Telefunken V72 Power Transformer

these modules are packed up tight and then packed even tighter into the console,

have you ever seen the rotisserie chickens down at the super market?

after a while the crust kind of turns into some type of weird mix between the spice rub and the Goodyear rubber company, you with me so far?

anyway i think you get the point, lams are processed with sulfer sometimes,
well if you keep lams in slow bake mode for too long, the sulfer starts to gas out and have you ever heard of sulfuric acid? long>short, you need to look at these transformers for that and also the real nightmare, transformer leads that are just hanging by a thread and coated with some funkenhauser cloth and varnish that hardens into a cracklin mess if you can dig where i'm comin from,

so even if the lams are ok, the leads are sure to be ready to snap, so we are gonna rebuild this thing,

there is a mu can around this pwr trans, there are mu cans everywhere, and the chassis itself is mu metal i believe.

ok, ref pick for puting humpty dumpty back together again,

these lead plates get real brittle after a while, one of the sec leads fell off when i was taking it apart, here is a better shot of the lead plares>

they use the same style snap in lams as they do in the plate choke, only since a power transformer  does not hande much dc, there is no need for a gap, so the tongue on this lam pretty much touches the opposing I bar. manny I-bar. place kicker for the miami dolphins, right?

the lams in the middle are cooked. nuthin but rust left. this is because the center of the transformer is like the center of the earth, real warm, so since we have enough problems with heat, we might need to prevent eddy currents from continuing there path into self destruction,

as luck would have it, the 625EI  is  almost a dead ringer (same width but different height),for replacing the bombed out stuff,and it fits nicely inside the mu box,,

and we can take control of the transformer design parameters by using different core materials and stack heights and bobbin sizes to dial in the flux and DCR and all the good stuff,

first wind to come off is the

heater winding, 1.5 ohms DCR, #25, 112 turns,

which means this thing will probably go

pri-sec-heater

as far as winding order.

notice the dislexic numbering system, hmmmm.......  8)

here is the heater start leading,

just another unwind shot,

we will be able to adjust down the turns on the sec, so the B+ will stay the same with major reduction in fwd drop from the new diodes, the old selenium  rectifier can really chomp down on the bit,

and a nylon bobbin will be much easier to wind than this old school jig saw puzzle made from phenolic and permali,
we do not care about the sound of insulation materials in the pwr transformer,

FYI, the best book in the business for insl is:

Handbook of Electrical and Electronic Insulating Materials
2nd ed.
W. T. Shugg

"Covering virtually all classes of insulating materials for electrical and electronic applications, this handbook offers immediate access to detailed information in one easy-to-use source. Included are major producers, technologies, methods of manufacture, trades, applicable standards and specifications, properties, uses, development programs, and market trends. Complete with a wealth of data and lacking in technical jargon, this book will be invaluable to electrical and electronics engineers who need to make informed choices about dielectric and electrical insulation materials as well as electrical engineering students in need of a comprehensive reference."


ok back to the sec, which is a real nightmare unwind, baked on tape that gets harder with the heat gun, and #43 wire, 

at turn 3367 on the sec unwind, took the sides off,

probably a few thousand to go, have to use the heat gun,

works as a vaporizer also,  8)

looks like 4600 Turns Sec.and 3200 Pri, which means either the sec voltage will be 318 peak instead of 280 like the label, or that if you run it at 200 VAC you get 280 out,

there will be a 10 to 15 percent loss on that 318 when you load it, so then it looks closer to the label.


got some more pics in a while,


.

back at it, a few pics and then we build the new one,

there is aCu  shield between the pri and sec,

got the label aunt mabel,

here is the print for the original V72 pwr trans,

forgot to check the inductance of the original transformer before the tear down,

however Zebra has my back on that over at the plate choke thread,

and, as long as you have the turns and the core, you can get the AL number of the core and then calculate the inductance, that is how we got the figures for the print posted up above,

what is AL? it is a constant for a core that can be used to calculate inductance for any number of turns,

how to do you extrapolate the AL from a core? depending on the size and frequency of the inductor or transformer, you use a turns quotient, i use 1000 turns as it is kind of a standard in a lot of core catalogs, the formula goes:

AL = Henries * 10^9 / N^2, where N is turns,

so we wind a 1000 turn coil with leads and place it onto the core and read inductance,

then we will do the simple math to get AL.
in fact, we can simplify the above equation to a really easy one now that we have the turns,

AL = Henries * 10^9 / 1,000,000.

and finally

AL = Henries * 1000

and it gets even easier for power transformers, we only have 2 freqs to hassle, 50 and 60 hz.

so now all we do is check inductance at max level possible from the ol HP,
as this is a pwr trans, so just check excitation current with a series ammeter and we are done,




here is the test coil for testing the lams from the original xfmr>

wow, we have some disturbing news from the lamination studies, 

these old Tele pwr xfmr lams only have 25% of the Henries as an equal core of 625 EI,

i get 3.6 Henries per 1000 turns at 50 Hz, and 3.8 Henries at 60 Hz. 26 volts applied,

and i get 30 and 32 Henries from the 625 EI !  so those lams might be another reason that these thngs get baked, all the eddy loss from the low perm lams and the sardine can construction.

so if AL = Henries * 1000 for a 1000 turn coil,

then original snap in lams have an AL of 3.6 * 1000 = 3600 @ 50 Hz and 3800 @ 60.

the 625 core has an AL value of Henries * 1000 = 32,000.

or it has about 8 times the perm,  ???

so with the stock turns and the stock lams of the power trans, we got the AL constant, now we can estimate the pri and sec inductance of the orig.xfmr:

we have to fix the formula up to make it easier to use,

AL = Henries * 10^9 / N^2,      so multiply out the N^2 in the denominator,

AL * N^2 = Henries * 10^9,      then divide out the 10^9 term,

Henries = (AL * N^2)  / 1 * 10^9

so if we run the original xfmr in Europe at 50 Hz, we use AL = 3600,

so pri Henries will be

3600 * 3200^2 / 1 * 10^9 = 37 Henries Primary Inductance @ 50 Hz,

and for the sec:

3600 * 4600^2 / 1 * 10^9 = 77 Henries Secondary Inductance

Heater can be done too, (3200) * ( 112^2) / 1 * 10*9 = 46 milli Henries.

using the same turns on the new 625 EI we get

328 Henries Pri and 677 Henries Sec Inductance, quite a difference,

this means we can probably use less turns = bigger wire = less DCR = less resistance heat,


excitation current due to the drop in pri inductance is 220 volts / 11.6 k = 19 ma,
excitation for a good core will be 220 volts / 102,992 = 0.2 ma,

if pri DCR is 344 ohms, we add an additional 19 ma * 344 = 6.536 watts! 

so less core loss heat and less DC R heat means this thing ought to run as cold as a brick of ice compared to the original,  are you with me so far? 

that old Tele pwr trans has some weird turns that don't match the label,

we better design this thing from the ground up to see what's shakin,

now if you multiply the cross sectional area of a  core by the total window area where the copper goes, you get what is called the area product. it turns out that this AP number relates to current density, temperature rise, regulation, the surface area, the volume,  weight and power handling ability.

by using the Area Product approach to xfmr design, you get a better match the first time out, where as if you design a xfmr with the RDH4 method, you get a very conservative design that can be wasteful as far as size, core material, copper wire and weight. so we will use the AP approach.

1. Calculate Power Output,

9 ma * 280 = 2.52 VA  sec and 6.3 * 0.5 = 3.15 VA heater,

P-o = 2.52 + 3.15,

P-o = 5.67 VA

2. Calculate the Electrical Conditions - K-e:

this is where all the work on Area Product went into, this neat formula here:

K-e = 0.145 * K^2 * f^2 * B-max^2 * 10^-4

K is stacking factor, usually 4 is used,
f = frequency
B-max is the maximum flux level you want the core to run at, in Teslas
so

K-e = 0.145 * 4^2 * 50 Hz * 1.6 T * 10^-4

K-e = 1.4848

3. Calculate the core weight

K-g = VA / K-e * a    a is your regulation constant in percent,

K-g = 5.67 / 1.4848 * 4  (lets use 4 percent regulation for a little sag factor)

K-g = 5.67 / 5.932 , so K-g = 0.9558

4. Select the Core from the Table of Core Constants

in this case, we select 625 EI, only weighs 411 grams for a sq stack but the next size up, 75 EI is way to big, and we already saw how much better the 625 EI is against the stock lam,

Core: Sq Stack - 625 EI

5. Calculate the number of primary turns using Faraday's Law

N = E-p * 10^4 /  (4 * A-c * B-max * f)

N = 220 VAC * 10^4 / (4 * 2.52 cm^2 * 1.6 T * 50 Hz

N-p = 2728

6. Calculate the Effective Window Area

W-a(eff) = W-a * S3,  (typical value for S3 is 0.75)  W-a for 625 EI is 1.89 cm^2,

W-a(eff) = 1.89 * 0.75 ,  W-a(eff) = 1.4175

7. Calculate thr Primary Winding Area

Primary Winding Area = W-a(eff) / 2  (since power in = power out, the area required by the primary will be the same as that required by the secondary,)

W-a = 1.4175 / 2

Primary Winding Area = 0.70875

8. Calculate the wire area A-w with Insulation, Use a fill factor of S2 = 0.6

A-w = (W-a(pri) * S2) / N

A-w = (0.70875) * (0.6) / N,  A-w = 0.42525 / N,  N =  2728 Turns, so

A-w =  0.42525 / 2728 , so

A-w = 0.0001558

9. Select the wire area A-w with insulation from the wire table

they give this spec as Cm^2 * 10^-3, so we will shift that A-w number 3 places to make it easier to look up,

so 0.0001558 > 0.1558, which is close to #37, but will go to the next largest gauge just to be safe, so

Wire Size = #34

10. Calculate the resistance of the winding using the Winding Table

R = MLT (mean turn length) * N * Resistance Constant * 10^-6

R-p = 8.84 cm * 2728 * 8572 * 10^-6

R-p = 206 Ohms

11. Calculate the primary copper loss P-cu

I-p = VA / E-p

I-p = 5.67 / 220,

I-p = 0.0257 Amps

P-cu = I-p^2 * R-p,  P-cu = 0.0257^2 * 206,  P-cu = 0.000661 * 206,

P-cu = 136 milli-watts

12. Calculate the Secondary Turns

N-s = (N-p * E-s) / E-p

N-s = (2728) * (280) / 220, N-s = 763840 / 220,

N-s = 3472 T

13. Calculate the secondary winding area A-w

uis with insulation using a fill factor of S2 = 0.6

A-w = W-a(sec) * S2 / N,  A-w = 0.70875 * 0.6 / 3472, A-w = 0.42525 / 3472,

A-w = 0.00012248

14. Select the secondary wire area A-w with insulation in the Winding Table for an equivalent AWG size,

bump 3 decimal places again to match up with the chart,

0.00012248 > 0.12248,

Sec Wire Size = AWG #38

almost there!

15. Calculate the resistance of the secondary winding using the Wire Table

R-s = MLT * N * Resistance Constant * 10^-6

R-s = 8.84 * 3472 T * 21,266,

R-s = 653 Ohms

16. Calculate the copper loss P-cu:

P-cu = I-s^2 * R-s,  P-cu = 0.009^2 * 653, P-cu = 0.000081 * 653,

P-cu = 53 milli-watts

17. Calculate the regulation:

a = P-p + P-s / (P-out +  P-p + P-s)                * 100

a = (0.136 + 0.053) / (2.52 + 0.136 + 0.053)        * 100

a = 0.189 / 2.709  * 100,    a = 0.0964 * 100,

a  7%

there ya go! 
so a ball park set of specs to get you going would be

625 ei  - sq stack
pri - 2,728 turns of #34  DCR = 206 Ohms
sec - 3,472 turns of #38  DCR = 653 Ohms

ratio is 3472/2728 = 1.27 : 1

original ratio is 4600 / 3200 = 1.44:1,

voltage ratio is 280/220 = 1.27,

we will spool this baby up and see what gives, heater winding will get center tapped, and we might have room for a 208 tap on the 220 pri.

we have quite a bit fewer turns than the original and the better core will surely support this change,

and it looks like the 625 is gonna fit into the can,>

> i get 3.6 Henries per 1000 turns at 50 Hz....26 volts applied

In a power transformer, we _know_ it will never see less than 230, 220, 204V.

And inductance generally rises with voltage. Up to a point. And we don't work power iron much past that point.

Ah, Hmmmm. Some generic data says from 26V to 230v, the inductance may barely double?

Which is still mighty low. Say 0.1A primary current un-loaded. 344 primary DCR. Around 3.5 Watts of heat when just doing nothing (a bit less due to my slop-math, but considerable more when juicing tubes too). Out in free air, not bad, but cramped up in small rooms it's warm in there.

These are Posts of the highest quality and they are an example of why this is such a great forum.
best
DaveP 8)

thanks for the props, i like this place too!

electronics has to be the best hobby out there,

indoors, no heavy lifting, 


on this xfmr rebuild we may try out a pri-heater-sec winding scheme,

we can bring out a CT on the heater winding, which, when grounded, will act as a shield,

this way we save installing the Cu shield, which takes up valuable winding area,

and we should have less hum from the tubes with the 6.3 CT grounded,

this means we can use bigger wire and save some some DCR heating too,



regulation is not spectacular with this xfmr as the power output is pretty light,

so winding DCR and current are a larger percentage of the total power delivered,

and there are no bleeder resistors in the V72 pwr supply,

we could shave a few percent of the regulation spec by adding a bleeder resistor?

or maybe Telefunken wanted low regulation as part of the sound, hard to tell,

and, i just found out that you can fry an egg in the microwave, 

let's roll one for the road, not a burrito, a transformer, 

did some prototype turns testing and we have it wired,

put the heater winding after the pri as a shield since we brought out a CT,

this saves having the mega wire on the outside, which saves squishing the sec all to heck,

and we get a quiet heater circuit as the original heater was were floating,

and we can add a few layers of Nomex for better pri-sec breakdown voltage,

coil ready for sec: