Interstage Transformer Loading

[Steven-H]

Gents,
How does one 'load' an interstage transformer ? When I had ordered my IT's from Electraprint - I was told that 'A grid is not a load'...Yet almost every schematic I see; the one lead of the IT's secondary goes to the bias rail, and the other lead goes directly to the grid. Am I missing something here ? On the primary side, one lead would go from the B+, and the second direct to a 45's Anode.

Thanks !
Steve

[Sheldon]

The grid is not an effective load because the impedance for that node is very high, as long as you have no grid current.

Sheldon

[Miles Prower]

Originally posted by Steven-H
How does one 'load' an interstage transformer ?

Connect a resistor across the secondary.

When I had ordered my IT's from Electraprint - I was told that 'A grid is not a load'...Yet almost every schematic I see; the one lead of the IT's secondary goes to the bias rail, and the other lead goes directly to the grid. Am I missing something here ? On the primary side, one lead would go from the B+, and the second direct to a 45's Anode.

Some ISTs might work better with some sort of permanent load (~100K across the secondary, or between the ends and center tap if it has a balanced secondary) but almost all of them will work without one. This pretty much comes down to a "give it a try" scenario.

[Shoog]

What will happen with an unloaded secondary is the capacitances of the transformer and the miller capacitance will combine with the inductances to produce a resonant tank. This will usually be somewhere above 25Khz. This is also likely to induce ringing which in turn may produce very noticeable intermodulation distortion. End result will be a zingy bright sound which fatigues very quickly.

The secondary loading damps this down. On the Lundahl web site it gives recommended values for each of their transformers. I believe they recommend zobels on the secondaries for best effect. Unfortunately without this information you will need to do trial and error or tune with a scope. You may be lucky and find that the grid is an adequate load.

Shoog

[Steven-H]

Gents,
Thanks a bunch for all of your help - while I understand in theory what I am now trying to achieve; In practice I am still a little foggy.
I imagine that I should put a square wave through the primary of the IT; and see how various loadings on the secondary look from a ringing perspective. For the Zobel, I am trying to match the 5K of the secondary - right ?
I have looked through the Lundahl transformer website; and all of the schematics I've seen do not show any loading. Some examples I looked at were the PX25 amp, Andrea Ciuffoli's 845 (which closely resembles what I am trying to achieve), and also Andrea's SV572-3 amp. All of these are fixed bias without anything in the circuit to load the IT.

[Gluca]

Remeber to put a resistor in series to your signal generator to simulate the impedance of the driver (your tube's Rp). Things change and resonances get larger with a low driving impedance.

You can also look at the whole freq responce of the IT.

All that said, you might or might not need to load the secondary with a resistor: only a test can tell.

ciao
Gianluca

[analog_sa]

Originally posted by Steven-H
All of these are fixed bias without anything in the circuit to load the IT.

I can only offer an anecdotal observation. My pp amp using a Lundahl ITs used to produce a really ugly square wave unloaded.

Any improvement in the shape of the square wave brought by loading the secondary turned out to be detrimentary to the sound. A snubber worked better than a resistor but a lot of the magic still disappered. It turns out this is pretty much the consensus on Lundahl ITs. It is probably better to find an IT which can work well unloaded.

[Steven-H]

Guys,
Thanks again. My ITs came from Electraprint....I guess this is sort of an area that would benefit from some research - once I get down to testing I will be sure to post some results and findings.

Thanks again,
STeve

[Gluca]

Ugly? I used LL1660PP and I found nice and sharp square waves and a peak resonance (+6dB I believe) at 40kHz or somewhere around there. OK, there were small ringings at the edges, but that is NOT a problem.

Ciao
Gianluca

[Sheldon]

This site has some practical information on how to get values for the Zobel with a little testing.

http://www.siteswithstyle.com/VoltSe..._in_xfmrs.html

Sheldon