Modifying the SC480 for 2 ohm load

[guitar_joe]

Hey guys i've been mucking around with a switchmode in class and i wanna build it into a car amp.

I've built an sc480 and was thinkin i could add a few more output transistors and make the thing stable into 2 ohms.
I've already read in silicon chip running higher rails is not a good idea as the transistors max voltage will be quickly exceeded.

Obviously i realise with power amplifiers the way to increase the output power is to run a higher supply voltage OR run lower impedance loudspeakers yeh?

So just as long as i have enough output devices so a 2 ohm load doesnt exceed the SOA of the transistors and the drivers can supply enough current there should not be any problem yeah?

Currently the amp has 4 output devices 2 pairs of tip3055 and tip2955 complementaries

[guitar_joe]

So..... will it just.... melt???
Should i give it a go???

[pinkmouse]

Okay here's a few things to think about.

You will need a much higher current supply, so as well as adding more outputs you will need to redo the PSU and up-current all the PCB traces. If the amp has protection, you will need to recalculate and possibly redesign the circuit. Will the driver/Vas stage cope with the extra output required to drive the extra output trannies? If not, that will need to be reworked as well.

These are just the first things that come to mind, I'm sure if I thought about it for longer I could come up with more. If you are capable of dealing with the above, then go for it, otherwise pick a more suitable amp if you don't like the smell of charred silicon.

[AndrewT]

Hi,
the SOA of the output devices has to exceed the loadline of the speakers. Similarly the SOA of the drivers has to exceed the loadline of the speakers reflected by the outputs to the driver emitters. It's not just current but peak output current times a significant percentage of the supply rail. This percentage is already past halfway when the load presents a 45 degree phase angle and is approaching 75% of Vrail when the load is 60degree phase angle.

I suspect that 2pair of those puny output devices cannot cope with a 4ohm reactive speaker load. Doubling up the output devices might be a great way just to drive 4ohm properly.
This will very slightly reduce the loadline presented to the drivers, so you get a second advantage.

[Stuart Easson]

Hi,

The output transistors you mention are relatively weak, you can use the same number of far more capable devices and get much closer, if not all the way to your target.

For instance, the mj15003 & 4 pairs are more than twice as capable from a power, current and voltage perspective. They are also close enough in terms of their other characteristics to think they'd work without changing the rest of the circuit. If the to3 package is an issue consider the mjl21193 & 4, probably better transistors, slightly lower power handling.

Depending on the rail voltage present while trying to drive a 2ohm load two pairs of these transistors may be all you need...on a big heatsink...

Doubling up the 3055s will definitely help, but probably not enough.

HTH

Stuart

[guitar_joe]

Thanks for the replies guys.

For those unfamiliar with the circuit it uses a +/- 40 v rail.
Input stage is a diff pair of 557's with a CCS.
VAS is a diff pair with a current mirror.
Bias is supplied via a vbe multiplier
Drivers are mje340 / 350
Output stages are tip3055 / 2955 ( 2 complementary pairs )


Redesigning the pcb was in the back of my mind and increasing the track width for the high current tracks was definately already considered.

As far as the power supply is concerned it will not be a restriction to output power.

In the article for the amplifier it says the output into 8 ohms is 50w RMS and output into 4 ohms is 70wRMS. What restricts it from being closer to 100w RMS, is it the power supply or the current gain in the output, or something else?

As far as output devices go Mjl21193/4 or 15003/4 seem like a good idea, but i'd say 15003/4 are better cause im only going to use +/- 40 rails. Using mje 340 / 350 drivers does this seem more feasible?

I have read about the difference of a resistive load and a real (inductive/capacitive) load and i've seen the difference between the two. If anyone's got a link to more info it would be great as i want to learn how to draw all the complex impedances etc onto a load line etc.

thakyou for replies.

[Stuart Easson]

Hi,

Without looking it's hard to know, it could be the drivers and outputs current limiting (or a protection circuit), but I'd guess the power supply is the limiting factor. The current required for a 4ohm load is within the short term capabilities of the transistors, if not the long term, the TIP package is quite small...

Ignoring SOA issues: With a perfect 40v power supply etc your maximum peak current is 20A into a resistive 2ohm load, with 3 pairs of mjl21193 you get ~6 amps each. On-semi show a typical gain of 40 at 6A, so you'd need 500ma peak from the drivers, probably too much for an mje340, you might consider the mje15028 and it's partner.

There is a spreadsheet floating around that, suitably primed, performs SOA calculations, and will show you what safety margin a particular set of transistors offers with different loads, both resistive and reactive.

HTH

Stuart