Forums

Guesstimate A Transformer

Started by Unknown November 25, 2018
One of my useless things I think I'll throw together. An audio amp. 

I have transformers, audio output transformers from a tube amp that have a certain feature I need for this. This is a mobile audio amplifier. The basic configuration is shown at the end of this post in Spice. The coils, L1 and L2 are actually a center tapped winding. 

What I want to know is how much power dare I push through the thing. I don't have much to go on. The core is 3.757" X 3.158" X 1.931". The DCR across the whole thing is about 0.5 ohms. 

This is not a new configuration. With 12, or 14.4 volts Vcc is should be about 30 watts into 8 ohms. Theoretical about 50 watts but with losses 30 is more like it. I have used much smaller transformers for this at 30 watts. Now I intend to multiply the voltage. I could double it or triple it. Doubling it should get me 4 times the power, theoretically but we know that is not going to ha[pen. Tripling it should get me 9 times the power but we KNOW that's not going to happen. 

There will be junction losses in the outputs of course, but most of the loss is going to be in the transformer. That means heat. Well now just how much heat ? I am sure there is a way to figure it out but I don't know it, and am not an expert on transformers. I don't know if I even have the right parameters. I guess I got total surface area, and I have DCR which is where much of the loss will go, what else is a significant factor ? 

This is just one of those "for the hell of it" things. I have a tripler built, all I have to do is populate it with transistors and drive it but I am pretty sure it will be too much. I might be able to regulate it down, a bit of inductance and maybe synchronous rectifiers or something. I'll burn that bridge when I come to it. If I do that I might go with a tracking power supply, hey, it beats commutators. 

I figure it is logical to not include the shielded parts of the transformer in the area as it is pretty insulated, so just on the sides I got 26.7"^2 right ? Should I stick some fins on it ? I could probably find enough scrap heatsinks around here to do it on all sides, but one might be inaccessible. 

Also, is there something I can soak these things in to enhance the insulation of the windings ? I am sure even going to 24 volts Vcc is going to be more than this thing was ever made for. 

Past experience though has shown them to be quite tolerant. I have even used 24 CT power transformers and they worked, though performance was not stellar to say the least. I know I was inducing a bunch of voltage back into the primary, and that will happen with these as well. I will have to clamp it somehow because if it ever gets run without a load I don't want it blowing smoke rings. that's also why I would like to enhance the insulation. 

So, any idea ? Or should I just get out the old powersupplyzilla and see what it can take ? If anyone like smoke I think I can make regulated about 50 volts at about 8 amps. If necessary I can get drastic but it wouldn't be regulated. (what would be the point ? lol)

To wit:

Version 4
SHEET 1 880 680
WIRE 496 -16 224 -16
WIRE 160 32 16 32
WIRE 288 80 224 80
WIRE 400 80 288 80
WIRE 400 144 400 80
WIRE 288 160 80 160
WIRE 80 176 80 160
WIRE 288 240 224 240
WIRE 400 240 400 192
WIRE 400 240 288 240
WIRE 160 288 16 288
WIRE 432 336 432 304
WIRE 432 336 224 336
WIRE 496 336 496 -16
WIRE 496 336 432 336
FLAG 80 176 0
FLAG 432 304 0
SYMBOL ind 272 64 R0
SYMATTR InstName L1
SYMBOL ind 272 144 R0
SYMATTR InstName L2
SYMBOL npn 160 -16 R0
SYMATTR InstName Q1
SYMBOL npn 160 336 M180
SYMATTR InstName Q2
TEXT 16 56 Left 2 ;+ DRIVE
TEXT 16 320 Left 2 ;- DRIVE
TEXT 392 168 Left 2 ;SPK.
TEXT 400 264 Left 2 ;+ Vcc
On 11/25/2018 08:09 AM, jurb6006@gmail.com wrote:
> One of my useless things I think I'll throw together. An audio amp. > > I have transformers, audio output transformers from a tube amp that have a certain feature I need for this. This is a mobile audio amplifier. The basic configuration is shown at the end of this post in Spice. The coils, L1 and L2 are actually a center tapped winding. > > What I want to know is how much power dare I push through the thing. I don't have much to go on. The core is 3.757" X 3.158" X 1.931". The DCR across the whole thing is about 0.5 ohms.
It would help if you can find out the primary inductance or get an approximate read on that figure. I like the energy-perspective for transformer/inductor design; the core geometry you need is proportional to how much instantaneous magnetic energy the core has to hold without saturating (the peak AC component plus any DC component of the magnetic field if the amp is single-ended output.) The following set of slides outline a procedure to derive a ballpark core geometrical constant/"effective electrical size" from design requirements like the max saturation flux density of the core material, peak primary current, required primary inductance, etc. You know what the required primary inductance will need to be because it forms a high pass filter with the tube(s) output impedance and determines bass response. You know what the required peak primary current will be for your desired power output into the load on the secondary. You can ballpark the rest of the required parameters to get a geometrical constant that will let you infer the required size of the core. Conversely if you already have a core you can run the equation on page 11 "backwards" to infer what kind of energy handling ability it might be able to provide. <http://ecee.colorado.edu/~ecen5797/course_material/Ch14slides.pdf>
> This is not a new configuration. With 12, or 14.4 volts Vcc is should be about 30 watts into 8 ohms. Theoretical about 50 watts but with losses 30 is more like it. I have used much smaller transformers for this at 30 watts. Now I intend to multiply the voltage. I could double it or triple it. Doubling it should get me 4 times the power, theoretically but we know that is not going to ha[pen. Tripling it should get me 9 times the power but we KNOW that's not going to happen. > > There will be junction losses in the outputs of course, but most of the loss is going to be in the transformer. That means heat. Well now just how much heat ? I am sure there is a way to figure it out but I don't know it, and am not an expert on transformers. I don't know if I even have the right parameters. I guess I got total surface area, and I have DCR which is where much of the loss will go, what else is a significant factor ? > > This is just one of those "for the hell of it" things. I have a tripler built, all I have to do is populate it with transistors and drive it but I am pretty sure it will be too much. I might be able to regulate it down, a bit of inductance and maybe synchronous rectifiers or something. I'll burn that bridge when I come to it. If I do that I might go with a tracking power supply, hey, it beats commutators. > > I figure it is logical to not include the shielded parts of the transformer in the area as it is pretty insulated, so just on the sides I got 26.7"^2 right ? Should I stick some fins on it ? I could probably find enough scrap heatsinks around here to do it on all sides, but one might be inaccessible. > > Also, is there something I can soak these things in to enhance the insulation of the windings ? I am sure even going to 24 volts Vcc is going to be more than this thing was ever made for. > > Past experience though has shown them to be quite tolerant. I have even used 24 CT power transformers and they worked, though performance was not stellar to say the least. I know I was inducing a bunch of voltage back into the primary, and that will happen with these as well. I will have to clamp it somehow because if it ever gets run without a load I don't want it blowing smoke rings. that's also why I would like to enhance the insulation. > > So, any idea ? Or should I just get out the old powersupplyzilla and see what it can take ? If anyone like smoke I think I can make regulated about 50 volts at about 8 amps. If necessary I can get drastic but it wouldn't be regulated. (what would be the point ? lol) > > To wit: > > Version 4 > SHEET 1 880 680 > WIRE 496 -16 224 -16 > WIRE 160 32 16 32 > WIRE 288 80 224 80 > WIRE 400 80 288 80 > WIRE 400 144 400 80 > WIRE 288 160 80 160 > WIRE 80 176 80 160 > WIRE 288 240 224 240 > WIRE 400 240 400 192 > WIRE 400 240 288 240 > WIRE 160 288 16 288 > WIRE 432 336 432 304 > WIRE 432 336 224 336 > WIRE 496 336 496 -16 > WIRE 496 336 432 336 > FLAG 80 176 0 > FLAG 432 304 0 > SYMBOL ind 272 64 R0 > SYMATTR InstName L1 > SYMBOL ind 272 144 R0 > SYMATTR InstName L2 > SYMBOL npn 160 -16 R0 > SYMATTR InstName Q1 > SYMBOL npn 160 336 M180 > SYMATTR InstName Q2 > TEXT 16 56 Left 2 ;+ DRIVE > TEXT 16 320 Left 2 ;- DRIVE > TEXT 392 168 Left 2 ;SPK. > TEXT 400 264 Left 2 ;+ Vcc >
On Sunday, 25 November 2018 13:09:26 UTC, jurb...@gmail.com  wrote:

> One of my useless things I think I'll throw together. An audio amp. > > I have transformers, audio output transformers from a tube amp that have a certain feature I need for this. This is a mobile audio amplifier. The basic configuration is shown at the end of this post in Spice. The coils, L1 and L2 are actually a center tapped winding. > > What I want to know is how much power dare I push through the thing. I don't have much to go on. The core is 3.757" X 3.158" X 1.931". The DCR across the whole thing is about 0.5 ohms. > > This is not a new configuration. With 12, or 14.4 volts Vcc is should be about 30 watts into 8 ohms. Theoretical about 50 watts but with losses 30 is more like it. I have used much smaller transformers for this at 30 watts. Now I intend to multiply the voltage. I could double it or triple it. Doubling it should get me 4 times the power, theoretically but we know that is not going to ha[pen. Tripling it should get me 9 times the power but we KNOW that's not going to happen. > > There will be junction losses in the outputs of course, but most of the loss is going to be in the transformer. That means heat. Well now just how much heat ? I am sure there is a way to figure it out but I don't know it, and am not an expert on transformers. I don't know if I even have the right parameters. I guess I got total surface area, and I have DCR which is where much of the loss will go, what else is a significant factor ? > > This is just one of those "for the hell of it" things. I have a tripler built, all I have to do is populate it with transistors and drive it but I am pretty sure it will be too much. I might be able to regulate it down, a bit of inductance and maybe synchronous rectifiers or something. I'll burn that bridge when I come to it. If I do that I might go with a tracking power supply, hey, it beats commutators. > > I figure it is logical to not include the shielded parts of the transformer in the area as it is pretty insulated, so just on the sides I got 26.7"^2 right ? Should I stick some fins on it ? I could probably find enough scrap heatsinks around here to do it on all sides, but one might be inaccessible. > > Also, is there something I can soak these things in to enhance the insulation of the windings ? I am sure even going to 24 volts Vcc is going to be more than this thing was ever made for. > > Past experience though has shown them to be quite tolerant. I have even used 24 CT power transformers and they worked, though performance was not stellar to say the least. I know I was inducing a bunch of voltage back into the primary, and that will happen with these as well. I will have to clamp it somehow because if it ever gets run without a load I don't want it blowing smoke rings. that's also why I would like to enhance the insulation. > > So, any idea ? Or should I just get out the old powersupplyzilla and see what it can take ? If anyone like smoke I think I can make regulated about 50 volts at about 8 amps. If necessary I can get drastic but it wouldn't be regulated. (what would be the point ? lol) > > To wit: > > Version 4 > SHEET 1 880 680 > WIRE 496 -16 224 -16 > WIRE 160 32 16 32 > WIRE 288 80 224 80 > WIRE 400 80 288 80 > WIRE 400 144 400 80 > WIRE 288 160 80 160 > WIRE 80 176 80 160 > WIRE 288 240 224 240 > WIRE 400 240 400 192 > WIRE 400 240 288 240 > WIRE 160 288 16 288 > WIRE 432 336 432 304 > WIRE 432 336 224 336 > WIRE 496 336 496 -16 > WIRE 496 336 432 336 > FLAG 80 176 0 > FLAG 432 304 0 > SYMBOL ind 272 64 R0 > SYMATTR InstName L1 > SYMBOL ind 272 144 R0 > SYMATTR InstName L2 > SYMBOL npn 160 -16 R0 > SYMATTR InstName Q1 > SYMBOL npn 160 336 M180 > SYMATTR InstName Q2 > TEXT 16 56 Left 2 ;+ DRIVE > TEXT 16 320 Left 2 ;- DRIVE > TEXT 392 168 Left 2 ;SPK. > TEXT 400 264 Left 2 ;+ Vcc
If you look up mains transformers with similar dimensions you'll get a rough idea on power. NT
On Sun, 25 Nov 2018 05:09:22 -0800 (PST), jurb6006@gmail.com wrote:

>One of my useless things I think I'll throw together. An audio amp. > >I have transformers, audio output transformers from a tube amp that have a certain feature I need for this. This is a mobile audio amplifier. The basic configuration is shown at the end of this post in Spice. The coils, L1 and L2 are actually a center tapped winding. > >What I want to know is how much power dare I push through the thing. I don't have much to go on. The core is 3.757" X 3.158" X 1.931". The DCR across the whole thing is about 0.5 ohms. > >This is not a new configuration. With 12, or 14.4 volts Vcc is should be about 30 watts into 8 ohms. Theoretical about 50 watts but with losses 30 is more like it. I have used much smaller transformers for this at 30 watts. Now I intend to multiply the voltage. I could double it or triple it. Doubling it should get me 4 times the power, theoretically but we know that is not going to ha[pen. Tripling it should get me 9 times the power but we KNOW that's not going to happen. > >There will be junction losses in the outputs of course, but most of the loss is going to be in the transformer. That means heat. Well now just how much heat ? I am sure there is a way to figure it out but I don't know it, and am not an expert on transformers. I don't know if I even have the right parameters. I guess I got total surface area, and I have DCR which is where much of the loss will go, what else is a significant factor ?
This is for 50/60 Hz: https://www.dropbox.com/s/jnb3jmw8rcmdeir/XfmrScatter.JPG?dl=0 -- John Larkin Highland Technology, Inc lunatic fringe electronics
 

It depends on the method of cooling ( free air, confined, forced ) and acceptance criteria for hot-spot temp rise in the windings e.g. 50'C.  

I would expect for conventional cold-rolled Grain Oriented Steel (CRGOS) ( not Amorphous type) 

- core losses about 1W/kg +/-30%  better with thinner Laminate. 
- core capacity < 100 VA/kg depending winding efficiency, winding design, copper losses, cooling design
- for resistive load only 
- if bulk unregulated load with 10% ripple V then I peak:RMS= 10:1 
    - then derate max VA by at least 35% 
the required core size is a strong function of the frequency.
As an audio output transformer for HiFi you may ant to go down to 30 Hz.
If its only PA system voice then maybe 100Hz.

m



Thanks folks. It'll take some time to sift through all that. At this point I think I would be alright feeding it 24 volts but not 36. 

It seems a couple do not realize that I am not using the primary.I am using the secondary as a split primary, which is why I need the 4 and 16 ohm taps. It must be balanced. 

On that Spice print, that IS the secondary. I am also going to incorporate Shottky diodes at the collectors of the outputs to emulate tube sound best as I can. 

I am really thinking about a tracking regulator now, that way I can cut back the peak power when the temperature rises. I also will incorporate a sabotage proof protection system, like another one, if you hook the speaker wires to say, the AC line it just burns the wires off and still works after being turned off for a few seconds and turned back on. Any amp I build that costs more than a couple of cases of beer will have that. Simply indestructible, in fact I jst might make a challenge. You can connect anything to anything but you can't go inside the case. Reward money in escrow. 

I just might not...
On Mon, 26 Nov 2018 12:24:23 -0800 (PST), jurb6006@gmail.com wrote:

>Thanks folks. It'll take some time to sift through all that. At this point I think I would be alright feeding it 24 volts but not 36. > >It seems a couple do not realize that I am not using the primary.I am using the secondary as a split primary, which is why I need the 4 and 16 ohm taps. It must be balanced. > >On that Spice print, that IS the secondary. I am also going to incorporate Shottky diodes at the collectors of the outputs to emulate tube sound best as I can. >
Does that circuit actually work? I guess it should, barring any DC imbalance issues. -- John Larkin Highland Technology, Inc picosecond timing precision measurement jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
On 26.11.18 22:36, John Larkin wrote:
> On Mon, 26 Nov 2018 12:24:23 -0800 (PST), jurb6006@gmail.com wrote: > >> Thanks folks. It'll take some time to sift through all that. At this point I think I would be alright feeding it 24 volts but not 36. >> >> It seems a couple do not realize that I am not using the primary.I am using the secondary as a split primary, which is why I need the 4 and 16 ohm taps. It must be balanced. >> >> On that Spice print, that IS the secondary. I am also going to incorporate Shottky diodes at the collectors of the outputs to emulate tube sound best as I can. >> > > Does that circuit actually work? I guess it should, barring any DC > imbalance issues.
I just wonder why on earth the transformer. In tube designs, it was a mandatory part, to get a ballpark match to the low impedance of the speaker(s), but a complementary transistor amplifier can be made to work directly to the speaker. I've made a bunch of tube amplifiers (about 5 decades ago), and the output transformer was a PITA, honestly. -- -TV
On Tue, 27 Nov 2018 10:21:13 +0200, Tauno Voipio
<tauno.voipio@notused.fi.invalid> wrote:

>On 26.11.18 22:36, John Larkin wrote: >> On Mon, 26 Nov 2018 12:24:23 -0800 (PST), jurb6006@gmail.com wrote: >> >>> Thanks folks. It'll take some time to sift through all that. At this point I think I would be alright feeding it 24 volts but not 36. >>> >>> It seems a couple do not realize that I am not using the primary.I am using the secondary as a split primary, which is why I need the 4 and 16 ohm taps. It must be balanced. >>> >>> On that Spice print, that IS the secondary. I am also going to incorporate Shottky diodes at the collectors of the outputs to emulate tube sound best as I can. >>> >> >> Does that circuit actually work? I guess it should, barring any DC >> imbalance issues. > > >I just wonder why on earth the transformer. In tube designs, it >was a mandatory part, to get a ballpark match to the low impedance >of the speaker(s), but a complementary transistor amplifier can >be made to work directly to the speaker. > >I've made a bunch of tube amplifiers (about 5 decades ago), and the >output transformer was a PITA, honestly.
That transformer allows the outputs to be emitter followers, and to get more voltage into the speaker. True, there are simpler ways to drive speakers... like buy a class D full-bridge chip. -- John Larkin Highland Technology, Inc lunatic fringe electronics