a matter of tiny differences

sometimes it happens that a decision has to be made about very small improvements. and it may not be clear, in terms of some absolute value, what to do next. the significance of this can easily be blown out of proportion, but it is a situation that still has to be dealt with. in post modern culture, a great deal of emphasis is placed on the idea of incremental improvement in the commercial sense… that, and a rapid planned for obsolescence. we tend to ignore the second part of that dance. instead, we quickly jerk our chins expectantly towards the next snack. it better be there. if it doesn’t show up, we reach for the remote. the process surrounds us and can’t be avoided anywhere. 3 year old children know this… perhaps the species is evolving? but we, as crafts people, are dealing with something completely different. for us, there is also the matter of the idea becoming real.

i understand that it can be confusing. the marketing of “high end” audio and musical electronics is based entirely on this premise… that small differences are what make product X higher end, or “better sounding” than product Y. this can then be graded and filed away for eventual disposal. as you probably know by now, in my not at all humble opinion, audiophiles are least qualified of all to chime in on small differences, and are of course (conveniently enough) clinically obsessed with them too. that is a shitstorm to be avoided. we need a reference that sits a small distance away from all this.

supposedly, we have one. an “objective” standard based on another idea: a “linear” transfer characteristic. ok. but, it turns out that even this idea is subject to the process of incremental improvement and disposal that reality undergoes… not strangely, it has changed over time. it’s not as “objective” as one perhaps would want (i don’t have a problem here). more troubling to me is a sense that the moral aesthetic behind the definition “linear” seems a little resentful. it is resentment against change, resentment against difference, and resentment against living things. “non-linear” things. everything we hackers do, often falls under the criticism that we are endorsing the non-linear. must be the distortion. the rolled off fakey tube thing. i think they are pointing at something else. for them, the messy world of the homo sapien shouldn’t be allowed to interfere and we won’t leave well enough alone. they have a point! recordings are NOT LIVE. they should remain dead, and unchanging. and well behaved because “now they are mine”. you can’t let those michael boulton CDs get too free, man… something’s gonna blow!

if you think about the cable or treatments side of the business, ALL of the power is invested in tiny differences. “it does something”! was it good? but audiophiles have some control over the choice and direction with that one, at a cost. so genius, actually…. because none of them can make or design a gadget, the gear can seem like a black box. a flight recorder. but they can change speaker cables or glue “dots” all over everything and suddenly they are powerful. they made something happen! who needs to know anything when there’s monster cable! i think of clark johnson running around with a bag of weird smelling juju goop waving a green marker… i wish i could smoke pot sometimes. but it would quickly go bad. i do like that guy, though!

in some ways, the entire high end business, and this little corner of it that shouldn’t make a difference to anyone, is an anathema. and DIYers, we are all a bunch of dr. frankensteins at heart. merrily sewing together corpses and playing god (elevate me!). there are prestigious institutions housed in tall buildings (to help represent the distance needed) who are supposed to be there for us. but, the universities, and the industrial “standard makers”, the ieee and the aes are actually working their hardest (along with the wrinkled freaks in belgium) to do away with this business altogether. that isn’t just a matter of the economy. we still smell like 63/37 (tin/lead) and shellac. we don’t buy and throw away, or at least not new stuff, in fact the opposite. for us, some of their garbage is our gold! we almost can’t be CE approved anymore. at least we can call it art for the time being… we’re supposed to be obsolete, so why don’t just stay that way? and i think they don’t approve at all of the playing god thing… that is their job, after all.

why? is it not enough with miley cyrus? do they miss madonna? tupac? what’s at stake? what would happen if music were reproduced in a truly subjective emotion laden torrent of EMDA-like engineering? would everyone get nekked and just fuck each other? would a wormhole open and beezlebub himself climb out and set us all on unholy fire?  fear of a black planet? or, god forbid, “HAIR” would come back, although somewhat dehydrated. i WANT to be there for all that! where do i sign up? can i help make it happen?

i’m getting sidetracked. a LOT of lip gets wagged in the name of tiny differences that are also somehow life threatening, career changing, and liable to create a rift in the fabric of the universe…  between the audio inclined. what they actually need is a script for zoloft. or, a real hobby that gets them out of the house. maybe a girlfriend or a boyfriend? let’s be clear: we are talking about SMALL differences. even though those differences can be like the red thread of life… pull it out, and it isn’t anymore.

as an engineer or a hacker, the intent is justifiably aimed at the big stuff. the direction, obstacles, the raw materials and the fit and finish. still, the devil is in the details, and he has work for idle hands… the last 100 yards of any design (i hate the metric system. fuck you very much) is often involved in decisions about tiny differences. and yes, that can be what separates the wheat from the “chaff”. but this is about a very personal human thing. it’s about taste. about love, or hate. it’s more like cooking than anything else. on the other hand, you should also be aware that no matter what you do, your work is thrown to the retarded lions once it leaves your hands. it never comes back the same. and it doesn’t even belong to you anyway. chances are that it will be ignored, sold immediately or misused entirely… and it still may end up beloved! or thrown out or covered in “dots”.. but, it won’t be up to you. you are incidental. ahh… is it beginning to sink in?

what does a great chef need most of all? good eaters. without them a great chef is nothing but ALONE, with the cooking wine and all that butter fat… it’s about forming a true partnership with another human being. we live in a bad time for that. but how do you know? the more you push your respective envelope, the more outland and frontier you go… when there isn’t much precedent, it helps a great deal if you can lean on someone else who appreciates the effort, knows the vocabulary and has a sense of right and wrong. even if you end up not listening to them, it is a sign of maturity to be willing to give up your own will in the service of a joint enterprise. not a mathematical model. in this case, that’s what i do. i can recommend it.

in roughing out a new idea, even with the inevitable mistakes one makes, there is always a way to muddle through at this level. you make discoveries and learn more. even into areas that are not particularly familiar: a body of work, a history of pattern recognition and the internet can all contribute to good adventures… and the more adventures you have, the more interesting it gets. for me, it’s worked out so many times now, it’s tempting to let go of the memories when it unpredictably crapped out. i don’t though. for the final spit and shine, i have some close friends now who i trust more than myself. it’s the most valuable resource i have, as a designer. my tribe. you should get one of your own, if you can. think of it as a guild of your own.

i think can name them all, the close ones… joe roberts, mj chung, herb reichert, john pisani, kato hideki, dr. bae, takeshi kawana, gary poulsen (although he isn’t speaking to me at the moment), stewart ferrell and noriyasu komuro. almost forgot blackie pagano! he moved to LA and the fog of war did the rest…. these are my peers. they all have a piece of the puzzle. i don’t and can’t talk to them all or even regularly. at any given moment, one of them is there. and i am really grateful. there is also the kindness of strangers, which cannot be ruled out. if you are open minded, you never know who suddenly appears to lift you out of an uncertain rut. i am probably fundamentally incapable of being my best as a hacker and an artist without these people. they all give and have given me a perspective into the importance of small differences i wouldn’t and cannot have on my own. there are probably a few out there who can do it alone. i don’t know any, and haven’t met any, but statistically i know there should be…?

so how do you even discern these things without throwing out the dog with the bathwater…?

when i was working for dick sequerra, the first five months or so barely went by without him calling me “stupid” at least once a day. one day finally came, when i said to him, after a great deal of conversation with my shrink and some other friends, “you fucking grumpy old man, call me stupid one more time and you can make this shit on your own”. this kind of measured response is not characteristic of me and it was an effort, i assure you. dick laughed. and he apologized. and then he told me a story… about him and sid smith. how they fought (verbally) and swore at each other often from the beginning of the day until the end. each bout of abuse more savage and character assassinating then the previous. at the end of the day, they often hung out and enjoyed life and there was very little of the “friction” that played out at the job. it was their process. the messy working out of the issues and the solutions. dick never said anything about what sol marantz or his coworkers thought about this… i know joe grado and tommy cadawas worked there with them, i never heard their side of the story. he described this as a valuable process and that sid was his other half and his best friend. but, dick then said it was a sign of respect that he included me in his raging… and that i should get back to work and fuck off too. he told me to fuck off regularly from then on… hahahhahaahahah! i think i was with him for three years?!

ok, now to the practical stuff. i want to present you with 2 circuits. i have built both of them. the 6688 was way before the cascode 417A. the cascode Gm amp was very interesting to me. it is harder to find the right tubes for this circuit. not all cascodes make good Gm amps. not all pentodes either! in any case, i was able to set up a cascode with almost the same basic operating conditions as the pentode. and lo! the cascode had 1/10 the distortion and 6dB less noise. wow! i thought i was really on to something. that’s what the books say should happen, too. but upon listening, the situation reversed it self. the cascode sounded brighter and more “high end”. still fabulous, but the 6688 was more right. i started to measure as well as i could. have a look.

so first you see your basic Gm amp. it’s a 300 volt supply because that’s what i have for my test supply. it could be lower for preamp work. 200 volts on the plate. frank blöhbaum’s suggested device for the screen buffer, the MPSA44 is great. this has a gain of 63dB and a measured distortion of 1.2% for 10mV in (HP334). the lissajoux image image in a nulled X/Y test shows this to be almost entirely 2nd harmonic in nature (i will take pix of this and put it here…). i don’t have a spectrum analyser or fancy distortion meter yet, although i am working on it. noise is very low… shorted out input puts it at 1-2 mV for 63 dB of gain (HP410EL). this is not precise enough for audio magazine publication and they might laugh at my gear… but i know from experience that this is in the ballpark. this is baddass.

now that i have LTspice and stephie bench has so graciously provided reasonable models to work with, i can also look at what the models say. here you go:

fab. here you can see that stephie’s model is pretty damn good. the fft shows a 1% bump of 2nd harmonic, and everything else is really excellent… 3rd and the crap is -70dB and falling from there. in the phono stage it goes into, it sounds that way, which is even more important.

next you see the 417A cascode Gm amp. this is a marvel. all the operating points match the 6688. gain is 63 dB, and the measured distortion is at the bottom of my equipment. that doesn’t say that much, but it is at least that good. that is just under 0.1% and fluctuating another .03% for 10mV in. shorted noise is 1mV.

i like this circuit! i used a DN2540 for the current source, but you can use anything. the voltage reference was a divider and not the ideal you see here on the schemat but it is just for explanation. the same X/Y test with the HP334 used to null out a 1KHz input signal showed a completely undifferentiated wiggly lissajoux, no particular harmonic stood out. like i said, i don’t know what is the opamps in the HP334 or what is my cascode? i need better gear. next take a look at what Ltspice has to say…

fascinating. the harmonic distortion is clearly less than the pentode, but a careful look at the upper range shows more crap than with the 6688. things don’t fall off until above 20KHz. admittedly this is small. really small. but even i with my crappy ears and brain damaged head i hear the pentode doing something better. now i am using both of these open loop. no feedback. it’s for phono and the 10mV measurement is meant to study the clips and pops. we will have to wait and see what my buds think… but i find this stuff fascinating. i hope you do too?

Leave a Reply