Transformer Reversal LoZ_HiZ to HiZ_LoZ


2010-02-19 11:23 pm
1. i promised a needy acoustical guitar worship musician i'd construct an outfit for his guitar's signal to the church's sound engineer.
2. i already built the 24v lithium .05ma jfet and barium titanate transducer, pristine signal authenticated on scope.
3. trouble is the cost of expensive DI tranny to obtain balanced output for engineer.
4. purchase the opposite tranny cheaply, one dollar, because it's male 1/4" TS was bent severely (still useable). the other gender was female XLR.
5. can i extract the internal tranny and reverse it, arbitrarily assigning XLR (using XLR male gender) polarized earth ground to paragraph two's ground?
6. or is this unworkable? we do not know what this prehistoric radioshack japanese tranny windings ratio is. we do not know whether windings compensation was incorporated to optimize fidelity and signal strength presevation for the specific intention of converting (larger ampere) balanced to (lower ampere) unbalanced.
7. thanks for your assist in advance.
8. that other diyaudio forum discussion was a great success (pilot's mouthpiece used for paper speaker signal generation, jfet buffer, DI tranny for sound engineer's balanced signal. placement of sensor onto paper speaker grilling determines appeal signature!


2007-05-30 1:49 am
The Impedance would reverse, possibly putting a strain on OP/IP devices... There is no reason it wouldn't work other than the Amperage/Voltage changes.

Do the math, if all is within range.....Give it a try. Start low and check for Distortion, Heat, etc.
COnsider this option. Build your preamp for the guitar, and make it with an ordinary High-Z output like you would plug into a guitar amp. Then use a commercial direct box. For example this at $30.
Live Wire SPDI Passive Direct Box with Attenuation Pad | Musician's Friend

Behringer ULTRA-DI DI600P Passive Direct Box | Musician's Friend

And if you don;t care about ground lifts and other features, there is this really basic model for $14:
Rapco Horizon DBBLOXF Lo to Hi Z Transformer Direct Box | Musician's Friend
When I invented the DI box in the sixties (hey, I didn't say I was the only person to invent it, or even the first; I just wasn't in contact with anyone else who was using them, and the problem of impedance conversion and balancing were fairly obvious, as was the solution) I used a lot of tube input transformers salvaged from mixers and public address systems that had died. After all, there weren't any dedicated DI transformers around. Generally they worked very well. I wouldn't try and remount the transformer you've got, if the jack still functions; just build a male to male XLR lead, with an in line ground lift switch.

You could even, since you have a power supply (batteries), build an active balanced output, remembering to put output capacitors high enough voltage to avoid phantom power damage. A micropower dual op amp and a few resistors has to be cheaper than a transformer. Hey, if those batteries are rechargeable I'll dig out a charging circuit from phantom I've got around somewhere.

If the guitar's only going into the PA, there's no need for ground lift; that's for when he goes into an amp as well. I'd include it, and the extra output, anyway; with live sound it's better to be more flexible (as long as the guitarist understands what a ground lift switch is.)

If the


2010-02-19 11:23 pm
reversing loZ_to_hiZ tranny

thankyou jasonvh_enzo_chrispenycate_harleyjon

1. the japan tranny was extracted from radioshack casing, dcr was 1k and 48 for windings
2. the tranny is different: it is not constructed like the ones contemporary
3. instead it is flat low-profile, wrapped in mu-metal wrapper, just like a piece of candy wrapper
4. the modern tranny's meant to go from hiZ_to_loZ have much different dcr windings
5. this means the dcr 1k windings will load down the jfet's output ac signal, since the 24v lithium energy positive feeds thru a 220K (like a vacuum tube plate resistor)
6. normally the loading of the jfet's output should be in the ball park of 10K
7. should i proceed and hook up this reversed tranny?
thanks in advance for your help


2010-02-19 11:23 pm
thankyou JMFahey Reversing LoZ_HiZ_Tranny

1. the 24dcv lithium energy in conjunct with the 220K plate loading resistor in conjunct with the (must revisit actual value) cathode bias resistor determine current flow thru the toshiba n-channel jfet
2. i built several and necked current down to a minimum intent on saving chemical energy (extend life) without compromising the waveform (too cold bias bad, too hot bias bad)
3. i recall the gain was between three and five, so the cathode resistor value must be ball parked at 68K to 82K
4. because the transducer diameter (capacitive barium oxide titanate type) determines energy input, i necked down this diameter too, to preserve accomodation headroom. i recall the diameter selected final was nine millimeter, pretty tiny size (zero mass just about for soundboard bridge mass neutrality)
5. i can't scan requested hand drawn schematic (home has scanner but immediate neighbor floods my breathing air potent nerve agent smoke emissions, debilitating!). anyway, the schematic is just like vacuum tube preamp simplistic. grid tie_down resistor 10M, plate resistor 220K, cathode resistor 82K, coupling tantalum 1uf 35wvdc, output tie_down resistor 100K drains dc appearing on tantalum's negative. the japan jfet drain (plate side) voltage was centered midway of the lithium. a micro micro switch is soldered onto the positive lithium to gate energy release.
6. the previous musician complained of insufficient balls (throaty bass spectrum) even with the titanate stuck on the bridge rosewood bass side. so i parallelled a secondary titanate on the bridge rosewood treble side, epic response satisfaction (into a vacuum tube guitar amp 1M input impedance). the parallelled capacitive type of titanate impedance changes, for the better, the created signal is not going to be loaded (affected) by buffer circuitry which you are concerned of it not being low enough on its output impedance to drive typical cable lengths and musical amplifiers.
7. worse yet is the capability to drive the dcr 1K windings of reversed loZ_hiZ_tranny to provide the soundman's balanced XLR
8. usually the typical DI tranny has labeling to indicate 50K:600 impedance ohms but this reversed japan tranny has no impedance indication. the dcr 1K windings seems a bit low to throw the 220K impedance jfet's output into. this is your concern. i'll have to append this reversed tranny and scope the waveform legitimacy


2010-02-19 11:23 pm

darn, there's one more but my email search engine can't get it from historical but basically the design critereon was class A, low energy consumpt, and longevity of the channel from deterioration, yes, even jfet's wear like heaters and in the glass bottle.
also tried nte458/j201 but they drink too much energy
Ok, thanks.
From your verbal description I "see" :( you built a basic "tube type" gain stage, using a FET.

Fine in a preamp housed inside a head, connected to the next stage by at most a couple inches wire, but troublesome when 20 feet (or much more more) away from the amp or board.

And it demands a very expensive transformer (think Jensen or better) for a proper match.

You also have gain, which you had to pad down by trimming the Piezo pickup.
You are using a 24V battery, and the preamp needs 0.5mA , meaning a 12mW power consumption.

Not bad, but you still have too high output impedance and adding a (necessary) low impedance buffer stage will multiply that power consumption.

You may add a buffer , of course, but I suggest you go all the way and make a new 10M impedance Buffer, may be unity gain or have a little, if you happen to need it, using a TL061 ..... and which won't need the extra FET input .... because it already has its own :p

Instant high input impedance (10M if you wish), low output impedance (less than 100 ohms) and it will drive either a *long* cable or the transformer you have now.:)

You can feed it from an alkaline 9V battery which will last over a year (the one in my Bass is over 2 or 3 years, lost count, still works like a charm).

Current consumption: 0.25mA
Power consumption: 9V * 0.25mA= 2.25 mW .

Best is that if you use the double version, the TL062, you can have balanced output straight from the guitar, no transformer needed :eek:

Of course, you'll need to add a suitable balanced output connector, either (if you dare) an XLR connector straight at the guitar, or at least a stereo jack.

Breadboard one and give it a try.


2010-02-19 11:23 pm
thankyou JMFahey

1. mine is not 0.5ma consumpt, instead .08mA maybe upwards to .2mA
2. hammond sells the potted miniature tranny 107X, 300K:600 impedance without bankruptcy
3. isn't their internal supporting circuitry bjt's inside the texas instruments opamps recommended? yes they are fet input for fantastic decoupling
4. i built craig anderton's cd4049 buffers (was it six fet's inside?) but somehow the delicious wet sound signature just did not voice, i don't know why
5. my class A se tube amps have NE5532 and TL072 fet inputs because the 1st stages need to be clean as opposed to 12ax7 heater hum else dc filament support circuitry. they sound good so i put sockets to replace when their channels corrupted over time (take baseline data to show this)
6. so i must review your tl061/for any bjt's internally. they may work well in this bariuto XLR boundaries


2003-09-03 12:17 pm
With the kind of thing you have built, you really cannot drive anything passive (there is simply not enough current available). So you either have to build an external active DI (something like the Bo Hansen's one with an edcor xsm 10K:600 would be pretty cheap if you can rely on phantom power) or rethink your current preamp.

Is this what you built ? I had to correct your 82K to 8.2K to get working points (output at 12VDC and current at about 60µa).


Using the parts you already have, it probably would work better if you used a "cathode follower" kind of preamp, even at 0.5ma.


Btw, isn't the 107X 100K:1K ? Furthermore, the frequency response is pretty bad in the lower frequencies.


  • jfet.JPG
    25.6 KB · Views: 73
  • jfet2.JPG
    20.7 KB · Views: 59


2010-02-19 11:23 pm
thankyou belgium 00940

1. yes the 1st schem is what i built (i'll post cellphone pics, and the various modules of iterations, variant japan jfets, epoxied assemblage onto lithium dispose
2. yes i will investigate cathode derived ac output signal, floating input on dc bias, but increasing noise-prone discrete components of two (input dc blocking cap c2 and input dc bias float resistor r6
3. 0.5mA is quoted by the j201 jfet buffer literature, boasting years of battery duty. maybe i am compulsively overly energy consciencious. the energy law says you can't derive energy from nothingness, so you are correct, what we are striving for on energy conserve has a rock bottom minimum pursuant to constraints, endpoint boundaries (barium to XLR)
4. attached is the pdf of hammond's 107X and it clearly shows 100K input impedance on ONLY one half of the windings owing to CT topology meaning if the CT is unused the input impedance should be 200K
5. long back in public schooling i stumbled upon nazi germany application of magnetic transformer amplification, of course for war use. permanent magnets and transformers are PASSIVE but somehow the nazi german scientists achieved imparting energy and resultant ACTIVE amplification. so there must be something missing, such as guitar string vibration cutting fields of flux emanating out of the p90 for example, where the addition of energy clearly is the relative motion between the flux and vibrating ferrite


  • 107X.pdf
    357.4 KB · Views: 17


2010-02-19 11:23 pm

pl find attached pics
camera is not the greatest
so try zooming for values
i noticed some AAAA batteries so that's the 2sk596 operating at lowest headroom


  • IMG_1964.jpg
    643.7 KB · Views: 32
  • IMG_1965.jpg
    675.2 KB · Views: 15
  • IMG_1966.jpg
    742.5 KB · Views: 13
  • IMG_1967.jpg
    638.1 KB · Views: 12
  • IMG_1968.jpg
    519.1 KB · Views: 12