ST-35 with Tango Output Transformers

for the DIY ST35, the Dynakit and every other PP EL84

ST-35 with Tango Output Transformers

Postby Brik » Sun May 16, 2010 9:30 pm

Hi all,

I have a pair of Tango FE-25-8 transformers that I bought more than 3 years ago in the hope of building a Dynaco Stereo 35 clone around them.
Now that She Who Must Be Obeyed expressed her preference for not having idle items (i.e. the items in my precious Nice to Have These Parts for Some Non-specific Future Projects pile) in the house, I was compelled to get started on the project.

Tango FE-25-8s are potted, Ultra Linear, push-pull, 8K Ohm primary, 4-8-16 Ohm secondary, 25W capacity output transformers with silver-gray Hammertone finish:
Image
The "owner's manual" is a single page data sheet. These transformers have a good wide frequency response:
Image

One vital piece of information I could not find in the data sheet, however, was the screen grid winding ratio.
Since it is easy to measure the turns ratio of the screen grid winding vs. the total number of turns in the plate winding, I decided to measure the ratios using the trusty old Heathkit Model IG-18 signal generator.

While I was at it, I decided to measure the plate-to-plate inductance of the primary as well, since the only additional piece of "equipment" I needed was a reasonably precise resistor.
(See http://www.daycounter.com/Articles/How- ... ance.phtml )

Then, I decided to measure the same set of output transformer parameters of the units in some of the tube amplifiers I had on hand.

In all I have the following transformers measured:
1. Tango FE-25-8s
2. Dynaco Z565s in an old Dynaco SCA-35 : Measured in circuit, with speaker load "open", NFB connected.
3. Dynaclone Z565-48 in a Dynaco ST-35 clone : Measured in circuit, with speaker load "open", NFB connected.
4. Triode TF110-48 Deluxe in the "Poor Man's ST-35" : Measured in circuit, with speaker load "open", NFB connected.

Some scenes from the measurements:

Image
Image
Image

Here are the tabulated results:

Image

I have the following observations:

    1) The Tango transformers have SG winding of 40-42% which is the standard value I expected to see.
    These Tangos are very well matched in all respects: DC resistance, Sg winding ratio, and inductance.
    2) Both the Dynaco Z565s and the Triode Dynaclone Z565s have SG winding ratio of 22-24% which I did not expect. Are these optimized for 6BQ5s / EL84s?
    Would a "standard" output transformer with SG of 40% give an optimal performance with EL84s?
    3) Both the Dynaco Z565s and the Triode Dynaclone Z565s have exceptionally high inductance that is unmatched by the Tangos.
    Would Tangos sound as good as the Z565s with EL84s


All in all, I renewed my respect for the Dynaco Z565 designs, both old and new.

Note: I had to use the fairly high resistor value of 330K Ohms since the Z656s have such a high inductance that I had trouble measuring with lower valued resistors. As a consequence, the inductance was measured using miniscule currents which may result in higher inductance values.
Last edited by Brik on Mon May 17, 2010 7:39 pm, edited 1 time in total.
/b
User avatar
Brik
 
Posts: 193
Joined: Fri Apr 07, 2006 5:30 pm
Location: West of Boston

Postby Shannon Parks » Mon May 17, 2010 6:00 am

Nice tests, Brik. My Rustiest ST70 shop amp uses the TF110 outputs and I'm always amazed they work as well as they do! Regarding your new respect, I have a pair of original Z565s to swap for your Tangos. LOL
User avatar
Shannon Parks
Site Admin
 
Posts: 3478
Joined: Tue Mar 18, 2003 5:40 pm
Location: Mahomet, Illinois

Postby Ty_Bower » Mon May 17, 2010 6:33 am

Neat results. I once measured the DC resistances on the primary of a genuine Z-565, and found results similar to yours.

Dynaco Z-565
P1 = 208
S1 = 53
P2 = 172
S2 = 44
"It's a different experience; the noise occlusion, crisp, clear sound, and defined powerful bass. Strong bass does not corrupt the higher frequencies, giving a very different overall feel of the sound, one that is, in my opinion, quite unique."
User avatar
Ty_Bower
KT88
 
Posts: 1490
Joined: Wed Mar 21, 2007 2:50 pm
Location: Newark, DE

Postby Brik » Mon May 17, 2010 7:42 am

Hi Shannon,
Thank you for the reply!
As much as I respect the original Z565s, I'll pass on your generous offer at this time... Image, unless, you still have those Magnequest MQ-565s you tested
in the outstanding Transformer Shootout- Hammond vs Magnequest vs Dynaco vs HW article...

Hi Ty,
Thank you for your measurements! More data points, the better.

I have been wondering why the Dynaco Z-565 #1 has the measured inductance that is less than half of the other one's.
I wonder if there is a single-turn short or some such in it.
/b
User avatar
Brik
 
Posts: 193
Joined: Fri Apr 07, 2006 5:30 pm
Location: West of Boston

Postby Ty_Bower » Mon May 17, 2010 9:15 am

Brik wrote:I have been wondering why the Dynaco Z-565 #1 has the measured inductance that is less than half of the other one's.


I'm going to go back and read the page you linked with methods for measuring inductance. Maybe I can repeat your experiment on my sample. Inductance in transformers is a funny thing. While I won't argue that inductance is necessary to the operation of the transformer, I'm not convinced that what you've got is really the same under all conditions. I've read some articles that suggest the working inductance of the primary winding (in the actual amp circuit) may not be what you measure on the bench. Still, bench measurements do provide an interesting point for comparison.

I noticed with the sample of Z-565 that I saw, the laminations were fairly loose. They were not tightly varnished together into a solid block. If you loosened the bolts holding them together, you could easily shift the stack around a bit. In fact, this set came to me a little slanted and cock-eyed. I gently tried to square and snug the stack, then re-tightened the bolts. I bet if you had a few tenths of a millimeter gap between the ends of the E lams where they meet the I lams, you might take a big hit in inductance.
"It's a different experience; the noise occlusion, crisp, clear sound, and defined powerful bass. Strong bass does not corrupt the higher frequencies, giving a very different overall feel of the sound, one that is, in my opinion, quite unique."
User avatar
Ty_Bower
KT88
 
Posts: 1490
Joined: Wed Mar 21, 2007 2:50 pm
Location: Newark, DE

Postby dcgillespie » Mon May 17, 2010 4:57 pm

Nice work Brik! The screen tap % you were not expecting is one of the most overlooked issues when using a generic transformer to replace an OEM one.

Virtually all transformer manufacturers today think that "UL" operation is defined as 40%, so their entire line is often offered with that tap point, with little understanding as to how the tap %, the plate-to-plate load, and indeed the tube itself are all intertwined. The 40% point got it's start from the original UL experiments that were done on 6L6 class tubes, that respond most to the UL influence at 43.0% of the winding. But Hafler (and Keroes) quickly learned that the optimum tap % was not only tube specific, but also loading specific as well, such that the optimum tap % is hardly 40% for all tubes or conditions. Early on, much of the information was derived empirically, as was certainly the case for Hafler's original transformer, the A-430. This transformer was specifically designed for EL34 tubes, under conditions of maximum power output with minimum distortion. With a 4.3K primary, it employs taps at 33%. But raise the load to 6K plate to plate or greater, and the optimum point becomes 40% for these tubes. The difference is not insignificant. EL84 and 6V6 class tubes work best with a 25% tap. Using a 40% tap produces very compromised performance with these tubes. Overall power output suffers, and particularly so at the HF end of the scale. 6550s generally do like 40%. Real KT88 are good with anything from about 20% to 50%, because this tube was designed specifically with UL operation in mind, and the list goes on.

I'm not saying you can't use the Tango transformers. But they will not produce the superb power bandwidth performance from 20-20 kHz that the original Dyna, or the new Dynaclone transformers will produce. Thanks for sharing the results of your tests. It certainly shows that the Dynaclone transformers are indeed every bit as good as the originals.

Dave
dcgillespie
KT88
 
Posts: 399
Joined: Sun Mar 08, 2009 5:34 am
Location: Ball Ground, GA

Postby Brik » Tue May 18, 2010 7:00 am

Hi Dave,

Thank you for the informative post.
...But they (the Tangos) will not produce the superb power bandwidth performance from 20-20 kHz that the original Dyna, or the new Dynaclone transformers will produce.

That's my current sentiment also. I might go ahead and build an ST-35 clone anyways, but with dual, switchable UL/Triode modes. The problem is that the EL84s may want to see 10K Ohm P-P or greater load impedance at 330V-350V plate voltage levels. The Tangos would not be optimal for UL since the Sg ratio is wrong, and they would not be optimal for triode operation since the impedance is on the low side.
Image
I will go back to the drawing board.

Hi Ty,
I used the following scheme to measure the inductance.
Image

Unlike capacitance, inductance is not a "constant" or "inherent" attribute of a coil, and the measured value depends on the current level that is used - the greater the current, the less the value of inductance.

I started out with a 22K Ohm resistor as a reference.
For 10V RMS input, the current through the set up is:

10V / (2 * 22K Ohm) = .22 mA

At these current levels, for the Tango FE-25-8, the frequency at which the inductor/resistor voltages balanced was about 25Hz.
But for the Dynaclone Z-565, I could not equalize the voltage levels of the inductor and the resistor even at 3Hz.
I had to use a resistor with a significantly higher value, 330K Ohms, to be able to equalize the voltages across the coil and the reference resistor so that the frequency of the AC signal was more manageable at 25Hz - 190Hz rather than the lowish 1Hz - 25Hz range.

The problem was the current through the set up became even smaller:

10V / (2 * 330K Ohm) = .015 mA

Due to the miniscule current levels in the measurement set up, the computed inductance would be on the high side.
Sure enough, the computed inductance for the Tango was 280H which is exactly what the Tango spec sheet lists as the maximum inductance of the primary winding.
By the way this measurement technique assumes that there is no significant load on the secondary side. The ~20K Ohm NF connection on the secondary has a negligible effect on the measurements, however.

Re: Z-565s having the Sg ratio of 22-24%, I did not realize that Gregg had an excellent post about it already:
http://diytube.com/phpBB2/viewtopic.php?t=3871
Geek wrote:Thanks to Ed, I was able to bring you this for your info :-)

Image

This little tranny was a heckuva lot better built than the couple A-470's I unwound... winding-to-winding consistency was within two turns of each other.

Different manufacturers or years or something?


Here, Gregg found out that there are 1600 turns in the B+ to Anode winding, and 400 turns in the Sg winding which comes out to 400 / 1600 = 25%.

Belated thanks to Gregg!
/b
User avatar
Brik
 
Posts: 193
Joined: Fri Apr 07, 2006 5:30 pm
Location: West of Boston

Postby cedricb » Fri Sep 24, 2010 1:18 am

dcgillespie wrote:Nice work Brik! The screen tap % you were not expecting is one of the most overlooked issues when using a generic transformer to replace an OEM one.

Virtually all transformer manufacturers today think that "UL" operation is defined as 40%, so their entire line is often offered with that tap point, with little understanding as to how the tap %, the plate-to-plate load, and indeed the tube itself are all intertwined. The 40% point got it's start from the original UL experiments that were done on 6L6 class tubes, that respond most to the UL influence at 43.0% of the winding. But Hafler (and Keroes) quickly learned that the optimum tap % was not only tube specific, but also loading specific as well, such that the optimum tap % is hardly 40% for all tubes or conditions. Early on, much of the information was derived empirically, as was certainly the case for Hafler's original transformer, the A-430. This transformer was specifically designed for EL34 tubes, under conditions of maximum power output with minimum distortion. With a 4.3K primary, it employs taps at 33%. But raise the load to 6K plate to plate or greater, and the optimum point becomes 40% for these tubes. The difference is not insignificant. EL84 and 6V6 class tubes work best with a 25% tap. Using a 40% tap produces very compromised performance with these tubes. Overall power output suffers, and particularly so at the HF end of the scale. 6550s generally do like 40%. Real KT88 are good with anything from about 20% to 50%, because this tube was designed specifically with UL operation in mind, and the list goes on.

I'm not saying you can't use the Tango transformers. But they will not produce the superb power bandwidth performance from 20-20 kHz that the original Dyna, or the new Dynaclone transformers will produce. Thanks for sharing the results of your tests. It certainly shows that the Dynaclone transformers are indeed every bit as good as the originals.

Dave


Sorry to resurrect this thread... I'm currently in discussion with Trafomatic to build some custom toroid output transformers (at last this time I'm going to order). Should I ask them to use a UL tap of 25% instead of the standard 43% ?
cedricb
 
Posts: 89
Joined: Mon Jan 28, 2008 4:49 am

Postby cedricb » Fri Sep 24, 2010 2:50 am

Reply from Sasa @ Trafomatic

For a pentode in UL connection, upping the winding percentage for the UL tap means (everything else unchanged) more triode like behavior (100% UL means just a triode-strapped pentode ),
--> less sensitivity,
--> less Po,
--> less THD,
--> higher input capacitance

So, in your case going from 43% to 25% essentially means:

--> power stage is easier to drive due to higher sensitivity and less CMiller,
--> higher Po and THD
--> higher Zout,

You can choice between 43 and 25 % in depend what you wanted of amp . Both are very good


This is obvouisly chineese to me... :/
cedricb
 
Posts: 89
Joined: Mon Jan 28, 2008 4:49 am

Postby EWBrown » Fri Sep 24, 2010 3:25 pm

What would be REALLY interesting, is if he could produce custom OPTs which had BOTH 43% and 25% UL taps on the primary windings. Then this would open a whole new world of UL possibilities.

/ed B
Real Radios Glow in the Dark
User avatar
EWBrown
Insulator & Iron Magnate
 
Posts: 6076
Joined: Wed Mar 19, 2003 6:03 am
Location: Now located in Clay County, NC !

Postby dcgillespie » Fri Sep 24, 2010 7:23 pm

I assume you are also building an ST-35 clone. If so, you will want to copy the original Dynaco specs, which include an 8000 ohm primary winding, with the screen taps placed at 25% of the winding. The difference in performance between 25% and 40% taps is notable with EL84 type tubes.

Dave
dcgillespie
KT88
 
Posts: 399
Joined: Sun Mar 08, 2009 5:34 am
Location: Ball Ground, GA

Postby apurcell22 » Mon Sep 27, 2010 9:31 am

Hey guys,
I just wanted to say hey, and tell you that I just got off the phone with Edcor and ordered a pair of their 8k 25w'ers with a screen tap @ 23%. It was not listed, but goes for the same price as their CXPP25-MS-8K. I don't know if they will be as good as the Dynaclones, but I saved ~$75 for 2 OPT's and a power TX. I like the Edcors in Pete Millet's Engineers amp, so I'm sure I wont be disappointed. No affiliation
Al
apurcell22
 
Posts: 57
Joined: Mon Apr 12, 2010 7:17 pm

Postby cedricb » Mon Sep 27, 2010 9:46 am

EWBrown wrote:What would be REALLY interesting, is if he could produce custom OPTs which had BOTH 43% and 25% UL taps on the primary windings. Then this would open a whole new world of UL possibilities.

/ed B


He's going to add both UL taps to my output transformers. I think, if you are from Europe you can deal directly with Trafomatic.
cedricb
 
Posts: 89
Joined: Mon Jan 28, 2008 4:49 am

Postby EWBrown » Mon Sep 27, 2010 10:00 am

I'm in the USA, but I've bought various kinds of Trafomatic iron from Boris Sasic before - he is located near "Queens" borough of NYC, and close to the "Little Russia" or "Little Odessa" section, near Brighton Beach.

If you've ever seen the second movie of "Men in Black", MIB II, the restaurtant scene , with "the killer bug waiter" is in that neighborhood of NYC.

I do have a set of the Trafomatic "ST35" trannies (as does Shannon) but I never measured to determine if the SG tap is at 43% or 25% of the B+ to Plate windings.

FWIW, try measuring the AC signal voltage across the primary winding, as well as across the 330K resistor, in your test setup.

There will be an approximately 90 degree phase shift in the primary winding, due to its inductance.

Just remember the mnemonic ELI the ICE man,

Voltage Leads Current in an inductor; and Current leads Voltage in a Capacitor.



/ed B in NC
Real Radios Glow in the Dark
User avatar
EWBrown
Insulator & Iron Magnate
 
Posts: 6076
Joined: Wed Mar 19, 2003 6:03 am
Location: Now located in Clay County, NC !

Postby 20to20 » Tue Aug 30, 2011 9:57 pm

dcgillespie wrote: Early on, much of the information was derived empirically, as was certainly the case for Hafler's original transformer, the A-430. This transformer was specifically designed for EL34 tubes, under conditions of maximum power output with minimum distortion. With a 4.3K primary, it employs taps at 33%. But raise the load to 6K plate to plate or greater, and the optimum point becomes 40% for these tubes. The difference is not insignificant. EL84 and 6V6 class tubes work best with a 25% tap. Using a 40% tap produces very compromised performance with these tubes. Overall power output suffers, and particularly so at the HF end of the scale. 6550s generally do like 40%. Real KT88 are good with anything from about 20% to 50%, because this tube was designed specifically with UL operation in mind, and the list goes on.

Dave



Dave, ...and All

I came across an article by Ed Laurent in the May '61 RE where he offers an amp circuit that was probably his forerunner to the ST-35. The amp used 7199's and 7189's with the A-410 tranny. It was interesting that he states after trying UL mode he decided it had no significant affect on sound quality so he instead used the screen taps for cathode feedback to the 7199's. He also used feedback from the outputs.

This raises a question about the A-410 screen taps point. Do you know if it was at 40% or 25%? If the A-410 was developed before it was understood that a 6BQ5/7189 responded better at 25% then that would explain why Laurent would try the UL mode and find it had no affect, if the A-410 had the tap at 40%.

On the other hand, if the A-410 did have the tap at 25%, then that might say something about the 7189 and the UL tap it prefers. Is the 7189 that much different from a 6BQ5?

My guess is the A-410 had 40% taps. I couldn't find any specs for it.

Perhaps that all led to the development of the Z-565 for the ST-35 with 25% taps.

Any thoughts? Thanks.

20
Last edited by 20to20 on Fri Sep 02, 2011 7:18 pm, edited 1 time in total.
Headed for Tishomingo to sing in a can...
User avatar
20to20
KT88
 
Posts: 527
Joined: Wed Apr 21, 2010 1:48 pm
Location: W-S, NC

Next

Return to stereo 35

Who is online

Users browsing this forum: No registered users and 0 guests