When you rectify 120volts you get the Peak output voltage which is some 147 volts.
Essentially (if your in america) you can take two 1:1 transformers, run the primaries in parallel, on the secondaries tie the negative of the first to the positive of the second. where the secondaries connect is your ground. the other two leads go to the bridge rectifier and the +/- of the bridge is your +/- 147. The rest of the circuit for the powersupply is the float/filtering/smoothing/protection/regulation
Essentially (if your in america) you can take two 1:1 transformers, run the primaries in parallel, on the secondaries tie the negative of the first to the positive of the second. where the secondaries connect is your ground. the other two leads go to the bridge rectifier and the +/- of the bridge is your +/- 147. The rest of the circuit for the powersupply is the float/filtering/smoothing/protection/regulation
So far, it's not a case of "close, but no cigar"; instead, we miss by a wide margin. For the Class-A2 output stage, we need to target inductors that can, at an absolute minimum, reliably pass a continuous 500mA, so we need something rated for at least 1A in order to conservatively operate the components. If we're going to design power-supplies that can sustain the rated 60-watt (RMS) output specification, the steady-state current requirement becomes something closer to 2.4A. Of course, we don't want too large of a voltage-drop across the inductors or the power-output levels will start to modulate against the power-supplies; low DCR values will be important.
So we need moderately high-inductance value inductors with a low DCR and a high heat-dissipation rating. A possible candidate would be something like the Jantzen copper-foil air-core inductors (DCR: 0.85-ohms, power handling: 350 watts RMS for 16 gauge, 500 watts RMS for 14 gauge and 650 watts RMS for 12 gauge). The problem is that the largest inductance value in that product line is 4.7mH, which when flanked by a pair of ~3000uF capacitor banks (C-L-C filter configuration) only gives us a -3dB corner-frequency of 42Hz. To best serve the M-60 signal path, we really need to have a corner-frequency that is well below the audio frequency band. The more traditional iron-core choke designs can certainly provide higher inductance values, but their current/heat-dissipation capacity is limited and they can easily reach core saturation (degrading performance). Before we start stacking-up multiple off-the-shelf inductors, I wanted to see if some custom inductors might offer us an alternative that saves space/weight in the chassis at a decent price-point.
Let's remember there are a lot of us guys out here with 240VAC.
Not to worry; that's why I had targeted power-transformers that have primaries which can be wired for either 120VAC or 240VAC...
Its nice to have the windings on three different transformers. The windings talk to each other so if one is loaded it will affect the others. That's why we use a separate B+ transformer for the output, a separate filament transformer and a separate driver B+/B- transformer.
We don't use a choke in the output section, just a 1 ohm 5-watt resistor. If the output section is really balanced, the sawtooth isn't a problem. We do use a choke in the B- of the driver, solely for V4 cathodes. The B- for V3 is fed by a 1K 1 watt resistor that comes directly off of the rectifiers, in parallel with the choke. We don't have any chokes in the B+; the 3K device that feeds V4's plate comes off the rectifiers, and is in parallel with another 1K unit for the plates of V2. We bypass the plates of V4 w/200uf, same for its cathodes. 100uf bypasses the plates and cathodes of the voltage amplifier.
The output section supplies are typically about 6600uf each.
We don't use a choke in the output section, just a 1 ohm 5-watt resistor. If the output section is really balanced, the sawtooth isn't a problem. We do use a choke in the B- of the driver, solely for V4 cathodes. The B- for V3 is fed by a 1K 1 watt resistor that comes directly off of the rectifiers, in parallel with the choke. We don't have any chokes in the B+; the 3K device that feeds V4's plate comes off the rectifiers, and is in parallel with another 1K unit for the plates of V2. We bypass the plates of V4 w/200uf, same for its cathodes. 100uf bypasses the plates and cathodes of the voltage amplifier.
The output section supplies are typically about 6600uf each.
I am currently selling some tubes on ePay. I collected some sweep tubes, but since they are unique and not in current production I decided to sell them for those who want to use them. I will stick with GU-50 output tubes, there are still plenty of them. But 6LW6 and another similar tubes will go.
NOS 6LW6 vacuum tube - Admiral - eBay (item 180476307383 end time Mar-09-10 16:57:59 PST)
NOS 6LW6 vacuum tube - Admiral - eBay (item 180476307383 end time Mar-09-10 16:57:59 PST)
Those 120/240 1:1 ct's are good for america but if your on 240 all you need is a regular center tap. half of 240 gives you +/- 120 RMS. In this sense it *SHOULD* be cheaper for those of you on 240 by a few units (dollars euros pounds or what have you (rubels maybe))
as for the inductors triad might be your best bet if your trying to get your corner frequency way down.
Inductor; Common Mode; Ind 8mH; Cur 3.2A; Thru-Hole; DCR 0.12 Ohms $4.17 might be steep. how many inductors are we talking about here 2,4,8?
I'm confused or maybe I should say surprised about the PSU's?
I like simple but I expected more to say the least. Does this really work?
I'm just having a hard time accepting the exteme simplicity.
Will you get stable voltages from this? Or isn't stability all that necessary?
How about noise?
Won't the VAC driving the heaters carry over to the audiosignal?
What voltage is suitable for the transformers secondaries to get the proper voltages?
Is there a nice little formula to go from VAC to VDC?
I like simple but I expected more to say the least. Does this really work?
I'm just having a hard time accepting the exteme simplicity.
Will you get stable voltages from this? Or isn't stability all that necessary?
How about noise?
Won't the VAC driving the heaters carry over to the audiosignal?
What voltage is suitable for the transformers secondaries to get the proper voltages?
Is there a nice little formula to go from VAC to VDC?
Supous you are begginer in tube DIY.That Atmasphere schematic is 100% OK.Circlotron output stage have excelent common mode hum& noise rejektion, so from dual floathing ouput PSU and filament to cathode hum & noise induction is automaticly canceled in the Load(speaker).that is not case in Futterman topology.Input dual diferencial cascode stage have same rock stable nature,driver stage too.Important is to conect Psu trafo secondary`` in phase`` with SS rectifier both for output stage and filament.Anyway this OTL amps are rock stable and sounds supreme.If you are not expirienced Diy guy maybe solution for you is to by M60 Diy Kit from Atmasphere(dont know did sell this Kit anymore) or finished amps.Think for output 2 phase bridge is 150v/1A(two time) and for input and drive is bipolar +- 300v/100ma(for precision answer ask Atmasphere).In general VDC is VACx1,41.I made some circlotron multiphase Otl amps and thats the best souding amps I ever made.Salute
You could use a choke in the output section power supplies. I like to use the one ohm 5 watt devices because I want immediate current availability from the power transformer. BTW, the better the transformer the better the voltage stability.
Noise induced by the AC filaments in the driver circuit can be an issue if the power transformer has poor or no electrostatic shielding. We usually run the filaments off of the output tube supply. The 6SN7s are quite rugged when it comes to filament/cathode arc over and so this means has proven quite reliable!
The driver supply should be HEXFREDs or similar for best results. OTOH a good quality silicon bridge is all we use in the output section. Cheesy ones will require a fair amount of neutralization. We have used HEXFREDs in the output too, but you will have to provide heatsinks for them and use the highest current ones you can find. Silicon can survive up to 10X its rating at surge; whatever the HEXFREDs are rated for is also their surge current.
If properly set up the amp is effective on horns, usually quieter than most SETs. BTW you can reduce the gain by eliminating the top 6SN7 of the differential cascode (V2) without any other changes. This can be useful if you are running high efficiency speakers. This amp is extremely effective on horns!!
Noise induced by the AC filaments in the driver circuit can be an issue if the power transformer has poor or no electrostatic shielding. We usually run the filaments off of the output tube supply. The 6SN7s are quite rugged when it comes to filament/cathode arc over and so this means has proven quite reliable!
The driver supply should be HEXFREDs or similar for best results. OTOH a good quality silicon bridge is all we use in the output section. Cheesy ones will require a fair amount of neutralization. We have used HEXFREDs in the output too, but you will have to provide heatsinks for them and use the highest current ones you can find. Silicon can survive up to 10X its rating at surge; whatever the HEXFREDs are rated for is also their surge current.
If properly set up the amp is effective on horns, usually quieter than most SETs. BTW you can reduce the gain by eliminating the top 6SN7 of the differential cascode (V2) without any other changes. This can be useful if you are running high efficiency speakers. This amp is extremely effective on horns!!
atmasphere,
Thank you for providing this information along with the power-supply schematic! I had assumed that the power-supply underlying the M-60 would share the traits of the signal-path topology --- a simple, rational design that has been refined by listening to the circuit over an extended period of time. The M-60 power-supply is the semi-massive, yet straightforward, support system that I had assumed would be necessary to provide the robust power characteristics necessary to support the 60-watt configuration of the Class-A2 M-60.
I am uncertain of the power-supply sub-circuit centered around the three-pin device, U1, on the B+(2), largely because I cannot discern the part designation on U1. Is this an AC-ripple regulator?
The M-60 power-supply design posted by "atmasphere" is closely aligned with my expectations; the balanced/differential nature of the M-60 signal-path topology would tend to provide a fair degree of immunity to power-supply disturbances. So the simple, rectifier-fed "pi" filter sections make perfect sense; we just need substantial, well-filtered energy-storage capacity to meet the power-supply requirements.
The inherent simplicity of the power-supply design places a premium on the quality of the power-supply components themselves, so we need to carefully select the specific components employed in the M-60 power-supply implementation in order to avoid compromising the audio characteristics. I imagine that a significant portion of the M-60 circuit refinement over the past three decades has been focused on optimizing precisely these type of parameters.
The M-60 is configured for Class-A operations throughout, so as long as we have sufficient energy-storage in the power-supply (that is well-filtered, combined with reasonably low ESR across the audio-band), the resulting power-supply rail voltages will be quite stable. The balanced/differential nature of the M-60 signal-path will largely cancel-out the residual power-supply noise.
The power-supply transformers that I had suggested for the M-60 were selected to meet the voltage/current demands of the various sections of the power-amplifier. The input/driver stage power-supply transformer will produce a +/-325VDC set of power-supply rails while the transformers suggested for each "phase" of the Circlotron/bridge output stage will produce a set of +/-104VDC power-supply rails.
The input/driver stage isn't especially sensitive to the slight increase in B+/B- voltages (325VDC vs. 300VDC). However, I will need to recalculate the resistor values for the voltage-divider that sets the output-stage fixed-bias current. The output stage can easily accommodate the somewhat lower power-supply voltages that result from the suggested power-transformers (104VDC vs. 140VDC) and actually benefits from the reduced 6AS7 plate dissipation.
The formula for converting transformer secondary VAC ratings to the resulting DC-voltage is (VAC rating) x SQRT(2), then subtract the voltage-drop across the rectifier diodes.
Last edited:
banat,
Thanks for your scrutiny of the power-supply schematic! I'll be certain to include the correction in the M-60 power-supply that I plan to post to this thread.
bacon665,
Thanks for the suggestions! Now that "atmaspehere" has posted the M-60 Mk.II.3 power-supply schematic, that will accelerate my efforts to post an equivalent power-supply schematic based on off-the-shelf components.
Once we have our baseline power-supply design, we can then assess the component costs associated with building the power-supply for our M-60 thread-based design. The M-60 is what I refer to as an "iceberg" design; an inherently simple signal-path circuit ("tip of the iceberg") supported by a massive, out-of-band power-supply ("the bulk of the iceberg, hidden beneath the waves"). So, once we can characterize the power-supply costs, we'll be in a good position to estimate the overall costs for an M-60 implementation. At that juncture, all of us can individually determine if we can self-justify pursuing an M-60 implementation.