Multiplier car amp

As I have said earlier, I need to start making car amps to not only get extra money for my projects, which are many and diverse, but all very expensive, but to serve my community, to offer all the "bumpin' dudes" something that can thump and also has high fidelity and reliability at a fairly low cost.

I have not yet attemted to make a switching supply, as I have tought about doing. It is not easy to aquire toroidal cores for switching trannies, and they are tedious to wind, also, they inject nasty stuff into the rails, and compromise reliability.

But perhaps I could do it this way: First run the 12V through a bridge, thereby making polarity irrelevent to the user, then through a large inductive filter and capacitor to filter the noisy incoming power. This would then run through an H-bridge controlled by an oscillator at a relatively low frequency, less than 1kHz, perhaps only 100Hz or something. The output from this H-bridge would then go to a simple Cockroft-Walton multiplier with maybe 1000uF capacitors. This output would then be filtered by large reservoir caps like in any home audio amp (but not really big ones, maybe just 5600uF or something, because the power would be a square wave and would have relatively little ripple).

Theoretically, this power scheme would far outdo a standard switching supply, though it may be somewhat larger and cost more.
One thing tough, it would not have isolation in the same manner that a transformer does. The multiplier center tap would not be the automobile ground though, the - rail would, this might be acceptable though, I'm not sure. I realize the input of the amp would definately have to be isolated. There could be a large voltage at the speaker output relative to the vehicle ground anyway. But I believe that is the same way in a normal car amp right? Isn't the center tap usually vehicle-grounded?
They step up much like a transformer. If you multiply the voltage by 6, you will get 1/6 the current. Ripple does go up a lot with each stage though, so I'd have to use fairly large caps in the stages. If my math is right, I calculated the ripple @ 0.02V @20A @1kHz with 6 5000uF stages. Not only could the stages easily have 10,000uF or more, and possibly higher frequency, but they would not be run at 20A anyway.

Even if I can't get it to work well for audio, I can just use a transformer, it would the be just a low frequency switching supply. It would be really heavy and expensive though.
Don't rule out switchers

If you are going to use a chip like the LM3524 SG3524 then you will have a switcher with plenty of noise, but if you have been following there are quite a number of newer switcher chips with very low noise, and these are (ANATHEMA) making their way into instrumentation applications.

Switcher transformers aren't tedious to wind either -- the manufacturers have design software which specifies the inductance of the primary -- see National Semi's web site for examples -- and since the frequencies used these days are 50kHz and up.

I've built several switchers for various applications from busted computer power supply torroids -- not an original thought to my, some young engineering student at CalTech pointed it out to me.

Figure this too -- what's the ambient noise level in the quietest car sold in the world (hint, it's a Lexus)?
You don't have to use a toroidal core. There are plenty of EI cores with excellent specs for switching supplies.

They are not that expensive either.

A switching frequency of 100 Hz or so are more likely to be heard through the system than a switching frequency of 50 kHz. Not to mension that you don't get the benefit of smaller caps with 100 Hz that you get with higher frequency. Also a real switching power supply is regulating its output - something you would miss in a multiplier.

I know what I'd put my money on :D

The multiplier circuit can be ruled out because it will be too lossy and will not be able to respond fast enough. Don't think about running it at a low frequency because the transformer will be huge. The cross section of the core will need to be large. I have some toroidal ferrites, Kool-mu, MPP, and high flux MPP which can be used for transformers and inductors. A switcher will have to be regulated to provide power for an audio amplifer. This mean a PWM supply of some kind. The hardest part of about switchers is avoiding flux imbalance. Don't think that switching supplies automatically put crap into the supply rails. If done properly the rails can be pretty good looking. It is all a matter of sound engineering practices. You can email me if you need some help or have specific questions.

It is good that you brought that topic up because sometimes I take that stuff for granted. When you use a switching power supply, the seconaries must be rectified (whatever scheme is selected half, full-wave) with fast switching diodes. Regular diodes cannot switch off fast enough and will cause noise and spikes and more importantly will not block current in the reverse direction. It takes a finite amount of time for the charge carriers to be swept across the depletion region of the diode (the E-field region between the P and the N material.) The diode is still in forward conduction mode until all of these carrier have been forced out of this region by the existing E-Field. Once it has, it is blocking reverse current and hence rectifying. Slower diodes take a longer amount of time to sweep these carriers out of the region than the period of the switching frequency. A good diode for the application will have a recovery time 'trr' in the lower ns region. There are some diodes available in three leaded packages,such like transistors (TO-247, TO-220), that contain a pair of connected diodes for a full wave rectification scheme. Intersil, OnSemi, InternationalRectifier make these types of devices. The blocking voltage should be more than 2.5 times the peak secondary voltage.


I'm currently experimenting with a multiplier circuit for a portable 3way active sound system.
My intention is to build a stereo system that works a whole night from a single car battery for our up on the hill parties.
At the moment it works at a switching frequencies of approx. 10khz, coupled by 1,000uF and 6,800uF reservoir capacitors.
It is setup as a bridge providing +-12V symmetrical to ground.I used Schottky diodes for the rectifiers and BUZ 11 Mosfets for the bridges,where the upper two are bootstrapped to get them fully on.
The quiescent current without load is 15 mA ,wich is pretty low compared to the 1 - 3 A idling current of a normal car amp supply.
The stability at 5A drawn across both outputs is more than sufficient.
Of course I looked at that project. ESP is always the first place to turn:D There is no PCB available though. I might just try it anyway though (design my own PCB) if I can find toroid cores. I'm sure it would help if I read a book on switchers first. One has been recommended in another thread, but it costs like $90 or something.

BeanZ, where did you get your toroids?
They manufactured by Magnetics Inc. You can also buy them from AllStar Magnetics like I did.

There may be a minimum order for some pieces. I have a few that you could buy if you have a design in which they will work.

The ferrites I have are P-material and are coated with grey 1000V GVB . The part number is ZP44813-TC (the important thing is 44813. Z=grey coating, P=material, ). The cross sectional area is 1.15 cm^2. The window area is 2.85 cm^2. The dimensions are 1.5 inch OD, 0.75 inch ID, 0.5 inch height.

If you want to try and design your transformer around this core you can use 800-1100 Gauss. Good Luck

It can be enough, 300W is not a whole lot. The issue will be with determining if you have enough window area or not. Since the primary is about 10V-16V, you use the Vin(min). For 300W you will need a lot of cross sectional area for the wire on the primary side for 30A. This may take up most of the winding area for the toroid. I can run some numbers, let me get back with you.....

Learning about switchers

There are 3 great web-tutorials on switchers -- look at National Semi, Texas Instruments and Linear Tech. Also, <em>QEX</em>, one of the ARRL publications, is running a series of articles on Switch Mode PS. The series just commenced with the current issue.

I don't know that On-Semi (was Motorola's discrete's biz) has a tutorial, but they provide PCB layouts right within the product folders. On-Semi has Switchers must be built on PCB's -- you can't even prototype them on protoboard, vectorboard etc.

Lastly, what you don't (necessarily) see in the ap-notes are the chokes etc used to reduce radiation. For this reason the construction articles in QST and The Radio Amateur's Handbook are very helpful. (They will also tell you how to wind a torroid.)