Jensen input transformers

[pooge]

Anyone ever use a Jensen input transformer in a preamp? If so, where should it be located? The Jensen site is unclear about source impedance requirements. Assuming source impedance should be small, it would seem that the transformer should be placed before the volume control. This way, the volume control could always provide a load impedance greater than the required 10K ohms. However, the load would vary with volume selection.

If, on the other hand, the transformer is placed between the volume control and the amp, the idea load of 10k could be had, and it wouldn't change. However, the source impedance would be higher and variable. My hunch is that the first layout would be preferable, but not sure of the tradeoffs. Anyone know? I'm not a big enough customer for Jensen to answer my question.

[Steve Eddy]

Originally posted by pooge
Anyone ever use a Jensen input transformer in a preamp? If so, where should it be located? The Jensen site is unclear about source impedance requirements. Assuming source impedance should be small, it would seem that the transformer should be placed before the volume control. This way, the volume control could always provide a load impedance greater than the required 10K ohms. However, the load would vary with volume selection.

That depends on the type of volume control you're using. Unless you're using an attenuator which uses a fixed series resistance and a switched shunt resistance, the load will stay the same unless the input impedance to the preamp itself is unusually low.

And you wouldn't want the pot to provide a load greater than 10k, at least not for the 11P-1 and 11P-1HPC if you want to avoid using the RC damping network. These two are both ideally loaded with a 10k resistive load so using either of these with a 10k pot or attenuator would be ideal.

You might also want to consider the 11P4-1. It's ideally resistively loaded at 20k and also provides a wee bit of voltage gain compared to the 11P-1's 3dB loss.

If, on the other hand, the transformer is placed between the volume control and the amp, the idea load of 10k could be had, and it wouldn't change. However, the source impedance would be higher and variable. My hunch is that the first layout would be preferable, but not sure of the tradeoffs. Anyone know?

Again, the ideal load can be had using a 10k pot or attenuator. As long as it's not of the fixed series/adjustable shunt type, it will present a 10k load to the transformer.

I'm not a big enough customer for Jensen to answer my question.

Why do you say this? Have you asked them already and they haven't answered?

se

[pooge]

I am using a fixed series/switched shunt. That is why I'm having questions. I anticipated having to use the RC damper if the transformer is placed before the volume control.

The control will always be above 10K load.

As I said, I can get an ideal load if placed after the lineamp, but I'm unsure of the variable source impedance from the volume control. Can't find any info on source impedance issues.

I do not need any more gain with horn loudspeakers, but the 20K load of the 11P4 looks interesting.

I wrote them over a week ago, and got no reply.

[Steve Eddy]

Originally posted by pooge
I am using a fixed series/switched shunt. That is why I'm having questions. I anticipated having to use the RC damper if the transformer is placed before the volume control.

The control will always be above 10K load.

Ah, ok.

As I said, I can get an ideal load if placed after the lineamp, but I'm unsure of the variable source impedance from the volume control. Can't find any info on source impedance issues.

It's not so much the variability of the source impedance as it is how high your source impedance is. What's the value of the series resistor in your volume control?

Basically, increasing source impedance will result in increased distortion (primarily at the lowest frequencies), reduced low frequency bandwidth and higher losses.

I think you'd be better off with the transformer feeding your volume control. That's really where you'd want it anyway. If you put it after your volume control, then you lose the advantage of ground isolation.

I do not need any more gain with horn loudspeakers, but the 20K load of the 11P4 looks interesting.

Yes. I mentioned the 11P4-1 not so much for its voltage gain (which is just 1.5dB) but rather for its reduced voltage loss. With the volume control you're using now, you've already got a fair amount of loss built-in. No need to add to it if you don't have to.

I wrote them over a week ago, and got no reply.

Strange. In my experience (and I'm hardly their biggest customer) they've always been very responsive. And not just to me but to those I've sent their way who weren't customers at all.

If you'd like, EMail me the name and EMail address you used to contact them and I'll look into it for you.

se

[pooge]

I looked at the 11P4-1. The specs aren't as good. Even with a 20K load, the input impedance of the transformer is not as high as with 11P1 with a 10K load. Also, the CMMR specs aren't as good.

I deduced from the specs that a higher source impedance degrades CMRR. Thus, it would appear to be better to place xformer before volume control to benefit from the low output impedance of a source, and use the RC damper.

I made a stepped volume control designed with a 12K series impedance (before contemplating a xformer). This can be either single-ended or balanced. I will probably use two 6K series resistors on either side of the switched resistor to feed a balanced lineamp. (However, I read somewhere that the secondary of the transformer is not balanced, so I don't know if this would be that useful. Perhaps that is why all the application notes show the secondary grounded at one end.)

BTW, what is your take on the transparency of these transformers?

[fdegrove]

Hi,

(However, I read somewhere that the secondary of the transformer is not balanced, so I don't know if this would be that useful. Perhaps that is why all the application notes show the secondary grounded at one end.)

They're not unbalanced unless the secundary is grounded on one end...in that scenario the xformer serves a balun to the next stage which is often ahem...SE.

If you want to take further advantage of the balanced properties of the xformer you'll need a balanced stage to feed into.

Cheers,

[Steve Eddy]

Originally posted by pooge
I looked at the 11P4-1. The specs aren't as good. Even with a 20K load, the input impedance of the transformer is not as high as with 11P1 with a 10K load. Also, the CMMR specs aren't as good.

No, the absolute specs aren't as good. But then they're just specs and won't necessarily tell you how it will sound.

I deduced from the specs that a higher source impedance degrades CMRR. Thus, it would appear to be better to place xformer before volume control to benefit from the low output impedance of a source, and use the RC damper.

Yes. Though perhaps more importantly, you lose the galvanic isolation between components that transformers can afford.

I made a stepped volume control designed with a 12K series impedance (before contemplating a xformer). This can be either single-ended or balanced. I will probably use two 6K series resistors on either side of the switched resistor to feed a balanced lineamp. (However, I read somewhere that the secondary of the transformer is not balanced, so I don't know if this would be that useful. Perhaps that is why all the application notes show the secondary grounded at one end.)

The secondaries aren't balanced so if you want to feed a balanced input with them, you'll need to create a pseudo center tap using a pair of resistors for the load. A pair of 5k resistors for the 11P-1. Then use the node between the two resistors as the ground reference for the shield/case as well as the amplifier's reference ground.

BTW, what is your take on the transparency of these transformers?

Let's just say that I've been using them for nearly 20 years and have been very happy with them.

se

[Steve Eddy]

Originally posted by fdegrove
There not unbalanced unless the secundary is grounded on one end...in that scenario the xformer serves a balun to the next stage which is often ahem...SE.

The secondaries of input transformers such as the 11P-1 are not balanced. That's due to the winding techniques used to get the level of performance out of them that they offer, primarily keeping the winding capacitance as low as possible in order to achieve 100kHz bandwidths.

If you want to take further advantage of the balanced properties of the xformer you'll need a balanced stage to feed into.

Yes. But the neat thing about a good quality input transformer is that they can give you outstanding common-mode rejection even if they're fed from a wholly unbalanced source. The 11P-1 for example will give you about 100dB of common-mode rejection when fed from an unbalanced source. Which is far greater than most electronically balanced inputs will give you even when they're fed from a balanced source.

For an electronically balanced input to get the same performance from a balanced source that a transformer can give you from an unbalanced source, you pretty much have to match the source impedances to a rather high degree of precision.

se

[fdegrove]

Hi,

The secondaries of input transformers such as the 11P-1 are not balanced. That's due to the winding techniques used to get the level of performance out of them that they offer, primarily keeping the winding capacitance as low as possible in order to achieve 100kHz bandwidths.

Sounds to me like they're limited use fancy gadgets then...

Maybe Jensen should broaden their range and include some real xformers?

Ok, they have their uses...nothing there that appeals to a difficult customer like me however.

No offense to you,

Cheers,

[pooge]

No, the absolute specs aren't as good. But then they're just specs and won't necessarily tell you how it will sound.

Are you saying the 11P4s sound better?

Yes. Though perhaps more importantly, you lose the galvanic isolation between components that transformers can afford.

Not sure I understand this comment. How does source impedance affect galvanic isolation? Since this is done by the transformer, what does the source impedance matter?

The secondaries aren't balanced so if you want to feed a balanced input with them, you'll need to create a pseudo center tap using a pair of resistors for the load. A pair of 5k resistors for the 11P-1. Then use the node between the two resistors as the ground reference for the shield/case as well as the amplifier's reference ground.

Although I can configure my lineamp and/or volume control to be balanced or unbalanced, I'm wondering if it's worth the bother at that point. I'm primarily looking to break ground loops (damn those horns are sensitive suckers!), and I doubt there's much to worry about in the short distance between the volume control and lineamp.

Yes. But the neat thing about a good quality input transformer is that they can give you outstanding common-mode rejection even if they're fed from a wholly unbalanced source. The 11P-1 for example will give you about 100dB of common-mode rejection when fed from an unbalanced source. Which is far greater than most electronically balanced inputs will give you even when they're fed from a balanced source.

Exactly! And the fact that you can connect a balanced or unbalanced input without any further fuss. And Jensen shows an easy way of impedance balancing the output of an unbalanced source in Figure 4 of AN-003. Although I am confused by the different recommendations for the ground connection of the case in AS002 and AS089. In one, it is connected to the output signal ground. In the other, it is connected to chassis ground. All is made more confusing in that the case is conductive, and there is no recommendation other than the warning that a ground loop may be created with the case lead if the case is not insulated from ground. The two schematics appear to be for somewhat different applications, though. One appears to be an outboard balanced to unbalanced converter, while the other appears to be inboard.

For inboard, like I will be using, I guess you just leave the white wire disconnected if the case is conductively mounted to the chassis, such as with the new case style that can more easily mount the transformer in a hole in the back of an amp case, with the wiring easily feeding through the mounting hole to the connectors. (This conductive mounting doesn't square with AS089, but all in all, probably doesn't matter.)