Designing a Class A/B Push Pull Power Amp

Hi,

It's been a while since I've been in here. After a weird career shift into computer repair as a result of COVID-19 I've kept sane by building some Paraflex 8th order bandpass boxes for the club sound. For anyone who is interested in building some please search Facebook for "High Order Quarterwave Society (DIY Paraflex & Super Planar dev community)". These projects are very rewarding and mine only need some aesthetic touches to be completed. Let's get to my point though..

The 'it' power amps right now are commonly class D which is a can of worms I'm not going into. I prefer class A/B and do not care for switching transistors, weight, or the fractionally less distortion. I could go on but my question is -

How might one lay out a board for experimenting simply with two transistors for the crest and two for the trough? I'm looking for a fundamental building block for an output section only, that can be paralleled to the moon. I need a new rabbit hole.

Thanks!

buildafriend said:

Hi,

It's been a while since I've been in here. After a weird career shift into computer repair as a result of COVID-19 I've kept sane by building some Paraflex 8th order bandpass boxes for the club sound. For anyone who is interested in building some please search Facebook for "High Order Quarterwave Society (DIY Paraflex & Super Planar dev community)". These projects are very rewarding and mine only need some aesthetic touches to be completed. Let's get to my point though..

The 'it' power amps right now are commonly class D which is a can of worms I'm not going into. I prefer class A/B and do not care for switching transistors, weight, or the fractionally less distortion. I could go on but my question is -

How might one lay out a board for experimenting simply with two transistors for the crest and two for the trough? I'm looking for a fundamental building block for an output section only, that can be paralleled to the moon. I need a new rabbit hole.

Thanks!

Click to expand...

There are tons of published schematics of existing (proven) designs. While some posted here for troubleshooting are also proved to fail.

Back decades ago there were popular designs using a single pair of metal (TO3) power transistors that could deliver 75W reliably.

Peavey made truckloads of Class A/B amps rated up to 300W, at higher power levels it becomes cost effective to use Class G/H (smaller heat sink, smaller power transformer, etc).

JR

PS: I am now a fan of class D, but what would i know.

some 5-6 years ago I stumbled upon the LM3886 - up unto that point I always kinda wanted to roll my own amplifier, because I want to cut my own compromizes in parameters

But after evaluating what I could do with this simple chip, I realized the era of homegrown poweramps is probably over.

Since then used and abused this chip for many other power purposes like 3-phase sync spindle drivers and such - it seems these chips are not only good-sounding, but also very, very forgiving when it comes to abuse and heavy use..

/Jakob E.

Yes I could easily search out documents and I own Douglas Self's power amp design book!

I'm reaching into your brains here and no one is questioning your mental agility John! Your experience shows with every key strike Those old TO-3 cans ( I like the motorolla brand ones ) are very nice. My father used to cook the wafers for said types of Q's.

Thank you as well Jakob!

I'm talking about reaching about 3-5kW per channel 2Ohm stable and I'm well aware of why they use class D to get there. I just subjectively enjoy class A/B more.

buildafriend said:

Yes I could easily search out documents and I own Douglas Self's power amp design book!

I'm reaching into your brains here and no one is questioning your mental agility John! Your experience shows with every key strike Those old TO-3 cans ( I like the motorolla brand ones ) are very nice. My father used to cook the wafers for said types of Q's.

Thank you as well Jakob!

I'm talking about reaching about 3-5kW per channel 2Ohm stable and I'm well aware of why they use class D to get there. I just subjectively enjoy class A/B more.

Click to expand...

I was product manager over all power amps at Peavey back in the 90s.

As I already shared Class A/B is uneconomic at high power levels.  The biggest class A/B we sold at Peavey was the CS1200x 600W @2 ohms, we also sold a class A/B PV2000 cheaper design with less output devices similar power criteria.

To cost effectively deliver high power you need to use class G or Class H with multi-level rails.

Back last century switching device technology was not up to the task to deliver serious power... The biggest class D I had in my stable was nominally 1kw (500+500) but just barely.

This century technology has made high power chip sets and devices available inexpensively ( I am jealous... you don't want to drop a CS1200x on your foot).

JR 

JohnRoberts said:

I was product manager over all power amps at Peavey back in the 90s.

As I already shared Class A/B is uneconomic at high power levels.  The biggest class A/B we sold at Peavey was the CS1200x 600W @2 ohms, we also sold a class A/B PV2000 cheaper design with less output devices similar power criteria.

To cost effectively deliver high power you need to use class G or Class H with multi-level rails.

Back last century switching device technology was not up to the task to deliver serious power... The biggest class D I had in my stable was nominally 1kw (500+500) but just barely.

This century technology has made high power chip sets and devices available inexpensively ( I am jealous... you don't want to drop a CS1200x on your foot).

JR

Click to expand...

I would really like to focus on Class A/B. I understand the basic differences

buildafriend said:

I would really like to focus on Class A/B. I understand the basic differences

Click to expand...

OK one difference is that class A/B is impractical for 3-5kW power level.

Have fun.

JR

JohnRoberts said:

OK one difference is that class A/B is impractical for 3-5kW power level.

Have fun.

JR

Click to expand...

Yes, it draws more current, uses less consistent components, and is less temperature stable.

Just putting Sziklai pair's on the output of an op amp makes a great power amp. From there it's all about heat sinking and managing offsets and protection.

For quick and easy, just get an MX50 SE kit for $16 on Ebay. I've been using those for years and they're solid and clean. They're just a discrete Sziklai current feedback design IIRC.

As for Class D, I think like a lot of switching technology like SMPS, my impression is they have come a long way. I would not be surprised in the slightest to discover that a modern Class D chip is superior to just about eveything else and the only thing holding it back are people who are emotionally invested in "classic" designs.

buildafriend said:

Yes, it draws more current, uses less consistent components, and is less temperature stable.

Click to expand...

Just as a mental exercise try to find a 3-5kW class A/B power amp made by any manufacturer. If you can find one copy that, but I wouldn't get your hopes up.

At Peavey last century I sold a 3kW class G/H.  UL wanted us to put a camper plug on that one because of mains current draw, but we negotiated them down. 

A 5kW class A/B will need its own tie in.. (exaggeration but you won't be able to use a common IEC plug set).

JR

squarewave said:

Just putting Sziklai pair's on the output of an op amp makes a great power amp. From there it's all about heat sinking and managing offsets and protection.

Click to expand...

The problem with that approach is that you can only go so high voltage wise with opamps, usually not more than +/-24V like the OPA604, you can raise the voltage capability by using bootstrapped transistors a'la MCI JH500, with an OPA604 you can virtually run it on +/-48V power supplies with bootstrapped transistors, power disspation of the chip may be an issue.

Regarding the OP question: 3-5KW on an AB amp is quite a challenge, you would have to parallel a lot of output transistors, massive power supplies and some massive heatsinks, and don't forget protection, you wouldn't want 200V DC at your speaker terminals if one transistor shorts out.

The way to go is class D, I've seen 150W per channel class D amps using GaN FETs that do not even require a heatsink at all, the FETs are tiny and SMD, I couldn't believe my eyes when I first saw them.

There are a lot of class D amp chips, you just need to add the inductor(s) (in some cases they are inductor-less), and the supporting circuitry around them. Only downside is that many of those chips come in non hand soldering friendly packages.

My advice is to start with low power at first, a 20-ish Watt amp is a great learning experience.

If you haven't joined diyaudio.com, now its a time to do it.

Ah, I missed the 3-5kW part. For that anything I would think Class D would be the way to go. And the power supply is probably even more important than the amp.

JohnRoberts said:

Just as a mental exercise try to find a 3-5kW class A/B power amp made by any manufacturer. If you can find one copy that, but I wouldn't get your hopes up.

At Peavey last century I sold a 3kW class G/H.  UL wanted us to put a camper plug on that one because of mains current draw, but we negotiated them down. 

A 5kW class A/B will need its own tie in.. (exaggeration but you won't be able to use a common IEC plug set).

JR

Click to expand...

The Crown Macrotech 5000VZ. See reference manual pg 36

buildafriend said:

The Crown Macrotech 5000VZ. See reference manual pg 36

Click to expand...

well you got the power level correct.

Not a basic class A/B, more like a driven rail with multi-level power supply... Not literally a class G/H but functionally similar.

Have fun DIY one of those beasts.

JR

buildafriend said:

Yes, it draws more current, uses less consistent components, and is less temperature stable.

Click to expand...

In addition to what's been already said, what transistors are you going to use that handle 300V Vceo?

abbey road d enfer said:

In addition to what's been already said, what transistors are you going to use that handle 300V Vceo?

Click to expand...

A whole lot of these https://www.mouser.com/datasheet/2/308/FDA24N40F-D-1806446.pdf ?

buildafriend said:

I'm talking about reaching about 3-5kW per channel 2Ohm stable and I'm well aware of why they use class D to get there. I just subjectively enjoy class A/B more.

Click to expand...

Why on earth would you need that much power?

Cheers

Ian

user 37518 said:

A whole lot of these https://www.mouser.com/datasheet/2/308/FDA24N40F-D-1806446.pdf ?

Click to expand...

Exactly: MOSFET's.

abbey road d enfer said:

In addition to what's been already said, what transistors are you going to use that handle 300V Vceo?

Click to expand...

Something like this?

https://www.st.com/resource/en/datasheet/bu508aw.pdf

Matador said:

Something like this?

https://www.st.com/resource/en/datasheet/bu508aw.pdf

Click to expand...

with 1'C theta Junction to case, plan on using a lot of those on a forced air heatsink. Of course you'll need a bunch in parallel just to get the current needed.

JR