. Happy holidays!

[Folsom]

Input series resistors are normal to help with RF. However I found an LDR was vastly, crazy, superior to a Caddock TF series. The question is if the small amount of capacitance that they have, has something to do with it.

A capacitor multiplier feeding the phono section/s would be extremely helpful. A lot of the things currently played with won't be as important if you can keep the noise down. For example a super low noise OPAMP doesn't necessarily sound good at all, and some discrete transistors are better but neither go over well in a higher noise environment where you're trading off quality for proper noise voltage division. 

[noviygera]

Years pass, and I practice smd soldering. I am back to this phono stage mod!

post#13 was a recommended modification to original Ecler phono stage. Is this still holding true today? Any other recommendations to reduce the noise and improve the cheapo original circuit?

thanks!
Herman