Is one winding producing more heat than another?

[Jamier]

Transformers get hot. Some run hotter than others. If you have a transformer with a single HV winding and a single, say, 12V heater winding, is one of those producing more heat than the other? Also , would not having the heater winding connected, and using a separate filament transformer reduce the heat produced in the main transformer?

Jamie

[Paul Birkeland]

When we have a new power transformer design wound, one of the things that PJ does is to take some measurements to estimate the core temperature of the transformer.  You haven't provided enough information in your post to really know whether the HV or LV winding would get hotter, but it is true that there is some give and take in terms of loading one winding a bit more and unloading another winding.  In terms of using a separate filament transformer, it would help to quantify the VA rating of each winding, then consider how much reduction you're actually providing by using a separate filament transformer. 

[Jamier]

PB, I don’t have enough experience to say what the VA rating of each winding is, unless it can be derived from the figures I have: HV: 170 V @ 300 mA / Heater: 5.9-0-5.9 CT @ 2A. If there is something I can measure please let me know and I will.

Jamie

[Paul Birkeland]

VA is just volts times amps.  170V * 0.3A=51VA.  12V * 2A=24VA.  You also have to factor in who designed the transformer and how conservative they are.  Certain transformer designers are more careful than others!

[Jamier]

I am not an expert on transformers, but let’s just say that I’m pretty certain it is not in the same league as Bottlehead iron. So, assuming that it is probably working on the ragged edge of it’s capabilities, will relieving it of heater duty help the temperature performance?

Jamie

[Paul Birkeland]

The transformer is rated for 75VA and you'd be relieving about 1/3 of that, which should indeed reduce the operating temperature noticeably.

[Jamier]

PB, thank you for that. That will help a lot! I will get a filament trans that has enough headroom as to not run too close to it’s limits and I think I’ll be good.

Jamie

[EricS]

I don't know if this is "standard" practice or not, but I do know that some transformer manufacturers specific their ratings assuming a full current draw which also comes with a final transformer temp of somewhere near 50-60c.  In my new amp, my PGP8.1 transformer hits a temp of about 50c within 30-40 mins.

The general rule in the Class-A solid state world is to not exceed 30% of the transformer's VA rating for long term viability of the transformer.   For example, the PSU for one of my amps draws 100wpc, so double this for two channels, then multiply by three: I need a minimum size of 600VA for this transformer. 

In my brief experience, the tube world seems to be much more comfortable running current draw right up to the VA rating of the transformer.  I am wondering why this dichotomy in design/implementation exists across these two domains.  Perhaps it is merely a philosophical difference in comfort level of the individual designers?

[Paul Birkeland]

The general rule in the Class-A solid state world is to not exceed 30% of the transformer's VA rating for long term viability of the transformer. 

In your example, a 100W per channel stereo class A amp could have a maximum efficiency of 50%, so your bare bones minimum transformer rating would be 400VA, but such an amp will also have multiple stages that also consume power and won't be 50% efficient, so you'd move up to the next size which is 500VA or 600VA.  Which I would select between those two would depend on the amp design and the transformer manufacturer.

Sizing transformers for a tube amp is not nearly that simple, as there are tons of windings.

Some examples of tube amp transformers:

Dynaco ST-70 is about 300VA for a 75W amplifier.
Kaiju 300 is about 265VA for a 16W amplifier.
Paramour 2A3 was about 55VA for a 3.5W amplifier.

As you can see, these are all way, way, way above 3x output power.

[EricS]

Ah... Those are interesting examples, thanks for sharing them!  I guess I was not fully accounting for the variety of windings that power different amplification stages.

As you point out, efficiency also has a big impact here.  With my previous example, a 100w draw on the PSU per channel is from a push-pull design that operates at 40% efficiency (which is pretty high), so I get 40wpc of output power out of a 100w quiescent current draw.  Using a 3x safety/longevity factor, that results in 300VA for 40w of speaker power (7.5:1 ratio of VA to output power).  Some single ended topologies run more like 10-25% efficient, which comes closer to matching the ratios you provided for tube transformers.  The SS amps I've built all feature two gain stages.  For one design that I use, the voltage gain stage runs at ~20mA on a bipolar 22v supply (next to no power at all) but the current gain stage pulls 9A out of +/-22v rails, which creates some cooling challenges...  I get ~125w of output power from this design that generates ~400w of heat that is pulled from a 1500VA toroid (12:1 ratio).

It seems these designs are all closer in overall PSU spec than I thought...  I knew there had to be something that I was missing, this perceived disparity was confusing me!  Thanks for the details!

[Paul Birkeland]

It seems these designs are all closer in overall PSU spec than I thought...  I knew there had to be something that I was missing, this perceived disparity was confusing me!  Thanks for the details!

The bigger point is that you don't really need to derate things that severely unless you don't trust your transformer winder.  If you triple the VA of each winding in a tube amp compared to what you need, you'll end up with poor regulation (voltages that are higher than they are supposed to be) and a much larger transformer than is ideal.  Though it may look really cool to have an enormous transformer, having high heater voltage can reduce tube lifetime.   

PJ may decide to weigh in on this, as he estimates temperature rise, then measures it as part of the power transformer design process.  Modern wire insulation is ridiculously good at coping with heat, and the limiting factor nowadays is keeping the temperature of the transformer "finger safe".

[Paul Joppa]

Well, I tried to write something, but it is just too complicated to fit in a post. I'll just summarize that transformers are designed to a specific requirement, including voltages, currents, and what rectifier circuits are used. Anything different will affect all the other parameters in ways that are difficult to generalize.

[EricS]

The bigger point is that you don't really need to derate things that severely unless you don't trust your transformer winder. 

This could very well be the point in the DIY arena where people around the globe are copying a specific design using whatever supplier they have easiest access to.  It just simplifies things to aim for a higher rating than is necessary rather than end up being under powered. 

One potential challenge that I'm just starting to realize with a super large donut transformer is the total weight of the amp.  One of my big monoblocks is in need of some maintenance after a decade of service.  Hefting that thing upstairs so that I can work on it is less than a trivial task.  In another 2 or 3 decades, I may not be able to lift it... 

As for certain parameters affecting others, I've been quite surprised over the past few weeks at just how much this happens!  In some parts of the tube circuit, merely adding a DMM (with a presumed 10M ohm load) to make a measurement weighs down the voltage by a noticeable amount.