Marantz 8 B : Output Transformer

[refference]

Hi all ,

I'm planing to build a Marantz 8 B CLONE , for my own use .

Does anyone knows the impedance ( PLATE TO PLATE ) of
an original Marantz 8 B output transformer ???

And the TURNS RATIO of the screen grids taps ( 10% , 33% ,
43% or what ... ??? )

I already have the original Service Manual , and this is the
only data that I do not have yet .

Thanks in advance for any input .

Regards for all ,

Carlos

[rdf]

Have you seen this? Top article.

http://www.plitron.com/audio_articles.asp

[refference]

Thanks rdf ,

I already had seen , this article , BUT S. Kobayashi , has
adopted the primary impedance as 4,400 ohms plate to plate ,
and I do not know if this is the correct value ( neither him ) , like the original Marantz 8 B .

What I need to know , is the exactly value of an ORIGINAL
Marantz 8 B , and the turns ratio of screen grids taps .

I can not believe that anybody on this forum , has not measured
this value , yet .

Please guys , my question is still "on the air" , does anyone
know the answer ???

Thanks in advance ,

Carlos

[Brian Beck]

Carlos,

John Atwood had measured a number of commercial output transformers many years ago. I can no longer find his results on the web, but I kept the data. With an 8-ohm load on the 8-ohm tap of the Marantz 8B’s OPT, Atwood measured a primary impedance of 6.2K plate-to-plate. He measured the ultralinear taps at a 44% ratio. By the way, with a 4-ohm load on the 4-ohm output tap, he measured 7.3K ohm on the primary, showing that the turns ratio of the secondary windings was not quite perfectly set at 1.414:1, and/or that DCR was having an impact. I would use the 6.2K primary impedance as a target.

[JohnAtwood]

Here is the page with my transformer test results. Read the attached letter to see some of the limitations and to avoid misunderstandings. If I were to do the tests today, the results would be more sophisticated, but I believe the 1990 results are still valid, if you realize that they are really based on a comparison in a P-P 6550 amp.

Page: http://www.one-electron.com/Trans_Tests.html

- John Atwood

[Brian Beck]

Thanks John. I had perused your site this morning intending to post a link to these tests, but I must have overlooked where you had placed them (but there they are, under "Tech Library").

BTW, do you attribute the slight difference in primary impedance between the 4-ohm loading and 8-ohm loading to the secondary turns ratio not being 1.414:1, or to a secondary DCR-related difference?

[refference]

Hello Brian ,

THANK YOU very much for your reply , and for your
comments too . Great !!!
Talking about it , I think that my guess is 6600 ohms
plate to plate , with 43% screen grids taps , because
this value was “more or less” a standard at that time ,
for EL34’s push-pull output . But the target of 6200
ohms is good , too .
I am not completely sure about that , as I said , only my
guess . About the difference I think you are right , the
DC resistance , may have some influence . Every time
I had measured , any kind of output transformer I
NEVER could read the same primary impedance value
with 4 , with 8 or with 16 ohms secondary load .

Hello John ,

CONGRATULATIONS for the excellent job . It is really
amazing and complete .
This kind of job does not become older , it will be always
up-to-date , for sure .

Best Regards ,

Carlos

[Brian Beck]

Carlos,

As you probably know, the primary impedance need not be very precise. Most speakers will vary wildly from either 4 or 8 ohms across the band, and the plates will see a stepped-up version of this varying impedance (both in magnitude and in phase). I would think that a good wideband OPT with a 6600-ohm primary and 43% taps will work fine. The only catch is that the Marantz 8 OPT used separate windings to return feedback to the first stage cathode, rather than taking feedback directly off the speaker windings. John’s table doesn’t address those windings. Note that there are several passive components used to compensate the response for good closed-loop stability. All of this would be different with a different OPT. You would have to redesign the feedback circuit compensation like Kobayashi did in the AudioXpress article. A faithful clone will not be easy. Good luck!

[JohnAtwood]

I've been traveling, so couldn't answer your question right away. The inexact nature of secondary impedances is most likely due to the quantum effects of keeping a constant number of turns per layer. I've been getting into transformer design, and have found that breaking a layer in mid-winding, to get the exact impedance, is asking for trouble - mainly high leakage inductance and lousy HF response. From dissecting commercial transformers, it is common for the secondary impeances to be only approximate (although consistent). I suspect the "43%" ultra-linear tap came out the same way - if the primary was made up of 7 layers, picking the tap off of the end of the 3rd layer gives almost exactly 43%.

- John

[Brian Beck]

Thanks for the insight. I'll remember that 3/7 rule of thumb.