BA4560 or LM833N

[Smoky]

I'm noticing a slight "gain" difference between the left and right channels of an equalizer that I'm working on.

The two and only types of IC's in that circuit are a BA4560 and a LM4136CN.

I wanted to buy a couple of new ones to see if it they make any improvement since the originals are over thirty years old.

I first typed in BA4560 on Digikey and they have them in stock but they also recommend an LM883N as an alternative.

Are these two IC's interchangeable and is an LM833N an improvement?

As for the LM4136CN IC's, they appear to be discontinued. NOS on Ebay will probably be my only route to purchase.

Thank you 

[wraper]

I wanted to buy a couple of new ones to see if it they make any improvement since the originals are over thirty years old.

What improvements do you expect, making gain equal? It's not an opamp issue to begin with.

[Smoky]

Yes, I'm trying to make the gain equal.

I'm feeding the Left and Right channels the same signal simultaneously but the response is different when measured. The Right channel sags somewhat compared to the left.

What I'm afraid of is the slide potentiometers may be the problem. They are a "ganged" type (one slider controls both channels for each frequency). I do the testing while the sliders rest in their center detents.

I'm trying to post pictures of the graphs but EEVblog gives me an error message. What's up with that???

[wraper]

Gain is not determined by opamp (assume it has infinite gain) but by surrounding components in feedback loop. If there are any electrolytic capacitors, first check those. Then measure if resistors did not drift.

[Smoky]

Btw, there are six sliders in the equalizer section. I measured the resistance of each slider while they rested in their center detents too.

Here are the results:

Slider Freq / Lch / Rch

22Hz / 15.68K / 15.78K

31Hz / 14.58K / 15.16K

45Hz / 13.19K / 14.73K

63Hz / 14.89K / 14.96K

90Hz / 15.28K / 16.08K

125Hz / 14.10K /14.78K

I imagine this difference between channels alone affects the frequency response too.

[shakalnokturn]

That doesn't look too bad. 20kOhm pots?

If there are balance and volume controls those pots would be my first suspects.
You have checked solders first, haven't you?

Feeding the EQ with a function generator and sweeping through the audio range with the outputs subtracted on a scope could give you a whole load more clues on what the problem may be. (A bad capacitor should show quite fast ...)

Edit: As for the LM833 IMHO it's a good Op-Amp for audio.

[Smoky]

Here's a section of the schematic that deals with the equalizer.

At BA4560 #9, pins 1 and 7 are equal.

At BA4560 #8, pins 1 and 7 are not equal.

I have matched to 1% and installed all capacitors in the equalizer section.

I just got in all of the Vishay 1% MRS25 resistors which I'll match and install next.

Thanks Wraper and Shakalnokturn 

[floobydust]

Although the potentiometers are never perfect, these older EQ's using tantalums/electrolytics in the gyrator circuit, I never found they sound good due to dielectric absorption and the leakage current can upset things. Once the op-amps age, their offset voltage can change (polarity) and some (tantalum) capacitors end up with a reverse bias and then go leaky and fail. Noisy, scratchy pots that can't seem to be cleaned or fixed are a giveaway there.
So I would be checking the voltages the capacitors are seeing. The EQ has a DC-coupled input stage, so if it was run with some DC offset coming in, it could damage things.

For sonic improvements, swapping in film capacitors and low noise audio op-amps does help a lot. Look at LM837.

[Smoky]

Imgur seems to work 

Here is the disparity between the Left and Right channels of the equalizer during a frequency response test by a HP35660A (identical signal and amplitude fed to both channels):







Thank you FloobyDust!

I found this guy on Ebay in Arizona selling the proper package type needed of the LM837. They are supposedly NOS National parts. I bought his last four.



I also swapped all Polyester caps with PP caps a few weeks ago. Some of the electrolytics were replaced with film of the values that I could find/fit.

*I also forgot to mention that the equalizer section only covers the low frequencies going to the sub-woofers. The internal cross-over module limits it even further since it's set at 75Hz.

The Sub-Sonic/Rumble/Cross-over graphs are perfect as long as the EQ isn't activated. I spent a lot of time in those other areas to make that happen. The EQ section is the last portion to deal with and I want it as good as I can get it. But remember, it's just the lower frequencies that are affected. Everything above 75Hz will be passed through to separate mono amplifiers un-touched. So the sliders for 90Hz and 125Hz will probably be set close to the "flat" position.

The signals below 75Hz are sent to a large integrated stereo amplifier dedicated to the L & R sub-woofers.

[floobydust]

I see -4.2% on the R channel vs L, with a little bump around 5Hz that the other channel does not show.
It's the same as the difference in potentiometer values, for a set of 6 in parallel. So adding a 56k resistor across the pot's end-end on the R-ch would match them, but it would affect the boost/cut amounts I think so say +3dB boost might end up being +3.4dB for the R-ch but I'm not sure on all the math.

[Smoky]

Thanks for spending the time to contrast the numbers FloobyDust

I can't focus on it too much tonight but wait until you see how simple the PC board is where the slide potentiometers are mounted.

After watching Feedback Loop's video on how he designed and made a new PC board for his Fluke handheld scope, I'm thinking of making one to replace what's in this equalizer, just for fun and the experience.

New 30mm slider pots are available but they have  a slightly different footprint but can easily be incorporated into a new layout.

A ribbon cable is all that connects the slider's PC board to the main board and only one other trace is printed on it and it goes to the DB meter.

With these equalizer's breaking the $800 mark, there may be a demand.

[wraper]

For purity of experiment you can replace pots with fixed 1% resistors and see how it works out.