Still using the 8 gauge ultra pure copper wire in place of the fuse. Though I very much like the peace of mind of a SR Blue fuse (the next best thing according to the crude tests I did) the difference with the copper wire is addictive. I am trapped with this unwise decision. I am not as much concerned about equipment damage. What concerns me most is it being a fire hazard. I take my precautions as with always being around. Always being around will not prevent equipment damage as it happens fast - if a fault develops- that only a poised safety device like a fuse can prevent but I can easily prevent a fire.

Needless to say while testing e.g. after every upgrade I first start with a fuse in and after ensuring stable performance for a few hours usually the next day or two I switch back to the 8 gauge wire. Same when I switched to the KT120s for the first time just in case the they developed a internal short in transportation.

I don't think I need a fire extinguisher poised and ready as the probability of a short and a fire is very low after the initial test phase.

Only if I leave the area and the initiating small spark propagates (there is very little highly combustible material internally in the ZMA) then there could be a fire hazard. Since I am around at the smallest sign of trouble I just need to turn it off.

Learning 5:

Having done many changes over a short period of time, I have refined a a method for judging effectiveness of an upgrade.

When trying to ascertain whether an upgrade was effective one has to make a concerted effort to remember that change by itself in not an indication of greater accuracy and to try to objectively judge if it is just change that one is perceiving. Very often because we psychologically want improvement to show up being humans it is possible to perceive change as improvement and pat ourselves on the back thinking that money and time was well spent. In a way I am thankful to Bybee "quantum" products for teaching me this until it finally dawned on me a year or so later. E.g. A/B testing soon after buying the Bybees compared to how I perceive them now is different having lost the initial expectation from the hype by the manufacturer. This is more likely to happen if one has no reference point to gauge with. Because after an improvement is also a change. I have over time realized that, that reference point for accuracy is live music and not another higher end system.

While comparing between systems or before/after an upgrade, it is often helpful to isolate each instrument mentally and decide if a saxophone's sound is close to a real saxophone, likewise for other instruments and voices. To this point I very often find judging the sound of an applause, catcalls and all (of all the things!) to be a better indicator for judging accuracy. Live applause seems to be more layered being densely packed with humans shoulder to shoulder and front to back with each point in the soundscape producing a different frequency and all in a tight midrange.

This is my own technique developed over time about dealing with judging if an upgrade is effective. It will be great to hear and learn from you everybody and especially Steve as he has been voicing his builds for donkey’s years now.

Expanding on contact quality mentioned in the above post.

This is the reason for my resulting in dumping all attenuator bar one in the entire audio chain (have one at source equipment only) as for me they are the single largest cause of electric pathway degradation. The wiper leaf spring force attached to the knob spindle has to be deliberately designed to be inadequate (from conduction point of view) to allow easy turning of the knob by weakest of fingers in a given population (male, female, adult, teenagers, children) i.e. if I were designing one I would design it such that a kid (just because an adult may not want to be separated from the cozy chair and askes his kid to turn it for him). Because as a designer I know lay people are not thinking of signal conduction but "who is that moron that designed it to make it so hard to turn". The effect of inadequate spring tension is like not tightening the screws tightly on the capacitor connections. Compounding this weakness in contact integrity is that the attenuators are in the signal pathway - the most delicate, fragile and sensitive of all the electrical currents flowing within an amp. Had they been in the power supply circuits instead the shear magnitude of the current would have bamboozled through this weak connection but alas it has to be in the weakest and the most sensitive pathway.

PCB that had micro cracks on the solder pins on both the top edge and bottom edge pin lines. Soldering these without shorting the adjacent pins was something else. On the other side of the PCB are the micro LEDs (dot matrix) that gives information on the front display so they heat up a little bit causing the pins to expand while running and contracts when turned off. The many on/off thermal cycling over 19 years caused the solders to fatigue and crack.

Your burn in take is cynical but I suspect correct.  For myself, I lose my reference super fast and despite things being improved, I start to think it doesn't sound as good as time goes by as I get used to the new normal.  Capacitors ought to be different than tubes since the materials are more "organic" and presumably flow with heat.  Do you find the same 5 to 6 hour window for them?

Even if you were being cynical, I didn't mean it as a fault.  There is more snake oil than snake these days.

I personally concur with your short burn-in findings.  But I always figured it was my lack of discernment.  The metallurgy explanation makes a lot of sense to me.  Do you feel the same way about ICs and other low heat devices?  Do they burn-in at all?  I don't personally hear a change with time for these kids of things.

Sorry, I used IC to mean interconnect cable and not integrated circuit.  

Bonny,
I definitely feel the burn-in phenomenon is "type-specific."  You raise some interesting points of discussion with regard to voltage and burn-in effect.  I can't say that I disagree that  the long stated burn-in periods  from vendors are a product of sales/marketing to exceed the customer's trial period.  I have often been suspicious of these prolific claims.  

One thing I do want to comment on has to do discussion on oxidation.  Copper Oxide (CuO) is a very poor conductor.  If you have in your system, raw exposed copper, then you must be aware of CuO and rid it as it develops.  It is not a great conductor, so as you consider skin effect, CuO is the enemy.  

Silver on the other hand does not share the same inherent handicap.  Silver Oxide (Ag₂O) is in fact a great conductor.  That said, one need not worry about their raw and exposed silver cabling oxidizing like is the case with copper.  Ag₂O is a great conductor and scientifically proven to conduct as well as Ag.  

What does this mean?  Well, that depends.  If you are using interconnects or speaker cabling with raw metals that are exposed to oxygen, then it is important to know if the Oxide of that metal type will have a detracting effect on the signal, or not  Ag₂0 is nothing to worry about.  

Jeff, aging is measured in years anyways v/s burn-in which is in hours so aging is less of a consideration.

I have not checked the resistivity values of Ag v/s Ag2O I'll take your word that they are equal. To me they should be different perhaps not by much but I being a nitty gritty person and wanting to squeeze every ounce of performance that even a small difference between the two will drive me to do something about it. That is why I buy 6N or 7N silver wire instead of 5N or less. The difference in conductivity here is even less I would think than oxide/no oxide on silver, but I still do it because for interconnects and internal signal wires from the RCAs to the input tubes (mentioned in one of my earlier post on this thread) I like to preserve the delicate input signal as much as I can.

Having said that both gold and silver being noble metals do not oxidize unlike copper. The tarnish you see on silver is more of a sulfide than oxide. Sulfide I believe is a bad conductor therefore e.g. RCA connectors are not usually plated with silver but gold (gold being a more noble of the two metals resists reaction with chemicals in the air).

You made a good point on metal inside of insulation being oxidation free. That is a good discussion point. My take is oxidation happens not only to exposed wire but also to the insulated portion (at a much slower rate) of the wire. This is because of permeability of the insulation. If you have any spare 10 years or more old wire strip off some insulation and you will see that there is an oxide/sulfide layer or what ever have you with the pollution in the air these days. If you use a sand paper on it you'll than expose the parent metal.

What we all experience with exposed silver is Oxide. Oxide is a chemical compound formed when silver and oxygen react. That is the tarnish on our untreated, exposed silver cabling and interconnects.

And you are correct Archie, research has shown that the oxide of silver is actually a better conductor than pure silver.

As far as silver sulfide, that would be extremely abnormal. Sulfur is not generally present in ones home. If it was, you would smell it!  :-[  Silver Sulfide can only form when silver is exposed to sulfur and they react, forming the compound Silver Sulfide.

Ha! I stand corrected.  Bonnie is right in that the tarnish is in fact a Sulfide.  I just learned that it does not take much environmental sulfur concentration to form the tarnish on silver. Tarnish on  copper is in fact a product of oxidation.  

I guess there has been too many decades since my last college chemistry course  ;D

Thanks Bonnie and Archie.

Thank you guys!  What's amazing is that every site I've seen about using silver wire sights the fact that silver oxide conducts better than silver.  Yeah, so ... ?

Jeff, Archie:

In the bargain we (well you both) might have discovered a brand new method for enhancing audio performance further. If silver oxide is a better electrical conductor than pure silver what if we deliberately accelerate the oxidization of silver? Because silver is resistant to oxidation in air what if it is exposed to higher concentrations of oxygen, say the oxygen used by medical patients or if it is available from somewhere else?

But for the life of me, I am finding it difficult to find the actual value of resistivity of Silver Oxide unlike you Jeff, Archie.

For pure silver it is 1.59 × 10−8 Ωm at 20C what did you find the value is for Silver oxide?

If the value is lower as quoted by a accredited lab or research paper I might try exposing a pair of silver cables to higher levels of oxygen if I can find it easily.

I mean we buy cryo treated cables and all kinds of things so why not expose silver wire to concentrated oxygen if that will squeeze more performance?

Sounds crazy now, even far fetched, but you never know it might be the next rage just like cryo treatment has now firmly established itself.

Sorry to report that, that research paper does not apply to our discussion. It applies to electron beams and their propagation (conduction) in optical media with varying thicknesses of Ag2O applied to the surface not really to transmission of electricity in conductors such as wires.

Learning 8:

I have found that turning ON the ZMA about 15 to 20 minutes before actually starting to play music is the quickest way the amp reaches its sweet spot.

I have consistently found that if I start playing music right away it takes longer to get to the sweet spot and the louder I start playing music from the very initial time the power is turned on the longer it took to reach a point where the SQ stabilizes to that fantastic level.

This phenomenon may be unique or more appropriately exaggerated in my ZMA because I have an inordinate amount of power capacitance: to remind every one it is 13500 uF per side.

On another note the experience with adding capacitance has affected my belief in it so much that I have now ordered extremely fast oil capacitors that I am going to add to the 13500 uF. I am limiting these oil caps to only 70 uF (even at 70 uF they are huge in size) as the objective here is not to add further storage of power (diminishing returns on that front) rather to compliment the already large capacitance currently present. By adding even more speed to the “power package” to deliver the very first micro burst of supply power as and when demanded before the other larger capacitors take over.

Learning 9:

Obligatto Film Oil Power caps install:

These are true metal (aluminum) film capacitors filled with oil. Not pseudo film caps such as metalized polypropylene. Copper or silver film would be better but the price increases substantially and for power supply silver would perhaps be an overkill unless you have money to burn. I must confess buying the Obligatto caps was by chance. I bought these only because these 70 uF ones were 50% off in price and thought why not try them out.

For small 70 uF capacitors they are pulling above their weight. I am tired of extolling the benefits of this and that, but these definitely are making the highly modded ZMA even better. Had I experienced what I am now, I would have bought the largest Film oil cap(s) that would fit in the capacitor opening in the ZMA top plate when I first replaced the large red Mallory 2000 uF power capacitors with Mundorf Mlytic 4700 uF HP+ 4 pole capacitors. These turned out to be even better than the Mundorf Mlytic’s 4 poles for imparting added dynamism.

I am still wrapping my head around as to how the relatively small value 70 uF Film Oil caps brought more air, depth, timing and agility to an already fantastic output.

Which led me to do a bit of experimentation:
To zero in on the contribution from the film oil caps and get acquainted with their properties I disconnected all of the 13500 uF of capacitance and only left the 70 uF oil film caps in the power supply. Intent was to enable comparison to the 13500 uF of electrolytic capacitor aggregate setup that I was well familiar with by now. I can say that this was perhaps the biggest learning for me thus far. For almost all of the other upgrades mentioned in earlier posts above I would say they came out as expected with no major surprises as they were careful calculated buys (except perhaps the Bybee purifiers).

These oil film caps at only 70 uF improved dynamism better than all the 13500 uF electrolytic caps put together even considering the fact that mine were the top grade 4 pole low ESR/ESL 4 pole Mundorfs Mlytics. These relatively small oil film caps resulted in the timing being more to the point. There was a loss of air and extension and noise floor increased. I think this is because of the low value of capacitance them being alone. I am certain if I had oil film caps with a total of 13500 uF or even 3000 uF the results will outstrip the 13,500 uF electrolytic caps.

I than installed the 13500 uF electrolytic capacitor aggregate back along with the 70 uF oil film capacitors. This setup was the best, better than either alone. Together the background got lowered considerably also besides enhanced airiness and that all important timing improved palpably.

As mentioned earlier but worth repeating with perhaps more detail and perspective. I wish I had discovered these earlier, before I purchased the Mundorf caps (6800 uF and 4400 uF costing about $800 in all) I would have instead bought the largest oil film caps that I could fit. The oil film caps are 3 times the physical size compared to an equivalent capacitance electrolytic cap. I do not think I could fit an oil film cap greater than 600 uF in the same place as the original 2000 uF Red caps. But I believe 600 uF of oil Film caps would be awesome.

In hindsight, I now think there is no point in having large capacitance caps if they are slow (high ESR). Music is all about timing. Fast caps do not allow the circuit voltage to sag during instantaneous demand spikes while the music is being played (a deep bass note here and there, rhythmic bass notes and other demands such at another micro instant more than normal number instruments / voices output higher demand notes together). When the amount of instantaneous required power supply is not forthcoming at the instance it is needed that is a lost opportunity and is fleeting and timing of the music is affected due to circuit voltage droop. By the time the slower electrolytic caps (even though they may be very large) respond the train has already left and having missed the train the power supply now is always playing catch up thus affecting timing. Mind you the Mundorf Mlytic 4 Pole caps are way faster than the original stock ZMA red Mallory caps and resulted in improvements mentioned earlier in this thread but these oil film caps are in another league in terms of speed. The very best would to install very large fast oil film caps and the larger the capacity and fast the caps are gives greater isolation from the grid (the noise floor becomes eerily quiet) while not causing a depression in circuit supply when demand increases due to faster charging/discharging cycles.

Now I can say that ideally all of the power capacitors should be oil film type. But that would mean capacitors physically the size of the whole footprint of the ZMA amp (I think) since film oil capacitors size  to power ratio is very large. So given that space is a constraint, a hybrid mix of oil film and electrolytic caps would be the solution. This way the faster oil film caps will serve the initial micro burst of power demand while gargantuan electrolytic caps can than take their sweet time continuing to maintain the downstream circuit supply.

I’ll slowly be adding more film oil caps to my capacitor aggregate in the power supply, perhaps even remove the red Malory 2000uF caps to accommodate the new caps.

Sometimes good things happen by chance.

I do not have as much experience with swapping different types and sizes of capacitors as you Steve (not even close), but now that you mention it I can recall feeling the same: physically larger film caps have more whoomph than smaller ones of the same capacitance value e.g. in the above case of the Obligatto Film oil cap install and earlier when I installed the 47uF Mundorf Tube caps in place of equivalent capacitance F&T electrolytic caps.

In the power circuit of an amp I generally tend to go for the largest capacitor value based on what will fit physically and quote the size in terms of capacitance i.e. the largest value possible in order to achieve battery like power supply conditions. But also the Time Constant T (tau) of a capacitor is perhaps more important in the power and signal circuits. Where as very specific capacitance values are required in the noise filtration circuit (which of course is a subset of the power circuit) to filter out specific frequencies of noise (e.g. 120 Hz noise from a fast full wave solid state diode based rectifiers or very high frequencies from switching power supplies from other equipment in the house or neighbors or an electric arc furnace in a foundry in the town somewhere).

A lot of experimentation and expense can be avoided if manufactures of capacitors meant for house hold equipment give extensive specification details including the time constant and ESR/ESL values. Very few if any do that. They generally only specify the capacitance value, the temperature rating v/s life in hours and some times if ever the ESR/ESL values and almost never the Time Constant (Tau) of a capacitor and talk in terms of sweetness, mellowness, warmness and other adjectives. I must admit these adjectives do not help me at all in buying decisions as these are subjective even to the extent that someone's sweetness is another person's mellowness. I also have a hard time with what sweet , warm sound etc,.is.

An attempt to lower the internal temperature for longevity of capacitors.

Patterned holes drilled in the bottom panel - see picture.

Was not comfortable with the fact there was no ventilation of the internal space. Especially as the base is made of wood which is a poor conductor of heat compared to a metal enclosure. Most audio equipment has metal enclosures, as well as ventilation slots all around.  The metal enclosure dissipates heat through conduction (along with internal heat sinks) and the ventilation slots dissipates heat by convection.

Those look like they are too few holes to make a substantial difference but there was no point as the openings at the top were limited to those around the two large capacitors where they protrude through the top plate. The cross-sectional area of the sum of the holes in the bottom plate is about 2 times the circumferential gap around the two large capacitors at the top plate. The 4 pole 6800 uF Mundorf caps are about 2 mm smaller in diameter as compared to the original red capacitors which helps with the ventilation.

I do not think this will decrease the internal temperature by much (have not taken before and after temperature measurements) but it is better than being sealed, I thought.

Finally glad to report the last upgrade that I had envisaged for this multi-year upgrade: KT150 power tubes, the highest wattage KTs available and the highest wattage tubes I can install on this ZMA.

These tubes are tall: 5 inches in height not including the pins.

Setting the bias to 95 mA seems to be the sweet spot in terms of musicality and transformer loading.

There is a lot of talk of Tungsol KT150s on various forums and none negative. I am glad to say that I concur.

The benefits of additional available headroom provided by the KT150s   over all the rest in the KT genre even the KT120s that I had before is immediately evident. All the positives mentioned all along in this thread earlier have improved.

I believe this ZMA is now is showing the true colors of the music that is being played on my 90dB speakers without rolling off the extreme frequencies at either end and playing it with unhindered aplomb wringing every nuances from the source. Details have surfaced, very subtle details but important to make the listening experience even more pleasurable.

Have a Merry Christmas folks.

Archie, I'll second that!

Due to a recent experience I'm wondering about heat input into the top plate from the power tubes.  I actually melted the plastic scales on my meters -- don't ask!  lol  I assume the KT 150s put out a lot more heat.  Are you seeing any adverse affects?  Or maybe they don't?

Sorry for the late reply.

No heat related issues. The meters, the top plate, the nearby capacitors and output transformers are as good as new.

Speaking of heat and the related current draw. I am still not using a fuse but a 8mm dia. 7N copper wire mentioned before. Having tasted the pill, simple can't go back. It's a risky proposition with the wire. One day when a tube decides to internally short I am in trouble. But I am enjoying the music so much that I am not even thinking of it.

The initial inrush current draw is between 14 to 15 amps which settles to 9 amps after all the capacitors charge to some extent in about 5 minutes. Remember my power circuit has about 30,000 uF of capacitance besides the large inductance from the larger input and output transformers. This may be wishful thinking but the fuse rating that I would need is so close (within 3A) to my 20A circuit breaker at the electrical panel that the CB should work about as fast considering the I would need a slow blow fuse.

Did you type that wrong or am I just misunderstanding?  If your ZMA is drawing 9 amps, that's about 1100W.  I think the stock ZMA draws 150W.  Is that possible?

I suspect the stock ZMA draws about 4 amps after settling down. When mine was stock before the mods, I needed at that time to install at least a 7 amp fuse in it. Anything below would blow during start up even though the recommended fuse was 6.3 A.

As is well known the ZMA being a Class A amp is very in-efficient w.r.t power consumption. Being a push pull amp helps a bit in this regard but is still very in-efficient. The benefit is the simplicity in design resulting in awesomeness in terms of sound production and fidelity to the original music.

Regarding the modified ZMA. Around 9 amps is correctly typed. Most likely the caps are not yet fully charged yet at the 5 minute point because they are also discharging a significant amount in powering the amp at the same time so still drawing more current than steady state current. Besides the massive amount of capacitors I also have large chokes in the power filtering circuit.

It takes about 30 minutes for this ZMA to start sounding its pristine best possibly the caps are fully charged by then and reach an equilibrium between the charge/discharge rate. I usually have to lower the volume after 30 – 40 minutes as the volume goes up with time. After I toggle the power switch off all the tubes still glow fully for about 10s solely powered by the caps before they flicker and turn off.