Well , it started long ago …. when I was really young(st)er, full with naivete and BS , came to pretty novel idea to make xformer loaded tube line stage …
Result of first tests , hardly optimal ( using two halves of 6SN7 in parallel and loaded with ancient xformer from vertical section amplifier of ancient B-W tube TV set) , I was blown with ‘till then unknown dimension of sound – everything was so alive ….
Later I made several iterations of line stage with same approach , being so proud for reinventing boiled water , until I got some xerox copies of Sound Practices , where I found that even in boiled water reinventing I wasn’t first ….. boyz from Italia already made Euridice preamp ; anyway , intermediate iteration was with two halves of 6DJ8 in parallel , latest one being with ubiquitous EC8010 and custom made xformer . That being even published on my old site – WOT preamp and battery charger
Result of first tests , hardly optimal ( using two halves of 6SN7 in parallel and loaded with ancient xformer from vertical section amplifier of ancient B-W tube TV set) , I was blown with ‘till then unknown dimension of sound – everything was so alive ….
Later I made several iterations of line stage with same approach , being so proud for reinventing boiled water , until I got some xerox copies of Sound Practices , where I found that even in boiled water reinventing I wasn’t first ….. boyz from Italia already made Euridice preamp ; anyway , intermediate iteration was with two halves of 6DJ8 in parallel , latest one being with ubiquitous EC8010 and custom made xformer . That being even published on my old site – WOT preamp and battery charger
Attachments
Anyway , what was main quality of (then) newly explored territory for Mighty ZM …. I couldn’t easy describe what that was , and only after few years I came with explanation how I understood it – what I was hearing ( and I hear that even after all these years) is result of important phenomena – Signal Energy Preservation.
Simple – with resistive loading – efficiency of stage is always low , while with inductive loading it is way higher …… energy of signal isn’t dissipated along the chain (as with resistor) , it is being transformed in ratio of intrinsic values (voltage , current , thus impedance).
Remember that I’m not formally educated in electronics , so I’m pretty much free to invent new postulates whenever I want .
During the years , having direct contact with plenty of pro gear (in radio production), I started noticing signal iron in plenty of positions , my Guru discover to me how versatile can be all sorts of so called repeater coils , scavenged from old telephone equipment …
To cut (some , not all ) story short – I started collecting all sorts of signal Iron , thinking how one day I should/could/would use them in something ( that behavior is , I believe , familiar to every Greedy Boy around). Besides my (very successful) experiments with xformers as main DAC I/V conversion elements , especially when used fully differential ,and numerous more or less successful attempts to include Iron everywhere in audio chain , I was (not worthy) given by pair of (so called basic ) Intact Audio 200$ AVCs; my own personal German ( known around as Golden Ear-ed Generg ) sent them as Geschenk .
They’re ( realized as basic passive box) pretty good sounding , combined with papa’s Koan M2 : AVC & Input Selector for Papa’s Koan M2 ; temporary case ? | Zen Mod Blog
So , What is Gain structure > What is Gain Structure? , I’m pretty happy with recippe – efficient and sensitive speakers ( not necessary synonims!) , good amp with power in regular FW bracket , few or none db of gain in box intended ffor selecting and adjusting signal source .
However , during the years , playing with all sorts of signal attenuating devices ……resistive (P&G, Noble, TKD, ALPS,all sorts of stepped attenuators,blahblah) , LDR based Dropbox - PSM LS Cook Book.pdf
inductive (Intact Audio , Sowter) , I came to conclusion that – besides principle of Signal Energy Preservation , impedances are also very critical , so buffer is conditio sine qua non ….. one step back , but that one step is what is necessary to achieve further two steps in front.
It’s not easy to make both cheap and good soundiung buffer ; complementary Toshiba JFet buffer (see PSM LS CookBook) is becoming more and more PITA to make (money not being crucial , but hassle in obtaining them ) , while all other ones are either crude or bulky in comparison .
To latest idea about no gain/small gain preamp I came by chance , already written here : What's wrong with the kiss,boy ? and Iron Pre | Zen Mod Blog , so I’m not going to repeat that.
Simple – with resistive loading – efficiency of stage is always low , while with inductive loading it is way higher …… energy of signal isn’t dissipated along the chain (as with resistor) , it is being transformed in ratio of intrinsic values (voltage , current , thus impedance).
Remember that I’m not formally educated in electronics , so I’m pretty much free to invent new postulates whenever I want
.During the years , having direct contact with plenty of pro gear (in radio production), I started noticing signal iron in plenty of positions , my Guru discover to me how versatile can be all sorts of so called repeater coils , scavenged from old telephone equipment …
To cut (some , not all ) story short – I started collecting all sorts of signal Iron , thinking how one day I should/could/would use them in something ( that behavior is , I believe , familiar to every Greedy Boy around). Besides my (very successful) experiments with xformers as main DAC I/V conversion elements , especially when used fully differential ,and numerous more or less successful attempts to include Iron everywhere in audio chain , I was (not worthy) given by pair of (so called basic ) Intact Audio 200$ AVCs; my own personal German ( known around as Golden Ear-ed Generg
) sent them as Geschenk .They’re ( realized as basic passive box) pretty good sounding , combined with papa’s Koan M2 : AVC & Input Selector for Papa’s Koan M2 ; temporary case ? | Zen Mod Blog
So , What is Gain structure > What is Gain Structure? , I’m pretty happy with recippe – efficient and sensitive speakers ( not necessary synonims!) , good amp with power in regular FW bracket , few or none db of gain in box intended ffor selecting and adjusting signal source .
However , during the years , playing with all sorts of signal attenuating devices ……resistive (P&G, Noble, TKD, ALPS,all sorts of stepped attenuators,blahblah) , LDR based Dropbox - PSM LS Cook Book.pdf
inductive (Intact Audio , Sowter) , I came to conclusion that – besides principle of Signal Energy Preservation , impedances are also very critical , so buffer is conditio sine qua non ….. one step back , but that one step is what is necessary to achieve further two steps in front.
It’s not easy to make both cheap and good soundiung buffer ; complementary Toshiba JFet buffer (see PSM LS CookBook) is becoming more and more PITA to make (money not being crucial , but hassle in obtaining them ) , while all other ones are either crude or bulky in comparison .
To latest idea about no gain/small gain preamp I came by chance , already written here : What's wrong with the kiss,boy ? and Iron Pre | Zen Mod Blog , so I’m not going to repeat that.
Attachments
Last edited:
After successful realization of concept , I knew that few shortcomings of Iron Pre aren’t in necessity for Preputium made of parts , or Iron choice …… it is in fact of still having resistive attenuator …. and always tempting fact that somehow is possible to dispense scarce Toshibas for something you really can buy in nest Tobacco Store.
So ,as a result of my day(night)dreaming, it came to prototyping of Iron Pumpkin , both SE and Bal iteration . Last one simply without suffix, while former obviously having SE as part of name.
Due to intrinsic nature of balanced xformer ( fed ballanced , outputting signal also ballanced ) with necessary involvement of GND on either or both sides , it’s also having virtues of SUSY , as Papa patented it years ago. (Old Bull/Devil , always tongue in cheeks ).
So , conclusion – to my ears – inductive attenuator is best sounding of all alternatives , and I really succeeded in dream of achieving (settable) gain with same device and , last but not least , natural conversion of SE to Bal , or other way around , in case of Iron Pumpkin (Bal) .
So ,as a result of my day(night)dreaming, it came to prototyping of Iron Pumpkin , both SE and Bal iteration . Last one simply without suffix, while former obviously having SE as part of name.
Due to intrinsic nature of balanced xformer ( fed ballanced , outputting signal also ballanced ) with necessary involvement of GND on either or both sides , it’s also having virtues of SUSY , as Papa patented it years ago. (Old Bull/Devil , always tongue in cheeks ).
So , conclusion – to my ears – inductive attenuator is best sounding of all alternatives , and I really succeeded in dream of achieving (settable) gain with same device and , last but not least , natural conversion of SE to Bal , or other way around , in case of Iron Pumpkin (Bal) .
Attachments
Time for few block schematics ; first Iron Pumpkin SE ( chinch, RCA , whatever)
What’s shown are basic blocks – input selector , single buffer ( single phase , yup) fed with Good Gemini shunt reg , then AVC , with selectable input tap ; when fed to top – possible gain is odb , with chooice of tapping in to -2 to -8 db , respectively , so possible gain from 2 to 8db.
Selector – simple 5-position rotary switch operating with 5 Takamisawa relays ; they’re good for industrial and medical instruments , so ….
Good Gemini – simplified but also hot-rodded iteration of well known Shunty.
Buffer – current mirror biased Diamond Buffer …… nothing new , but painstakingly to achieve same THD spectra as with Toshibas , luckily with even beter behavior under heavier loads.
AVC – called Iron Turtle SE ,based on Hitachi Metglas (AMCC008) cores. Coil – 10H/100Rdc jobbie , as (always!!) compromise between needed Good Inductivity and several Bad Factors (known as Bad Capacitances , Bad Inductivity etc. …. in fact , using AVC instead of TVC is Bad Inductivity non existent
); insulation between layers is Teflon based , bobbins are/will be custom ones , everything PVC coated after assembly.
Disclaimer – as for difference to almost all AVCs on market , these aren't intended to be Swiss Knife universal jobbies , they are constructed to be driven with buffer or , at least , pretty potent signal source apparatus.
Two AVC Irons on one Iron Turtle pcb , so that’s enough for stereo.
Switching again through Takamisawa relays ; flat cable going from Turtle pcb to front plate mounted 24 position Chinese Switch .
Chinese Switch – even if made by carefully chosen group of Chinese Virgins , only during full moon nights , each one is gladly disassembled then greased with graphite grease then assembled again , by Mighty ZM himself.
What’s shown are basic blocks – input selector , single buffer ( single phase , yup) fed with Good Gemini shunt reg , then AVC , with selectable input tap ; when fed to top – possible gain is odb , with chooice of tapping in to -2 to -8 db , respectively , so possible gain from 2 to 8db.
Selector – simple 5-position rotary switch operating with 5 Takamisawa relays ; they’re good for industrial and medical instruments , so ….
Good Gemini – simplified but also hot-rodded iteration of well known Shunty.
Buffer – current mirror biased Diamond Buffer …… nothing new , but painstakingly to achieve same THD spectra as with Toshibas , luckily with even beter behavior under heavier loads.
AVC – called Iron Turtle SE ,based on Hitachi Metglas (AMCC008) cores. Coil – 10H/100Rdc jobbie , as (always!!) compromise between needed Good Inductivity and several Bad Factors (known as Bad Capacitances , Bad Inductivity etc. …. in fact , using AVC instead of TVC is Bad Inductivity non existent
); insulation between layers is Teflon based , bobbins are/will be custom ones , everything PVC coated after assembly.Disclaimer – as for difference to almost all AVCs on market , these aren't intended to be Swiss Knife universal jobbies , they are constructed to be driven with buffer or , at least , pretty potent signal source apparatus.
Two AVC Irons on one Iron Turtle pcb , so that’s enough for stereo.
Switching again through Takamisawa relays ; flat cable going from Turtle pcb to front plate mounted 24 position Chinese Switch .
Chinese Switch – even if made by carefully chosen group of Chinese Virgins , only during full moon nights , each one is gladly disassembled then greased with graphite grease then assembled again , by Mighty ZM himself.
Attachments
Now , on to Iron Pumpkin (Bal) ; block schematic similar , except few facts :
-two buffers per channel
-AVC/autoformer connected in differential manner ; again – there is intrinsic SUSY -error correction , whatever you choose to call it . Facts about construction written for SE, applies here also .
-One AVC Iron on one pcb , so two pcbs for stereo ; Newest iteration of I. Turtle pcbs optimized for stacking , so footprint stays the same as for SE.
-There is option for using any of 5 inputs as SE ; put RCA instead of XLR , mount 1N4148 diode in adequate position for energizing one additional relay and that’s it . Trick is to energize additional relay , to break connection between negative side buffer ( now being without signal ) and negative side of AVC; that way buffer’s output is not acting as virtual zero to negative VAC winding , thus not choking it. AVC is now fed solely through positive side winding , producing ( naturally) output on both sides/phases . So – you can use SE inputs from none to 4 …….. in case that you want all 5 of them being SE inputs , having Bal output …….. then either use SE amp , so choose Iron Pumpkin SE instead , or use Iron Pumpkin with your Bal input amp , but skimp on neg. buffer elements .
-two buffers per channel
-AVC/autoformer connected in differential manner ; again – there is intrinsic SUSY -error correction , whatever you choose to call it . Facts about construction written for SE, applies here also .
-One AVC Iron on one pcb , so two pcbs for stereo ; Newest iteration of I. Turtle pcbs optimized for stacking , so footprint stays the same as for SE.
-There is option for using any of 5 inputs as SE ; put RCA instead of XLR , mount 1N4148 diode in adequate position for energizing one additional relay and that’s it . Trick is to energize additional relay , to break connection between negative side buffer ( now being without signal ) and negative side of AVC; that way buffer’s output is not acting as virtual zero to negative VAC winding , thus not choking it. AVC is now fed solely through positive side winding , producing ( naturally) output on both sides/phases . So – you can use SE inputs from none to 4 …….. in case that you want all 5 of them being SE inputs , having Bal output …….. then either use SE amp , so choose Iron Pumpkin SE instead , or use Iron Pumpkin with your Bal input amp , but skimp on neg. buffer elements
.Attachments
Last edited:
Attachments
Last edited:
Iron Turtle details .......
if you see two Irons (pcb sandwiched in between) that's stereo pack Iron Turtle SE ;
if you see one Iron on pcb , that's mono (Balanced) Iron Turtle
if you see two Irons (pcb sandwiched in between) that's stereo pack Iron Turtle SE ;
if you see one Iron on pcb , that's mono (Balanced) Iron Turtle
Attachments
so , schm are here
both Pumpkins ( meaning - what's on electronic pcbs ) are completely compatible to be used as (still MIA) Iron Pre , not yet visible in DiyA Store
so , selector - pot - buffer - autoformer ( be it everything SE or Bal) is good and certainly cheaper way , even if definitely few notches down.
schematics are proven , so there is one (little more complicated ) possible substitution for Toshiba/LS approach.
feel free to try it , on protoboard or PtP , your choice
how it sounds ?
well - foot-tapping , music is there ..... no thoughts about spectrum dividing , music is simply there
and - I'm not responsible for that ( even if Genius) ..... that's simply nature of Good Iron
....... remember M2, F6 , Ultrasimple BAF Sony amp ?
both Pumpkins ( meaning - what's on electronic pcbs ) are completely compatible to be used as (still MIA) Iron Pre , not yet visible in DiyA Store
so , selector - pot - buffer - autoformer ( be it everything SE or Bal) is good and certainly cheaper way , even if definitely few notches down.
schematics are proven , so there is one (little more complicated ) possible substitution for Toshiba/LS approach.
feel free to try it , on protoboard or PtP , your choice
how it sounds ?
well - foot-tapping , music is there ..... no thoughts about spectrum dividing , music is simply there
and - I'm not responsible for that ( even if Genius) ..... that's simply nature of Good Iron
....... remember M2, F6 , Ultrasimple BAF Sony amp ?
Attachments
Last edited:
Store got Iron Pre (selector - pot - buffer - Jensen autoformer ) , both iterations - SE and Balanced
you can see details in What's wrong with the kiss,boy ? thread ....... for all details regarding time frame and availability, 6L6 is man to ask
regarding Iron Pumpkin(s) - all details are here ... if somebody is willing to make own pcbs and use whatever AVC available , that's OK with me ;
if not , everything is already available from me , as more than elaborate packages .... elaborate meaning - everything critical and tedious to obtain will be packed together
that includes AVCs , those shipped as fully assembled (on pcb) and tested units .
not cheap , but cheaper than anything on market
details to follow , not here , but in trading thread
here we are to learn and share
you can see details in What's wrong with the kiss,boy ? thread ....... for all details regarding time frame and availability, 6L6 is man to ask
regarding Iron Pumpkin(s) - all details are here ... if somebody is willing to make own pcbs and use whatever AVC available , that's OK with me ;
if not , everything is already available from me , as more than elaborate packages .... elaborate meaning - everything critical and tedious to obtain will be packed together
that includes AVCs , those shipped as fully assembled (on pcb) and tested units .
not cheap , but cheaper than anything on market
details to follow , not here , but in trading thread
here we are to learn and share
I forgot to write that down - Turtle is having 24+1 taps , where 24 being for 2db increments from -46 to 0db ,and +1 being lowest tap ,obviously for GND
frankly - I was content even with Intact's Audio 15 position/3db increment 200$ jobbies, for my needs , not being nit-picky for extra-fine steps ;
however , what I wanted -when starting from scratch- is not just finer increments and more steps,but also solution with really decent (superior?) switching , so I chose this sort of relays ....... too much Elna and Grayhill and whatever pro switches passed through my hands , needing cleaning after few years of daily usage.
smaller Takamisawa cousins are used even in input switching of MC/MM signal , in gadgets of some reputable production
I chose bigger ones , as better for task , even with price of slightly larger pcb
frankly - I was content even with Intact's Audio 15 position/3db increment 200$ jobbies, for my needs , not being nit-picky for extra-fine steps ;
however , what I wanted -when starting from scratch- is not just finer increments and more steps,but also solution with really decent (superior?) switching , so I chose this sort of relays ....... too much Elna and Grayhill and whatever pro switches passed through my hands , needing cleaning after few years of daily usage.
smaller Takamisawa cousins are used even in input switching of MC/MM signal , in gadgets of some reputable production
I chose bigger ones , as better for task , even with price of slightly larger pcb
Last edited:
wow! i also love the relay solutions everywhere.
nice reading also, more than perfect for a sunday morning coffee. (my brain was too weak to really get the genious gnd-management in balanced iteration, but i´m SE-boy anyway).
nicely done!
nice reading also, more than perfect for a sunday morning coffee. (my brain was too weak to really get the genious gnd-management in balanced iteration, but i´m SE-boy anyway).
nicely done!
An externally hosted image should be here but it was not working when we last tested it.
