Find me a tube amp input transformer that is 600 Ohms from DC to 20kHz.
And no, I do not want to spend more for that transformer than for the rest of the amplifier.
DCR, Low frequency inductive reactance, Distributed capacitance at high frequencies.
All 600 Ohms, Huh?
For all we know, the NE5532 output is directly connected to the RCA output jack.
Or, at best, it has a series resistor.
If there is any DC offset of the NE5532, the current may saturate the otherwise pristine input transformer.
A cheaply engineered NE5532 circuit probably does have DC offset.
And, who says every NE5532 circuit drives 3V, 5mA, and 600 Ohms at low distortion?
A high gain input tube with 60pF of Miller Effect capacitance, and a 1:16 transformer will reflect 256 times the Miller Effect capacitance back to the NE5532 (518 Ohms capacitive reactance).
(Of course, we would never specify that large of a step up ratio).
This is no more complex, than any other stereo system.
But it is a System.
Some systems are well designed from beginning to end.
Other systems are giving their owners grief, they can not understand why they do not want to listen to them.
Again, just my experience and my opinion.
And no, I do not want to spend more for that transformer than for the rest of the amplifier.
DCR, Low frequency inductive reactance, Distributed capacitance at high frequencies.
All 600 Ohms, Huh?
For all we know, the NE5532 output is directly connected to the RCA output jack.
Or, at best, it has a series resistor.
If there is any DC offset of the NE5532, the current may saturate the otherwise pristine input transformer.
A cheaply engineered NE5532 circuit probably does have DC offset.
And, who says every NE5532 circuit drives 3V, 5mA, and 600 Ohms at low distortion?
A high gain input tube with 60pF of Miller Effect capacitance, and a 1:16 transformer will reflect 256 times the Miller Effect capacitance back to the NE5532 (518 Ohms capacitive reactance).
(Of course, we would never specify that large of a step up ratio).
This is no more complex, than any other stereo system.
But it is a System.
Some systems are well designed from beginning to end.
Other systems are giving their owners grief, they can not understand why they do not want to listen to them.
Again, just my experience and my opinion.
Last edited:
Why from DC? 20 Hz is the lowest frequency of interest.
The transformer that I recommended, O-26, meets the 20 - 20,000 requirement +/- 1 dB in the proposed application.
As I posted, it allows a little bit of unbalanced DC, so can be connected directly to op-amp output.
The transformer that I recommended, O-26, meets the 20 - 20,000 requirement +/- 1 dB in the proposed application.
As I posted, it allows a little bit of unbalanced DC, so can be connected directly to op-amp output.
Last edited:
6a3: Thank you as ever for your long and careful response. You give us a lot on this forum. I could try to get more measurements on the DAC with a multimeter on the pins (I don't have a scope) but that's not the point in the end. The point is it could be any DAC with an op-amp in the output. At some point I'll replace it. So the amp itself has to be designed to work with "a typical DAC", not one particular one. The amp itself needs more gain since the choice of driver is one with a mu of around 20. This is deliberate because after trying several input tubes I most liked the EL33. So it needs x2 or x3 step-up to work in a typical small-scale system.
So how do you choose a step-up transformer that will accommodate a large range of possible sources? I presume that what you do is use a transformer that is specified as "10K into..." whatever. In reality one specified as "600 into..." may work or it may cause problems. I know very little about transformer construction but it's clearly a question of impedance matching which is what you'd expect from a transformer.
I emailed Per Lundahl and he recommended the LL7903, which is what Ale Moglia uses. This is a larger transformer and so capable of handling larger signals. So presumably that's also a factor.
A lot of this thread is going to go in the direction of "it depends..." which is understandable. However, there must be a way of defining a "preferred part" that will give 1:2 or 1:3 stepup. It's something any commercial designer would have to confront. This kind of impedance matching is commonplace in pro-audio, studio and broadcast equipment. A rare thing in hifi, however, where the added expense of an input transformer would be hard to justify when you could use something simple like a 12A*7 input tube with a resistor load. But that's not a choice I want to make.
If I were to order a custom step-up transformer, for instance, I would have to have some means of knowing what to specify. I know of at least two winders who would be happy to make me a good one but what would I specify? A clone of LL7903 but with amorphous or nickel or Finemet? There must be some way of defining such a part within a reasonable range of options.
So how do you choose a step-up transformer that will accommodate a large range of possible sources? I presume that what you do is use a transformer that is specified as "10K into..." whatever. In reality one specified as "600 into..." may work or it may cause problems. I know very little about transformer construction but it's clearly a question of impedance matching which is what you'd expect from a transformer.
I emailed Per Lundahl and he recommended the LL7903, which is what Ale Moglia uses. This is a larger transformer and so capable of handling larger signals. So presumably that's also a factor.
A lot of this thread is going to go in the direction of "it depends..." which is understandable. However, there must be a way of defining a "preferred part" that will give 1:2 or 1:3 stepup. It's something any commercial designer would have to confront. This kind of impedance matching is commonplace in pro-audio, studio and broadcast equipment. A rare thing in hifi, however, where the added expense of an input transformer would be hard to justify when you could use something simple like a 12A*7 input tube with a resistor load. But that's not a choice I want to make.
If I were to order a custom step-up transformer, for instance, I would have to have some means of knowing what to specify. I know of at least two winders who would be happy to make me a good one but what would I specify? A clone of LL7903 but with amorphous or nickel or Finemet? There must be some way of defining such a part within a reasonable range of options.
Last edited:
600r pro transformers tend to be better because they are designed to be driven with 600r preamps which tend to squash parasitics which cause peaking. So 600r is the way to go. However is your DAC output capable of driving a a pro line ? Many will not be as they will have to high an output impedance.
I tend to agree that transformer can often ruin a good system as they are very exacting in component matching. This is especially true when people attempt to drive them with weedy high output impedance tube preamps.
Shoog
I tend to agree that transformer can often ruin a good system as they are very exacting in component matching. This is especially true when people attempt to drive them with weedy high output impedance tube preamps.
Shoog
Yes, thanks for that. My intention isn't to use a "weedy high output impedance tube preamp" however nice an 01A or 26 stage can sound. It's to use a "typical good DAC". If this means driving a typical pro audio 600 ohm input transformer, then it's possible to specify that "the DAC must be designed with an output impedance of e.g. no more than 100-200 ohms". This then throws the issue back to the DAC. I haven't looked at the specifications of typical good DACs and maybe that's the next step.
Last edited:
Andy: transformer's performance is specified for certain driving impedance. However, it doesn't mean that you have to use exactly the same source impedance - in fact, the lower the driving impedance, the better will be transformer's performance, bandwidth and distortion wise. For example, a 600:600 transformer does not have to be driven with exactly 600 Ohm source; it's 600 Ohm rating only means that it's performance (undistorted maximum signal level and frequency range) is specified for a 600 Ohm source. If you drive this transformer with a 100 Ohm source, it's performance will be much better. So, 600 Ohm is basically a maximum recommended source impedance. By the same token, 600 Ohm secondary rating does not mean that secondary load must be exactly 600 Ohm. It means that transformer performance will be adversely affected if it drives a load less than 600 Ohms. For the secondary, the higher the driven impedance, the better.
So, a 600:600 transformer will perform much much better with 100 Ohm source and 100 K load than with 600 Ohm source and 600 Ohm load.
You can order a custom transformer specifying primary inductance (something like 50 H or more), ability to handle a couple of mA DC, and power level of about 5 mW at 20 Hz into 10 K load.
So, a 600:600 transformer will perform much much better with 100 Ohm source and 100 K load than with 600 Ohm source and 600 Ohm load.
You can order a custom transformer specifying primary inductance (something like 50 H or more), ability to handle a couple of mA DC, and power level of about 5 mW at 20 Hz into 10 K load.
Btw, any op-amp's output impedance is very low, it is largely determined by series output resistor, which is used to block RF oscillation. If this resistor is replaced with an RF choke, the output impedance could be about 10 Ohms.
Lundahl don't specify pri/sec inductance. Taking the LL7903 as their recommendation, they provide a + dBU rating and a DC resistance. So choosing another input transformer from their range looks like a question of determining the required + dBU rating and DC resistance.
"The LL7903 is a large, high level, high performance audio transformer, made for extraordinary requirements. The transformer combines high level capability (+28 dBU @ 50 Hz primary level) with low copper resistance and is designed for the most demanding applications. The core is a high permeability mu metal lamination core."
"The LL7903 is a large, high level, high performance audio transformer, made for extraordinary requirements. The transformer combines high level capability (+28 dBU @ 50 Hz primary level) with low copper resistance and is designed for the most demanding applications. The core is a high permeability mu metal lamination core."
It will work, but it will be like shooting sparrows with a cannon. If you are determined to use Lundahl, I can look at their catalog and see what will work for you.
A typical CD player may specify a minimum load impedance of 10k.
That would include the input impedance of a tube amp step up transformer.
(even though that same CD player's output impedance may be only 1k or 500 Ohms).
We are not talking about RF circuits that match impedances for maximum power transfer.
Generally, the circuit is narrowband, like an AM station at 1500kHz, that only has 10kHz of bandwidth. 1500k/10k = 150:1
A CD player is nothing like that 20-20000Hz, 1000:1.
And we are not talking about the old 600 Ohm standards for many audio production and broadcast systems.
Most CD players are not like that either.
Just like tube preamp outputs, solid state op amps can have more distortion when having to drive lower load impedances.
A CD player that specifies a minimum 10k load, may not meet its distortion specifications, even though it essentially meets its voltage output specs.
A 10k primary of a 1:3 step-up transformer, has a 90k secondary.
No step-up comes totally free of some tradeoff or other.
Find a transformer that matches the exact application including signal source parameters, and the input tube impedance (including Miller Effect Capacitance), and minimize the tradeoff.
There are many good step-up transformers, but not all of them meet all the various application needs, which do have a wide range of parameters.
Find the ones that will fit your application (I hope you are not designing a universal amplifier for commercial applications, to be used with everybody's different CD players, and wide range of stand-alone outboard DACs.
That would include the input impedance of a tube amp step up transformer.
(even though that same CD player's output impedance may be only 1k or 500 Ohms).
We are not talking about RF circuits that match impedances for maximum power transfer.
Generally, the circuit is narrowband, like an AM station at 1500kHz, that only has 10kHz of bandwidth. 1500k/10k = 150:1
A CD player is nothing like that 20-20000Hz, 1000:1.
And we are not talking about the old 600 Ohm standards for many audio production and broadcast systems.
Most CD players are not like that either.
Just like tube preamp outputs, solid state op amps can have more distortion when having to drive lower load impedances.
A CD player that specifies a minimum 10k load, may not meet its distortion specifications, even though it essentially meets its voltage output specs.
A 10k primary of a 1:3 step-up transformer, has a 90k secondary.
No step-up comes totally free of some tradeoff or other.
Find a transformer that matches the exact application including signal source parameters, and the input tube impedance (including Miller Effect Capacitance), and minimize the tradeoff.
There are many good step-up transformers, but not all of them meet all the various application needs, which do have a wide range of parameters.
Find the ones that will fit your application (I hope you are not designing a universal amplifier for commercial applications, to be used with everybody's different CD players, and wide range of stand-alone outboard DACs.
Last edited:
"but it will be like shooting sparrows with a cannon"
I disagree.
Small LL SUTs are a strong trade-offs, without compensation most of them has lesser or greater HF bump.
7903 is acceptable device, not the HF champion, but enough good for ordinary tube amplifier.
I'm not a fan of Lundahl, several manufacturer produce better devices.
In England Sowter is available (Brian retired and sold the firm, but "business as usual") and -in many ways- better than LL.
I disagree.
Small LL SUTs are a strong trade-offs, without compensation most of them has lesser or greater HF bump.
7903 is acceptable device, not the HF champion, but enough good for ordinary tube amplifier.
I'm not a fan of Lundahl, several manufacturer produce better devices.
In England Sowter is available (Brian retired and sold the firm, but "business as usual") and -in many ways- better than LL.
Yes - I was looking at that, and I think it would do the job. Cheaper too and mu-metal.
As for 10K input this is what I have right now with the Hammond 124B which is 10K:90K and 49% nickel. I assume mu-metal is 80% nickel though it may vary. I also have LL1540 which is probably around 10K:50K and mu-metal. So I'm in the ballpark. When I connect my amorphous LL1544A I'll know if amorphous is a better way to go.
Incidentally, I don't have a CD player, it's a DAC I have which may or may not be the same. The input to the DAC is all on my Mac Pro.
Mu metal or 80% Ni are the way to go here. They have practically zero remanence, and, unlike other materials, will not be permanently magnetized by accidental DC spikes.
Looks the right kind of transformer. What's the core material - I didn't get that. Also who actually makes these?
How do you buy from them? Via email?
They have some interesting ITs I might need for the future projects.
Generally, bigger transformers have more HF problems than smaller ones. Because they require interleaving, they cannot use Faraday shield, so there is quite substantial capacitive coupling, not a good thing.
Sowter better in what ways?
Last edited: