The Xtal Set Society
The Xtal Set Society
profile | register | search | main site


email us
Forums | Crystal Radio Think Tank ** Advanced Forum | High Impedance Transformer Post Reply Send Topic To a Friend
Author Topic
gzimmerPosted - 21 September 2012 2:36  Show Profile  Email Poster  Edit Message
I need to find a higher impedance audio transformer.

Dick Kleijer has built an interesting transformer at
http://www.crystal-radio.eu/entrafounit2.htm

He took 6 x Adastra transformers, disassembled them, and re-assembled them on a new core with the primary windings connected in series.

It seems to me that he could achieve the same results by leaving the transformers intact, but paralleling the secondaries so that the cores all shared the same excitation.

Presumably there is a good reason why Dick didn't do this, perhaps the losses would be greater? (a big single core being more efficient than multiple small cores?).

Dick has specifically requested that people not send him questions, so I hesitate to ask him directly.

Thoughts?


........ Zim

Edited by - gzimmer on 9/21/2012 2:36:46 AM

Electronic TechnicianPosted - 21 September 2012 22:8  Show Profile  Email Poster  Edit Message
Hi Zim.

I think my transformer assembly works as you suggest. I notice the expected change in tone quality as the optimum impedance is found and passed. I have not done extensive testing with it. I would suggest building the transformer simulator to see how the ideal transformer might work for you. When you find something you really like, you can configure your transformers to try to match that performance. I think that would take away much of the guess work. If you find a really good high impedance transformer that doesn't break the bank please let us know. I'm still looking.

Electronic Technician.

golfguruPosted - 21 September 2012 22:36  Show Profile  Email Poster  Edit Message  
Would paralleling the windings give you the same impedance as sharing a core would?
2 windings on the same core would give you 4x the impedance whereas 2 isolated windings give you 2x impedance?

In Dick's case - 36x impedance (for 6 windings) versus 6x separately?

Dunno - might have the concept misconstrued.

.......................

Edited by - golfguru on 9/21/2012 10:38:13 PM

gzimmerPosted - 21 September 2012 22:57  Show Profile  Email Poster  Edit Message
Golf,

Well you would put all the secondaries in parallel (which would link the cores) then put all the primaries in series.

So it should end up the same as Dick's ?

......... Zim

Edited by - gzimmer on 9/21/2012 10:57:55 PM

golfguruPosted - 21 September 2012 23:12  Show Profile  Email Poster  Edit Message  
You are probably correct, Zim.

I thought you would have to share the magnetic flux. "Mutual inductance" and all that (to achieve the "squaring" effect).

Quote Dick: >>>>>>>>
If we place several coils (number=N) on the same core, and connect the coils in series, the impedance will increase with a factor N² .
In addition we can load the transformer output with a higher load resistance then where it was designed for, this will increase the input impedance even further.
>>>>>>

..............

Edited by - golfguru on 9/21/2012 11:57:22 PM

Garry NicholsPosted - 22 September 2012 8:2  Show Profile  Email Poster  Edit Message
Hi to All;

I don't know about efficiency, but as Zim suggests, putting all the primaries in series and all the secondaries in parallel will get the transformation he describes.

I was tutored by Ben on this some years ago when I was completely befuddled on using mulitiple transformers.

Putting two windings in series doubles the turns count on that side. Putting two windings in parallel maintains the same turns count for that side. ("Side" being input or output side of a multiple transformer combination.)

Then apply Z transformation = square of the turns ratio to get the transformation.

Garry

golfguruPosted - 22 September 2012 16:31  Show Profile  Email Poster  Edit Message  
I suppose, put another way:

Paralleling the secondaries would reduce the individual impedance of each secondary by a factor of "n" (n = no. of transformers) while the impedance of each primary (seriesed) would stay the same?

Or:
Tot. secondary impedance = Zs/n
Tot. primary impedance = Zp.n
Transformation ratio = Zp.n:(Zs/n)
=Zp.n^2:Zs

Which gives the same transformation ratio as the common core version.

...........

Just wondering:
This approach might give the same impedance *ratio* as the "common core" method but the *actual impedance* values on either side will be "n" times less than the common core Z's?

This would change the "impedance" to "copper loss" ratio?

How would that affect the efficiency?

(Dick increases his tranformater's "apparent" impedance by connecting higher value, source/load Z's so could that work here, as well?

But ....... , how low can you go?
One would think that half a dozen honking great power transformers would not fit the bill?

Responses please.

.........................

Edited by - golfguru on 9/22/2012 4:52:08 PM

golfguruPosted - 24 September 2012 17:27  Show Profile  Email Poster  Edit Message  
golfguru asked:

>>>>>>>>>
Just wondering:
This approach might give the same impedance *ratio* as the "common core" method but the *actual impedance* values on either side will be "n" times less than the common core Z's?

This would change the "impedance" to "copper loss" ratio?

How would that affect the efficiency?

(Dick increases his tranformater's "apparent" impedance by connecting higher value, source/load Z's so could that work here, as well?

But ....... , how low can you go?
One would think that half a dozen honking great power transformers would not fit the bill?
>>>>>>>>>>>>

Any takers? Somebody must have an opinion?

.......................

Edited by - golfguru on 9/24/2012 5:29:10 PM

gzimmerPosted - 25 September 2012 0:5  Show Profile  Email Poster  Edit Message
I still think they will act as one big core.

However it seems that large cores are more efficient than small cores. Why I'm not sure, less flux leakage perhaps?

I do have a few of the Adastra transformers. Will try hooking them up shortly.

Have also purchased one of these from Ming
http://www.ebay.com/itm/-/150622666151?item=150622666151&ViewItem=&ssPageName=ADME:X:AAQ:US:1123


......... Zim


Edited by - gzimmer on 9/25/2012 12:08:23 AM

golfguruPosted - 25 September 2012 0:35  Show Profile  Email Poster  Edit Message  
>>>>>>> Have also purchased one of these from Ming >>>>>

They seem to have some good feedback at TRB.

......................

gzimmerPosted - 25 September 2012 2:42  Show Profile  Email Poster  Edit Message
This article has a good explanation of core losses
http://www.allaboutcircuits.com/vol_2/chpt_9/8.html

There are various factors, but larger transformers have less loss because they have lower flux density and therefore lower Hysteresis loss.

In the parallel transformer version I would have though that the flux density would be shared across all the cores (each primary having 1/6 the excitation voltage). So the total losses would be the same as one big core?


............ Zim

Edited by - gzimmer on 9/25/2012 2:45:18 AM

golfguruPosted - 25 September 2012 19:25  Show Profile  Email Poster  Edit Message  
All sounds good to me - be interested in your findings.

.................

gzimmerPosted - 6 October 2012 21:54  Show Profile  Email Poster  Edit Message
I put up a very modest antenna and had a play with the ebay matching transformer.

It works very well. It gives about the same sensitivity as my 0-15 Ouncer (which supposedly has 1M Ohm input).

The big difference is having lots of taps, it's obvious when switching them around which taps are the correct ones.

Definitely recommended.

later ......... Zim

Click Here To Close Thread, Administrators & Moderators Only.

Show All Forums | Post Reply