Removing coupling capacitors in signal path

I've got an older (c. 1983) NAD 3150 integrated amplifier that I'm working on that has a slew of 1 uF electrolytic capacitors in the signal path (I've got a copy of the service manual with schematics). The NAD sounds a bit "grungy" to me compared to listening to the headphone output on my Echo Gina24. I've recently replaced the stock NE5532 with an LM4562, which has already helped improve fidelity somewhat.

I'm wondering if I might be able to clean up the sound a bit more by removing or replacing the interstage coupling capacitors.Would it be advantageous to remove some of them if there is sufficiently low DC offset present? Are there places where coupling capacitors are absolutely necessary and places where they can safely be removed? How much DC offset is acceptable before a coupling capacitor must be used?

Thanks!

Highly unlikely that a cost-sensitive design such as would be coming out of NAD would have gratuitous coupling C's!

To your other question, it depends on the point in the circuit as to when a C might be replaced by a wire. If there is negligible gain thereafter then a few tens of millivolts probably wouldn't matter. But one needs to look at what's going on, case by case.

Having said that, I have seen some d.c.-block-happy designers in my day. I won't mention any names because some are still friends.

Thanks for your reply!

Is it acceptable to remove the coupling capacitors if the DC offset is less than 10 mV? It won't be too much trouble to measure the offset at the various stages. What I was thinking is that I could just leave one coupling capacitor before the power amp stage, ensuring the no DC goes through the power amp to the speakers. Probably replace the electrolytics with 1 uF film caps....

Switching... BANG!

Volume... Crackle.

Tone... Rustle.

And as Brad mentioned, you have to consider the GAIN involved.

Nobody else can tell you what something will taste like, you have to "suck it and see".

But they are there for a reason, and my money says that there are penalites, otherwise they'd have saved some of their own money and used fewer.

Keith

Probably replace the electrolytics with 1 uF film caps....

Click to expand...

This may be your best (& safest) bet.

Depending on voltage... the Panasonic stacked metal films are nice & small.
They're available to at least 1uf.

You could also try some better 1uf's.

Like Keith said... you really have to listen for yourself to see.

[quote author="12Bass"]I've recently replaced the stock NE5532 with an LM4562, which has already helped improve fidelity somewhat.[/quote]
Interesting, please tell more.

[quote author="clintrubber"]Interesting, please tell more.[/quote]

In the process of making extensive modifications to my Eden WT-500 bass amplifier, I've found the LM4562 to be one of my favorite high performance op amps. The LM4562 is now in several sockets in the WT-500, replacing the original TLO72s which I found somewhat fuzzy and murky sounding. Two of the sockets are presently filled with the TLE2072 due to the the high impedance in that part of the preamp circuit (I'm still looking for a better FET input replacement). The addition of the LM4562s plus numerous capacitor and resistor upgrades has resulted in a less noisy amplifier which sounds much more transparent with considerably extended frequency response (from 12 kHz to 20 kHz) when compared to stock.

When I was testing op amps, I tried a variety of well-regarded devices including OP275, OP2604, TLE2072, TLO72, OP2134, AD822, AD823, LF441, JRC4558, NE5532, and AD8599. Of all of the op amps tested so far, I've liked the LM4562 the best - it seems to have the least noticeable characteristic "sound" when compared to the others, i.e. more transparent reproduction of the source. From what I can tell, it is a good substitute for the NE5532.

Thanks for adding. All I had seen so far about the '4562 opamp were the discussions here & the test-report from Self ( http://www.dself.dsl.pipex.com/ampins/webbop/LM4562.htm ), so all I recalled was in the vein of '5532 still not really beaten' (as far things like this need to be a contest of course).
As for more things, it sounds like it all depends on the circumstances & preferences, so interesting info you added.

FWIW, w.r.t. bass, the SWR-heads I use are full of TL072 as well (SM-400), but I look for colour in a bass-amp, so we're in fact looking for different things (even stopped using the 'sterile'/'to clean' (imho) pre's from those amps in favour of another pre-amp).

Regards,

Peter

Hmmmmmm......

I don't find the modified Eden to sound more sterile. It still sounds like an Eden, but better IMO - clearer, deeper, more articulate. Eden uses a 7025 tube at the input of the preamp for a small amount of color, so that hasn't changed. Besides, the TL072's don't add a color that I particularly care for, except maybe when used in a distortion pedal....

BTW, I definitely prefer the LM4562 to the NE5532. Clearer sounding, lower distortion, and less trouble with DC offset. Actually, based on my recommendation, a certain bass amplifier manufacturer is now using the LM4562 in their amps....

Well, the parts in from Digi-Key arrived and have been installed in the NAD 3150. Before replacing the stock electrolytics I measured the DC voltage across each one. The output of the LM4562 was the only place where the DC offset was negligible, (one channel was around 4~5 mV, the other less than 1 mV). Given the voltage across the caps, I'm wondering if perhaps they were intentionally DC biased?

Anyway, I replaced all of the 1 uF electrolytic coupling capacitors with Panasonic ECQ-V stacked metal films, while the 100 uF electrolytics were replaced with Panasonic FM electrolytics with a 0.01 uF polypropylene bypass across each electrolytic.

Just now I finished modifying my Echo Gina24 and am now giving it a listen through the NAD and my Energy Pro 22 monitors. All of the NJM2068 op amps in the Gina's input path and channel 1-2 output path were replaced with SOIC LM4562s (output channels 3-8 were unmodified). During the process I recorded some 96/24 audio for comparison purposes.

Overall, I'm quite happy with the upgraded sound of both the NAD and the Gina24. The NAD seems a bit clearer sounding than before. I can also notice a definite improvement in transparency in the Gina24, like you can hear "more into the music". Cymbals sound more defined and the overall sound is natural and less veiled. With the NJM2068, the highs sound a bit ragged or distorted. I may post more info on the Gina modifications later....

Thanks for all of the help!

[quote author="SSLtech"]

Nobody else can tell you what something will taste like, you have to "suck it and see".
Keith[/quote]

love it :thumb:

you need the old siemens 55xx stuff.
its out there.

don't do an A/B unless it's against the best 5532 out there.

[quote author="CJ"]you need the old Siemens 55xx stuff.
its out there.

don't do an A/B unless it's against the best 5532 out there.[/quote]


Are you sure... Siemens or do you mean Signetics?

I don't mean to question you, I never heard of Siemens 5532's.

FWIW, I've got a Signetics NE5532N and XR5532N for comparison. IMO, the LM4562 sounds better than both - more open and detailed sounding.

err...signetics.

Siemens made the 55SS.
:twisted:

Before you use the LM4562 to replace any 5532's, note that the absolute maximum operating voltage is +/- 18 Volts, but +/- 22 V for the 5532...

Working on a Harrison console with 18 V rails, so that's why I thought about it.

Martin

[quote author="Martin B. Kantola"]Before you use the LM4562 to replace any 5532's, note that the absolute maximum operating voltage is +/- 18 Volts, but +/- 22 V for the 5532...[/quote]

Good point! I measured the rails at 17.5 Volts with the LM4562s installed.... a little high, but below the absolute maximum. The new LME49860 appears to be very similar, but supports 22 Volt rails....

http://www.national.com/pf/LM/LME49860.html

So far, this is my favorite sounding op amp. I'd like to find a FET input dual op amp with a similar transparent sound. Any ideas?