A dream came true - R2R.

[Atmasphere]

Thanks for the heads-up that tape EQ is available for/in your preamps - PLUS, if I'm not mistaken, your low-level inputs are all BALANCED (electronic not transformer) - as all magnetic inputs should be!

Yes- we are certainly on the same page with that!
One of the things that annoyed me about the Ampex tape head cables was their complexity. They went through great lengths to make them sound right- including Teflon insulation (rare in the 1950s), double shielded and double insulated. It would have been so much easier if they were simply balanced.

BTW I used Nortronics heads on my machines. I knew the head engineer at Nortronics, Joe Dundavic, who helped me set them up in the head nests. They performed quite a bit better than the original heads. For mastering I liked to use 1/2" tape at 15 ips.

 

[Foxbat]

BALANCED (electronic not transformer) - as all magnetic inputs should be!

Charles, many engineers will not agree with that.

 

[stellavox]

Charles, many engineers will not agree with that.

Why not?

 

[adrianywu]

Why not?

Charles, do you mean a fully balanced topology rather than transformer balance at the input and output with single ended amplification circuit in between ?

 

[Atmasphere]

Charles, many engineers will not agree with that.

And many will.

Transformers are not the only way to drive or receive a balanced line. Its been possible to do it without transformers for a long time. A floating output is preferred to drive the balanced line if you want to take maximum advantage of the CMRR at the input of the next stage in the equipment chain. That is why transformers are still common in the recording industry.

But obviously That Semiconductor would not be making specialized chips for balanced drive if that was the only way; generally speaking their technique represents a second method of a floating balanced line drive.

We patented a 3rd method which is floating and allows tubes (or transistors) to be direct coupled to drive a balanced line.

WRT internally balanced differential operation, this method reduces noise that obtains from stage to stage by about 6dB (ideally) with each stage of gain. It also reduces the sensitivity of the active circuits to noise in the power supplies and tends to induce less noise (due to being a constant load) in those supplies (which reduces intermodulations). Finally, even ordered harmonics are cancelled from stage to stage, not just at the output (in the case of an output transformer driven PP); this results in a lower distortion circuit overall (since distortion is compounded less from stage to stage) with reduced amounts of higher ordered harmonics due to the circuit's non-linearity being cubic in nature as opposed to the quadratic non-linearity expressed in a single-ended circuit.

 

[Foxbat]

Why not?

Because when it comes to connecting a floating transducer there is no simple answer. Some people like twisted pairs, others prefer double-shielded coaxes. There is no such ambiguity when connecting things such as preamps to power amps, there balanced connection is pretty much universally better. Things get even more interesting when you have limited hardware resources, say, a single tube, and you are building a circuit with no feedback. Then the single-ended connection might rule.

 

[Atmasphere]

Because when it comes to connecting a floating transducer there is no simple answer. Some people like twisted pairs, others prefer double-shielded coaxes. There is no such ambiguity when connecting things such as preamps to power amps, there balanced connection is pretty much universally better. Things get even more interesting when you have limited hardware resources, say, a single tube, and you are building a circuit with no feedback. Then the single-ended connection might rule.

FWIW the latter connection isn't likely to be used in a studio machine or one intended for pro use in the field.

@stellavox 's exuberance aside, all magnetic audio transducers (magnetic cartridge of any kind, tape head, dynamic microphone) are all balanced sources. Any engineer will agree with that.

 

[Foxbat]

Hmmm... I kinda thought Studer A820 was a studio machine...

 

[Atmasphere]

Hmmm... I kinda thought Studer A820 was a studio machine...

It is. Not all studio machines take advantage of the fact that tape heads are a balanced source. If you look though, you'll see they have to jump through some hoops to avoid buzz by treating the balanced source as single-ended.

However, that machine does not use a tube input with zero feedback to the best of my knowledge. Yes, I was being that specific. I tend to be pretty literal and sometimes that gets me in trouble...

 

[stellavox]

Charles, do you mean a fully balanced topology rather than transformer balance at the input and output with single ended amplification circuit in between ?

Hadn't thought that thru when I posted .

IMHO - have a fully balanced topology, like what Ralph offers (I believe). If not, AT LEAST accept a balanced INPUT. My K/C pre has an unbalanced input (as did all the Cello low-level preamps) and without using a third ground wire, from a headblock ground to the ground binding post on the pre (like are present with all tone arm setups), some customers experienced hum from ground loops. Don't know (never asked) if Nick's balanced input eliminated the ground loop problems.

Realize that all our external preamp customers had the heads wired out "by others".

AND MCI and XX (Whomever made the Frieda - senior moment here) had differential, balanced INPUTS in their playback circuits

 

[Foxbat]

Charles, the designers have the right to offer what they think is right. There is no dogma. And the internal architecture has nothing to do with interface in this case with a floating transducer. I already mentioned earlier that there are clear tradeoffs. In some cases balanced connection can be up to 6dB noisier, and if you are working with tubes every last dB can be valuable. So all my head and phono preamps are fully balanced internally, which is a given, but some are connected to the heads the single-ended way. I go by what is best in each particular case, not by what some dogma tells me.

When customers call me and ask that question - how to connect their phono cartridges, I tell them to try it both ways and form their opinion. I am not going to dictate them.

So basically I am suggesting opposite of what you are saying - keep the internal architecture balanced, but be flexible with the interface.

https://www.ni.com/docs/en-US/bundle/ni-daqmx/page/floatsig.html?srsltid=AfmBOoo-Lge5CF7fxhmtgqp8A7QvL-IJ8UI1P9UQMgO2Qo8KQ2KPfXIw

 

[adrianywu]

Charles, the designers have the right to offer what they think is right. There is no dogma. And the internal architecture has nothing to do with interface in this case with a floating transducer. I already mentioned earlier that there are clear tradeoffs. In some cases balanced connection can be up to 6dB noisier, and if you are working with tubes every last dB can be valuable. So all my head and phono preamps are fully balanced internally, which is a given, but some are connected to the heads the single-ended way. I go by what is best in each particular case, not by what some dogma tells me.

When customers call me and ask that question - how to connect their phono cartridges, I tell them to try it both ways and form their opinion. I am not going to dictate them.

So basically I am suggesting opposite of what you are saying - keep the internal architecture balanced, but be flexible with the interface.

https://www.ni.com/docs/en-US/bundle/ni-daqmx/page/floatsig.html?srsltid=AfmBOoo-Lge5CF7fxhmtgqp8A7QvL-IJ8UI1P9UQMgO2Qo8KQ2KPfXIw

This is a good discussion. My current tape head pre is balanced differential. When I tried wiring the tape head single ended, there was hum. I changed to balanced and the noise disappeared. I am only using a twisted pair of teflon sheathed solid silver wires from the head to the preamp. I am building the Marantz 7 tape head preamp as recommended by Victor. I am still trying to decide whether to connect the signal ground to the chassis ground, since the circuit is single ended and ground loop is always a risk. I guess I will try both ways and see.

 

[dminches]

This is a good discussion. My current tape head pre is balanced differential. When I tried wiring the tape head single ended, there was hum. I changed to balanced and the noise disappeared. I am only using a twisted pair of teflon sheathed solid silver wires from the head to the preamp. I am building the Marantz 7 tape head preamp as recommended by Victor. I am still trying to decide whether to connect the signal ground to the chassis ground, since the circuit is single ended and ground loop is always a risk. I guess I will try both ways and see.

Where did you source the Marantz 7? Did it need a lot of work before starting the conversion to being a tape preamp?

 

[Atmasphere]

Charles, the designers have the right to offer what they think is right. There is no dogma. And the internal architecture has nothing to do with interface in this case with a floating transducer. I already mentioned earlier that there are clear tradeoffs. In some cases balanced connection can be up to 6dB noisier, and if you are working with tubes every last dB can be valuable. So all my head and phono preamps are fully balanced internally, which is a given, but some are connected to the heads the single-ended way. I go by what is best in each particular case, not by what some dogma tells me.

When customers call me and ask that question - how to connect their phono cartridges, I tell them to try it both ways and form their opinion. I am not going to dictate them.

So basically I am suggesting opposite of what you are saying - keep the internal architecture balanced, but be flexible with the interface.

https://www.ni.com/docs/en-US/bundle/ni-daqmx/page/floatsig.html?srsltid=AfmBOoo-Lge5CF7fxhmtgqp8A7QvL-IJ8UI1P9UQMgO2Qo8KQ2KPfXIw

I'd like to see a case where a balanced connection is actually noisier! I've never heard of that before and no-one has ever mentioned that to me- this is the first I've heard of such a thing. Normally its the other way 'round. My surmise if such a thing actually occurs, its due to a connection or design flaw, one that should be quite obvious (such as ignoring the connections standards).

The advantage of balanced line, if the tenants of the balanced standards (AES48 and low impedance) are observed, are generally lower noise in the interconnect cable, immunity to artifacts of the cable itself and immunity to ground loops. If you've ever auditioned interconnect cables and found an audible difference, this is what I'm talking about.

If ever there is a place to get the interconnect cable right, its at the beginning of the audio chain; regardless of how good the equipment is downstream it will be unable to make up for a deficiency at the input. IME a distinct advantage of this practice is the interconnect cable (whether tonearm or tape head cable) need not be exotic or expensive; merely a twisted pair within a shield with low capacitance.

 

[Foxbat]

I can't speak for why anyone has or has not mentioned something, but that is standard textbook stuff. As I mentioned, it does not apply to every case, only to tube input stages with zero global feedback - the only type I care about.

In this case your input noise is pretty much just your tube noise, and with that in mind you probably reach for the quietest tube among the readily available ones - the 6922. Its noise resistance is the incredibly low 200 Ohm.

200 Ohm corresponds to noise voltage of .25uV in a 20,000KHz band. And this is what you get when you connect your signal source to one such tube.

In balanced configuration you have two noise sources connected in series, so your noise resistance is now 400 Ohm, with the resulting noise voltage of .35uF - a 3dB increase.

But this is not the end of the story, because the 6922 has two triodes, so you can connect the second one in parallel with the first, for the total noise resistance of just 100 Ohm. That is how you gain 6dB compared to balanced connection.

This is how I I have done it in some of my phono preamps, and the result is incredibly low noise input.

All this is totally separate from the effect of lines, but there are some cases there. One case is long lines, and they do pickup noise, but tube circuits really have no CMRR to speak of, so any hope of a balanced section significantly reducing the common mode noise is way too optimistic. The noise reduction benefit is usually obtained when precisely trimmed instrumentation amplifiers are on the receiving end... or a transformer.

However, this is not our case, as the interconnects between the tape heads and head preamp are usually short.

As result, in the case of my Marantz 7 projects the tape drive used there, the Teac 25-2, has RCA jacks connected to its head, and I use RCA interconnects to send the signal to the active stage. The interconnects are 3' long and there is absolutely no audible pickup in that configuration.

There is also the subject of ground loops... but if everything is done right there will be no ground loops with a floating source, unconnected to anything on its end. Yes, you will have some stray capacitance there, but that is not what causes ground loop.

So no, Charles, the balanced input will not eliminate ground loops, unless it is a transformer input.

 

[Atmasphere]

I can't speak for why anyone has or has not mentioned something, but that is standard textbook stuff. As I mentioned, it does not apply to every case, only to tube input stages with zero global feedback - the only type I care about.

In this case your input noise is pretty much just your tube noise, and with that in mind you probably reach for the quietest tube among the readily available ones - the 6922. Its noise resistance is the incredibly low 200 Ohm.

200 Ohm corresponds to noise voltage of .25uV in a 20,000KHz band. And this is what you get when you connect your signal source to one such tube.

In balanced configuration you have two noise sources connected in series, so your noise resistance is now 400 Ohm, with the resulting noise voltage of .35uF - a 3dB increase.

But this is not the end of the story, because the 6922 has two triodes, so you can connect the second one in parallel with the first, for the total noise resistance of just 100 Ohm. That is how you gain 6dB compared to balanced connection.

This is how I I have done it in some of my phono preamps, and the result is incredibly low noise input.

All this is totally separate from the effect of lines, but there are some cases there. One case is long lines, and they do pickup noise, but tube circuits really have no CMRR to speak of, so any hope of a balanced section significantly reducing the common mode noise is way too optimistic. The noise reduction benefit is usually obtained when precisely trimmed instrumentation amplifiers are on the receiving end... or a transformer.

However, this is not our case, as the interconnects between the tape heads and head preamp are usually short.

As result, in the case of my Marantz 7 projects the tape drive used there, the Teac 25-2, has RCA jacks connected to its head, and I use RCA interconnects to send the signal to the active stage. The interconnects are 3' long and there is absolutely no audible pickup in that configuration.

There is also the subject of ground loops... but if everything is done right there will be no ground loops with a floating source, unconnected to anything on its end. Yes, you will have some stray capacitance there, but that is not what causes ground loop.

So no, Charles, the balanced input will not eliminate ground loops, unless it is a transformer input.

If you are paralleling tube sections to reduce noise in a single-ended circuit, to compare to a differential input circuit the latter should have the same number of paralleled tube sections of the same type as the single-ended circuit. Once you eliminate those variables you'll find that the differential gain stage can be up to 6dB less noise on account of cancellation, depending largely on how effective the CCS circuit is.

Zero feedback can be common to both circuits under test.

If you use a good constant current source, just like with solid state, tubes benefit from that with regard to CMRR so that value can be decent- well into the 90dB region.

We routinely run interconnect cables 50 feet or more and have done so for decades. IME your comment about ''way too optimistic' doesn't hold water.

 

[Foxbat]

In a balanced input the noise sources are connected in series, not in parallel. Even if you made each leg of it with two triodes in parallel (haven't seen that ever done), you would still have a 3dB penalty. Balanced input always lose on noise.

Furthermore, if you go to the extreme of two triodes per leg, then my advice would be to connect them ALL FOUR in parallel for even less noise.

The constant current source you can make as ideal as you wish, your CMRR is still going to be lousy because tubes are not stable enough to even get close to the tight trimming required for high common mode rejection. 90dB CMRR requires astronomically tightly trimmed resistors and tubes that don't have their parameter exhibit ANY drift, stuff that doesn't exist.

 

[Atmasphere]

In a balanced input the noise sources are connected in series, not in parallel.

IME this statement appears to be a red herring. The idea that balanced inputs 'lose on noise' seems to be presented here for the first time.

Its generally agreed in most circles that balanced lines offer lower noise. We'll have to agree to disagree on your surmise as I don't see it worth arguing. I suggest you google this one. Wikipedia might be a good place to start.

FWIW we've been paralleling triodes in differential amplifiers for decades now. In a nutshell it works quite well.

About 30 years ago I was at our local dealer when the Mark Levinson rep was showing off their new ML29 preamp, which was their first balanced line entry in the marketplace. I asked the rep why the phono section was not also balanced and the answer was they were concerned about how the RIAA equalization would work out, which told me instantly they had not actually tried it. I'm having that feeling again during this conversation.

 

[Foxbat]

Well, like I said, it is all text book stuff, anyone can find it, I am surprised anyone would struggle with it. "Works quite well" is not a technical term.

BTW - I would like to see some proof of that totally improbable number of 90dB CMRR in a tube circuit. That number is several orders of magnitude higher, than is achievable, so some proof would be nice.

 

[adrianywu]

Where did you source the Marantz 7? Did it need a lot of work before starting the conversion to being a tape preamp?

I bought everything from Taobao, the Chinese equivalent of AliExpress. The seller of the board is also active on Ebay. https://www.ebay.com/itm/3532273047...uid=k1rzcijyqes&widget_ver=artemis&media=COPY
There are manufacturers of PCBs, chassis, transformers etc. on Taobao that can do custom orders. Only 2 to 3 days turnaround. Very convenient. I spent about US$150 for the whole thing.