[PLMONROE]

This has been driving me nuts for over a year! One of my several amps (I really need to clean house) is a Parasound HCA-3500 which has been modded to the gills by Steve Nugent of Empirical Audio. It is one of my nicest amps (and ugliest) which Steve claims sounds better than the Parasound JC-1s except in the area of dynamics. A year or so ago I changed speakers to VMPS RM-V60s only to find that every time I would hook up the amplifier to them it would shut itself off. Other amps worked fine with the speakers so I ended up shipping the heavy beast back to Parasound, not once but twice. Each time it came back with a clean bill of health. I switched interconnects, preamplifier's, power cords, and everyting I could think of still the amp would shut itself off whenever I applied power.  The other night I was dozing off and had one of those periods when things pop into your head. Yes I had tried changing everything in the would even fuses, well ALMOST everything. I had not changed speaker cables and I had gotten new cables at the same time I had changed speakers! But what the heck can be the effect of a piece of wire? Well I unhooked the speaker cables - Acoustic Zen Holograph IIs which are nice cables and replaced them with a set of Kimber Kable 8TCs. The amp loved them !!! So the amplifier is OK with Kimber Kables, Audioquest Pike Peaks, plain lamp cord and everthing else except the Acoustic Zens. I checked resistance to see if there was an internal partial short but everything was fine, besides the cables work fine with ALL my other amps, both tubed and solid state. John Casler thought that perhaps  it might be some sort of capacitance issue. Con someone please suggest what might be going on before I lose my last visages of sanity?

[jules]

Someone more tech savvy than I might well be able to improve in this but ....

it's possible that capacitance is the issue. To test, you could try using some very basic non-paired wire. Some amps don't like capacitance it seems. The separate wires will reduce [massively] capacitance but at the same time they will raise inductance, or more precisely, they'll remove the cancellation of self inductance that occurs with two proximate wires carrying the same flow in opposite directions. You win in one direction, you lose in the other BUT ... it might be worth doing if that's what your amp likes

jules

[art]

It is the cable capacitance. The designer of that amp does not believe in using output inductors. He feels that he has sufficient phase margin that the amp is unconditionally stable. (You can go over to one of the hobbyist forums and look for his rantings if you don't believe me. I know the guy, so I know how he feels.)

Some don't believe him. Other people muck around with it, and no telling what that has done to the phase margin.

Get some standard speaker cables. And don't let people mod your gear. (Well, maybe me, but I am not in the mod business..............!) Not all are competent.

On the general subject of cable inductance and capacitance.........any time you raise (or lower) one, you lower (or raise) the other. The Laws of Physics. Can't get around them.

Although I did have some guy who makes cables tell me that he found a way. I think the letters "BS" were in his product name. Coincidence? Anyway, I told him he was nutz and never talked to him. (A mutual dealer insisted that I contact him. He never made that mistake again.)

Pat

[PLMONROE]

Jules and Pat, I REALLY appreciate your inputs, I really do. However I don't feel I have completely solved the mystery yet for several reasons so I guess I  need to expand on this a bit.

It is the cable capacitance. The designer of that amp does not believe in using output inductors. He feels that he has sufficient phase margin that the amp is unconditionally stable. (You can go over to one of the hobbyist forums and look for his rantings if you don't believe me. I know the guy, so I know how he feels.)

.

Parasound's HCA-3500 was designed by Gordon Curl who is pretty far up the respect "food chain " of many. It was the parent for the design of his JC-1 which is still a Stereophile A rated amp and was Sam Tellig and Stereiphile's joint Amplification Product of the Year in 2003. So, in theory at least, it should not have any design defects.

Get some standard speaker cables. And don't let people mod your gear. (Well, maybe me, but I am not in the mod business..............!) Not all are competent.

AMEN! Although many on this forum have far more expensive and exotic sound systems than I, I am not about to let Joe Soldergun  of the local "TV Repair a Furniture Outlet" touch my system in which I have over 50k invested! But please don't pass that number on to my wife.   Only two people have ever laid a hand on any pieces of my equipment and those are Dan Wright of Modright and Steve Nugent of Empirical Audio both of whom are here on Audio Circle. Steve, who did modify my HCA-3500, is an Electrical Engineer and is one of the top four or five experts in the country on jitter. I think about all he did was upgrade parts and wiring. Anyhow remember that it was AFTER his work that I twice sent the unit back to Parasounds main facility in San Francisco and twice they sent it back with a clean bill of health. Had any errors been made in modification I am sure they they would have surfaced on one inspection or the other. In fact, Parasound ran the unit on its last visit for almost one solid week with several speakers and several speaker cords and could not duplicate my problem. So, I discount that any of my difficulties were the result of Steve's modifications.

As for standard cables --  dunno?  Bill Gates might term Nordost Vahalllas as standard cables while my neighbor might lay that label on Rat Shack's 7 cents per foot speaker wire. The Acoustic Zen Holograph IIs cost several hundred dollars a pair so do not exactly fit in the category of either of the two previous. The point is that they are cables which have received good reviews (else I wouldn't have them) and which must be owned by at least several hundred Audiophiles. So I am loath to lay blame on them. Besides they work superbly in my music room with other combination of amplifiers and speakers.

The obvious solution is simply not to use the Acoustic Zens which is the path I will take. This leaves two lose ends, however. One is my curiosity as to the root of my problem. The second is that these are too expensive to just "throw away", I would like to sell them however how can I ethically list them on Audigon if there might in fact be some hidden defect there??????

To take the mystery a bit further, my amp works fine with my speakers as long as I do not try to bi-am them  If I bi-amp and hook my amp and Acoustic Zen cables to the mid and treble portion of the speakers (and put my tube amp on the bass) everything is fine, if I just use my HCA-3500 and Acoustic Zen cables to drive the whole speaker system every thing is fine, BUT when I be-amp (with tubes on the top)  and power the bass with the HCA-3500 and Acoustic Zen cables the durn thing shuts off!  OH yes, one other thing -- if I insert one of those two-bit (actually $2.95) in line -6db attenuators into the RCA input of the HVA-3500 two things happen . One, I get a lot of hum and I can assure you that otherwise my system is absolutely humless. And two, the amp will NOT kick off line when the Acoustic Zen cables are attached! Go figure! (I don't know why but I think this is a different issue). Bottom line, this is driving me nuts!!! 

[mgalusha]

Pat is dead right, it's the cable capacitance. I have a friend with JC1's and RM40's. I'm not 100% sure which cables it was but I think they were Cardas and it did this exact thing. The amps would shut down almost immediately. He could use the same cables with JM Labs Mini Utopia and the Parasound's and they were fine but it was something in the combination of the cables with the VMPS speakers that would cause the amps to oscillate and activate their protective circuitry.

Perhaps the crossover in the VMPS offer a high capacitance load which when combined with the cable capacitance was enough to send the amps into lala land. Oh yeah, his was the same way as yours, they would only shut down when we tried to biamp them, when driving the speaker full range it was OK. From what I have read the HCA-3500 and the JC1 are pretty similar in design.

BTW, not using and output inductor isn't a design defect, it's a design choice and a trade off like any other design choice. Lots of SS amps have them to provide stability into any load. Some don't, which makes them more prone to instability. I can tell you from experience to be happy the protection circuity kicked in, without it things can get very expensive when an amp goes unstable.

[Gordy]

What I find odd is that the TC'8's worked well.  I've been under the impression that the heavily braided Kimber Kable and the Alpha Core ribbon cables were among highest capacitance cables around.

[denjo]

Just a suggestion, you might want to email Acoustic Zen to let them know what you have been experiencing with their cables. and I am sure they will be able to shed some light to this "mystery".

Best Regards
Dennis

[PLMONROE]

Good idea! Will check to see if they will be at RMAF and if so will button hole them there.

Paul

[audioengr]

Paul - I agree with the other assessments: It's the combination of the speaker load and the cable capacitive load.  Many folks have had this problem with Goertz cables because they are essentially linear capacitors.

John Curl does have a Zoebel network on the output of both the 3500 and the JC-1's, but I dont think this is enough to counteract a large capacitance.  It's probably many thousands of pF.  It may also have something to do with the amount of negative feedback that he is using and the really fast output transistors.  It's a double-edged sword really.

Steve N.
Empirical Audio

[Daryl]

The problem is with the amplifier design.

I remember a stereophile review from a long time ago where they were having stability issues with a certain speaker and Parasound amplifier.

The stability issue isn't hard to design around, especially when your amplifier doesn't have a particularly low output impedance anyway.

The reason why your speaker wires cause the amplifier to lose stability with the woofer section only but not when driving the whole speaker or just the mid/hi section has to do with the impedance the wires are terminated with.

Speaker wires have both capacitance and inductance but they cannot have both at the same time at the same frequency.

The load at the amplifier can be resistive and reactive at the same time but the reactive portion of the load must be either capacitive or inductive or in the middle (zero reactance).

You're probably wondering now whether your speaker wire will have an inductance or capacitance and why.

The answer is related to the characteristic impedance of your speaker wire.

If the load impedance is higher than the characteristic impedance of your speaker wire then your speaker wire will be capacitive.

If the load impedance is lower than the characteristic impedance of your speaker wire then your speaker wire will be inductive.

This situation will change with frequency (load impedance and the wires characteristic impedance).

Essentially when you connect the low frequency section by itself you then have no load upon the speaker wire at high frequency (high impedance).

Because the high frequency load upon the speaker wire is a very high impedance the speaker wire loads the amplifier with capacitance at high frequency.

With the mid/hi section connected to the speaker wires the high frequency load upon the speaker wire is below it's characteristic impedance and instead of adding capacitance the speaker wire adds inductance.

Adding a conjugate network to the woofer section to stabilize high frequency impedance to a value below the characteristic impedance of the speaker wire would fix the problem.

Or simply get an amplifier desgned with more rigorous performance criteria.  

[art]

The problem is with the amplifier design. 

No, the problem is self-appointed experts like you bloviating on what you think is wrong. Correct me if I am wrong, but just how many commercially available products have you made? I bet the answer is zero.

Go start your own company. Figure out how to come up with money to do it, the technical and manufacturing expertise needed to do it, as well as the marketing expertise, and then maybe your constant pious remarks might have validity.

And for the record, it is John Curl.

As I mentioned, some think some of the Parasound amps might have a problem with a certain type of load, wrt to impedance level and phase angle. Yes, it certainly is possible. And it can happen, even with rigourous testing and design.

But anyone who really knows anything about this subject already knows that.

As for Steve N:

Don't know the guy. May have exchanged messages with him over the years. Probably knows what he is doing.

Should we hold him accountable if his work does absolutely nothing to the amp's phase margin? After all, judging by the remarks of some (ok, one guy) we should tar and feather him because he did not find that one obscure case that could possibly cause the amp to go nuts.

Of course we shouldn't. But maybe next time he will think twice about it.

Maybe. Maybe not.

In any case, none of us outside of Steve knows what he did to it, and since the o/p did not tell us, let's not run to conclusions that he erred. But for his sake, I will assume that he has taken notice of the situation, and will due his best to keep his reputation clean.

Translation: don't work on amps without a series inductor. Odds are it is stable, but once your name gets stuck on it...............the chickens might come home to roost.

BTW, I assume that you are including me in that number of jitter experts. If not, you haven't been paying attention.

Pat

[Daryl]

The problem is exactly how I explained it, and I went into great detail.

Whether someone has a company or makes products is unrelated.

I know of John Curl and as detailed in Stereophile in the past as well as the two cases above these stability problems with Parasound amps are not uncommon.

Some capacitance is to be expected at the output of an amplifier and to not design the amplifier to tolerate it is incompetent.

[Niteshade]

A good web site:

http://www.audioholics.com/reviews/cables/speaker-cable-reviews-faceoff-2/speaker-cable-reviews-faceoff-2-page-7

I would like to point something out: I can't find in the study what frequency they were testing the cables at to derive their inductance and capacitance. I assume it's 1khz. I also couldn't locate the amount of cable being tested, to be exact. It's either 1 meter measurments or 1 foot measurements.

From what I gather about solid state amps, a few (even thousands) of picofarads shouldn't bother them. A picofarad is a very,very extremely small amount of capacitance. It takes 500,000pf to equal 1/2uf. It's my estimate that even 1uf of capacitance shouldn't bother things that much. There will be high frequency roll off at 1uf, but not that much, maybe not enough to tell.

An amplifier should not be so sensitive to cables.

Another web site to look at:

http://books.google.com/books?id=d7ft6F8ZUdcC&pg=PA288&lpg=PA288&dq=speaker+cable+sensitivity&source=web&ots=1gWMVxF63s&sig=wGMNqanIhyJd9ZEbjTB_S5HyBLY&hl=en&sa=X&oi=book_result&resnum=8&ct=result

[avahifi]

The problem is indeed excess speaker cable capacitance.

Do any of you remember the old scope photos of amplifier square wave tests in the mainstream audio magazines?  Almost always the amplifier's response fell apart when a capacitor was connected across the speaker load.  The same thing is happening here, no matter how much you admire the designer of the amplifier and the speaker cable.  Mother Nature does not care.

What is really happening is the fact that a capacitive load slows down the output stage.  This causes the feedback signal to arrive late at the summing node.  This causes the negative feedback to go positive, causing high frequency signal peaking at best, and full bore instability at the worst.  The instability (ultra high frequency oscillations) generate excess heat in the output stage. The amp in question is obviously well designed to protect itself, a lesser amp often will simply melt down under this condition.

Excessively capacitive cables are always bad news, no matter how much glowing purple prose has been written about them by people who never use a capacitor meter.

Regards,

Frank Van Alstine

[Niteshade]

I regards to amplifiers sensitive to capacitive loads, what is the maximum capacitance the amp can see before it becomes unstable? I realize it must vary between models. Is there a median value?

 

[richidoo]

Since this is the lab, I'll ask:

How does one correctly measure the capacitance of a speaker cable?  One conductor, I guess you need access to the shield then measure just as any capacitor?

Anticable is an example of no shield. How to measure capacitance of that?
Thanks
Rich

[jneutron]

Speaker wires have both capacitance and inductance but they cannot have both at the same time at the same frequency.

I think you meant to say..""they will present an I/V relationship which will be either capacitive or inductive, but not at the same time.""  A speaker wire will of course have both present at the same time, but how the amplifier sees the combined load is what you are actually referring to.

The answer is related to the characteristic impedance of your speaker wire.

If the load impedance is higher than the characteristic impedance of your speaker wire then your speaker wire will be capacitive.

If the load impedance is lower than the characteristic impedance of your speaker wire then your speaker wire will be inductive.

This situation will change with frequency (load impedance and the wires characteristic impedance).

Essentially when you connect the low frequency section by itself you then have no load upon the speaker wire at high frequency (high impedance).

Because the high frequency load upon the speaker wire is a very high impedance the speaker wire loads the amplifier with capacitance at high frequency.

With the mid/hi section connected to the speaker wires the high frequency load upon the speaker wire is below it's characteristic impedance and instead of adding capacitance the speaker wire adds inductance.

Eloquently stated.

It is nice to see another who understands T-line theory..

Cheers, John

[jneutron]

A good web site:

http://www.audioholics.com/reviews/cables/speaker-cable-reviews-faceoff-2/speaker-cable-reviews-faceoff-2-page-7

I would like to point something out: I can't find in the study what frequency they were testing the cables at to derive their inductance and capacitance. I assume it's 1khz. I also couldn't locate the amount of cable being tested, to be exact. It's either 1 meter measurments or 1 foot measurements.

If you go to the previous page in the link, you will find a table of L and C values for the cables.  The measurement frequency ranges from 100 hz to 1 Meg.  The higher frequency is a result of my request to Gene, as when I origionally proofed his article, he did not go that high.

Running the tests out to a Mhz allows one to view how the meter is interacting with the cable setup.  The data remained well behaved out to that frequency, the LC product remained within an acceptable range.

A question:  I do not seem to be able to post a jpeg of the effective dielectric constant for that article for all the cables through the frequency range tested..  It was not included in the origional article as the information was not considered of value to the article (I concurred with that thought.)  But I haven't posted here in a while, I thought I knew how to load pics..

Cheers, John

[opaqueice]

Essentially when you connect the low frequency section by itself you then have no load upon the speaker wire at high frequency (high impedance).

Because the high frequency load upon the speaker wire is a very high impedance the speaker wire loads the amplifier with capacitance at high frequency.

What is really happening is the fact that a capacitive load slows down the output stage.  This causes the feedback signal to arrive late at the summing node.  This causes the negative feedback to go positive, causing high frequency signal peaking at best, and full bore instability at the worst.  The instability (ultra high frequency oscillations) generate excess heat in the output stage. The amp in question is obviously well designed to protect itself, a lesser amp often will simply melt down under this condition.

Excellent explanations - thank you both. 

Question:  following this logic, if I attached a very low impedance load and the inductance of the speaker cable was relatively high, it seems the same thing could happen - the phase delay caused by the cable inductance could be large enough to put the amp into positive feedback.  But I've never heard of that happening - is that because the load impedance is rarely low enough for that to be a problem?

[Niteshade]

I would believe capacitive sensitive amplifiers will have problems in the higher frequency range. Ideally, amplifiers want to see a purely resistive load. I would believe loads that are overly inductive in nature can cause problems as well.

Three scenarios:
1) Resistive: Looks like a non-reactive load, neither voltage or current are out of phase.
2) Capacitive: Current leads voltage
3) Inductive: Voltage leads current

An amplifier should be made to accommodate imperfect (reactive) loads as long as the faults are not unreasonable. I would like to know what others believe is unreasonable. A few picofarads shouldn't upset an amplifier nor should a few microhenrys. If it does, the amp has issues. Not all amps are this sensitive.