Hello Jason:

My setup is similar to the one you describe in your article. Head unit up front, amps, XO and signal processors (Carver Cathedral and Audio Control Epicenter) all in the trunk.

Plus a nuance as follows. I have a 93 Maxima SE, and kept the factory Bose head unit with CD. I got an old Powerdriver unit that is designed to change the "funny" Bose interface signal to plain RCA.

This setup had noise from day one. The typical Aahrrr response after two years of building, but I could get the system quiet by removing the signal ground wire. This power driver unit has two ground connections to the radio: One to the case, and one to the shielding of the signal wires that go to the Bose speakers. I only kept the radio for looks, everything downstream has been changed. So curiously, if I disconnected the signal ground, so the unit only had one ground connection, it used to be quiet, however at the cost of a fairly loud turn off pop which I could not get rid off. It the price I was willing to pay.

The tech helping me with the noise indicated a new Powerdriver might work better. I tried, but no, the new unit is noisier under all circumstances. Playing with the different ground wires I could get different levels of noise, but never zero.

The noise is definitely louder under high load. But it seems, under similar circumstances, sometimes louder than others.

I just cleaned all the battery connections, and the battery ground cable connection to the chassis, and to the engine. To no avail.

I don't have a scope, so I can't tell if there is excessive AC on the alternator wire.

My core question is: Do the symptoms of a quiet system GOING back to noisy make any sense?

Could this be an indicator of one of the diode in the alternator having gone? The new stereo would have put the alternator to work much more that in its first 50,000 miles of life.

The only other thing I can think of is a wore that loosened, but I have not found any yet. The ground in the trunk is a two foot four gauge going to one of the seat belt anchors and to bare ground around it and then to a distribution block. Every component in the trunk goes to this block. Drawing a wire from this block to the power driver up front makes no difference in noise level.

Your trick disconnecting one channel, leaves one noisy channel. A channel connected without a ground is noisier. The 4 gauge power to the trunk goes along with most the other wiring along the driver's side. The RCAs go back over the center tunnel where there is no other wiring.

ANY insight or advise that you could share on this would be most welcome. Thank you in advance.

NOW,

If I haven't overstayed my welcome or lost you earlier in this diatribe, I have a power supply puzzle that I have not found the answer to yet.

Someone in the UK that I have been in touch with has a unique power conditioner for his car. It is about 10 amps, so it only "cleans" signal processors and the head unit. It was built for him by a Ph.D. in electronics friend of his. I have asked to be put in touch with him to see if he would share the design or sell it, but he has lost contact with him. In respect to his work and his help, he did not want to open the unit and reverse engineer it. The only clues I have is that it only took $30 dollars or so in parts (in the UK).

The unit had an adjustable voltage up to 18V or so, and had its own ground OUTPUT, so all the devices connected are isolated from the chassis ground, including the head unit. Can a switching power supply be made this quiet this inexpensively, or do you think something else was built here.

If I can solve this puzzle, then this is a venue I would try on my project. What do you think?

Jason, like I said before, I would welcome any comment or insight and I appreciate the time you took to even read my mail. Thank you.

Best regards,

Hi John,

Seasons greetings. Sorry I only responded now - I check my hotmail only occasionally because I don't use it much anymore, it's so full of junk mail. I'm forwarding this to my main email address, let's discuss from there.

That's quite a problem you have. Here's some stuff off the top of my head I would suggest you buy or borrow to help troubleshoot your problem.

1) A separate 12V lead acid battery (a car battery or UPS battery will do)

2) audio transformers from Radio Shack (4 pcs)

3) shielded cable with 2 "hot" conductors (have you already tried using this to replace all your stereo rca cables in your car?)

4) assorted RCA jacks and plugs to make cables with

5) DVM

Other things to do:

Have you tried eliminating your system components one by one until the noise disappears?

One basic requirement in the design of audio components in a car environment: Take an ohmmeter and try to measure continuity between the input RCA jack ground and car ground, there should be a relatively high resistance. If you find a few ohms or maybe even 100 ohms, this component will be very susceptible to "conducted noise" or "common impedance noise" otherwise misnomered as "ground loops". I personally consider this bad design.

The above problem is alleviated by a separate power supply such as a separate 12V battery or that custom isolated power supply your friend has. BTW a 10A one would cost the company I work for about <$7 in parts - but those are mass production prices.

Take an ohmmeter again and measure the continuity between the input RCA jack

grounds of the left and right inputs, and an audio component with

"differential" inputs will again show no continuity or high resistance

(upwards of 1 kilo-ohm). A component with differential inputs will be less

susceptible to "magnetically coupled noise", which is alleviated by my trick

of using shielded cable whose left and right channel signals are surrounded

by a single shield.

The 2 problems above (ground loops and non-differential inputs) could be alleviated by a signal conditioning circuit that is a differential input to balanced driver (I'd have to think about this one). Such a circuit would sit close to the badly-designed component. An audio transformer will do the same thing, but as we know, usually have poor frequency response. If you're handy with building circuits I can sketch you a schematic. CTTOI I wonder why such noise-killing circuits aren't sold...

The above 2 ohmmeter tests are worth doing.

If your alternator has a bad diode, its max output will not be up to full rated current. If there is some way to check that - like maybe turning on all electrical loads at idle, and measure that battery voltage is still 14V.

If you still have a problem sketch me a detailed diagram of your system.

Hello Jason:

Seasons Greetings to you as well, and my sincere thanks for your willingness to share your experience and time with me, being a total stranger.

For ease of typing, the Power Driver, the interface from my Bose Radio to the RCAs up front, is called a PD-2 (for 2 channel unit). The new one I tested was a PD-4 (it will take the rear channels as well).

I have done a lot of reading on the internet on the subject. (That's how I found your page). Most of them rehash the same things one way or an other, your page introduces a totally new concept, which I am not sure would work in mine, but is still worth a try. I can't get one silent channel, though, so that solution may well not apply. One expert advised against just plopping in filters to mask the problem. He said find it and fix it. I tend to agree for two reasons. I am weary of the quality loss using audio transformers, and, secondly, I spent two years on a very high quality install. (quality in all the wiring, but not the over the top cosmetics as it is a daily driver, not a competition vehicle). Third, being a fixer, I would like to know what's wrong and fix it right, better solution, more result satisfaction. So with your help I may get there. Again my thanks in advance for your trouble.

I have already done the following.

With a long, and much cheaper RCA cable I have tested the front to rear connections. The cable gave me the same noise, and moving it around the interior gave no difference, so, most likely, the noise is not induced into the wire.

Connecting the front to ANY component in the rear gave me noise. Direct to amps, and skipping any of the signal processors, if it was connected to the front, there was noise. If it was NOT connected to the front: Silence. So, again deduced, I don't think there is noise between any of the processors and the amps.

At one point, after finishing the install, and depriving the PD-2 from the signal ground lead (at the cost of loud turn off pops), the system was totally silent. Ear to tweeter only revealed some amp hiss (cheapish Jensen amp). The noise CAME BACK. To my amateur brain, only four things could have gone bad: Ground connection deteriorated, alternator partially broken down, battery gone bad, radio gone bad.

The last one, the radio, I find least likely, but it is hard for me to test. The battery I put in because one author, after trying everything, solved his problem that way. The battery was new after the stereo was done. The old one had little capacity left and would not start the car after only minor music playing. I plan to go the Auto Zone, and ask them to test a new battery for me. Again I don't think that will be it, I have good voltage, and powerful cranking, but I want to eliminate it as a cause. Same way I will have the alternator tested by an expert who can measure ripple voltage, in case that is high, but the diodes are still all good.

Since my initial e-mail I have cleaned all the connections on the alternator. I would like to say it so slightly improved the situation. I also cleaned the battery terminals and the battery ground connections to the chassis and to the engine. That was to zero avail.

I hooked my wife's Honda up to my car with jumper cables. Honda running, Mine not: Initially slight noise, but then quiet, and no noise no matter how much load I put on.

My car running, the Honda not, I can not hear any noise on the Honda radio, but it is factory all the way.

Something else I discovered. I am sure it means something, but I don't know what: My car NOT running, I can hear noise of fuel pump, radiator fan and interior fan in my tweeters. I don't think that was the case when the car was quiet.

I am still leaning towards a ground problem somewhere.

In the rear I have a distribution block for ground with the lead going to a seat belt anchor.

I still want to inspect all the 8 and 10 gauges and the 4 gauge going into the block. I think they are OK. I sprayed WD-40 on all bare copper wires to prevent or at least slow down oxidation. Some electronics guy advised he did that for molex connectors in his car. Maybe that is coming to hunt me. But if O loosen all the wires and reconnect them, the disturbance may fix it. In that case I can wait for it to go bad again.

I have tried to connect the PD-2 ground to the rear block to no avail. I still want to try running a separate ground wore from the case of my radio to a new and good ground up front and to the block in the back. On the past with a thin wire it made no difference.

I still need to try to feed the radio from a separate power source. This will be tough as the power feed is through the harness, but I will do it soon, if nothing else fixes it first.

IMPORTANT thing for me to do is to measure RCA ground to chassis ground resistance. I think the PD is like that. ( as I can't eliminate that box in my current design, ANY work around? )

YOU WROTE : "a signal conditioning circuit that is a differential input to balanced driver (I'd have to think about this one). Such a circuit would sit close to the badly-designed component." I would love to try this. I can build complex circuits on a made for purpose PCB, and simple ones on a blank PCB. If you have a schematic I could try, I would. Especially when expected to function with no loss of signal quality, I could soothe my conscious more easily to live with such a patch.

BTW, do you have a schematic for an isolated power supply you could share? If it is $7.00 wholesale in components even at 3-4 times that cost it would still be a bargain if it fixed my problem. At this point I still don't know if I have bad power or bad ground. I suspect the latter.

I don't have software to draw a system diagram, but I think I can describe it to where your interpretation, or drawing on paper would be accurate. BTW all my RCAs are custom length with MCA Electronics' high resolution audio cables and the fitted for that cable gold metal case RCA connectors.

Wires between A and B not shielded. Audio cables tightly twisted (each channel separate, both channels separate, not twisted with one another) Two single RCAs to trunk connected to Carver Cathedral unit. ©. C. has an internal Y connection on its input. this Y goes to A Fosgate OEQ-1 equalizer. (D)

Thus output of B goes to both C and D.

Output of Cathedral C goes to rear channel inputs of A Carver SX/2-3 Active XO. (E)

Output of OEQ-1 (D) goes to front channel inputs of Carver (E)

Rear Mid outputs of XO (E) go to 2 channels of a Jensen 432BBE amp (F).

Corresponding outputs to rear fill speakers. (F) is 4 x 50W RMS at 4 ohm

Rear low and high outputs of xo (E) are not used)

Front high outputs of xo (E) go to the other 2 channels of the Jensen 432BBE amp (F)

Front mid outputs of xo (E) go to Carver 4150 amp (G). 2 x 75w RMS at 4 ohms Front low outputs of xo (E) go to an Audio Control Epi Center H, and then from (H) on to an Ultimate amp (I) rated at 2 x 150W rms at 4 ohms.

Power in the trunk is connected to 2 100,000 microFarad computer grade capacitors. These as I am writing this, I realize I must disconnect to assure they have not gone bad.

Plus the 12 v in the trunk trickle charges a small 12V battery through a diode and resistor. This is the backup for the alarm.

This is about as complete as I can describe my setup. I am looking forward to any other suggestions you may think of as you read this, and if you have any circuits you can share, that would be gravy on top of all you have already done. Thanks a million.

Happy between holiday period,

John

Inserting audio transformers inserted into your cables (you have to put them in both left and right channel) will be a troubleshooting aid, not necessarily a fix (yuck). This will eliminate magnetically induced noise and ground-conducted noise. It will help if you can identify between which components the noise is coming in.

Regarding the Honda test - an aging battery will indeed draw more current from the alternator - you didn't mention trying to swap in the Honda's battery into your car. An aging battery's higher internal resistance will also mean a higher ripple voltage for a given charging current.

There's a 4th way noise comes in which I haven't mentioned - inadequate "power supply noise rejection" of a component. That is, 12Vdc has ripple (form the charging system), makes its way inside the circuitry into the audio signal. The cure for this is to place a power line filter into the component. Basically an L and a C, the C right across that component's power leads, close to the component.

You said "Honda running, Mine not: Initially slight noise, but then quiet,

and no noise no matter how much load I put on"

Here's what I think - the initial slight noise is the ripple because the alternator is charging the honda battery to replace the drain from cranking.

I suspect the PD-2 is the dud in your system.

Placing a separate battery across the PD-2 will test this. If the noise goes away, and stays away even if the PD-2 power ground is left connected (IOW you separate only the PD-2's +12V), there's your problem (inadequate PSU noise rejection).

If what works is placing a separate battery as above but leaving the PD-2 power ground connected makes the noise stay (i.e. both power and ground have to be separated to kill the noise), then the PD-2's problem is a non-isolated input signal ground (it will fail the RCA to power gnd ohmmeter test). The cure here is to connect via a single wire, its ground to the ground of the transmitter (in this case the head unit). The fix is NOT to connect its ground to the common ground in the trunk, and don't put multiple grounds either. A single wire to the transmitter is the right solution. Also, this ground wire should be twisted along with the RCA signal cables between the head unit and the PDA-2. (to prevent magnetically coupled noise)

In summary, its the INPUT circuit design of components that's usually the problem.

Inadequate PSU noise rejection needs a power supply filter to fix. A ground conducted noise problem (fails input RCA gnd to power gnd ohmmeter test) means you have to connect its ground (the receiver) to the transmitter's ground.

You don't really need a single, massive ground point in a system. All a system needs is that all inputs are differential, and failing that, any components with badly designed inputs are grounded to the ground of the upstream component and NOWHERE ELSE. The idea is that the receiver sees the same ground as the transmitter, because the transmitter's output signal ground is almost always its box/case.

About isolated switching power supplies, many car audio components have them inside, for example, my Hifonics Plato EQ, and IIRC my Coustic XOver (which I'm replacing with a custom circuit, btw). This is a good idea, but IMO is overkill for controlling noise unless, but is good if the designer needs the voltage headroom.

I don't have a schematic offhand. Easy to make one though. The hardest part is sourcing the transformer core (toroids are easy to understand how to wind) and the layout. The layout of a switching power supply is absolutely critical - a bad layout can render a design inoperable, or a tremendous noise generator, at best. So unless you have a ready-to-follow PCB layout, you're stuffed.

A diff mode input circuit is easy. (diff =differential) A balanced current drive is not so easy. The latter, I would have to read up on in the literature if my ideas are correct - perhaps the Rane audio site has schematics of balanced line drivers as used in their pro gear. The former however may very well fix most if not all problems. I could give you a sketch from my work email address.

If the PD-2 takes the head unit spkr level outputs as its input, and has line-level RCA jacks as output, again a diff mode circuit will do the same.

No probs, I'm returning the favor from other strangers who've in turn helped me with other problems :-)

Good luck and keep me posted! I plan to add this conversation to my webpage, if that's OK with you.

Hi Jason:

First off, please feel free to add our dialogue to your web site. Your suggestions are both lucent and comprehensive; I have not found anything close to it anywhere else, so if you are building a collection of problems solved on your site, it will be amongst the best on the subject. Once My noise is gone, it will have come from a solution I have not found anywhere else.

The prime quality suppliers like Mouser list a 10A 12-12V unit for around $35. If they are so sensitive to being built right, this may be the way I will have to go if I can't find the right unit surplus. If you have any suggestions as to where to get one, I am listening. Does Astec have any cosmetic rejects that are internally 100% ?

I agree that my PD-2 is the most likely source of problems; it in combination with the head unit.

As long as I keep the Bose radio, I am married to a PD-2 or PD-4. Right now it gets its ground from the case of the radio. The hookup with the LEAST noise.

I know the Fosgate EQ has its own p/s inside, as may some of the others. If the DC-DC is the right route, I was thinking I could feed the PD-2 and all the processors by one DC-DC, or I could feed the processors by a biggish one and the PD-2 by a small one. This way they both have a ground that never touches the frame of the car and isolated 12V to boot.

The PD-2 does have its RCA ground shield isolated from the case, BTW. Also, the signal from the Bose is not speaker level, nor RCA, it's a hybrid in-between. It is symmetric and low level. That's why the PD-2 is a required piece. There would be numerous high quality speaker level to RCA converters to choose from, The PD-2 is the only one I have found that is specifically designed to be a Bose interface.

You described an LC noise filter. This too, I could try (Can you tell me some component values to use?) However, wouldn't this merely attenuate the noise? To my amateur observation, the DC-DC route would more likely eliminate the noise altogether if the electric system checks out to be in spec, and thus be a preferred route to the sought end.

What exactly is differential input? Isolated from case? Both channels not having their grounds connected? Both?

What is circuit?

Jason: Ongoing thanks!

John.

Hi John,

Yes most "DC-DC converters" in a catalog would be isolated. No an 18V input unit will not work on 12V. Astec doesn't have cosmetic rejects - they get generally get reworked, and there are no 12V-12V that I know of.

Interesting that the PD-2 has its ground isolated. But this doesn't mean it has adequate "power supply noise rejection".

Before jumping and buying any DCDC converters, do tests with a separate 12V battery first. Even 8 pcs of AA cells in series will do. You can buy holders in Radio Shack.

Here's what I would do:

Remove all components in the system to the bare minimum that produces noise, in your case, the head unit, PD-2, and amps. Isolate each component with a separate 12V power supply one at a time until you have no noise. I still think the PD-2 is it. Unfortunately the AA cells don't have enough capacity to power the head unit or the amps - you'll have to use a small lead-acid gel cell at least.

True a filter won't eliminate noise - but 50 dB of reduction would most definitely bury it in the noise floor of the system. A filter made of a 1 mH inductor and a 50,000 uF capacitor would provide 50dB of rejection at 500Hz, and more at higher frequencies: 12dB per octave. If you make a 4th order filter (an LCLC filter - 2 500uH inductors and 2 10,000uF capacitors) you get 50dB of rejection starting at 350 Hz and more, 24 dB per octave above that.

If you can put 100,000 uF of capacitance for your amps then you can put 10,000 uF for your PD-2. Capacitors for filtering the power into a low-power circuit like the PD2 need not be expensive - any generic 16V electrolytic capacitors will do.

A differential input circuit has 2 inputs, neither grounded, and the output is simply the difference between input1 and input2. A regular input circuit also has 2 inputs but one of them is ground. The problem there is that if you have say a signal carried on an RCA cable, the power ground to which a receiving circuit (for example an EQ) can have a noise voltage on it as compared to the ground on the RCA cable. The circuit then will think the signal ground is the power ground so the output will contain noise. A diff input circuit would only look at the signal on the RCA cable - that is the voltage between the RCA "hot" and RCA "ground".

I believe in car audio there are "line drivers" - my term, not theirs, which boost a head units output from 200mV to something like 4V rms, to send a large signal through RCA cable to increase noise rejection. Maybe these companies also sell "differential receivers".. again my term.

Let me know if the above test doesn't solve your problem.

I just realized that another trick that could work to get rid of power supply noise for the PD-2 is to build a small step down regulator for it - you could try an LM317 ( an "adjustable terminal regulator") with 2 resistors to set its output voltage to 11V - if the input voltage is 13 to 14V (when engine is running) then it can provide 11V output. This will provide your > 50dB of attenuation without the bulk of those inductors and capacitors. Those resistors to set its voltage would be something like 2.5 kohms and 20k.

Jason:

Your trouble isolating techniques have helped me what I believe is positively identify my problem. Poor old PD-2, got a lot of blame, but turned out not to be the problem at all. Today I gave the PD-2 an isolated 12V battery feed. Guess what? Noise. Interesting!. I ripped out all my RCAs and started measuring. The Cathedral, OEQ1 and Epicenter had infinity resistance from RCA shields to ground.

So did my sub and mid bass amps. The Jensen, the lowest ranked amp had about 200 ohms from shield to ground. But before I disconnected, I had a direct connect form shield to ground. Since I had noise, and it disappeared once I disconnected the PD-2, so it became an easy target for blame. BUT (drum roll please). the one piece of gear with its shields connected to ground is: The Carver active XO !!.

Now, that will still be a critical piece of equipment, so in order to silence it I connected the 12V battery to its terminals. Result: Dead silence. Beautiful, ecstatic, the way it was supposed to be. I could even hook up the signal shield to my PD-2 and get rid of the annoying switch off pops I had learned to live with.

Jason, just for confirmation, would the best, and possibly easiest solution be to only feed the XO with a 12V-12V DC-DC converter? It only draws about 100 mA, so I won't need a big one.

This was a good day in car stereo land. I thank you for your succinct, useful, and effective advise!

Happy New Decade!

Congratulations on your find. Why not try the following, and a diff amp circuit first, before a DCDC converter?

1) Connect the Carver active XO's power ground to the component that feeds it, through a THICK wire (gauge 8?) that is twisted or braided around the RCA cables that feed it. The Carver must NOT have any other ground connection other than To minimize the length of this wire, mount the Carver and the upstream component near each other.

2) Try the converse - connect the upstream component's power ground to the Carver through a THICK wire twisted or braided around the RCA cables that connect the two. That component's ground must not have any other connection other than this thick wire.

The above should minimize or eliminate the noise. If the noise is still bothersome, then build a diff amp input circuit for the Carver. I can't draw you the circuit today, but if you wanna see what it looks like, look in the pdf file I'll send you in another mail - it's a schematic of a Rane product. Aside from using 1 opamp per channel, you'll need a 3rd opamp to provide the "virtual ground". I'll give you a schematic in a few days. I would recommend using the TLC2274 opamp, which is a quad opamp from Texas with rail-to-rail inputs and outputs, with relatively low noise and distortion. This diff opamp will be much simpler to build than a DCDC converter.

Hi Jason:

I can try the ground cable to Carver trick. This will have to be a temporary try, though. The way I complexly routed the two rca inputs and 8 rca outputs to a sliding drawer would make modifying the cables a daunting task. Same goes for moving the Carver to the front of the car. However, maybe a ground cable from head unit to carver ground will silence the noise. If it does, I can still run an additional wire along the existing "snake" harness.

The DC-DC converter I would buy, rather than make. This setup will give me a great continuing test bed to experiment with the other suggestions you mentioned. I would like to make a differential circuit, so if you have a schematic, it will not go unused.

Thanks also for recommending an opamp. I have read that they vary greatly in quality, though not necessarily in cost. But I don't know which ones are the good ones.

I read the post of a guy who replaced all the opamps in a Jensen car amp and achieved notably better sound. I hate the dumb accounting type penny pinching you so often see in commercial designs. Genesis amps supposedly free of those compromises, and are truly a work of art, I have been told, but have a price tag that reflects that as well.

For future reference, I would love to have the ability to build a 10A or so DC-DC converter. (I have reasonable build skills), unless you say that even with few components it is really design sensitive and difficult to do right, I may abort that option. The one my friend is using has a "dialable" output voltage. I have seen that in some of the specs of existing finished units, as well as a range tolerance for input voltage.

I'll see if there is a book in the library that elaborates some more on the ins and outs of that particular application.

Thanks in advance for the schematics you have thought of. I'll keep you posted on progress and final outcome.

In review:

The (grounding method) cure:

You have to connect the grounds of the unit with the lousy input circuit, and the unit(s) feeding it, together. This is the key.

i.e. connect the following grounds together: the XO, the Cathedral and the EQ.

The head unit and PD-2 have nothing to do with it.

The fact that 2 units are feeding the XO will make it a bit more complicated.

As an experiment, 1st disable the rear channels - i.e. unplug the RCAs going into the XO rear channels. Connect the XO power ground to the EQ using a thick wire twisted with the RCAs connecting the front channels between the EQ and the XO. Remember there must be no other power ground connection for the XO other than this thick wire. The output RCAs of the XO don't count as a power ground because the components downstream of the XO (the amps) have RCA input grounds that are isolated from power ground.

Or try connecting the EQ power ground to the XO via a thick wire instead.

Noise should be substantially reduced or eliminated.

Then add in the rear channels - to get the 3 components working together (XO, EQ, Cathedral) noise-free given that the XO has no diff inputs, the 3 units need to have a single point ground (which can be a bit separate from the rest of the system). The EQ and Cathedral power grounds should each have a thick wire to the XO. Again, for the EQ and the Cathedral there should be no other power ground connection other than the thick wire. The EQ's ground wire to the XO should be twisted along with the RCA cables connecting the EQ to the XO.

The Cathedral's ground wire to the XO should be twisted along with the RCA cables connecting the Cathedral to the XO. The XO will act as the single point ground. The XO can then have a thick wire to your grounding block.

If the remaining noise is still not reduced enough, then you have to build a diff input circuit - IOW you're endowing the Carver XO with a diff circuit that the penny pinchers should have put in the first place.

I'll tell you about opamps later.

This circuit should also work.

It's still a power line filter but note that it has an inductor in the ground leg. The inductor in the ground leg isolates the ground to audio frequencies, but lets the power (dc current) thru.

Oh, and you can try the following experiment too:

Running only the front channels:

If the EQ shows 0 ohms from its OUTPUT RCA grounds to its power ground (unplug all cables 1st when doing this test), then it can get its ground from the shield of the RCAs going into the XO. Disconnect the power ground wire of the EQ.

The EQ uses little power so the power supply current flowing through the RCAs from

the XO will be small. Doing this gives you a single point ground between the EQ and the XO.

If this works, add in the Cathedral. Again if is shows 0 ohms between the output RCA ground and its power ground, it can get its ground through the RCA cable to the XO. Again remove or disconnect the Cathedral power ground wire.