Are you saying that NOT playing music is required?  Or will it catch up eventually if music is played from the start?

Got it.  Thanks.

Something huge in the works.

My drive to achieve more and more perfection or OCD in modern parlance has compelled me to place an order for input and output transformers made by the same manufacture as the original ones  on the ZMA but of substantially higher output.

Rationale:

My calculations and discussion with the manufacturer of the transformers viz. Hammond Manufacturing have led me to believe if driving higher wattage, lower efficiency wide frequency speakers the original transformers are close to their "core saturation" point at mid/high volumes levels. Definitely using KT88 tubes at higher outputs will exceed the manufactures specs for the transformers (even though the ZMA owners manual has listed KT88 as an alternate) entering the region of saturation of the transformer core. BTW the stock ZMA will easily drive speakers that are 90 dB or more efficient and that do not go lower than say 30 Hz or so. In this case these and all earlier posted upgrades will yield proportionally less results as the speaker demands on the amp are far less.

Thing is these new transformers are physically larger in size,I have no idea how I am going to fit them on the amplifier base. As with all the upgrades done so far if things do not work I have the option of going back.

While I am at it I decided to include a filtering section so also ordered a couple of 5H chokes and 47 uF Mundorf Tube caps. These definitely will have to be external to the ZMA base unit.

Both looking forward to it as well as not for the amount of work involved. Might be far more than I have done so far for any single upgrade task. To gain access to shoulder points I expect to remove a lot of the work already done and re-fixing it all back.

No rest for the wicked ...

The new input transformers are about 30% larger and weigh a ton. Original ones are rated at 150 VA and 173 mA while the corresponding figures for the new ones that I am installing are 243 VA, 288 mA.

The new output transformers are rated for 60 W v/s 40 W.

So size is a head spinner for this upgrade. See photo. The new one is to the left and the old removed input transformer is to the right. This was a case of getting the transformers of the power rating I wanted and thinking of how to fit them on the fly. There is very little excess real estate on the ZMA base to accommodate the new input transformers. The output transformers are only slightly larger, it looks like they are going to fit. Physical size must have been a consideration when Steve chose the transformers that he did. I feel especially the input transformer could have be higher rated, but 372 FX is the largest 600V transformer in the Hammond's line up that will fit on the ZMA base. The model I am installing is the 372 JX.

Can't wait to finish this project, only about 20 % done so far. If the original transformers are underpowered for my 99 dB, 300 W speakers than this upgrade will make a substantial difference on the other hand if they are not underpowered than the difference in sound quality will not be as much if any.

Bonny, impressed in an understatement!  You are my DIY hero brother!  

I have been following your journey the whole time.  Far have you come since our original collaboration on some of our early ZMA modifications.  You have taken the ball and ran 1000 yards beyond the point I reached with my modifications.  I am envious of the transformer upgrade as that was an area I researched tremendously but the fell short as the transformers I wanted to install would not fit on the plate.  Plus, that was a soldering venture I had not the time  or patience to commit.  

Your post hit home as there were many times during my upgrade process where I was up to my ass and elbows in work on my dining room table convincing my wife that it was all worth it.  While she may not truly grasp the ROI on the improvements, the time, money, and effort were ABSOLUTELY worth it.  My ZMA is Divine.  But.........   there is no doubt in my mind that it can't hold a candle to yours.  

That said, congratulations my friend and a fine single malt is such an appropriate way to reward yourself for enduring.  Bravo!  I might need to invite myself to your house when I am in your neck of the woods to have a listen.  Based on what I know and what you have done, it would be well worth the effort.  

Cheers,
Jeff

Mostly we should all congratulate you for still being alive with all the microfarads.  Also congratulations you've just built yourself a pair of TORII Monos : )

I'm tipping a beer to you right now.

Steve

I feel honored Steve.

You brought up a good safety point. I should have mentioned about the dangers of working with such high power reserves, just to put it out there. Normally I do as when I mentioned of the dangers when experimenting 8 gauge wire in place of the fuse, because I breath and live safety as I work for a reputed electric power generation company that supplies power at 500,000 V to the Ontario and by extension to the North East American electric grid. Looking back I got caught up with the single minded purpose of improving SQ so missed mentioning the safety aspects.

After the upgrade work and testing is done and once assembled with all raw contacts inside or insulated if exposed outside as in my case it should be safe for operation.

The effects of so much capacitance though is just giddy and needs to be experienced.

I must also humbly confess I learned a lot from you – directly from your white papers and posts but for the detail technical aspects indirectly by passively studying your amps at great length. While I have experience in electrical equipment and parts such as capacitors, resistors, diodes, rectifiers, solenoids, motors, generators etc. in many other different applications, I did not until a few years ago have much experience in audio circuitry and have used your amps to self study your designs. There are times I believe I read your mind through your designs and figured out why you chose a particular path or component quality or size.

This is the only forum I am active on as I find it to be an environment that checks all the right boxes for an open discussion designed to elevate each others audio experience and is the reason I decided to spend bit of my time chronicling my experiences to give back to you and others in this audio community.

Hi Bonny,
No active crossovers.  The drivers are 90dB so in the transmission line cabinets they are probably closer to 92dB.  Anyway, they are nott what I would consider high efficiency drivers and there are 10 of them!

The ZMA is handling them like a champ!  I am a proud father  :)

But, my ZMA got drafted after my Canary Audio monblocks went in for a major overhaul/upgrade.  My ZMA's are assigned to my Hegeman Model 1A's, but called to duty these past couple months as I wait for the Canary "Reference Two" mondoblocks to come back.  They are 300B based 80W monoblocks that are being upgraded to the new flagship, "Grand Reference" model.  

I definitely think that bi-amping could be a good option.  I am hesitant to bi-amp with different amps and these Grand References are cost prohibitive to obtain a second pair.  They retail for $42K.  

A ZMA tri-amp setup would be something I could get excited about!  

Time for some pictures of the development process.

This one shows the mod done on the right side with the old set up on the left all powered up and playing music. I decided to test it at this midway stage to enable me to discern the difference between the new and the old setup by listening to the difference on the right v/s the left channel.

Conclusion: the right channel was significantly less labored with the mid-range and woofers benefiting tremendously. More audio information on the right channel compared to the left.

It might be hard to tell from the pictures but the right set of tubes were more brightly lit than the left ones.

Picture of the temperature of the left old input transformer.

Good question.

It is the filament voltage that will affect the brightness of the tubes. So bias can be anything: low or high the brightness should not be affected unless biased way I mean way out of the allowable range which no one is going to do. Even than I think the brightness change will not be such as to be discernible (I think, I have not tried it nor do I want to).

So since we are talking of filament voltage and current a good way to look at it would be a normal old fashioned filament bulb in a house hold circuit. A 100 W bulb will consume 100W if given its designed voltage which in north America is 120V is maintained. Given that the voltage is fixed the current consumed by the bulb will be defined by the resistance of the filament in the bulb - which is also fixed. Now suppose one was to load that same circuit in the house on which this bulb is by adding other loads say a turning on an electric heater and other stuff the bulb will continue to be lit with the same brightness only if the supply can keep up with this extra load. If the supply is weak the bulb will dim as there is a drop in circuit voltage.

Similarly in an amp if the input transformer power rating is lowish such that its core is overfluxed or close to being overfluxed it will not keep up with the load demands of the various loads within the amp that rely on the input transformer and the brightness of the filaments - which is one of the loads - will be lowered and the behavior of the other things in that circuit will be affected. If say the original transformer was of substantially higher VI rating by my installing an even higher rated transformer will not increase the tube brightness as the load which is the tube will not consumer any more power than it's design. That would be a waste of money and time.

So what I am trying to say in this long winded explanation is that one can only suppress a load's (e.g tube) output and that it will not consume more power than its rated design. So by increasing the input transformers rating I have ensured that the filaments are being supplied by adequate enough power to operate the circuit components upto and no higher than design rating i.e. there is no suppression of that components output due to a paucity of power supply.

So yes the tubes are pushed harder but up to their rated design level. I am sure you are aware that it is the filaments that supply the electrons (indirectly by heating the cathode) that flow to the plate and the tube grids than control the flow to the plate to modulate the sound to match that of the music. So a copious supply of electrons will ensure that ease of producing music. Lack of supply electrons will take away portions of the music.

I hope the message is not lost in this long explanation: in short yes the tubes are driven harder but to their design rating which I like. Why have a race car and use it as a street car (if allowed) just to ensure longitivity.

To continue the analogy I am not upping the voltage to above 120V to cause the 100W bulb to burn brighter. I am bringing it to 120V i.e. bringing the filament voltage to the rated 6.3V under load conditions where as it appears the old transformers were loaded so much as to suppress the voltage hence dimmer tubes. In my case they are brighter because it is brought up to spec.

Yes it does appear Steve likes to design his stuff with longitivity in mind. Hence for the pre-amps such as the CSP3 he has deliberately staved the filament voltage supply ensuring longer than design life for the tubes and the components in the amp. In the case of the ZMA though I suspect it is because of space constraints. The physical size of the higher rated transformer caused the move to the current one. Only Steve will know the exact reason. In any case it fits his preference towards longitivity.

For him it is a balancing act getting the best bang for the buck in terms of longitivy only versus performance only criteria. As for me I am not constrained by the requirements imposed by others e.g. long life as I am doing it for myself and my personal bias is whole heartedly towards performance.

Mind you I do not over load anything beyond design but to design. So the life of my upgrades will still yield 20 years of operation. If I could get further performance improvement and it meant a life span of the equipment of 10 years I would still be happy. But going beyond design rating usually bring other undesirable artifacts into play so I like to stick to design knowing fully that the design specifications for industrial equipment (may not apply to audio equipment as it seems to be a free for all approach i.e. no national or international standards) is still conservative.

It does not appeal to me that I am using my equipment sub-optimally for 30 years when I can use it optimally for 20 years.

My statement in my earlier post needs elaboration as it is not so cut and dry. I said "It does not appeal to me that I am using my equipment sub-optimally for 30 years when I can use it optimally for 20 years"

I would not fret too much either way as I think Steve has found the right balance for him to give the lifetime warranty while maintaining very good performance.

Unfortunately in life one can't have everything, there are compromises: in this case a conservative design approach for long life warranty or a absolute performance driven approach not caring if it involves more maintenance along the way.

My high tech test bench.

At what point will you just bite the bullet and get a bigger plate?

Bonny, your pictures bring back memories...  My journey was pretty fun and very educational too, but I would be lying if I said there were no moments where expletives were blurted and tools thrown.  It was intense at times but once done, all that much more satisfying.  You made me laugh with your comments about the family and having your face implanted in the underside of the amplifier.  That, I can relate to.  

This ZMA is a great amp stock.  With the upgrades we have done (especially yours) it is a greater amp.  And despite having my crazy Canary monoblocks, I still love the sound of this amp!  While a very different experience with the ZMA in comparison, there are things I like about each and enjoy listening to them both with the different speakers I have in my stable.  

Do you have a pic you can send of the finished amp?

This picture is a good example for why my folks were tired of it all, cannot blame them. The whole kitchen table and surrounding area was taken over.

Note: I would evacuate everything after every 2 to 3 hour work phase for the day. Not an easy task with those now unbalanced heavy transformers.

I think I can see a bulge in the case!  

Jeff, did you mean the above finished pictures or a close up of the internals after finishing it all?

Also did you also mean send to you by PM or post here?

Very cool modifications to the ZMA Bonny. Way to take it to the max, then beyond.

What's next? Maybe a new exotic wood base?

HK

Bonny,

My ZMA with limited mods done by Steve sound great. Thanks to your suggestions, Parts Connexion's WBT sale, and the 25th Anniversary mods the amp has been taken to a another level.

The only weak link I had was an aging stylus on my Dynavector 20x2 low output cartridge. Now that it's been retipped it's game on. My Ray Brown Soular Energy album pressing (Pure Audiophile label) absolutely blew my mind this morning. Everything, including CD's and TIDAL streaming just sounds amazing. Enjoy.

HK

Folks, thoroughly enjoying the upgraded ZMA these past days, nicely burnt in by now.

The KT88s I have are back in business, putting on quite a show given the added power provided by the mods in the ZMA. Prior to the transformer mods the 7851As were the tubes that I settled on, now they seem puny.

Anyways here is a recap of the things I learnt from doing this project that I thought will benefit all. I'll post them one at time to reduce both the amount I'll have to type and you read in one go:

Learning 1:

Go the whole hog i.e. think about the whole audio chain – from source to speakers when accessing for weakness in your audio system. Minor improvements here or there may not result in the improvement you expect for those minor mods. Bottle necks in the rest of your system may choke this latest minor upgrade you did not allowing it to bloom.

The weakest of the bottle necks in your entire audio system will define the SQ not your most recent upgrade even if you have spent a lot on it. As I said previously the weakest chain link in a ships anchor chain will define the tensile force at which the chain breaks not the strongest link (say your latest upgrade) in that same chain. The strongest chain link if tested on its own may yield or even break at a much higher tensile load but put together in the chain along with that weak link, it is the weakest link that calls the shots causing the chain to break at a lower load and the ship to drift and get wrecked. Similarly for audio. This fact may lead you to be disappointed with the latest bit of upgrade you have done even thinking it did not work. In short if the upgrade is rooted in science and not in voodoo (e.g. silver v/s copper as a electrical conductor) it will work, it has to work, only that other things in your audio chain may be blocking it from expressing itself.

It's a judgement call as to what your weakest link(s) is/are and then eliminating those with good science not just conjecture and emotion. Hopefully some of the info in this thread helps in making your decisions.

The ease with which the ZMA is now driving KT88s got me to think of trying KT120 tubes. Upon checking the specifications of the KT120s I found that the input transformers I just installed (which are if I am not mistaken even more powerful the Torii Mono’s transformers) will be able to handle even KT150s however the new output transformers I put are the same as the Torrii Mono’s (except the Mono’s have the potted version) will saturate with KT150 tubes so the highest tube I can safely install in the modded ZMA are KT120s.

Just placed an order for a quad of Machlett KT120. Cannot wait to give them a swirl. On a side note : on doing some research it seems that Machlett tubes are better than Tung-Sol and more over as a company Machlett have better pedigree than Tung-Sol. Below is a direct quote from Machlett website regarding comparing their KT120 tubes with those of Tung-Sol.

“These (the KT120s) are our own design and not to be confused with the lower end Tung-Sol tubes which share too many similarities to a KT88”

I have now been using KT120 output tubes for 5 days.

After I installed the KT120s and no sooner I started playing music it bought an unconscious smile on my face even before I set about consciously trying to figure out the impact as one would when trying out anything new.

Because these tubes have a higher output, it is obvious that the bass and to a lesser extent the midrange drivers have benefited further in spite of all previous improvements. And all along I thought I had no more margin for improvement.

Particularly beneficial in my situation I think. As mentioned previously on this forum thread I have PSB T3 speakers, they each have 5 drivers, 3 of which are 7.5 inch in size. These 3 drivers separately operate from the lower midrange to 24Hz handing over from one to the next as the frequency lowers from 750Hz to 24Hz.  The T3s have quite a flat +/- 3 db window all the way from 24Hz to 24kHz and are 89 db efficient anechoic or 91 db real world. So installing the KT120s have improved the overall balance between high, mid and low range frequencies for me. Even though I thought it did I can see now with the installation of the KT120s that previously the KT66s and KT88s did not have the whomp to easily drive these 3 drivers with the felicity as is now.

When I was younger and inexperienced I used to vacillate between a preference for accentuating treble and at other times bass or both as the sizzle of the treble and the impact of bass were both enticing depending on which phase I went through. With maturity I have come to realise that what elevates music to the epitome is a very flat response across the widest possible range frequency wise: ideally from 1Hz to infinity but in reality to the extent possible from 18 Hz to say 100,000 Hz. Sounds outlandish but let me explain my thinking. The flatter and wider the range on both sides of 20 Hz to 20 KHz the more exquisite the music will be even if ones conscious hearing ability is nowhere near that range’s extremes. I think we hear or rather experience music with the ears and also other senses particularly the skin, no kidding. Here is why. I was in contact with the design engineer of TakeT the highend supertweeter manufacturer from Japan, he conducted blind A/B tests placing very mini supertweeters (very light and about 1 inch in size – I have 4 of them myself) on subjects cheeks while listening with headphones and he says that the candidates experienced elevated music experience with the real supertweeters v/s dummy ones that looked exactly the same. In our communications he came across as a true no nonsense audio research scientist. He says that TakeT Batmaster supertweeters (which b.t.w. I have) goes beyond 150,000Hz he does not know how much higher as the Japanese national laboratory (or it could have been a University lab - I forget which) instrumentations where he tested them had a limitation of 150,000Hz and his readings were offscale those instruments.

On this subject of frequency range I also think we will never ever be able to exactly produce natural real sound (but only try to approach it) because we are still talking of reproducing within +/- 3 dB of 20 Hz to 20 KHz and we are extremely pleased if we achieve this. But natural sound is 0.000000001 Hz to infinity within a tolerance of +/- 0 Hz impossible to achieve. That is why no matter how good ones system is if the exact same song is played by a live band at exactly where the plain of speakers are one will easily tell a live band even from a really expensive system. I realize I am kidding myself when I say my system sounds live. The more appropriate language would be "live like".

I have come to believe that the very low frequencies and the very high frequencies though not audible, influence the air in the room just like live music would which in turn influences the audible frequencies. The air pressure peaks and rarifications of all frequencies (humanly audible and beyond audible) act cumulatively to produce a different resultant of air patterns that reaches the ear.  What we hear is the summation of the whole range of air pressures and rarifications just like in live music. Any instrument be it a drum, a guitar etc. produces not only the frequencies we hear but also the extreme frequencies we do not hear. The question to ask is why would sound emanating from any device emit only the frequencies we hear. With age we hear less but that drum or guitar still produces the frequencies beyond our now limited hearing that it produced in our earlier years. Because of this influence super tweeters work to enhance the music similarly. Similarly as your speakers go below 20 Hz.

Based on these concepts when choosing speakers to buy, I look for the most efficient speakers that can deliver the most frequency range – off course budget limitations has to be factored in. In my opinion if speakers roll off at below 50 Hz or even 40Hz a lot of music information from the music recording is lost straight away even though the midrange will be beyond superlative. The missing bass in this case feels unnatural and likewise the super high frequencies.

Learning 4:

Reduce noise in the power supply as much as possible - to the nth degree. I myself have gone exceedingly overboard w.r.t cleaning the power supply. I use a Furman IT Reference 20i power conditioner combined with a Blue Circle BC 6000 conditioner both rated at 20A and a Akiko Corelli conditioner. While each of these would on its own be enough they each take care of different aspects of power line noise so work well together all noise from the lowest to the highest of frequencies. Additionally I have numerous mini noise suppressors by iFi power at each devices IEC connector, a iFi USB galvanic isolator and iFi Gemini USB 3.0 split cable that carries the signal and USB power separately at the Roon server, a Furutech Flow 28 at the DAC etc. and even some Bybee Quantum purifiers (of all the things I have done I think the Bybee’s are more on the hocus pocus side – I am not sure they work as claimed but at any rate they do not degenerate the sound). Also I use a dedicated 10 gauge power line directly from the breaker panel (not running closely parallel to power lines of other circuits) for audio with a Furutech NEC wall receptacle. In all I estimate I have spent about $12 to $13 grand only on power conditioning – I said nth degree after all.

The goal was to simulate 2 a.m. night like electrical grid conditions during the day. I selected the Furman/Blue Circle/Corelli combo instead of the PS Audio P10 as I thought it was not sufficient capacity wise (see learning 2) and it’s sibling the massive P20 was not available at that time. Besides, I am not a fan of regenerators per say. They produce a “pure” sine wave electronically by first chopping the AC and then reassembling to form the sine wave in minute steps which will never be the same as a true analogue sine wave produced by mechanical generators in a power plant just like a very expensive DAC will only approach the analogue nature of a turntable but never get there. Though the P10/P20 have multiple other benefits like voltage regulation the jagged re-assembly of the wave over-rides it for me. That is why I think PS Audio P10 specs talks about noise reduction by greater than 80 dB only above 100 kHz.

Additionally where possible eliminating power supply noise internally within equipment yields even better results as the noise is removed closest to the actual delivery point (input and output vacuum tubes) as was my experience with the modifications to include power supply filtration in the ZMA.

I found that having as clean a power supply as possible enables hearing of extreme subtleties in music and the overall tone improves such that the sound is more authentic meanings string instruments sound more like actual strings not simulated (electronic) strings: stings vibrations can not only be heard but also felt and imagined approaching that of real life. Like vise for percussion, wind instruments and voices. Music sounds natural, real and live. Noise reduction makes a profound difference to SQ. Reading Steve eloquently elaborated 2 a.m. music listening experiences which got me to think “why not try to get that experience during the day?”.

Also as I mentioned somewhere earlier but is worth repeating here: noise distorts music as the peaks and troughs of the noise frequencies when present concurrently with music frequencies produces a resultant outcome of accentuating or diminishing actual pure music just like noise cancelling headphones deliberately superimpose an opposing frequency to suppress noise thus cancelling each other out. For noise in the power line case the noise because it is random in amplitude and timing i.e not exactly superimposed w.r.t. music may result in partially opposing or partially accentuating music frequencies so it either partially cancels or accentuates certain music frequencies causing it to deviate from actual.