Triad HS-29 Innerstage Transformer

this transformer is found in the Pultec EQP-1a,

already covered it but we can do some additional tests,

here is the pri and sec inductance,

pri saturates at 4 volts, so we run the test again at 1 volt,

sec input level can be increased to 2 volts as there are twice the turns which means half the flux,

Nickel lams raises inductance over DCR, so phase shift is very low,

here are some leakage capacitance specs,
sec-pri-pri-sec coil structure means that C will be balanced well on the pri,

From Pin 1 to 3  220 pf
1-4 126 pf
1-5 206 pf
1-6 206 pf
1-7 116 pf
1-8 45 pf

5-7 46.7 pf
5-8 20.6 pf
6-7 90 pf
6-8 48 pf

1-shield 64 pf
2-shield 62.7 pf
3-shield 62.5 pf
4-shield 57.5 pf
5-shield 50 pf
6-shield 64.9 pf
7-shield 56 pf
8-shield 38.2 pf

As usual... Thanks for all your data on transformers CJ!

JS

What d'où ou mean by sec-pri-pri-sec coil structure?

this is a single coil job,

they wind
3600 turns
which is one secondary, then
1800 turns
which is one pri, then
1800
turns which is another pri, then
3600 turns
which is another secondary,

this puts the pri right in the center of the coil, so capacitance will be about the same from the leads to ground,

Thanks for explanation

no problema, we have a blueprint on the way,

here is the leakage inductance,

primary

1-2 with both sec coils shorted - 8 mH
1-2 with only 7-8 shorted  - 24.7 mH
1-2 with only 5-6 shorted - 11.4 mH

3-4 with both sec coils shorted - 9.6 mH
3-4 with only 7-8 shorted - 13.3 mH
3-4 with pnly 5-6 shorted - 22.6 mH

1-4 (2-3 connected) both sec coils shorted - 25.8 mH
1-4 (2-3 connected) with only 5-6 shorted - 58.1 mH
1-4 (2-3 connected) with only 7-8 shorted - 66.5 mH


secondary

5-6 with both pri coils shorted - 57.3 mH
5-6 with only 1-2 shorted - 25.9 mH
5-6 with only 3-4 shorted - 12.1 mH

7-8 with both pri coils shorted - 67.5 mH
7-8 with only 1-2 shorted - 41.9 mH
7-8 with only 3-4 shorted - meter gives negative inductance-bogus read

5-8 (6-7 connected) with both pri coils shorted - meter gives negative inductance-bogus read

do not know why the meter got funny with certain combinations, sometimes the ohmeter does weird things with nickel lam transformers also,

CJ said:

Nickel lams raises inductance over DCR, so phase shift is very low,

Click to expand...

Inductance does raise over DCR, but you probably meant inductance vs resistance at the same freq, since it's what matters. Most of the resistive effect are due to the losses on the core once you get above certain frequency, also skin effect may be part of that in high frequency, but we don't have as much of that on audio freq, at our higher frequencies is most because of core losses (hysteresis) the equivalent resistance we see.

I don't know how much it is on Ni lams, but in the 2503 for example the DCR (resistance at DC) is only 6.7Ω, but at 10Hz already 112Ω and at 100Hz 637Ω, measured about 1V or so, I don't remember exactly at 100Hz, 1.3V at 10Hz. I would guess the effect would be much lower in Ni lams since the hysteresis is much tighter. The best part of measuring that 2503 was that I did it with Saul Walker himself  8 ;D

JS

here is the lamination for this transformer>

Damn good job  are you planning building it ?

build one?  no way, toughest coil to wind in the world, takes a pro about 4 hours to do one,

here is the Pultec schemo just for ref>

here is the print>

here is the coil, remains that is,

wire looks as thin as #48 but ohms out more like #43 if you use the mean length turn method of computing resistance, so we will call it #44,

here is the EQP-1R schematic,

it could be that this was the first innerstage used and they liked the way it ran out of phase, so they ran the Triad like the Peerless,

Pultec EQ's were used on phone lines, not for recording, so they did not really care about phase,

you can elliminate the innerstage transformer if you use the MEQ 5 circuit which sounds just fine, a resistor goes from the filter section hot to the grid of a single ended makeup amp,

Thanks for your hard work.

lets roll out the freq chart,

bottom end pancaked,

high end gets wild,

graph has Zobel and not>

we still need low end Henries vs Level,

looks great, u-i (initial perm) is off the map,

this means low level signals do not disappear into the core,

excellent info. you rule.