Versatile Transformer Selection

[SoliS]

Hi everyone, its my first post!

I am a college student and thus have a budget. I would like to construct a power supply and gainclone circuit. However, I would like the power supply to be versatile. i.e. I will most likely construct a basic version and move on to more complex versions such as the snubberized and regulated circuits. Eventually I would like to design a discrete regulator.

Additionally I will likely want to run various amplifier chips and configurations from the supply. I expect my first gainclone to be basic, and then I will move onto more complex circuits. It would be good if I could use the power supply to at least prototype a wide variety of amplifier chips, from the high-power LM3886 so some of the lower-power TDA chips.

My first build will probably not be designed for extremely high power output. My speakers are some leftover onkyo surrounds rated for 8ohm, 130W... not sure what to expect from them loadwise but hopefully I can't blow them up

I would like to create a dual mono configuration, so for that I would need two toroidal transformers, each with dual secondary windings, correct?

I am not sure what voltage to get on the secondaries for maximum versatility. LM3886 says +-35V for max power into 8ohm speakers. Though I have 8ohms now, I could definitely see myself getting lower impedance speakers in the future and not wanting to re-buy expensive transformers (thats the heart of this post).

I was planning on buying 35V secondaries and regulating it down for use with different amplifiers. I have heard regulated supplies sound nice but I will not be able to experiment with unregulated ones if I choose this route. I am not sure what low power chips I want to use either, I was going to depend on the regulator to give me the power I want.

Anyone have insight into the most versatile transformer configuration? Are transformers (easily found, that is) available with, say twin secondaries at 0-28-35 or something like that? The problem is these things are expensive, dont want to buy new transformers for each new project!

if anyone has some insight it would be appreciated! is it feasible to be able to drop 10V through a regulator in a high-powered amplifier? the reg at that stage would be discrete and thus able to be heatsinked, but still, seems like a lot of power.

for now I am going to use all the leftover computer PSUs to try out amp designs on, but I will want a linear supply sooner or later.

thanks for putting up with my long post. I look forward to getting replies and starting my projects!

[AndrewT]

Hi,
keeping the cost down and regulating the supplies of a power amp do not meet.

Buy a 160VA to 300VA 20Vac to 25Vac transformer.
Buy 50V smoothing capacitors. 8off 4700uF to 10mF will do.
Buy bridge rectifiers 25A 200V or higher.
Buy some cheap but large heatsinks.

Try to find special offers and old stock that does not comply with ROHS regulations. Try to keep down the postage costs. It mounts up if you go shopping for a few components often or from different suppliers.

[tryonziess]

Solis,
Poor college student eh! Have you thought about a little visit to Goodwill or the local Mission outlet. You can find a plethora of used electronics at those places -- cheap. There are numerous posts on this forum where members canibalized old amps for projects.
Andrew's remark about the shipping is right on. That is one expense that will eat up a budget.
I have several of the Twisted Pear power modules laying around I could send you for parts if interested -- heck the entire module is complete and working. Google Twistedpear.com and you will see the lm3886 output module. Always willing to see someone start a new hobby. You do realize DIY Audio is addicting. It will not be long before you start building Aleph's and Leach amps and more. Your entire dorm room will be cluttered with boxes full of electronic hardware.
One little bit of advice. Try to learn to program in C. The small ic's from Atmel and Pic offer tremendous opportunities when designing control units. I sure wish I had learned this when my brain was younger. Long long ago.

Tad

[Redshift187]

If you go with a 18V AC transformer, you should be able to handle 4 ohm speakers if you decide to in the future, or a bridged design for 8 ohm speakers.

[AndrewT]

don't consider specialising in chipamps driving 4ohms and forget about bridging chipamps as well. Dead ends that will teach you nothing other than the severely limited current ability of chipamps.

[SoliS]

thanks for the replies. I appreciate it

regulated supplies may not fit with low budget but, being a college student, I have university labs at my disposal... so things like basic transistors, resistors, caps, zeners, and diodes, are essentially free. the university also throws out lots of old electronics which I scavenge. For example I found a pair of nichicon 10,000uF caps which retail about $5 apiece if i remember correctly, not counting shipping. I also have a pretty good relationship with the electrical engineering department's lab manager, so I can sometimes tack small orders (like for 5 transistors or something) onto a larger uni order to save shipping.

I also already have several bridge rectifiers, rectifier diodes, and heatsinks. which considerably reduces cost. and when I build a discrete regulator I'll write it off as studying

my idea is to get an amp *working* with cheap and/or scavenged components, and then I can start replacing parts with higher quality components as I can afford them. Though the uni throws out lots of electronics there arent many audio amps that they pitch, so I will take a look at goodwill. Thanks! and I am aware its addicting, I've been building other types of electronics for a while now, but they do not require expensive transformers/

and @ AndrewT, though they have limited current ability one of my goals is to design a high-powered output stage this year, so perhaps the limited current drive is a challenge for me to overcome.


anyway I want to supply enough power to several different projects including a possible room-sized amp (50-100W per ch) but still be able to supply a low enough voltage with regulation if necessary. Without buying many different sets of transformers.

[SoliS]

again with the university throwing things out, I appear to have found two SMPS which have 25V 3.7A rails!! (Also, some truly MASSIVE heatsinks)

I am thinking of wiring these two SMPS in series (well, "back to back") to get a +/-25V SMPS for my amps! According to Decibel Dungeon these sound quite nice. However I would also like to try linear supplies so your suggestions are still appreciated. At least I can get something working soon without 4 ATX supplies (the ones I got come from power mac G4s that were being scrapped. apparently the g4s need a higher voltage rail for something.)

[pacificblue]

92,5 VA SMPS and massive heatsinks are contradictions. Either the SMPS is very inefficient (=old and/or bad) or you have not seen a massive heatsink yet.

It is unlikely that you will overcome the limited current drive of a chipamp, because that is hardwired into them.

With regards to your idea of starting simple and then moving onto complex circuits. Those power op amps have the entire complex part of the circuit inside, so there is little you can usefully add in complexity to the outside. If you want to experiment with filters and feedback loops, you can do that with normal op amps just the same, but cheaper and easier. If you are after the secrets of building complete amplifiers, use transistors and go step by step from the simple one stage to the common three stage amplifier. But you might as well be patient and wait until you tackle that stuff at the university. That way you learn it right and avoid many mistakes as well as reinventing the wheel.

[SoliS]

Originally Posted by pacificblue

]92,5 VA SMPS and massive heatsinks are contradictions. Either the SMPS is very inefficient (=old and/or bad) or you have not seen a massive heatsink yet.

I'm not sure where you got 92.5 VA from (I never said that in my posts, could be be your computers interpretation of my typeset? We are from different countries.)

I meant that I found some SMPS units IN ADDITION TO some massive heatsinks (I was just excited). The heatsinks are from Power Mac G4 dual processor machines. The processors were horribly inefficient and thus the heatsinks were big - I think I can use them to heatsink a subwoofer amp I am thinking about building. (pictures here: http://www.findgreatstuff.com/359.html thats about a 7 inch cube or so. should be good for cooling chipamps

It is unlikely that you will overcome the limited current drive of a chipamp, because that is hardwired into them.

I mean to say that I will build a separate, discrete output stage (current amplifier) to increase current drive beyond the limits of the LM3886.

With regards to your idea of starting simple and then moving onto complex circuits. Those power op amps have the entire complex part of the circuit inside, so there is little you can usefully add in complexity to the outside. If you want to experiment with filters and feedback loops, you can do that with normal op amps just the same, but cheaper and easier. If you are after the secrets of building complete amplifiers, use transistors and go step by step from the simple one stage to the common three stage amplifier. But you might as well be patient and wait until you tackle that stuff at the university. That way you learn it right and avoid many mistakes as well as reinventing the wheel.

Well in addition to sounding good, I will probably want to play around with the thing. I know they are a very complex and complete circuit inside. As I said, I will add a preamp, output stage, and then work on designing an audio amp. I am currently tackling such subjects at school (op amp design) and as this is my hobby as well as my livelihood it will be a good learning experience to design in parallel with my courses.

[AndrewT]

Hi,
National sell three versions of a driver chip that couple to a discrete output stage.
lme498xx
These will make a far better job than trying to modify the operation of a chipamp.