Power Trans runs Cooler when Amp Driven Hard ?

[HaroldBrooks]

I apologize for the big gaps in my knowledge, but as always I appreciate any opportunity to learn.

Do tube amp transformers run a bit cooler in cathode biased amps when you are driving them hard, as opposed to at an idle ?

I have taken some IR measurements of a power transformer and I was surprised to find it a bit cooler when I was using it for the same time as idling it.

I may be off with this, as I had no controls. Once my wife is out of the house (LOL) I can run a fair test that would go like this : 1.) Room temp amp, start playing for 30 minutes or so, and then measure transformer at several pre-defined points.

2.) Room temp amp, just let idle for the same amount of time, check temps at all the same transformer points. Make sure room temp and wall voltage are very close in both tests, and amp is in same location.

Question : Is there any basis to support a Power transformer running cooler on a small push pull 6V6 amp ? If so, why is that the case ? If not, what happens generally and why ?

Thanks for your help !

[Helmholtz]

It might happen in a class A PP amp with cathode bias, caused by bias shifting at high power which makes the tubes run cooler.
Hardly conceivable in a class AB amp.

Remember that the PT has a very long thermal time constant. It may take an hour or more to reach thermal equilibrium.

[nevetslab]

I apologize for the big gaps in my knowledge, but as always I appreciate any opportunity to learn.

Do tube amp transformers run a bit cooler in cathode biased amps when you are driving them hard, as opposed to at an idle ?

I have taken some IR measurements of a power transformer and I was surprised to find it a bit cooler when I was using it for the same time as idling it.

I may be off with this, as I had no controls. Once my wife is out of the house (LOL) I can run a fair test that would go like this : 1.) Room temp amp, start playing for 30 minutes or so, and then measure transformer at several pre-defined points.

2.) Room temp amp, just let idle for the same amount of time, check temps at all the same transformer points. Make sure room temp and wall voltage are very close in both tests, and amp is in same location.

Question : Is there any basis to support a Power transformer running cooler on a small push pull 6V6 amp ? If so, why is that the case ? If not, what happens generally and why ?

Thanks for your help !

For any amplifier (tube or SS), just letting it sit at idle, and say it's pulling 1.3A/124W @ 120VAC at idle. In this case, a Fender Twin Reverb. Start at room temp of 23 deg C. In a few hours, that surface temperature on the core of the power xfmr will be around 60 deg C, too hot to leave your hand on it. The tubes being biased for about 33mA each, it heats up just from the current pulled thru the core. No load....excitation current on those is only a few watts (I'd have to dig up data recorded on that), but even at no load, that small amount of current will eventually increase the core temperature.

Drive output power thru the amp, and you'll find the AC mains power consumption near clip will be at least 200W. The transformer will heat up more under that condition for sure.

Now, at Cathode Bias, and Class A operation, such as a Morgan RCA 35, it's power tubes at idle are drawing around 60mA, regardless of tube type. I thought I had recorded the AC Mains current/wattage @ 120VAC, though don't find it in my notes. As I recall, though, the mains current/wattage remains about the same between idle and full power. Now, the power transformer will still increase in core temperature, simply because current is flowing thru the primary, and it produces heat. I think if you compared one amp of the same wattage, one being Class A, cathode bias, and one that's Class AB, both WILL for sure heat up over time. I can't see why a cathode biased amp would run cooler if driven hard. Not all cathode biased amps are Class A. Vox AC30's and their partners in crime aren't Class A. Their xfmrs will heat up like the rest of them.

To make a transformer run cooler, you can use a much larger core, select wire size to yield lower temp rise, but......physics always wins. current thru wire wrapped around a core & contained withing a core will produce heat, and will increase over time.

[HaroldBrooks]

Thanks for the responses !

Another question : Any issues with running a slight fan cooling over a transformer ? Not talking about a huge air current, just a small DC fan ? My worry is that I am only cooling the case, and not addressing the "Hot Spots" that might be in the core, but in all honesty the power transformer on this amp (old field coil Valco) only peaks around 140 degrees in hot room after a couple of hours, and it stays there. With the fan I get it down to a cool 112 degrees or so, but I want to make sure I am doing the right thing and not screwing something else up in the process.

Once again, thank you all.


Link to a pic of the actual amp (2nd pic) + just a small part of my pocket computer collection (I have way too many)... If you are interested !

https://www.flickr.com/photos/909242...th/8275738820/

[The Dude]

Have you checked bias? Maybe it's just overbiased.

[HaroldBrooks]

It is biased pretty hot currently.

The stock Supro (Valco) cathode resistor is 200 ohms, and I am sure that was ok for 115vac back in the 40's, but my outlet today is 125vac and sometimes a bit more !
I was thinking of soldering in a new 300ohm Ohmite wire wound, and I'm sure that would cut the current back a bit to both the 6V6 pair and the field coil, and give less stress to the power transformer in the process. I will order a 12 watt 300ohm "brown devil" and take care of it.

Seems like a lot of the amps from Valco were biased fairly hot, so with today's wall voltage is marginal. I was also thinking about a bucking transformer to reduce the wall voltage to around 115vac (filament voltage is currently 6 vac or I would even go a bit lower)

Thanks for the suggestion.

[SoulFetish]

It might happen in a class A PP amp with cathode bias, caused by bias shifting at high power which makes the tubes run cooler.
Hardly conceivable in a class AB amp.

Remember that the PT has a very long thermal time constant. It may take an hour or more to reach thermal equilibrium.

Well, if the flux density in a PT is highest with no load (or at idle?), then maybe the transformer is undersized and was designed to run close to saturation? From what I've read, this is why Microwave power supplies can get away with using smaller transformers than you otherwise might find in equally powered devices (in that they only operate near or at full load.)
- By the way, this is more of an inquiry than an answer. I'm still trying to better understand such things, so I invite any clarification or corrections.

[Helmholtz]

Well, if the flux density in a PT is highest with no load (or at idle?), then maybe the transformer is undersized and was designed to run close to saturation? From what I've read, this is why Microwave power supplies can get away with using smaller transformers than you otherwise might find in equally powered devices (in that they only operate near or at full load.)
- By the way, this is more of an inquiry than an answer. I'm still trying to better understand such things, so I invite any clarification or corrections.

PT heating is caused by its power dissipation/losses. Transformer dissipation is caused by core losses and copper (resistance) losses. Core losses in a given transformer mainly depend on primary voltage and frequency (which in turn determine magnetizing current and core flux density). In a PT these parameters are fairly constant and there is only little influence of the load.
Copper losses are caused by the currents, i.e. mainly by the secondary load and to a lesser extent by the fairly constant primary magnetizing current.

As the load current causes a voltage drop across the primary DCR, the effective primary voltage somewhat reduces with high load currents causing some decrease of core losses. But at the same time copper losses strongly increase.

I can't image a load situation with a PT where total dissipation is below the no load condition.

In other words PT dissipation/heating will always increase with total primary current.

[HaroldBrooks]

PT heating is caused by its power dissipation/losses. Transformer dissipation is caused by core losses and copper (resistance) losses. Core losses in a given transformer mainly depend on primary voltage and frequency (which in turn determine magnetizing current and core flux density). In a PT these parameters are fairly constant and there is only little influence of the load.
Copper losses are caused by the currents, i.e. mainly by the secondary load and to a lesser extent by the fairly constant primary magnetizing current.

As the load current causes a voltage drop across the primary DCR, the effective primary voltage somewhat reduces with high load currents causing some decrease of core losses. But at the same time copper losses strongly increase.

I can't image a load situation with a PT where total dissipation is below the no load condition.

In other words PT dissipation/heating will always increase with total primary current.

Just got a strange thought... The amp is a combo and this is not an ideal situation when it comes to cooling, as heat rises and can actual get caught in the pocket at the top of the amp, and get stuck there, particularly If the amp is idle and the speaker immobile.

I realize this arrangement is fairly common to all combo amps, but I was thinking a hard working amp might have some better air flow in an out of the cabinet space (open back) when you are playing hard with a lot of percussive notes. This is only a thought, and may not be that significant, but running the amp hard may contribute to air exchange a bit, and all it takes is a few degrees in the surrounding air exchange to net a noticeable change in the component temps.

I've noticed that a very small computer DC fan placed even a meter or more away has quite an effect, so perhaps the speakers in a combo can account for just a little more air exchange when running ? I'm just guessing here. I could run a fan all the time, but I don't want to start down that road if I can avoid it.

As a side note, the bias on the tubes is currently set too high (some of my other tubes red plated in this amp), and I will be changing out the cathode resistor from the stock 200ohm to a more tenable 350ohm I just ordered. That should help a bit.

Not too long ago I swapped in a 350ohm resistor on my Gretsch 6162, and the plate dissipation went down 5 watts in total for a pair of 6973 tubes to end up around 12.5 watts each, and that cooled down the tubes a noticeable amount. The transformer on that particular amp is huge, so it stays cool all the time an was never suspect. Again, Valco stretched the limit of hot tube bias on that amp right out of the factory, as they did with a lot of their amps.

[SoulFetish]

As the load current causes a voltage drop across the primary DCR, the effective primary voltage somewhat reduces with high load currents causing some decrease of core losses. But at the same time copper losses strongly increase.

Right. I suppose my question is in regards to the surface temperature on the stack. Would it be reasonable for a small transformer, with poor regulation, to idle near or at saturation? If so, in the scenario you describe above, how would the heat dissipation from the copper losses affect the surface temperature vs core losses?
Thanks.

Although, to your point – even if a PT was pushed to saturation at no load, I would think much of the heat would be from the increase in primary current (copper losses)

[Helmholtz]

Right. I suppose my question is in regards to the surface temperature on the stack. Would it be reasonable for a small transformer, with poor regulation, to idle near or at saturation? If so, in the scenario you describe above, how would the heat dissipation from the copper losses affect the surface temperature vs core losses?
Thanks.

Although, to your point – even if a PT was pushed to saturation at no load, I would think much of the heat would be from the increase in primary current (copper losses)

Not sure if I understand.
A mains transformer (or any transformer connected to a low impedance voltage source) must no be designed to idle near saturation (especially with a stacked core). Otherwise its behaviour (primary current, dissipation) would be uncontrollable. Tolerances in construction, core material and mains voltage fluctuations might lead to "full" saturation where input current is only limited by the primary DCR.

[Helmholtz]

I realize this arrangement is fairly common to all combo amps, but I was thinking a hard working amp might have some better air flow in an out of the cabinet space (open back) when you are playing hard with a lot of percussive notes. This is only a thought, and may not be that significant, but running the amp hard may contribute to air exchange a bit, and all it takes is a few degrees in the surrounding air exchange to net a noticeable change in the component temps.

I don't think the tiny cone excursions (which don't produce a net in-and-out air flow) could noticeably contribute to cooling. Don't forget that also speaker voice coils produce considerable amounts of heat. Typical speaker efficiency is around 5%, meaning that 95% of output power is converted to heat by the speakers.

[pdf64]

...I've noticed that a very small computer DC fan placed even a meter or more away has quite an effect, so perhaps the speakers in a combo can account for just a little more air exchange when running ? I'm just guessing here. I could run a fan all the time, but I don't want to start down that road if I can avoid it.

As a side note, the bias on the tubes is currently set too high (some of my other tubes red plated in this amp), and I will be changing out the cathode resistor from the stock 200ohm to a more tenable 350ohm I just ordered. That should help a bit...

It's not too big or invasive a job to fit such a fan on to the amp chassis, see http://www.blueglow.de/FenFan.html for how it might be done.

Nearly doubling a cathode bias resistor value is a big change; it will almost certainly result in a significant degree of bias shift at high signal levels - stick your volt meter across it and crank out a power chord.
A bit of bias shift, squish as Aiken calls it, is fine, but a lot will tend to thin the tone out.
Excessive idle dissipation in cathode bias is usually a result of the HT being too high; given your scenarios, I'd prefer to add 150 ohms into the HT supply side of the power tubes' current path, rather than their cathode return.

[Helmholtz]

I hope you don't run the amp enclosed in the white box (bedside locker) as shown in your pictures.
Nothing wrong with using a fan.
The PT is not necessarily the most temperature sensitive component in an amp.

[HaroldBrooks]

It's not too big or invasive a job to fit such a fan on to the amp chassis, see http://www.blueglow.de/FenFan.html for how it might be done.

Nearly doubling a cathode bias resistor value is a big change; it will almost certainly result in a significant degree of bias shift at high signal levels - stick your volt meter across it and crank out a power chord.
A bit of bias shift, squish as Aiken calls it, is fine, but a lot will tend to thin the tone out.
Excessive idle dissipation in cathode bias is usually a result of the HT being too high; given your scenarios, I'd prefer to add 150 ohms into the HT supply side of the power tubes' current path, rather than their cathode return.

LOL, I should have read your excellent suggestion before I took apart the amp today ! I found a suitable 350ohm cathode resistor in all my junk and put it in, but it's quite easy to take out again. I also have a cluster of Zener diodes running on a turret board that I could use. That idea came from Rob Robinette's fantastic site : https://robrobinette.com/images/Guit...iode_Stack.pdf I tested this on another amp and it works great at lowering the B+ for anything up to around 120v or in-between. Currently my plates are running at 355vdc, a bit high for an old amp as you intimated.

I think the idea of putting in a "semi-permanent" fan is a good one, as that small dc fan will follow the amp wherever I go. I think I will put the cathode resistor back around 250ohms, and use a short wood screw and a small DC fan. The idea will be to mount the fan with a short wood screw and a small aluminum bracket I can fashion easily from stock. I won't be using the amp's power supply for the 12v fan because in this amp it may be marginal do to it's age, so just a small DC power supply that I can tuck in the amp when I am transporting it. On an amp like my Gretsch 6162 with it's huge transformer, I wouldn't hesitate to tap into the power of the amp for a fan.

Do to my lack of experience with working on amps (about 2 years now) I still sometimes over-react to small things that can be solved rather easily, as you mentioned.

Thanks once again for the help pdf64, I will add the fan and be done with the time wasting and somewhat pointless analysis !

Still, I'm glad when guys like you and the rest chime in, You guys are all gentlemen and very smart and experienced, and I am climbing the learning curve non-stop as a result !!!