Lamination thickness / low eddy current has more to do with power efficiency. For HF bandwidth of output transformers, winding capacitance and leakage inductance is far more important. I compared the same coil with different core materials: 0.3mm and 0.1mm GOSS, amorphous alloy and even nanocrystalline (all easy to interchange c-cores with equal dimensions). There is no difference in HF bandwidth.
If low impedance, low mu tubes are used and (mandatory really) individual current sinks with the cathodes coupled with a single film cap, then there is resonant behaviour at the bottom end which may be beneficial to the overall performance. Also, this power differential stage hardly cares about the power supply.
Many people (myself included) have seen flat frequency response out to well above 20 kHz using power toroidals in real life working amplifiers. Shoog indicated that the way you connect up the 2 115V windings makes a big difference. Just go back a few pages in this thread for evidence.
Final measurement for today was a 750VA toroid with two 35V secondaries. Secondaries parallelled.
So some 7:1 voltage ratio and 50:1 impedance ratio. Not bad at all!
Loaded with 16 ohm and 700 ohm source impedance it showed LF loss, but HF was very acceptable.
A lower power one with some more inductance (this one has just 2H) and two 18V secondaries might suit a PP 6C33C. No primary DC unbalance allowed for this type of amplifier as the toroid will easily saturate.
So some 7:1 voltage ratio and 50:1 impedance ratio. Not bad at all!
Loaded with 16 ohm and 700 ohm source impedance it showed LF loss, but HF was very acceptable.
A lower power one with some more inductance (this one has just 2H) and two 18V secondaries might suit a PP 6C33C. No primary DC unbalance allowed for this type of amplifier as the toroid will easily saturate.
This is wrong, as pieter t pointed up, stray capacitance and leakage inductance dictates HF bandwidth on OPTs.
Eddy current loss is somewhat like
We ≈ ξ d² f² (Bmax)² x 10⁻¹¹
Bmax goes as 1/f, then the frequency dependence is canceled.
This makes sense, because distributed capacitance must be evaluated from +B (0V AC) to the anode of the valve.
It depends on the number of primary layers if the difference is big or not.
Garter bias works very well, just a bit more wasteful of supply voltage and more waste heat. Its a good way to go for moderate bias voltage tubes when using toroids.
Depending on how low you want to go on bass, a 150 or 200VA might work well. Current source or servo bias scheme might fit your needs.
This is why I use Triad xfmrs...
From Triad Magnetics:
Inductance depends on the drive level and the permeability of the core (which can vary significantly). Having said that, I have reviewed some history data from various lots and here are the results.
1. VPT230-110 = 213 to 439H @ 230V, 50Hz (primaries in series). 304H average from 15 samples.
2. VPT230-220 = 192 to 444H @ 230V, 50Hz (primaries in series). 283H average from 15 samples.
3. VPT230-430 = 93 to 214 @ 230V, 50Hz (primaries in series). 128H average from 15 samples.
From Triad Magnetics:
Inductance depends on the drive level and the permeability of the core (which can vary significantly). Having said that, I have reviewed some history data from various lots and here are the results.
1. VPT230-110 = 213 to 439H @ 230V, 50Hz (primaries in series). 304H average from 15 samples.
2. VPT230-220 = 192 to 444H @ 230V, 50Hz (primaries in series). 283H average from 15 samples.
3. VPT230-430 = 93 to 214 @ 230V, 50Hz (primaries in series). 128H average from 15 samples.
Hmmm… thinking about that, like that I guess it makes sense. What you're basically saying is that HF into a transformer doesn't change its instantaneous magnetization B much. Z = 2πFL … F high, Z high, and time scale also small due to 1/F.
My error-in-thinking comes from imaging a power transformer working at 10 kHz as a power transformer (i.e. engineered for the job). All of its windings and laminations would be completely different.
Thanks.
GoatGuy
B is the magnetic field, or "induction magnetic field" if you want, magnetization, M, is defined as
B = H + 4 π M
And I am basically saying that eddy current loss
We ≈ ξ d² f² (Bmax)² x 10⁻¹¹ = ξ d² f² [(Uac x 10⁸) / (√2 π f S Np)]² x 10⁻¹¹
= ξ d² [(Uac x 10⁸) / (√2 π S Np)]² x 10⁻¹¹
Does not depend on frequency.
B = H + 4 π M
And I am basically saying that eddy current loss
We ≈ ξ d² f² (Bmax)² x 10⁻¹¹ = ξ d² f² [(Uac x 10⁸) / (√2 π f S Np)]² x 10⁻¹¹
= ξ d² [(Uac x 10⁸) / (√2 π S Np)]² x 10⁻¹¹
Does not depend on frequency.
Last edited:
So this thread is about saving $20 on transformers by spending $200 extra on tubes 
What am I missing?
Thanks God I commercially design, wind and sell transformers, of course am fully equipped for that, and in that case "doing it right" costs about the same in $ and a little longer in time as "doing it wrong" .
Specially doing proper interleaving, both in copper winding and in lamination stacking.
Copper and iron cost is same in both cases.
What am I missing?
Thanks God I commercially design, wind and sell transformers, of course am fully equipped for that, and in that case "doing it right" costs about the same in $ and a little longer in time as "doing it wrong" .
Specially doing proper interleaving, both in copper winding and in lamination stacking.
Copper and iron cost is same in both cases.
$200 extra on tubes?
Where do you buy your tubes?
Look at tubes that end up with low impedance loads, low voltage pentodes, sweep tubes, regulation pass tubes, or even parallel output applications. Routinely found cheap even international.
Hell, even a parallel output 6V6 or EL84 setup can use these, or even better a parallel output 6L6GC AB2. ( man that was a fun build!)
Personally I want to grab a few 6C19P Russians and try them soon, they are cheap, and look like half of a 6AS7, just with nicer curves.
Where do you buy your tubes?
Look at tubes that end up with low impedance loads, low voltage pentodes, sweep tubes, regulation pass tubes, or even parallel output applications. Routinely found cheap even international.
Hell, even a parallel output 6V6 or EL84 setup can use these, or even better a parallel output 6L6GC AB2. ( man that was a fun build!)
Personally I want to grab a few 6C19P Russians and try them soon, they are cheap, and look like half of a 6AS7, just with nicer curves.
$200 extra on tubes?
Where do you buy your tubes?
8 to 10 EL84 pairs (and matched) got mentioned earlier.
Some high current TV or regulator tubes will certainly work better.
Last edited:
These are pretty useless data.
230V is about maximum voltage swing in practice.
What inductance remains when measured at voltage swings at typical listening levels?
You would know when you would take the effort to look into the differences between power supply transformers and output transformers. Your Triad data above apply to power supply transformers working at (constant) line voltages with core excitations approaching the max (around 1.7-1.8T).
Output transformers behave totally different; they handle audio signals, and core excitation does not exceed 1T at 20Hz to keep distortion low when you are looking for quality.
Output transformers behave totally different; they handle audio signals, and core excitation does not exceed 1T at 20Hz to keep distortion low when you are looking for quality.
Inductance does not depend on VP-P. Indeed… you'll get the same inductance reading from a transformer's coils whether you excite it with a 10 millivolt or a 10 volt signal. There are several ways to confirm this, tho' working at some convenient low frequency (such as 50 or 60 Hz, the line frequency), of a well-filtered, calibrated voltage and a series of well-measured (but not "standard") resistors does the trick. And an oscilloscope.
Down to 1 mH, 60 Hz … mΩ resistance
1 to 10 mH, Ω
10 to 100 mH, DΩ (deka-)
100 to 1000 mH, hΩ (hecta-)
1 H to 10 H, 1 kΩ
10 H to 100 H, 10 kΩ
100 H to 1000 H 100 kΩ.
1 to 10 mH, Ω
10 to 100 mH, DΩ (deka-)
100 to 1000 mH, hΩ (hecta-)
1 H to 10 H, 1 kΩ
10 H to 100 H, 10 kΩ
100 H to 1000 H 100 kΩ.
Below 1 mH, you need 1,000 Hz signal to many-kilohertz (for very low inductance chokes). Above 1 kHy, well ... the sky's the limit.
GoatGuy
