..is SSM2142 on its way out?

[gyraf]

Hi Group:

Question: is SSM2142 (balanced output driver) on its way out? I find it more and more difficult to source them, even in small quantities. Out of stock for longer and longer times, even at places like digikey and reichelt..

None of my usual suppliers can tell me what is going on - do any of you have additional insights?

I'm aware that it can be replaced by DRV134 and THAT1646, but I dislike change unless really really needed...

/Jakob E.

 

[john12ax7]

Haven't seen an official date but does seem on its way out as its not recommended for new designs.

https://www.analog.com/en/products/ssm2142.html#product-overview

https://ez.analog.com/audio/f/q-a/165838/replacements-for-ssm2142-and-ssm2143-parts

Best case scenario is you will get a chance to do a final buy.

 

[Bo Deadly]

Isn't DVR134 / DRV135 basically an interchangeable drop-in replacement for SSM2142? Just from glancing over the datasheets it sure looks like they are.

 

[analag]

Are they comparable to a cheap transformer.

 

[user 37518]

Are they comparable to a cheap transformer.

Much better IMHO, however, it depends on what "better" means to you.

 

[Bo Deadly]

I would argue that a line driver IC is probably superior to an output transformer. A line driver IC with current sensing output like THAT1646 is a pretty neat device. It's got good drive mostly because it's differential whereas with a transformer you're probably driving the primary single-ended and so you have to double the current and step-up to get the same output drive. Although with an output transformer you don't have to think about DC.

Input transformers OTOH are a LOT more useful. They don't have to handle a lot of power so they can be smaller, shielded, more sophisticated, etc. They can eliminate a lot of parts like blocking caps and protection diodes. Although the THAT1200 series work very well and are in fact "InGenius".

If I were designing a new device, I would use THAT1206/THAT1646 line in/out but use a line input transformer for a "premium component" and I would almost definitely use a mic transformer wherever possible unless it's for a mass-produced cost-conscious device.

 

[user 37518]

Although with an output transformer you don't have to think about DC.

Not entirely true, some transformers are not happy with DC, for example Carnhill sells gapped and un-gapped transformers depending on whether DC is present or not, a gapped transformer usually has worse freq. response than an un-gapped one.

To be honest, I am perfectly happy with what some people call "quasi" balanced, which means using an output single ended and adding a resistor on the cold side the same size as the isolating resistor used for the opamp side. Less circuitry ....

 

[Bo Deadly]

I mean you don't have t worry about DC on the output. With a line driver IC, there can be a small offset that will put DC on the line. And any DC offset on the input of the line driver makes this doubly worse. So it's rather important that you AC couple the line driver IC.

What you're referring to is called "impedance balanced" which is actually different from "quasi-floating".  Impedance balanced outputs only transmit signal on one wire so common mode noise induced in the line will not be cancelled as effectively as it would in a properly balanced line. But if the receiver is a properly debalancing circuit, an impedance balanced output still does cancel some common mode noise and it is very effective at limiting ground loop noise. For a variety of reasons, there should always be 50-100 ohms in series with line outputs. So adding another identical resistor to the other leg to ground is a very simple and highly effective way to cancel ground noise. For a device with many outputs for short runs, impedance balanced outputs would be a great choice.

 

[moamps]

.....Impedance balanced outputs only transmit signal on one wire so common mode noise induced in the line will not be cancelled as effectively as it would in a properly balanced line.....

Could you please point me to the source where did you found that info?
Aren't "properly balanced" lines also "impedance balanced" in the same time?

 

[Bo Deadly]

I said it would not be as effective. I didn't say it would not be cancelled at all. In fact the deficiency is minimal in that I think it would be limited to things like crosstalk from adjacent conductors because the signals currents won't cancel each other.

 

[moamps]

.... In fact the deficiency is minimal in that I think it would be limited to things like crosstalk from adjacent conductors because the signals currents won't cancel each other.

You lost me here.  How signal currents cancel each other in adjacent conductors? Which adjacent conductors? We are analyzing  here a single balanced line.

 

[user 37518]

What you're referring to is called "impedance balanced" which is actually different from "quasi-floating".  Impedance balanced outputs only transmit signal on one wire so common mode noise induced in the line will not be cancelled as effectively as it would in a properly balanced line. But if the receiver is a properly debalancing circuit, an impedance balanced output still does cancel some common mode noise and it is very effective at limiting ground loop noise. For a variety of reasons, there should always be 50-100 ohms in series with line outputs. So adding another identical resistor to the other leg to ground is a very simple and highly effective way to cancel ground noise. For a device with many outputs for short runs, impedance balanced outputs would be a great choice.

Read my post again, I never said "quasi floating" I said "quasi balanced" which according to wikipedia is the same as "impedance balanced", no need for the lecture.

 

[Bo Deadly]

If you have multiple balanced line pairs running next to each other (adjacent) in the same cable assembly, you could, in theory, get crosstalk between channels. With a proper balanced signal, the magnetic fields created by currents in the hot and cold conductors would exactly cancel each other out. But if the cold conductor does not carry the inverted signal, the magnetic fields will not cancel out and the unbalanced current in the hot conductor might induce currents in an adjacent channel.

But it would probably have to be like +20dBu into 600 ohms over a long line with no jacket around pairs just to measure crosstalk in an analzyer. So it's arguably an almost completely inconsequential issue.

 

[moamps]

If you have multiple balanced line pairs running next to each other (adjacent) in the same cable assembly, you could, in theory, get crosstalk between channels. With a proper balanced signal, the magnetic fields created by currents in the hot and cold conductors would exactly cancel each other out. But if the cold conductor does not carry the inverted signal, the magnetic fields will not cancel out and the unbalanced current in the hot conductor might induce currents in an adjacent channel.

I have no problem with that. That's old news. But you said:

Impedance balanced outputs only transmit signal on one wire so common mode noise induced in the line will not be cancelled as effectively as it would in a properly balanced line.

And I asked for a reference. Nevermind, let's move on.