Op-amp preamp discussion

[Dannyx]

True, that would be a pretty interesting test and since I have 3 boards, I could do that rather easily. That's why I also said we're talking from a purely subjective standpoint, relying on the human ear and each individual test subject, as I'm sure there would be some differences from a purely technical standpoint in a laboratory setting with high-end test equipment and number-crunching, maybe even batches of the same exact components (op-amp, cap, potentiometer).

[eccerr0r]

Put in a LM358... I'm curious. I've "heard" that it has problems with linearity.

[Dannyx]

Potentiometer imbalance is another thing I've always been bothered by and this project suffers from it as well. The left channel always come in hotter than the right at low volumes and sure enough, measuring the pot confirms the (often wild) difference in resistance between the 2 channels despite these not being the crummiest of pots out there. I THINK they're from a decent brand - Alps. Ironically, I happen to have a big-a$$ Alps pot in my parts box (RK50 series) I ordered of Ali a while ago, not even aware of its full potential (pun intended) and it reads NO imbalance at all indeed. The only reason I don't use it here is because I don't like the bumpy steps it's got - it's not a linear feel when you turn it, but rather an encoder-like feel.

Initially, I was planning to use some sort of volume control IC, but while that would solve the imbalance issue, I'd introduce unnecessary complexity and possible distortion of its own, so I scrapped the idea, at least for now...

[Dannyx]

I started wondering if there's a way to control the volume (gain) on two distinct op-amps with one single pot. Turns out there is, but doesn't look easy: VCA...haven't looked into whether it can be pulled off or is even the right thing for the job...

[eccerr0r]

well, all those radios wih no pots are doing somethingrather...

[Dannyx]

Originally Posted by eccerr0r

well, all those radios wih no pots are doing somethingrather...

I found some push-button volume control ICs a long while ago, but that's not really the style I want in this project...I also don't want to go crazy overboard with components and complexity either, but it looks like I'll HAVE to if I choose to go down this road....or just keep trying pots until I find one with least imbalance I guess

[sam_sam_sam]

I going to suggest something that might not be obvious at the moment but why not replace both pots with new ones with the same exact part number and see if you have the same imbalance

[eccerr0r]

or you can always add trimpots....

[lti]

Trimpots won't help if the imbalance constantly varies like the 100k dual gang pots that Radio Shack sold in their dying years.

[Dannyx]

All pots are new. Here they are. I got 5 of them. The problem is most obvious on the volume control for MAX4410 headphone amp, because headphones sit right next to your ear and this thing gets pretty loud pretty quickly, as soon as you turn the volume pot, so to get it to sound somewhat balanced, I'd have to turn it louder than what I'm comfortable with. The main amp on the other hand is harder to notice, since it's louder and I also had the speakers right next to each other during testing, plus I have those individual "downstream" pots which I can tweak to balance it out.

I guess at this point I should provide a schematic explaining what I want to achieve. It looks something like this...no inspiration - just came up with it on my own thinking how a car headunit works. It works in practice, save for the imbalance inconvenience. I'll use this setup to drive a subwoofer on the rear channels, hence why I needed a way to shut it off entirely if I want to. Normally that's where a balance control would go, which is NOT what this is, I'm aware of this. That would not be of use to cut off the sub entirely. This is more like individual volume controls for each of the 4 speakers, which is fine - I don't mind that. I'd have the last 4 pots all the way up, so they're basically out of the picture, adjust the main volume at the bottom as needed and simply trim the speaker that sounds louder if needed. The headphone amp is pre-master volume control.

Now that I think about it.....can I get away without the op-amps entirely and just parallel the pots on each side ? I think what would happen in that case is adjusting one of the pots, alters the volume on the other one as well. Haven't tried it. Will throw it on the simulator to see...

[Dannyx]

Out of curiosity, I went to my parts bin and dug out that big Alps pot I mentioned as being perfect in terms of balance and it turns out it's not really Alps and it's not really a "pot" either

First off, there's no info on this part whatsoever - RH2702 only gets me links to random sellers and references to the "real" thing which apparently is the RK271 series. RH is not even an official Alps series of its own, which leads me to believe it's a knock-off. Should come as no surprise given that I got it off Ali and it's common practice for sellers to spoof official names and models by altering a letter, such as Sqny instead of Sony and so on. In this case, it's RH instead of RK.

Backing this up is a discussion I found here where they talk about the same exact doubts about the authenticity of said pots. Someone said the real ones do not have a stepped feel and knurled shaft and also a serif font for the ALPS logo on the back....mine has both of the former and plain text as well, so it's deffo fake.

Looking past that, what IS this thing then ? I mentioned how it's got a stepped feel to it and how it's not really a pot. It's a "stepped attentuator" if you ask me. I found a picture of it dismantled here and the way it works is by moving the wiper across a series of fixed value SMD resistors, hence the stepped feel when you turn it. Out of curiosity, I took mine apart as well and sure enough, there's your 22 SMD resistors in series. It works fine, no complaints there and I guess one redeeming aspect is that it offers "true zero" when the wiper is at either end (why can't more pots do this ?), but the stepping is kind of rough and results in large jumps in volume. I also don't like the force required to turn it....

This got me thinking of getting a real RK one, since it seems to be praised by audio guys....

[petehall347]

loudness tap as well ..bargain ..
the ones pioneer used were pretty good . think its 60 clicks on my one .

[Dannyx]

Originally Posted by petehall347

loudness tap as well ..bargain ..
the ones pioneer used were pretty good . think its 60 clicks on my one .

This one is 22, counting the zero position. I had no idea what the fourth pin was for. Measuring it against the others just gave some random BS reading that made no sense. After reading about it, I learned it's called a loudness tap and it's connected somewhere in the middle of the resistors, so at one point the resistance between the wiper and this pin becomes zero indeed.

No idea what I could use that for though...

[eccerr0r]

I'm surprised, thought those pots with tons of little detents were mechanical detents on a continuously variable background; I'm surprised that it's cheaper to make a side switch on a PCB with tons of different little SMD resistors -- though getting a good long lasting log pot isn't easy.

[Dannyx]

Originally Posted by eccerr0r

I'm surprised, thought those pots with tons of little detents were mechanical detents on a continuously variable background;

I thought the same from the moment it landed in my parts bin years ago, until yesterday when I tried looking up a datasheet for it and got nothing other than that page which shows one open.

I tried it out as the master volume for the headphone part of my build, just out of curiosity and just as I predicted, after just one CW click (so 50 ohms "away" from GND), it gets louder than how I'd like it to be, at least for these particular earphones and laptop, as I'm sure it'd be adequate if I maybe had higher impedance drivers, a device with a weaker output and the content itself encoded at a lower volume. The channel balance on the other hand is, as I said and as you'd expect from fixed value resistors, BANG on, to the nearest digit ! I'd like to see something else like this but with more steps, hence more resistor values packed in there because the concept is great overall, albeit it's an overall fake part nonetheless !

[eccerr0r]

nothing is easy, all these workarounds...loudness, or end up just not caring when volume is low and LFE are lost with it.

I highly doubt there will be solutions with more detents because stocking so many resistor values would get kind of expensive.

Sigh audio, it's 100% analog... detents are a digital concept, and that's not analog thing. Or perhaps accurate and precise digital pots is the answer.