Variability of RIAA eq with tube characteristics?

[leadbelly]

Thinking out loud: for a given phono preamp design, couldn't the accuracy of the RIAA equalization vary significantly with individual tube characteristics, both for passive and feedback type? Although maybe not the case for certain NOS tube types, I was thinking this would almost certainly be the case for the ubiquitous 12AX7 preamp since there are so many variants of the 12AX7 (e.g. long plate vs short plate, NOS vs Russian, etc.).

Just for background, I was researching the idea of making an improved Marantz 7 type phono preamp by putting a tube other than 12AX7 in the first position for lower input capacitance.

[EC8010]

Yes. RIAA accuracy is always affected by the valve type. The only way to (almost) eliminate the variation in a passive type would be to use a cathode follower to drive each network and a cascode after the network. With a feedback network, changing valve characteristics will change the open loop gain and because it's not infinite, that will affect RIAA accuracy.

ECC88/6DJ8 is a favourite input valve.

[Jeb-D.]

If it is a composite passive RIAA like in the schematic linked below
http://www.geocities.com/rjm003.geo/.../diy_pho3.html

Switching to a tube other than a 12ax7, the plate and cathode resistor should be changed to optimize the operating point for the new tube type.

To get back the proper RIAA curve the only value you should have to tweak is the one listed as R5 in the above schematic. The rest of the network values shouldn't have to be changed.

If your new tube has a lower Z-out you will have to increase the value of R5. If it has a higher Z-out you will have to decrease the value of R5.

[SY]

The idea is to make the eq as independent as possible of what happens to the tube over time or when it's replaced. The cathode follower suggestion of EC8010 is certainly an excellent approach. Another way which is nearly as good is to make the build-out resistor in the RIAA circuit (that's R5 in the schematic Jeb linked to) very large compared to the expected source resistance. If the first stage is run undegenerated grounded-cathode, that will be the plate resistance in parallel with the plate load resistor. If R5 is ten or more times bigger, then the eq will be fairly insensitive to a range of expected tube variations.

The plate resistance of an ECC88, for example, will be something like 3-4k. If we choose 3k5 as a nominal for the sake of illustration, variances of 500 ohms up or down will cause less than 0.05dB of error if the build-out resistor is 100k or so.

[Jeb-D.]

The plate resistance of an ECC88, for example, will be something like 3-4k. If we choose 3k5 as a nominal for the sake of illustration, variances of 500 ohms up or down will cause less than 0.05dB of error if the build-out resistor is 100k or so.

Yeah, if he adapts and ECC88 into a circuit designed for a 12AX7. Chances are he will end up with a large build-out resistor by the time the RIAA curve is corrected.

A cathode follower is a good idea if building from scratch (low Cin and Zout), but would probably result in excessive gain loss if adapting it into an existing circuit (that isn't already using one of course).

[SY]

No, you should be able to minimize the gain loss- after all, an ECC88 CF will have a gain of about 0.97. It's just a pita to shove one into an existing scheme, especially when considering heater-to-cathode voltage issues.

[leadbelly]

Thanks for the thoughtful replies! That pretty much answers my question for passive eq.

Dare I press my luck and ask how it would be done for feedback eq, such as the attached schematic for the Marantz 7C phono stage?

[Eli Duttman]

LB,

The Marantz circuit uses filters in a NFB loop to achieve the goal. That's active EQ. In circuits with loop NFB, the feedback network dominates behavior, as long as open loop gain is sufficient.

You said you wanted to use a type with lower Miller capacitance in the 1st gain block. The 6GK5 should fill the requirement. Stage gain will easily be equivalent, even though 6GK5 mu is lower than 12AX7 mu. The 6GK5 is high mu, high gm, and low Rp. The smaller Rp makes up for the lower mu.

The use of a 12AX7 section in cathode follower leaves (IMO) much to be desired. Inside a full function preamp that included at line stage, the 'X7 CF was tolerable. No way is it going to drive both the NFB loop and a downstream external load that includes cable capacitance. Point blank, the 12AX7 is a WIMP. DC couple a ZVN0545A enhancement MOSFET to the 2nd gain triode's anode as the voltage follower. You eliminate a coupling cap. and improve drive capability. Id = 2 mA. should get the job done.

BTW, the O/P coupling cap. needs to be at least 3.3 muF., if you are going to drive the IHF "standard" load of 10 KOhms.

[SY]

A feedback RIAA is devilishly hard to do properly. It will take a much more complex circuit than that and unless you really know what you're doing, it will be difficult to stabilize and provide adequate overload margins. With all due respect to tradition and my elders, the Marantz phono circuit was mediocre at best.

I'd stick with a simple passive approach.

[oshifis]

Maybe a bit off topic, but I described my passive tube RIAA that gives exact equalization in this thread:

Vishay Audio Resistors