Black Freeday

 Not many thinks kicking around lately that need new homes but these might appeal to a few of you.

A customer needed some naked speakerformers that went form -3 to -7dB and I inadvertently calculated and wound a pair that were -3.5 to -7.5dB.  Rather than bin them for the copper I figure someone here can put them to good use.  Full details of the concept can be found here.

leave a comment with a way to identify you and random.org will pick the new owner mid next week and of course I'll cover shipping.

dave

New Math

 

THERE! I said it, and I have been trying to say it for years.  Now I am going to use my junkie pulpit to try to explain it in simple terms.

The basic rule of transformers is that the transform the load by the turns ratio squared.  This simple function allows an 8Ω speaker to present a 3200Ω load to a 300B by using a transformer with a 20:1 step-down ratio.  This makes the world an easy place to live in and keeps everyone sane UNTIL someone then asks a poor transformer winder for a 10K:600 linestage transformer and a hole is ripped into the space-time continuum.  The conversation typically goes like this:

Customer:  I need an 10K:600 lineout.

Dave:  What will this be driving?

Customer: Interconnects and the input of the amplifier.

Dave:  Do the interconnects have a 600Ω resistor soldered across them?

Customer: No.

Dave:  Does the amp have a 600Ω input impedance?

Customer:  Of course not.

Dave:  What is the amplifier input impedance?

Customer:  I dunno... 100KΩ... 10KΩ worst case.

Dave:  Then what is the purpose of the 600Ω secondary designation?

Customer:  A 5KΩ:10KΩ transformer would have a 1:1.4 step up ratio and I want a 4:1 step-down ratio to lower the output impedance.

Dave:  A 4:1 transformer loaded with 10K will load the tube with 160kΩ

Customer:  No I want 10K.

Dave:  Head hits table.... nappy time.

This is the thing that urban legends are made of.  Everyone treats transformers as if they all are used to deliver power to a load.  This is perfect for the vacuum tube output transformer since the entire purpose of the transformer is to take the high voltage high impedance anode of a tube and get it to a low voltage low impedance to drive a loudspeaker.  Universally applying this to every audio transformer simply leads to confusion and needless conversations like the one above.  I will categorically state that the worst thing you can do to a transformer sonically is to load it.  In the case of the tube-speaker interface this loading is a necessary evil. In just about every other case in audio, transformers are used to deliver voltage to a high impedance and not current to a low impedance.

I do not begrudge the really smart guys from Bell Labs who did the telephone grid at 600-900Ω.  I tolerate the audio engineers who borrowed the 600Ω impedance as a standard so everything played well together when any mismatch could prevent people from properly hearing Duane Allman at the Fillmore East.  Using those numbers today in audio where systems are meticulously curated is just begging for confusion and mediocrity.  

So back the the 10K:600 step down transformer.  The one place in audio where this transformer will really will be loaded with 600Ω is when used to drive the Pultec LCR Riaa module.  For everything else in audio, the 10K:600 has little if any relationship other than being a convoluted way to say the device has roughly a 4:1 turns ratio.  You will find many who insist that the transformer needs to be terminated with its characteristic impedance in order to work as designed.  With vintage stuff, that is often the case, but that is simply due to the fact that it was designed with the intent to be loaded with 600Ω.  

Lets assume a 4:1 ratio transformer is needed.  I can come up with two very different designs based on its intended use.  99 times out of 100 the actual secondary load will be the grid of a tube so the transformer will be operating essentially unloaded.  Rather than use an incorrect device designed for a 600Ω load, why not simply use a device designed to operate into the actual load? 

The proper (according to me) way to specify a transformer that operates in the voltage realm is to state the source impedance, turns ratio and expected load and then the proper device can be wound.  Using this approach takes what many call a black art back to the nicely ordered oblate spheroid we live on.

dave

Doc Hoyer

Audio lost another of its treasures last week when Dennis R. Hoyer passed away in Milwaukee.  If you had a vintage transformer that needed help, Doc was the person to call.  On a recent trip to Milwaukee where I grew up I had the chance to visit his lab as 60+ years of his work was being dismantled.  As I was admiring his winder his cousin Damian put his license plate from 1980 in my hand and suddenly I got all choked up and truly knew who Dennis was.  


His files are a Who's Who of tube audio history

His winder was CNC controlled and built by him

I can't imagine driving around Milwaukee as a teenager seeing that bright yellow TESLA on the back of a Ford station wagon and not realizing I was behind someone truly ahead of his time.

SUT Writeup.

I figured it was time to jot down some ideas on how step-up transformers are best applied to MC cartridges.  Hopefully the info at the link is easy to digest and not too much of a veiled ad.

Moving Coil Step-Up Transformer Design Considerations.

as a teaser here is the venerable peerless 4722 compared to a SUT that follows the practices I have laid out.  Please leave comments with any critiques.

SIT on it

Back in my younger days in Milwaukee I used to hang out with Arthur Fonzarelli drinking Shotz beer talking about music and all of the cool things in life.  Today we have the interwebs and these conversations now occur worldwide.  One of my good customers in Hong Kong was interested in some SIT (V-fet) amps and with curves like a triode I had to wonder why I have yet to hear a V-fet amp with a nice DHT sound.  I offered to do a pair of nanocrystalline cored transformers for the Tonkin 2SK180 and then things got a bit out of hand.  Here are three pair nano of outputs ready to be packed up.

I had the idea that maybe much of the lacking sound I have experienced from my limited exposure to amps with these "triode-esque" devices had to do more with the with the circuit than the devices. We figured the first place to start was the power supply which needed to be choke input with amorphous cored 250mhy 2.5A chokes.

Next came the issue of driving the ungodly 2.5nf of input capacitance these devices have.  Historically an active buffer and fixed bias is used and I figured it was best to go with what was known to work and suggested a 417A into a 49% nickel 4:1 stepdown transformer.  Then it was suggested that maybe a Telefunken RE404 could be used and of course 80% nickel too.  I had spice crunch some numbers and sure enough for this particular case the DHT would do it.  The problem became actually getting a wide bandwidth 200hy @10ma  4:1 done on an 80% core.   Since there will be Coleman regs on the filaments I figured why not run that 450ma of high impedance current through a tertiary winding on the 4:1 to null the current through the primary?

The resulting 4:1's have a 200hy inductance with a window of operation of 7-11ma and when all put together this amp should have -1dB points of just under 20hz to around 40Khz.   Bias for the 2A of output device current will be in the form of a 7-8 ohm source resistor and bypass cap.  In the future this can be compared to fixed bias through the transformer secondary.

Stay tuned...

Digital Sand Bomb

In keeping with the "Bomb" theme here is something I have been playing with.  It is a Step up device based on the Nelson Pass Zen  I/V converter for a Dac.

Of course I have no use for resistive loads, coupling caps or anything digital so I dressed up the 2SK170 / 2SJ74 combo with a good dose of 80% nickel and feed it with a MC cartridge.

Yes this is a stereo pair with both input and output transformers and it loads the cartridge with around 10 ohms. 

Initial tests are quite promising.

Tarragona Audio

Pictured below are 14 pieces of lovingly hand crafted nickel for an all out design of a microphone preamp by our good friend Peter in Berlin.

This device is unique because unlike the typical professional "buffer it all and let god find the music" approach of some, Peter harkens back to the days of yore when transformers and inductors were the proper tools to use and what something sounded like actually mattered.

The 14 permalloy pieces consist of 4 - DC carrying output transformers that load active devices, 2 - level control autoformers, 2 - tapped filter chokes, 4 - nickel power supply chokes, an input transformer and a phantom power choke.  No caps exist outside the power supply and at the next ETF I will lobby for some filament supplies.  As it stand, i think it is awesome and if given the choice between using transformer loaded Fet's and chokes or tubes and caps I think peter made a wise choice.

full details of this design can be found at Tarragona Audio

GM70, Silver and Nanocrystalline

Just finished these up for a GM70 amp Jeffrey is building for a customer.  They are wound with 0.9999 pure silver wire on big nanocrystalline cores.

The silver winding wire is brought out for direct connection to the circuit.  While I am a firm believer that it is the amp designers responsibility for the ultimate packaging aesthetic, I do have some 125mm cubed aluminum covers that will cover them if needed.

After many years of insisting that doing multiple taps only adds to the compromises I want to avoid in a design, I have finally come up with results that I deem acceptable.  Below are the plots of a similar transformer wound with the ability for 4, 8 and 16 ohm taps.

For behavior below 100hz I look at the inductance plots to make the best choice of inductance vs. current for the intended application and set the air gap accordingly.  The plots below were taken with the equivalent of a 16W excitation voltage @ 50hz.  Pink would be suitable for a parallel 300B,  Brown an 845 or GM-70 and something inbetween green and blue could be chosen for a 211.

dave

Nickel and Amorphous L1's

These just came out of the oven both for Choke Input power supplies.

The bigger one is on an amorphous core and weighs in at 14 pounds will find service in a GM-70 amp.  18Hy @ 250ma for a 1000V B+  The smaller pair are on 49% nickel and measure in at 100hy @ 20ma.

Autoformer for size reference:-)

Sut Selection (Why I don't do a one size fits all unit)

I just recently had a customer ask for a single SUT to match three different cartridges with a range of internal impedances from 1.5 to 15 ohms.  While that range is seemingly small one only needs to look at the frequency response plots to see why I like to match a impedance to a transformer.

These are the plots for 4 different cartridges (2r, 6r 9r and 15r) all feeding the input of one of our phono stages.  This particular SUT was designed for a 10 ohm cartridge and once that is known I was able to wind it so that the cartridge stayed well behaved to beyond 100Khz and only had a slight attenuation at 10hz.  I would be comfortable uning this cartridge in the 6-15 ohm range but if I were to drive this transformer with a low Z cartridge like a Lyra Olympos SL, suddenly a 10dB peak would appear above the audio band.  In this case, since this peak is happening at close to 200Khz one could argue it is sonically benign but armed with the knowledge that a 2 ohm cartridge will be the source I simply adjust things to minimize the out of band resonances like in the plots below.

I know there are those that will argue that adding a snubber network to the first transformer driven by a 2r source can eliminate the ringing and give a very beautiful frequency plot but somehow doing that also steals a fair bit of the music along the way.   For more of my thoughts on loading SUT's take a peek at this.

Filament Boards.

Hey... where have jeffrey and dave been???  Well the answer to that is busy.  To add at least a little content to this miniscule corner of the web, here is one of the small things I have recently made.

This is a "keeper" board that attaches to a pc mount bobbin and holds a schottky bridge and a pair of caps for a choke input filament supply.  The chokes are dual bobbin and wound with flat wire for choke input duty and have split windings so one leg goes in the + side of the supply and the other in the ground.  The PC mount bobbin allows for a nice way to hold the needed parts and the whole thing is supported by the mounted choke.

The ultimate use of these boards is for providing DC to some coleman regs and here is the resulting waveform when drawing an amp and a half of current.  Yes that is a 60mv p-p 120hz sine wave

Finally a picture of what is really going on.

Big (ger) Amorphous.

The old "Big Amorphous" cores have been downsized.  They were perfect for a 10W from a 300B or a 75TL but the requests for 20W at gobs of current keep coming in.  Sure I can gang up multiple cores or add an offset winding but the winding window gets small quick. 

 

 

The new option takes up a 6" cube and stay tuned to see what gets built from them.

Round Plates

This is a circa 1990 Alessa Vaic copy of a Marconi R-type and it has that beautiful round plate structure we have grown to love.

Unfortunately the tube has no emission and the real purpose for this post is behind the tube which is a set of iron for another round plate tube that does the R-type X 4... the 304TL.

The bigger ones are the SE outputs that can take 200ma of plate current while putting out 40 watts at 50hz.  They have bandwidth (-1dB) from 7hz to 49Khz and can do all of this due to the offset winding that runs the 12.5A filament current through the outputs in the opposite direction as the B+.  This brings the other two chokes into play which are the 20mhy 12.5A filament chokes that keeps the offset winding from loading down the tube.

The set of 4 weighs in at 46 pounds of copper and amorphous and boy am I glad the topic of silver never came up.

What The Vari-Load Batman.

What are all of those secondary wires coming out of that amorphous GM-70 output for??

75TL thrown in for size

I have made it well known over the years that I don't do multiple secondaries or taps because I feel they compromise the overall design. Well I guess I need to change that particular stance and I would like to introduce you all to the Vari-Load concept.

By adding a second strategicly places winding on the secondary I am able to get three different turns ratios which all have essentially the same frequency characteristics.  When you work the speaker impedance back to the load you get a number of options to play with.  Below is a screen grab of the spreadsheet showing how the numbers work out.

The three possible connections are:
Secondary 1 alone (Sec 2 remains floating)
Sec 1 in series with Sec 2 adding
Sec 1 in series with Sec 2 subtracting

since secondary two has a small number of turns and is never referenced to ground, it doesn't introduce the capacitive issues that plague other transformers causing varied response as you adjust the impedance.

To make things even worse (and I assure you i did not draft this on 1, April.... take a gander at this.

 Sure it is a box.... Big deal...   but the next three pictures may terrorize some of you.