100W peak 4Ohm

[zeus_threat]

Simple question am planing to build a near 100W peak 4Ohm amp. PSU is +-30V unloaded. Just to check if my calculations are correct for power dissipation, will 2 pairs of paralleled TIP2955/3055 be enough(180W total)?

Thanks

[AndrewT]

Hi,
80W might be possible from your +-30Vdc supply.
a guide for device power is around 2 to 3times output for each half polarity.
This indicates that around 160W to 240W of device power on each rail.
2pair might just about work.
Now do the full SOAR for the temperature and heatsink and smoothing and transformer and load reactance and see if two pair is really enough.

[Andy L. Francis]

Look for the SOA graph of the output BJT's! If you don't exceed 100mS SOA in case the load has 45...60° phase angle, and graph is derated to ~70°C case temperature (max. ~100°C junction temperature) then you're OK.
For your amplifier (max. ~44Vpp supply, ~60W-RMS on 4 Ohm load, max. ~6...7A output current) two pair of TIP2955/3055 is enough. Even for PA applications!

[Andy L. Francis]

Oh, sorry Andrew I sent my post too late so I couldn't read you comment.
Your post already includes all the necessary informations for Zeus Threat.

Anyway two helps are better than zero help.

[nitrate]

Hi

Welcome and glad to see you've choosen to build your own design.

You say that your aiming for 'near' 100 Watts. I'm glad you say near as your supply is going to be a little low. Assuming you end up with a design that is capable of rail to rail output and asuming your PSU voltage dose not droop much then your already cutting things close.

30V / 1.42 = 21.12Vrms
21.12Vrms / 4R ( load ) = 5.28A
21.12V * 5.28A = 111.5 Wrms

Even a slight drop in voltage of a few volts will bring this figure tumbling down. e.g a 5V drop already gets you down to 75Wrms

This in mind i would try and bump up the PSU by at least a few volts. Anyway about the trannies. According to the data sheet that i will attach below you are allowed 115W each.

Assuming that you build a class A/B thus rendering the bias current dissipation negligable then you will have:-

30V / 4R = 7.5Apk
7.5Apk / 2 ( shared current ) = 3.75Apk @ 30V

The above has to be checked for SAO operation limits.

Again:-
30V / 1.42 = 21.12Vrms
21.12Vrms / 4R = 5.28A
5.28A / 2 ( parallel pair share ) = 2.64A each
21.12Vrms * 2.64A = 55W rms power dissipation across each device. The RMS power can be pretty much considered to be what you are going to get if you drive your amp at max befor distortion.

30V * 3.75 = 112.5 Watts of potential power dissipation accross each device if the amp is abused. design for this level. According to the data sheet you are just about safe at this level. I personally would think that 2 devices are enough as long as you dont plan on driving the amp into hard cliping for extended periods. If you do intend to ensure abuse will not damage things then throw an extra pair in ( or two if you use a flimsy heat sink )

This of course is MY opinion and i apologise in advance for any errors but this is how i usally calc things.

Good luck!
Leigh

[zeus_threat]

Hi and thanks for all your replies. Well sorry i forgot to mention this amp will drive a music sub.My calculations i close to yours nitrate as i ended up with 360W total dissipation but let me explain this is in case i have an ideal power supply which will never collapse. Real life calculations for 2 tip 2955 and 2 tip 3055 is less than half of 360W ( sorry can';t tell the exact figure dunno where is put my papers again must be somewhere in my mess) Well nitrate you sent me datasheet for 2N3055/MJ2955 i know these are 115W am not using them cause i got counterfeits of these being sold here. The TIPs seem safer i've used these successfully elsewhere. So to be on the safe side i'll add another pair of output which means 3 tip2955 and 3 tip3055. Am attaching the schematic here i case you want to have a look and see anything wrong please do report.

Thanks

[Andy L. Francis]

Your amp should work, but I've got some comments if you don't mind!

C7 might cause oscillations, be careful with paralel caps in the FB path. Don't use if the amp is stable without.
For C6 first I'd try 47pF.
R20 could be a the standard 470 Ohm resistor.
About Q5: I'd use BD140 instead. TIP42C has low gain and low Ft, but high Cob.
C13 looks too low for me. I'd use 470pF.
VR1 sholuld be a 470 Ohm resistor in series with a 470/500 Ohm trimmpot. Safety reasons.

Nice circuit anyway! I like too the conventional Lin-topology.

[nitrate]

Hi,

I wasnt thinking straight after a boozey sunday afternoon LOL. My calculations are not correct for a class A/B. I was confused after spending most the day fiddling with my latest classA design were the high standing currents must be taken into account. As your going for a class A/B then disregard my previous calcs. The only thing you need to do is check the peak current against the SOA as i previously showed.

Your actual peak dissipation in the trannies is going to be

15v / 1.42 = 10.56Vrms
10.56 / 4R = 2.64A
2.64A * 10.56 = 27.88W
27.88W / 2 ( pair ) = approx 14W of heat dissipation each.

The reason for this is we forget that the closer to the power rails we get the less of a volt drop is seen across the tranny so although current is increasing the power actually falls after the half way point. This is why the likes of classD are so efficient and run cold.

Therefore under normal running i must conclude that the origanal parallel pair will be enough as long as the peak current shared between them is within the SOA for the given predicted temp of the heat sink ect.

I apologise for my errors, its only now i'm at work and sober that i see clearly once more LOL. P.S i like your choice of tip41/2 for drivers, I personally like to use them too, they are very linear.

Leigh

[zeus_threat]

Thanks leigh and andy. Well leigh no worries its mostly whem am in those states that am inspirational bout what i build lol but thats another story all by itself. I'll do the recommended changes you mentionned francis those you are mentionning for C7 and C13 were my initial design goals. I'll do the other mods as well and keep to the original 2 pair of output.

Just one more question guys before i jump into pcb design, regarding dc offset, would you recommend (1) a 100Ohm preset inserted between R1 and R2 with center preset going to Q13?
OR

(2)a series 100ohm preset in series with R20?

I know you will end recommend (1) but what would be wrong in doing (2)?

Thanks

[AndrewT]

Hi,
these dissipations apply to resistive loads.
The stress in the output devices is much higher when feeding a reactive load than driving twice as much power into half the intended load resistance.
You must check for temperature de-rated SOAR into a reactive load.