Poor transformer choice and dumping power in voltage reg

I've had a couple of transformers that just don't fit most if the amp/preamps I want to build. They are R core 30va with 24v secondaries. I've no idea when I bought them or why.

I know its less than ideal but other than heat, what would be the harm in dumping some of the excess power if it can be done? For example Im building Wayne's BA 2018 line stage and am aiming for +/- 18-22v. 24v secondaries means roughly 33.6v rectified. In a LM317/337 set up could I lose the excess 11.6v. And how would this be done?
IMG_6780.jpg

Alternatively what projects could use these transformers?

Yours,

An electrically challenged novice
 
There's no harm dumping heat as long as the regulator circuit is designed so that it doesn't over-heat. Usually this just means an adequately sized heatsink. Remember that depending on how much current you draw from this transformer the actual voltage will be different that what's stamped on the label.

Modern switching regulators might reduce the thermal challenge.
 
Last edited:
Another approach that I've used is to add series resistance - literally add power resistor(s) between the transformer and the rectifier. This has the benefit that it not only reduces the voltage at the output of the rectifier but also reduces the maximum current spikes through the transformer secondary winding, rectifier and capacitor that comes after the rectifier. The series resistance you need can be obtained from a simulation tool such as PSUD (you'll need the data sheet for your power transformer to get the winding resistances, or you'll need to directly measure these resistances with a meter).
 
Use a buck regulator to get rid of excess voltage on the positive side without turning it into excess heat. XL4015 is a cheap choice, it can tolerate up to 36V on the input. Negative bucks might well exist but I don't have a part number to hand.
 
  • Like
Reactions: 1 user
There's no harm dumping heat as long as the regulator circuit is designed so that it doesn't over-heat. Usually this just means an adequately sized heatsink.
Brilliant! So the LM317/337 can shift the 11.6v as heat. Wayne's BA 2018 line stage uses about 25ma per channel, so 50ma total. Using an online calc 50ma by 33.6v is only 1.68 watts. Is that right? Doesn't seem enough
Modern switching regulators might reduce the thermal challenge.

is that the DC to DC regulators?
 
Use a buck regulator to get rid of excess voltage on the positive side without turning it into excess heat. XL4015 is a cheap choice, it can tolerate up to 36V on the input. Negative bucks might well exist but I don't have a part number to hand.
A quick search for a negative buck brought up this +/- version using XL6019E1. seems fine for the voltage and amps. The switching frequency is nice and high at 220khz per the chip specs and 180khz per the board specs (I assume that's the same as switching power supplies).
IMG_38AA6FA2B791-1.jpeg
 
Last edited:
  • Like
Reactions: 1 user
Brilliant! So the LM317/337 can shift the 11.6v as heat. Wayne's BA 2018 line stage uses about 25ma per channel, so 50ma total. Using an online calc 50ma by 33.6v is only 1.68 watts. Is that right? Doesn't seem enough
That will be fine; fit the T0220 pattern regulators with hetsinks of 10degC/w or better, and they will be fine. For years.

Remember that these regs have thermal shutdown - if you do drop too much heat into them, they'll shut down fairly gracefully (the output voltage drops until the limited current flow resulting means acceptable dissipation i.e. they heavily 'brown-out' the regulated supply side.)

Read the datasheet too - the limit of dissipation on paper for these regs is about 1.5watts with no heatsink but mounting in free air; and, up to 20watts, at 25degc case temp - good luck finding a heatsink to do that reasonably!

But I would not go near either of those dissipation limits, for a few other reasons. With a modest heatsink, and ensuring it gets air circulation, 3-5W dissipation level , 2-3-4 x what you likely need is not an issue, at all. At 1watt without any heatsink measures, yes these things run too hot to touch - but remain reliable (e.g. on things I don't care about I've had a decade of continuous use while that crudely-implemented, without issue; tab temp might well be 65-75degC. : o
 
Last edited:
  • Like
Reactions: 1 user
Disabled Account
Joined 2002
Alternatively what projects could use these transformers?

Yours,

An electrically challenged novice
Any preamp project that runs on +/- 22…24V with classic regulators or +/- 26….30V with LDO regulators and this under 500 mA load. Of course the lower the load the less loss of energy to useless heat.

Any /up down conversion is adding complexity or useless heat which is not done nowadays. Adding switchers to one of the best transformer types is throwing away the child with the bath water.

We see that having a part and then design something around it is like tube devices are designed by many by choosing a random tube. It is often better to design a device and then calculate and pick a transformer. Things work out better that way.

Brilliant! So the LM317/337 can shift the 11.6v as heat. Wayne's BA 2018 line stage uses about 25ma per channel, so 50ma total. Using an online calc 50ma by 33.6v is only 1.68 watts. Is that right?
No. 33.6 - 22 = 11.6V

11.6 x 0.05 = 0.58W

0.58 x 2 = 1.16W

If you use 1 transformer and 1 pair 317/337 per channel for the Wayne device it is 0.58W per channel lost to heat in the PSU regulators or 0.3W per regulator which seems a perfect application for the transformers. And please: Volt, Watt, mA, kHz etc.
 
Last edited:
  • Like
Reactions: 1 user
Disabled Account
Joined 2002
True. Should be measured beforehand. My guess is that it will be below 37V but testing is mandatory. Recalculate power to heat loss with measured values under load.

One could also choose LM317HV and LM337HV. No experience with these but they’re OK till 60V input voltage.
 
Last edited:
Disabled Account
Joined 2002
Maybe the Wayne device you plan to build has only a footprint for smaller/lower heatsinks on its PCB!? There normally is a BOM with the required parts. Following the BOM is OK.

It will be in a casing I guess. Too high heatsinks are then quite annoying.
 
Last edited:
Thanks for the reply jean-paul. I jumbled things up and probably wasn't very clear earlier.

I'm building Wayne's BA 2018 line stage which requires a separate power supply board. LM317/337 looks to be the simplest option. I think I can extend the regs away from the board if necessary to accommodate the heatsink. Is bigger better in cooling?
 
Disabled Account
Joined 2002
Yes bigger is better as cooler is better for longevity but the calculations based on your data show 0.3W dissipation per regulator when using separate pairs LM317/337 per channel. 0.3W is almost negligable 🙂 I will not read up the details as it is not my cup of tea but please verify yourself. I was under the impression that an official PSU PCB exists and using that PCB with BOM heatsinks is advisable.

One can not be clear enough especially with native english speakers :D

Anyway the transformers are of the right type. They are only 2V higher than the required one but this is no problem. If you worry you could use CRC and drop a Volt. Reward is better filtering.
 
Last edited:
Disabled Account
Joined 2002
Calculate power to heat lows as shown and use real world numbers. Measure output voltage of the Rcore transformers with let’s say 100 mA load. Ohms Law is your best friend. Always.

The assumptions based on your data seem to be right. I don’t have the faintest idea what heatsink an ACA mini uses and I am definitely not into such gear. I deal with specifications and part numbers so sorry.

https://www.head-fi.org/threads/my-diy-bi-polar-crcrc-regulated-power-supply.937981/

Just an example, some statements are simply not right. Like LDOs having less ripple voltage and less noise. This is dependant of type and usage. It is the size of the heatsinks that is in our focus.

That’s about as much time as I wanted to spend so good luck with the project! BTW I took “couple” as 2 pieces and “R core 30va with 24v secondaries” as 30VA Rcore transformers with 2 x 24V secondary windings based on the picture.
 
Last edited:
In addition, the no-load/ light-load output of the rectified 24V (specified under max load conditions) will be much higher than the theoretical 33.6V.
It could even exceed the abs max rating of the 317/337
A good point, well made.

There is a very, very simple and reliable fix for that risk:

See the reverse-bias diodes the datasheet shows from output to input? *
Simply use 12 or 15v / 1.3w zener diodes in the same orientation.

At switch-on, or whenever the input grossly exceeds output voltage and risks the rated 37V differential, these will avalanche and limit the voltage across the regulator. And it is the voltage differential at 37V that is the limit -- it is not a 'maximum input voltage'.

Then, with a big zener - at switch on with say > 50v input, but the output held at 0v by an output capacitance of 10 +uF - the voltage differential would exceed the '37v' rating. Except that the zener now clamps the voltage difference across the 3-pin reg by avalanche until this avalanche current has charged the output cap enough that the voltage difference excess across the regulator diminishes below the Zener voltage. At that point the LM3x7 output regulation has control.

Very simple, very cheap, and very effective for this kind of application; and reliable, IME.


*which should be considered mandatory on higher volt use, esp for the 337s , which are rather more fragile about this kind of thing. In fact always implement the pair of recommended diodes for an LM3x7, regardless. Cheapest insurance you can have.
 
Last edited: