Continued sidetrack discussion from "Tiniest Portable amp I can build" - nikongod microtransformer-based impedance step-down box

[jdkJake]

Series a single source into the primaries of two transformers in series and parallel the secondaries of those transformers into the output?

Or do you mean output the secondaries of the first transformer into the primaries of a second transformer?

 

[jcx]

lots of options, depends on source V, load's sensitivity, impedance flatness as to what you want to do
 
if the base impedance ratio, DCR are OK and you just want a little more headroom and better LF distortion then I'd parallel windings on 2 pair and then series the pairs same order/pattern for pri and sec
 
(was easier draw series paralleled)
 

 
for really high sensitivity iem you may want to parallel all 4 secondaries getting lower cutoff frequency and lower series R
 
not "efficient" use of materials vs a custom design - but lots cheaper

 

[jogor]

Hi jcx,can you recommend a core material suitable for audio? What are the things that i need to look for in the specs when choosing the right material? Will toroid core do for such application? I would like to try my hands on this. Appreciate your help. Thanks.
 
And by the way thanks to jdkJake for starting this separate thread and of coarse to nikongod for bringing out this great idea. This should be a cool project.

 

[jcx]

not really - I have had to design specialty xfmr about 3 times in my career and was immersed in interface performance issues of tape wound core based sensors which I couldn't change
 
so I have some general knowledge but nothing to compare with actual xfmr designers at places like Jensen, Cinemag, Lundahl, Sowter.. or even Hammond for that matter http://en.wikipedia.org/wiki/Dunning%E2%80%93Kruger_effect
 
one of my employers had a in house xfmr manufacturing shop for sensor and smt ferrite core DC-DC converter products so I was able to get pro help and borrow machine and operator for a couple of prototypes
 
it is hard to appreciate just how fine the gauge of wire and its delicacy - it really has to be machine wound and skilled operators are helpful too when you try to get enough turns for audio xmfr at smaller sizes
 
this thread gives some pointers to audio xfmr companies: http://www.head-fi.org/forum/thread/453348/transformers-for-balanced-to-se-conversion#post_6126403

 

[jdkJake]

lots of options, depends on source V, load's sensitivity, impedance flatness as to what you want to do
 
if the base impedance ratio, DCR are OK and you just want a little more headroom and better LF distortion then I'd parallel windings on 2 pair and then series the pairs same order/pattern for pri and sec
 
(was easier draw series paralleled)
 

 
for really high sensitivity iem you may want to parallel all 4 secondaries getting lower cutoff frequency and lower series R
 
not "efficient" use of materials vs a custom design - but lots cheaper

Okay, that makes sense, thanks!

I will probably build it to try both topologies, however, with the 50ohm impedance of the image X5's, I suspect your layout will be a better first staring point. I have not been able to locate an impedance curve for the X5s, so, I will be kind and make no assumptions on it's flatness. Being a single armature, I suspect it is far from that, but, one never knows I suppose.

If you don't mind me asking, how does series resistance play into this scenario?

The hard part will be locating enough parts. I cleared out one RatShack of it's two remaining parts. Hopefully I will not need to visit four stores to fill out the material! Still, here in North America, there are more Radio Shacks than McDonalds, so, I will not have to travel far.

Thanks again for the insight jcx.

 

[jdkJake]

Round Two...
 
Barbecue got postponed due to rain, so, I had some time on my hands. I picked up the parts for "Round Two, the jcx Variations".
 
This one took a bit more than 10 minutes to construct:  

 


 
 
Now we are moving in the right direction! This is the build that jcx suggested in the posts above. Same input and output impedance (gotta love resistor math), slightly more complex implementation. 
 
Only had a few minutes to listen, but, far better sound than the first implementation as the load is being spread across multiple transformers. The high end that was missing in the first version is far more represented. Still a touch rolled-off, but, only a touch and not too shabby. Certainly a big step forward. I need to listen some more prior to posting a more insightful opinion, but, so far, I am impressed at how well it works. Ton's of signal getting through! Very little attenuation.
 
What a great learning experience and fun build! The Altoids box implementation is a trip. At times, like building a ship in a bottle.
 
In any case, I will post more later. One more before I go:
 

 
Gotta go, the pucks about to drop... GO BOLTS!!
 
 

 

[jackmccabe]

mmnmm... beautiful

 

[jdkJake]

So, I had a bit more time to listen to the "round two" version.

Using an iPod touch, it is best if volume is kept below the 3/4 mark (to the left of the fast forward button). Once you go past that, things get a bit more dicey. Whether that is the transformers saturating or just the iPod maxing out and getting a bit funky is unknown. I need to try another source with digital attenuation to be sure. In any case, anything above 3/4 volume is pretty loud for most material, even with the drop in output across the transformers. BTW, the Image X5's are rated at 115db/mW, so, YMMV.

In any case, the sound is pretty damn good considering the pedigree of the parts and my less than optimal, but extremely compact implementation. Certainly mid-centric with nice high-end extension, perhaps a tad rolled off at the very top, but far from objectionable. It really depends on how hard you push it (how loud you listen). In this case, lower volume yields better results. The low end is there, but, might not be quite as extended and deep as one might want, but, no slouch either. The transients seem fine, perhaps just a tad slow, if that really means anything. A pleasing sound nonetheless.

In any case, these are nits against a pretty basic design concept. Certainly a keeper for a different perspective. I like it.

Now how to figure out how to get it past TSA without them freaking out.

 

[nikongod]

Glad to hear you are still liking it, and mega-congrads for trying new stuff. That looks COOL!
 
Mine has been checked (and cleared) by TSA a few times.

 

[AudioCats]

the 8x transformer version is effectivly a 22:1 step down, if you are willing to go that far, might as well use the transformers from a Stax SRD6 or SRD7 box, I am pretty sure it will sound a lot better.  (yes, the evil-ness never ends!

Hunt down and gut all Stax boxes!).  
 
Won't fit into a candy box, but good for desktop use nevertherless.....

 

[jdkJake]

 
Quote:

the 8x transformer version is effectively a 22:1 step down, if you are willing to go that far, might as well use the transformers from a Stax SRD6 or SRD7 box, I am pretty sure it will sound a lot better.  (yes, the evil-ness never ends!

Hunt down and gut all Stax boxes!).  
 
Won't fit into a candy box, but good for desktop use nevertherless.....

I must admit, you are a man on a mission!
 
What did all those poor Stax adapters do to you to deserve such commitment to their ultimate disembowelment? 

 

 

[AudioCats]

 
 
now lets get serious. How about gutting a dead Stax transformer (if there is such thing.... of course, you might have to kill it first! ), pull out the silicon steel laminates, and cut them into tiny "C" & "I" pieces, roll your own mini bobbings, wind your own coil, and make your own mini Stax-ious "R" core step-down's, huh? Now that, is "old school"....

Many many many mini-Staxi-transfromer can rise from the carcass of just one SRD transformer, the dead SRD will go to Stax heaven with a big smile.....
 
(like I said, the evil-ness never stops!) ))
 
 
 
ok, that might be slightly over-board, but the "roll-your-own" mini transformer idea is still valid. If done right they can even fit into your candy tin. let me know if you want some M6 silicon laminate pieces (don't worry, not from a gutted Stax, they are from a shorted out JRM electrostatic step-up transformer).

 

[jdkJake]

Thanks for the offer, but, I have no plans on winding my own transformers.
 
At least not in the immediate future....

 

[AudioCats]

ok then. I guess I will wind a set of CI core transformers and maybe put them into an old empty iPod case I have.
 
 
done, here.
 
 

 

[samsquanch]

Thanks to a hang over and my love of going to the local mom and pop electronics shop (Baynesville in Towson, MD, whudup), I decided to do build a step down transformer box myself.
 
My conclusions, well they're transformers, they're by nature fairly transparent, so no coloration.  The highs are a touch rolled off, but only really noticeable if you're doing a frequency sweep, and I'm talking about around 15.5kHz, and even then, nothing dramatic.  I didn't notice any reduction in hiss, but it wasn't there before either.  I did notice some more detail in the music, but that could be attributed to me listening a bit more critically at the moment.
 
I'll have to play with it a bit more, but I think for now I'll mostly use this box for listening to iem's out of my beta22 when I finally finish building that...
 
Something that crossed my mind while building this, is a microphone pre I used once in a studio, it was essentially two pre's in a box, one tube pre and a step up transformer, and on the front was a mix knob that allowed you to blend the two together in what ever combination you wanted.  It made me start thinking about building a cmoy/transformer headphone amp to see how it sounded, add that to the list of projects I suppose...