Is 50 degrees C too hot?

[rdf]

For a power transformer's long term reliability? I'm prototyping a stereo Mullard 3-3, now with a Hammond 270HX (275-0-275, 200 ma max, 174 VA in) after watching my lovely NOS 270EZ (120 ma max) boil and short. The original Mullard article spec's 60 ma per channel, Duncan's PS designer software models the RMS current draw at closer to 70 ma per channel. That puts the 270HX running at 75% current draw on the HV lines, plus 2 EF86's and 2 EL84's on the AC filament. I know the circuit's OK, the EL84 cathode drops calculate to 45 ma per channel with 250 volts plate to cathode, a little low if anything but easy to trim at the screen. The original 100 uf caps were replaced with 50 uf's with little change. It's an RC supply providing around 300 VDC to the OT. The temperature was measured with a Fluke 65 Infrared Thermometer borrowed from work, it's an accurate figure.
Any input on operating temps, or properly sizing PS trannies, appreciated and thanks in advance.

[dhaen]

The temperature you quote of 50 deg C is quite ok IMO. However, factors to consider are temperature rise, and maximum ambient temperature. You can then check what the max transformer temperature is likely to be.
Also, consider the transformer manufacturers specification: "120mA max" probably means 120mA resistive load AC. The rectifier in a C R C or C L C circuit draws pulses of current near the crest of the sine wave. This causes extra heating in the transformer compared with a resistive load. So not only do you have to multiply your DC current by 1.4, but also include a "fudge factor". In my expeience *1.6 covers this.

If you want a rectifier / smoothing circuit that conducts for most of the wave angle, go for an L C, sometimes called "choke input" type. The disadvantage is that you have to start off with a higher AC voltage, and the choke can be noisy if not designed for this purpose.

[Eli Duttman]

When cap. I/P filtration is employed, the heating in the rectifier winding of the power trafo increases as the value of the 1st or only cap. increases. With really large cap. values, even 1.6X the DC draw may not be enough. My rule of thumb is 2X the DC draw when full wave rectification is employed and 4X the DC draw when a voltage doubler is used. That allows for both the Physics and ambitious manufacturer ratings.

[rdf]

Originally posted by dhaen
The temperature you quote of 50 deg C is quite ok IMO. However, factors to consider are temperature rise, and maximum ambient temperature...

Thanks dhaen, that's some comfort. After the first Chernobyl I started unplugging the amp to take a shower. It was as warm in the room as it ever gets when the readings were taken so they confidentally represent worst case. PSU2 models it as you describe, the 70 ma figure was RMS.

Originally posted by Eli Duttman
When cap. I/P filtration is employed, the heating in the rectifier winding of the power trafo increases as the value of the 1st or only cap. increases. With really large cap. values, even 1.6X the DC draw may not be enough.

Thanks Eli, I found the same thing. I should have been a little more specific, the PS circuit I was using is solid state D-R-C-R-C, something in the range (it's changed too many times to remember exactly) of 500 ohms - 100 uF - 250 ohms - 100 uf. The first R really helps reduce pulse magnitude. If the parts on order ever arrive the caps will become 60 uf and 50 uf ASC poly-in-oils and the resistance ratios increased a bit. It's still a bit of a question for me why the first one went. The original Mullard article specified a 300-0-300 60 ma (per channel) transformer with tube rectifiers. The 270 DX was 275 V 120 ma with solid state for a stereo amp, the lower voltage should result in less power required of the secondary. Have manufacturer's ratings become that 'optimistic'?
One other question, any recommendations for tube PS transformer manufacturers? Comparing the contruction of Hammond's new transformer to the NOS predecessor did them no favours, and for this circuit something a little over 250-0-250 would have been a better match but isn't in their line. Coming from sand-land it's still hard to wrap my head around a 275v x 200 ma supply being just enough for seven stereo watts.

[Sch3mat1c]

Also noteworthy is Hammond's confident ratings: Revision 3 runs around 65mA DC into a 50uF cap (SS FWB) off a 269AX. Last winter I left it on for weeks at a time but I'm keeping it off this summer, just because. Anyway, it's rated for 125-0-125 100mA which would be FWCT and only 160VDC, I get 330VDC so if it were FWCT it would be drawing 130mA. Note 50uF is conservative as input caps go for this current.

Tim