New Z for Input Transformer in Old Circuit

[Thee_Aldeen]

i need help finding the Zobel network to use an HA-100-x for the input on the redd47 line amp,

here is the schematic I'm using:  http://jbb.ru/schematics/47-shem.htm
UTC Bunker data (using H-100-X):  http://www.bunkerofdoom.com/xfm/UTCCHART/utc_DATA.pdf
output transformer is G.T.C. Ampex  output B-6300-1 — (i haven't found any data on it yet).

soon the better w/o being too pushy -- I'm doing the build today & the Zobel i know nothing about -- if anyone could teach me, that would be best. "give a man a meal he'll eat for a day -- teach I'm to fish & he'll never go hungry"

 

[jdbakker]

http://groupdiy.com/index.php?topic=1221.msg15714#msg15714 (EDIT: later in the same thread  http://groupdiy.com/index.php?topic=1221.msg15714#msg15714 is also relevant/helpful.)
http://groupdiy.com/index.php?topic=44421.msg556243#msg556243

Good luck,

JDB.

 

[abbey road d enfer]

i need help finding the Zobel network to use an HA-100-x for the input on the redd47 line amp,

here is the schematic I'm using:  http://jbb.ru/schematics/47-shem.htm
UTC Bunker data (using H-100-X):  http://www.bunkerofdoom.com/xfm/UTCCHART/utc_DATA.pdf
output transformer is G.T.C. Ampex  output B-6300-1 — (i haven't found any data on it yet).

soon the better w/o being too pushy -- I'm doing the build today & the Zobel i know nothing about -- if anyone could teach me, that would be best. "give a man a meal he'll eat for a day -- teach I'm to fish & he'll never go hungry"

Nothing in the documentation there is relevant to the prospective "calculation" of the Zobel component values.
For this calculation, one needs to know the nominal inductance at high frequency of the windings, the leakage inductance and the stray capacitance. Then, one needs to know the source impedance seen by the primary and the load impedance (both resistive and reactive) seen by the secondary. Finally, one needs to choose their preferred alignment (flat response, flat phase,...). Only when knowing all these parameters is it possible to either calculate or simulate the components.
In the absence of these informations, one has to experiment.
The reference given by JDB describes the most common methods.

 

[emrr]

....and no one ever built anything with a UTC and used a zobel.  Not needed, cart before the horse, etc etc. 

 

[Thee_Aldeen]

Awesome work guys, I want to add the option to switch between triode + pentode on the redd47... I think the buyer (whoever it might be) will find it a neat & special mod.  Here is what I'm working from, I need to know what I have to do to incorporate the mod into the design.

 

[Thee_Aldeen]

ok so using the ha-100 I don't need a Zobel at all? 
2) Why is that?
3) on the original schematic I posted "reed47" there is 680 piko cap + 100k resistor across the secondary -- this is the/a zobel network correct? Or no, this is something else?

*I appreciate you guys posting all the detailed information for finding future Zobels.  Knowing what unknowns I need is beyond helpful -- track Hugh you

 

[emrr]

Measure a UTC; does it look like it needs a Zobel? 

Triode/Pentode:  First, I gotta ask if you've used this preamp and have a problem with it.  Second, have you tried it in triode mode at all?  This sounds random and theoretical rather than practical without further descriptive.  If you go triode, it's no longer the same preamp, in which case there are easily all sorts of better preamps.  Thinking it will be neat and special is a long cry from actually being neat and special.  Not trying to harsh, just post some realism.  There are, uh, a few preamp designs to choose from over the last 85 years.  There's nothing particularly neat or special about the Redd47, it's just a tangentially famous preamp that was designed for a particular antiquated purpose.  It is what it is very much because of the exact original transformers, more than anything else about the circuit.  Build it with something else, it's something else. Change something, it's something else again. 

Build them all up and post your results and observations.  Someone has to do it. 

 

[abbey road d enfer]

3) on the original schematic I posted "reed47" there is 680 piko cap + 100k resistor across the secondary -- this is the/a zobel network correct? Or no, this is something else?

The original EMI schemo shows 100k + 100pF. That's the values the EMI engineers chose , based on the actual input transformer they used (a custom-built piece) and on their notion of what alignment would be the "average best", considering the various mics they had at the time.
The schemo you posted is for use with a Sowter 9970.
BTW, the original schemo has a fault, that has been copied everywhere, representing a junction between the R, C and an hypothetical center-tap on the secondary.

 

[CJ]

the pot idea is fine but ditch C2 and C3,
tie the top of the pot to B+ and the wiper to screen,

actually the V76 drops the screen of V1 by way of a 1 Meg resistor so i see no problem by having a 1M pot in there, interesting to see the effects on noise and gain,

100 X in the LA2a has a 68K resistor across the sec, so maybe fool around with dropping that 1 meg grid R to 100K and put some sq wave thru it both ways and compare,

 

[abbey road d enfer]

ok so using the ha-100 I don't need a Zobel at all? 
2) Why is that?

This comment has to be taken with a pinch of salt.
Remember that these transformers were designed about 50-60 years ago, when ultrasonic response was not as much a concern as it may be today. All transformers, including UTC's have a self-resonance. If it is enough above the audio spectrum, and not too sharp, one may consider leaving it alone. In fact, the typical circuits of the time, for which these transformers were designed, provided adequate loading of the secondary for an acceptable alignment. The typical 100-220k grid res and the typical 30-100pF of Millercapacitance would put the resonant frequency high enough and tame its amplitude adequately.
So emrr is correct in saying that a UTC does not "need" a Zobel in its typical applications. However, when measuring one with high input Z/low capacitance test equipment, the need for a Zobel may come forward.

 

[abbey road d enfer]

Awesome work guys, I want to add the option to switch between triode + pentode on the redd47... I think the buyer (whoever it might be) will find it a neat & special mod.  Here is what I'm working from, I need to know what I have to do to incorporate the mod into the design.

Interesting implementation. Indeed, the triode mode would not be 100% accurate but close enough and pentode mode would be perfect. However, there more to that. Pentode mode usually takes advantage of higher plate resistor for higher gain, so you would not experience its full potential. Anyway, that's the way the EMI engineers used their pentodes, so, who am I to say...
This reminds me of an experiment from a guy trying to eliminate pentode partition noise, by circulating the screen noise anti-phase; the circuit eventually operated like a triode.
Remember that the circuit is under NFB control, so the effects of triode/pentode may be mitigated.
What will not be mitigated is the reduded triode noise.

 

[emrr]

The major point of using a pentode is to allow use of heavy NFB while reaching target gain. 

 

[abbey road d enfer]

The major point of using a pentode is to allow use of heavy NFB while reaching target gain.

At the expense of noise, unfortunately.