Technical Question About Capacitive Coupling in Amps…

jbreezy5

Amplifier manufacturers make much ado about not having capacitive coupling occurring in their amplifiers.

I was looking at Parasound entry-level vs. Halo amps; particularly 2250 vs. A21+.

The 2250 clearly has crosstalk between channels due to capacitive coupling; however, it has the capability of being run mono at 750 watts.

When run mono, does capacitive coupling still have an effect?

Forgive my ignorance. 🤷🏼

audioluvr

Nope

mhardy6647

Capacitive coupling does not a priori only refer to DC-blocking capacitors on the output of a power amplifier circuit. If there are capacitors between amplifying sections ("stages") of an any amplifier, yes, the amp has capacitor coupling. Indeed, it wasn't uncommon to put a DC blocking capacitor on an amplifier's input, since a fully direct-coupled amplifier could dutifully amplify any DC "bias" that might be riding under the (AC) input signal ("music"!) -- with potentially disastrous results for one's loudspeakers.

The downside of capacitors in a signal path: low-frequency rolloff; phase shifts at low(er) frequencies, and the general sonic risks of any additional component in series with the signal path.
The upside: elimination of any DC (intrinsic or extrinsic) passing through the amplifier, which will wasting amp capability at best and put loudspeakers at extreme risk at worst.*

I haven't looked at the schematics of either amplifier, but if there are interstage coupling capacitors in an amp, it's hard for me to imagine that they're not still present if bridged to mono. If the outputs are cap-coupled to the loudspeaker (load) -- I would imagine that's still the case... but it is risky for me to speculate about these amps.

____________
* This is for solid state amplifiers. Essentially all vacuum tube amplifiers operate with high DC voltages (hundreds of volts) on the plates which must be removed from the signal path before passing on to the loudspeakers. This is usually done by the output transformer (DC will magnetize the transformer primary but not be induced into the secondary), or a choke, or sometimes a capacitor. It is possible to build balanced, output transformerless vacuum tube power amps, of course, but these are pretty specialized and somewhat finicky beasts -- with a huge downside possibility of wanton destruction if anything goes wrong.

jbreezy5

@mhardy6647 - so it seems that the type of capacitor coupling I’m referring to is the output stage type, as opposed to the inter-stage type.

Hypothetically speaking (no longer referring to the specific amps previously mentioned), if a stereo amplifier has output stage capacitor coupling, or interstage capacitor coupling, and is using its mono amplifier configuration capabilities, is there going to be an increase in L-R channel crosstalk?

Thanks for your earlier insight, btw!

mhardy6647

No, as far as I know.
To the best of my knowledge, the only reason to have a capacitor-coupled power output stage is to eliminate any DC in the signal from getting to your loudspeakers. There will be one such capacitor in series with the output of each channel. They should contribute nothing between or across channels -- but each signal's channel will have to pass through it.
The output capacitor essentially adds a high-pass filter to the signal coming from the output stage, thus attenuating (rolling off) bass response, with a hinge frequency determined by the value of the cap. Capacitors in the signal path of an amplifier are generally thought of as things to avoid unless absolutely necessary

Do you know why there is a capacitor on the output? In solid state amplifiers, cap coupled outputs were (are) usually only found on really low-end (think "Soundesign" and "Yorx") amplifiers that use single-rail power supplies, or in a transitional ss amp design topology called "quasicomplementary". In a quasicomplementary amp, the amplfier operates with, in essence a DC bias, so the amplified audio (AC) signal is offset from zero VDC. Quasicomplementary amps were sometimes found in the mid to late 1960s, due to limitations in the transistors that were commercially available in those days. "Direct Coupled" amps with dual-rail power supplies became pretty common starting in the first half of the 1970s.

The output caps are usually of a very high value to push the bass rolloff down to an acceptably low frequency. Size and cost pretty much demand the use of an electrolytic, AFAIK

Here is a quasi-intelligible discussion of quasicomplementary amplifier topology from a weird old (defunct) website called "Audion". I was never entirely sure if "Audion's" articles were meant seriously or tongue in cheek, but there is good information in them.
https://web.archive.org/web/20030306020231/http://www.lacieg2s.ca/w3terra/ols/scott260.htm

As an aside, I actually have a Scott 260B Well, the kit morph thereof, the LK-60B. It is indeed cap-coupled. The output caps are something like 2500 uF electrolytics (if memory serves).
It sounds... OK. It cost $10. I bought it mostly for its walnut HH Scott case -- which I put a nice Scott tube amp into

HH Scott LK60 DSC_0789 by Mark Hardy, on Flickr

EDIT: It may be that Parasound is doing something else entirely with capacitors in their outputs -- I have no idea. What I am describing is the classic mode of capacitor output coupling, found also in, e.g., the Marantz 1060 integrated amp and some of the older Sansui amps and receivers. Some folks hold these components in rather high regard.
Some folks.

mhardy6647

apropos of (almost) nothing

P1030127 by Mark Hardy, on Flickr
I believe that the two smaller grey caps in front of the output section in the 1060 photo I took above are the coupling caps for its outputs. The original coupling caps are electrolytic 3000 uF @ 55V; I guess the modern guidance is to up 'em to 6800 uF to improve the bass response. The larger grey cap is the "main" filter cap for the power supply (originally 4700 uF @ 80V).

https://theartofsound.net/forum/showthread.php?69812-Marantz-1060-Recap-amp-Upgrade

jbreezy5

Ok thanks for the education. Just being curious, I guess.

Sounds like my potential crosstalk concerns are unfounded.

Also thanks to @audioluvr, with the right between the eyes response. 😅

engie490

I'm sorry, but I think I'm confused. It appears the original post mixes 2 ideas.

I thought:

DC-coupled and capacitor-coupled circuits refers to whether DC is passed through the circuit (i.e., no cap in the signal path) or a capacitor filters out the DC component (causing some of the effects @mhardy6647 mentioned above).

Crosstalk (usually caused by capacitive coupling) is the presence of the signal from one channel in the other channel (and vice versa).

It happens especially across shared resources (e.g., circuit boards, volume controls, etc) but is not limited to the presence of a capacitor in the circuit. In other words, DC-coupled circuits can (and do) have crosstalk.

Usually crosstalk improves due to physical separation of channels (i.e., dual mono or mono) when sharing of components is minimized. It is obviously more expensive to have 2 of everything than it is to share some components.

Actually, the 2250 v2's (85 db @ 1KHz) crosstalk spec is better than the A21+'s (80 db @ 1KHz). But, you're unlikely to hear that difference.

Hopefully one of the big brains around here will clear up any misstatements on my part.

mhardy6647

DC = direct coupled; no capacitor or transformer in the signal path coupling stage to stage. Because these components are not present, the capability to pass "DC" (direct current) through a "DC" (direct coupled) circuit exists. You don't want to amplify and pass DC from source to loudspeaker, though. The end results are not pretty. BTDT, more than once, with malfunctioning ss amplifiers -- which is why, nowadays, I check for direct current on the outputs before attaching loudspeakers to direct-coupled ss amps... and I use cannon fodder loudspeakers for testing.

Excessive crosstalk usually arises from poor circuit design (layout) and/or power supply inadequacy AFAIK. Function selector switches are notorious for enhancing crosstalk. Yes, signals from one channel can be inductively coupled to another. In terms of these amps... I've got nothin'. 80 dB separation is, IMO, plenty. The folks at ASR would say it should be 100 dB (or more), but that's just gilding the lily.

jbreezy5

Thought I’d share a review w/measurements for those who like looking at such things. This integrated impressed JA @ Stereophile.

This did particularly well both on preamp/amp crosstalk measurements, among other things.

Clipdat

Did you forget the link?

I like the crosstalk specs on my Benchmark AHB2:

Crosstalk

Better than -115 dB at 1 kHz
Better than -92 dB at 20 kHz

And on my LA4 pre:

Crosstalk
< -133 dB @ 1 kHz, -116 dB @ 10 kHz

jbreezy5

My bad, here’s the link:

https://www.stereophile.com/integratedamps/407bryston/index.html

jbreezy5

Great numbers on your Benchmark stack!

Clipdat

Thanks for the link. Bryston's engineering is generally impressive.