Trafo choice for 15W, LM1875 based ss-amp

[Rutger]

Hi again,

It took a while but I've gathered all the components I need and started building the amp. I use a Talema toroidal 30VA 2x18V transformer. I measured the actual secundairy AC and that is ±20V, so I expected it to be like ±22V DC at max after the rectifier. A little higher than intended but that´s okay. Guess what... it is ±27,5V!

How is that possible? It´s like there are no losses in my powersupply. I use a standard diode rectifier, nothing fancy.

[Enzo]

18VAC rectified and filters should yield approximately 25vDC

Your AC was a couple volts higher, so 20VAC rectified and filtered should make about 28v, about what you got.

Your transformer will produce 18vAC under certain conditions. First, the mains voltage must be what is specified. If the transformer rating is for 230v mains and you have 240v, then the resulting outputs will be that much higher as well. Furthermore, the transformer windings have resistance in their wires, so the rated output voltage will be there onoly when the thing is loaded to its rated current - and that is whatever 30watts might be at 18v.

So that is why the AC voltange might be higher than expected.

But I think the fundamental problem is you don't understand how rectification affects the voltage. I don;t know why you might think 20VAC will make 22VDC. AC waveforms coming from the transformer are measured in RMS voltage. RMS is an average of sorts, it is .707 of the peak voltage - assuming a smooth sine wave. That means that the peak voltage is 1.414 times the RMS voltage.

When AC is rectified and filtered, the filter caps will charge uip to that peak voltage, not the RMS voltage. SO your DC will be about 1.414 times the RMS you started with. In the real world, your rectifier diodes will drop a half a volt. And the current drawn will interact with the transformer resistance, but in general, RMS x 1.414 is what to expect in DC once rectified and filtered.

And that is why your 20VAC (times 1.414) makes about 28VDC, less a bit for the diode. At full current output, the transformer AC should drop to about the 18v it is rated for, and the rectified and filtered DC then will drop to about 25v.

[Rutger]

Enzo, thank you very much for your explanation, it's clear to me now. Reading your post and this thread again, I see that I made a huge calculation mistake according to the losses in the rectifier. I calculated with a factor of 0,7 instead of a voltage of 0,7V. I thought I would get a railvoltage of 20x1,4x0,7=19,6V. I came up with 22V as a 'maximum' expected voltage.

So... do I have a real problem now?
The gainclone can handle it, and I just need a little correction on the preamp powersupply to make the railvoltage right for the preamp. As far as my knowledge goes, I just have a greater dynamic range and the max. output power is now limited by the power transformer.

I wonder what the max. output power will be in this situation? My power transformer can deliver 0,83A per rail @ 18V. Speakerload is 8 Ohm.

I'm sorry that I ask you to do the math, but calculating electronics is some time ago for me, and the technical english on the web makes it very hard for me to understand. But I would really like to learn how!

[J M Fahey]

Short all-in-one answer, no math.
18+18VAC trafo -> +/-25VDC rails -> 25W RMS into 8 ohm speaker (from datasheet)
Get a beefier speaker, rated at least 30/35W RMS; a 40/50W RMS would be even safer.

[Rutger]

@Fahey: thanks, but that's not exactly what I meant. It's allready been pointed out that I can find that in the datasheet
Maybe my question wasn't clear enough: can the transformer current keep up? According to the formula it can. But then again it confuses me that I'm told that I need a 24VA for 15W output and now it seems that 30VA is more than enough for 25-30W output...? Aren't there any loses then?

[Steve Conner]

The amount of power delivered by the amp is mostly determined by the transformer's voltage rating. The VA rating has only a minor effect: if you use a too small transformer, its output voltage will sag below the nameplate value, because you are overloading it. And so the amp will deliver a bit less than your calculations predict. Output power is notoriously hard to predict: I'm happy if I get within 20% of what I expected in my SS experiments.

What the VA rating mainly determines is whether the transformer will work reliably for years, or burn out at your first gig. It is very common to use too small transformers in audio amps: I commonly see them with a VA rating equal to the amp's output power, whereas from theory it should be 1.5x or more. The industry gets away with it because of speaker impedance curves, the intermittent nature of music signals, and so on.

Note this is for solid-state amps: tube amps need bigger ones because they burn huge amounts of power even at idle.

[Rutger]

Thanks Steve, I'm learning a lot! It's nice to get to know the theory behind all this.

So I build a 25W amp instead of 15W. Well... I wasn't planning on using it for gigs or bandrehearsals anyway, just for the guitarlessons I teach, so I won't blow the speaker . But in case it will come to that the 25W will be needed and I just need to swap the speaker. Would be fun to get my hands on a Celestion G10D-25 or G12L-35 that comes with the original Marshall Lead series

[J M Fahey]

You'll live on the edge, literally, with the G10L25
Remember Marshall uses it with a 12W amp.
The G12L35 is fine.
More modern equivalents, easier to find, are MOD 10-50 and MOD 12-50 (or a Celestion Rocket 50).
As said above, the transformer you have *will* do, but will sag and overheat if you play loud.
Maybe not enough to burn.
In a nutshell: for guitar classes or playing Jazz/Funk/Country live: no problem at all, even with your actual speaker.
Heavy/Metal/Punk .......
Good luck.

[Rutger]

Thanks Fahey. I don't think the small transformer will be a problem, chances are very small that I will ever use it at full power, and I never play punk/metal anyway so I'll never use it with max. distortion.
Thanks for the speaker tip, haven't thought about the Jensen Mod series before!

[Rutger]

New problem... I changed the powersupply for the preamp to get from +/-27,5V to +/-15V. In the original schematic there are a 1k Ohm/22 uF, I replaced them by 3k9 Ohm / 47 uF (had those lying around). According to my calculation with the Duncan tool 3k9 would be enough to get the voltage under 16V, but now I measure 27,3V, hardly any lower than the original voltage...

What did I do wrong?
I have a hard time dealing with the powersupply.

[loudthud]

Was the preamp connected when you made that measurement? The preamp must be connected and drawing current to get a voltage drop across the 3K9 resistors.

[Rutger]

Ah... no, it wasn't connected for I wanted to check the voltages first before I would burn something.

Well, problem solved, thanks!

[J M Fahey]

Ok then, *what* did you measure with the preamp connected?

[Rutger]

Still need to do that! I'm busy with other things at the moment (work...)

But it's clear that I didn't mess up things, I checked everything on my board after I measured the 'wrong' voltages and that seemed okay. Now I know I've been stupid, not crazy

[Jazz P Bass]

Do you have the 3K9 resistor connected to a circuit?
Or is the one end hanging in the air.
This type of voltage regulation needs a load.