Oscillation when used with Sowter output trafo

[owel]

I have a DIY preamp and I checked that it's output doesn't oscillate (checked via oscilloscope) even at max gain (no trafo in output).

Even If I added an old MCI 600:600 1:1 trafo, the output is also fine. No oscillation.

But when I use a Sowter 600:600 (Model 8751), it oscillates like crazy even before reaching the max gain... i.e. about 3Vrms

Any ideas why it would oscillate when used with a Sowter trafo but not with the other one? Would a zobel network help fix the oscillation problem?

 

[peter purpose]

Stick 1K5 and .01uF across the secondary and terminate with 600R and see what happens.

peter

 

[owel]

Thanks Peter, you beat me to it!

I just fixed the oscillation by terminating the secondary with a 604 ohm resistor.

 

[peter purpose]

Marvellous.. :thumb:

 

[owel]

Problem!

Okay... oscilloscope output looks good, no oscillation... but to my ears it sounds different when there's a 600ohm load on secondary.

Basically, at very low level volumes, it can't reproduce the high frequencies at the same quality level as without the resistor. At mid and higher levels, it's fine, but at fine low level work, I can notice a difference.

Since the resistor load only helps stop oscillation at max output (with RLoad, it can now go to 10.97Vrms output with no oscillation), and this preamp won't be running at max output everytime, do you think a front panel switch that switches on the 600ohm will be a good idea or not?

I think there's one commercial preamp that swithces their load resistor on or off.... is that right?

Hmmmm.... Maybe I should try a different output transformer. A vintage MCI transformer works better than this brand new Sowter. .. even if both are 600:600.

 

[mcs]

How about a different load resistor? A lot of transformer datasheets specify a recommended load network - the Jensens often do at least.

Perhaps loading the transformer with 6000 Ohms instead of 600 Ohms would work?

Best regards,

Mikkel C. Simonsen

 

[peter purpose]

[quote author="owel"]
I think there's one commercial preamp that swithces their load resistor on or off.... is that right?

[/quote]

Yep... Dan Kennedy's Great River MP2-NV

 

[Tekay]

If the unit after your pre has higher imp than 600 use a 604ohm for termination. Without the right load at the output you will have a HF boost starting at around 10kHz and up to around 1,5dB at 20kHz.
You'll probably also have a small boost from 10Hz -50Hz around 0,25-0,5dB with out the right load.

Btw, the 8751 is a gapped transformer maybe thats why you're having the problems?

 

[rafafredd]

There is an ungapped version of the same transformer. You should not be using the gapped neve output version, I think.

The inductance goes really low in a gapped core.

Personally, I would try any other 1:1 low impedance output transformer.

 

[Tekay]

I did a test with gapped LO1166 as output transformer in a 1081 pre, the only differency I could measure was a LF boost, like 2dB at 10Hz, down at zero again around 30-40Hz, no distorsion diffrency that I could measure!

Maybe CJ as something to add here?

 

[owel]

Okay, what is a "gapped" transformer? And how does it affect performance/freq response? It doesn't say anything about "gapped" in the 8751 spec sheet.

>Personally, I would try any other 1:1 low impedance output transformer.

Yes, I know my other MCI transformer doesn't exhibit this problem, only with the Sowters.

I should try some Edcors 600:600.

 

[peter purpose]

From the Sowter site...

"8751 Neve 1073 Line Output
Ratio 1:1.5. Sowter replacement design for Neve P/N L01166. Optimised for voltage drive circuit."

The Neve 1073 output amp has +24V running through the primary of the transformer and 'the gap' is there to deal with it. Quite how, I do not know, but I'm sure someone who knows what they're talking about will be along shortly.

If you are running a Neve 283 type circuit, then I believe that you're going to need a gapped XFMR.

peter

 

[rafafredd]

He is not running a 283 output circuit.

I think he uses a cap before the transformer.

A gap is an isolation between the lams sections (generally EI). Like this:

E gap I

Plastic films used here??? I don´t know...

As I understand it, it kind of break part of the magnetic continuity of the core, thus permiting you to feed a DC component throught the windings without messing with the magnectic orientation of the core. I don´t know how to tecnically describe it better, and it in fact is my intuitive perception of what goes within a gapped transformer. I´m probably totally wrong. Let´s wait for some of the known guys to explainit in technical and correct manners. Maybe someone can even poit us to some info about gapped audio transformers. never read anything about it.

Anyway, let me go on with my divagations. As it weakens the magnectic continuity, the gap also brings inductance down. So, you need more current at the output to drive gapped transformers, or you need more core and more wire on gapped cores for the same inductance and same maximum recommended driving impedance.

If you are feeding the audio transformer throught a cap, and the current that feeds the output stages doesn´t run throught the primary of the transformer, definitly, you don´t want a gapped core.

Try the 600:600 EDCORs. I bet it will work nice in there.

 

[owel]

Yeah, rafa... I think I'm going to ditch the Sowters and just get some Edcor to try. I'm thinking of using the WSM600:600. Have any of you guys tried the Edcors?

The oscillation problem with my circuit only happens with the 8751. It works fine without a transformer, or with a different 600:600 transformer.

Maybe the Sowters are for use only with a real Neve circuit.

BTW, the reason why I want a transformer in the output is to convert an unbalanced out to balanced.

Thanks to all....

 

[PRR]

> I have a DIY preamp

That does not tell us much. I have some thoughts but they depend on what this "DIY preamp" really is.

As a dumb answer: you often want a hundred ohms or so between any op-amp and a transformer. Transformers are not easy loads.

__________
now playing

 

[CJ]

Try filpping the pri. polarity just for the heck of it.
Also, you could play around with e-shield or mu can grounding.
What is the circuitt?
Gapping increases the magnetic path and lowers the perm of the core, but enables it to handle the dc.

 

[CJ]

probably won't help, but try grounding/ungrounding the case/core/electrosatatic shields if you have the two wires coming out for doing so.
try swapping the primary leads just for the heck of it. maybe there is a phase issue.
the gap just breaks the electrical continiuity of the core so it does not become one large magnet. this cuts down on permeabillity, which means more turns are added for the same core material. it also makes it harder to saturate the core.