Why 12 power transistors in a class AB (?)120 W (?) amplifier?

[solderdude]

It is to stay within the SOA of the transistors so their max. power ratings are never reached.
SOA stands for Safe Operating Area

At any voltage across the transistor a specific current is allowed. This is the SOA.
When amps must deliver high output power (so high voltages and high currents) then 1 pair of transistors won't cut it.
You will need more in parallel that share the current at the same voltage so the transistors aren't overloaded by current peaks.
The heat generated in the transistor 'chip' itself needs to be drained. That takes some time. Spreading this over a larger surface also helps.

 

[tomtoo]

I will fix that. Thanks for the tip.

Uhhps sry NO. Its why they used more "iron". Keep it thermal stabel.

 

[tomtoo]

Uhhps sry NO. Its why they used more "iron". Keep it thermal stabel.

The average hifi amp is not thermal stabel, they put out the rms power for some minutes, then they have to go into thermal protection. This older PA amps had to run at full power hours. If not days. So the overall design was more conservative.

 

[DanielT]

Uhhps sry NO. Its why they used more "iron". Keep it thermal stabel.

Ha ha, I thought you meant that more iron was needed on my Sela amp, that it was undersized while you meant that more iron is needed in general if an amp is to run like the train 24/7.

 

[tomtoo]

Ha ha, I thought you meant that more iron was needed on my Sela amp, that it was undersized while you meant that more iron is needed in general if an amp is to run like the train 24/7.

Exactly. Thats why they are heavy in relation to power. And get called ironpics. But that name shows some respect.
Modern class d amp with the same wight is in the multi kW range but with fans.
You know, reliabel and heavy.

 

[sq225917]

Let's see how many class d PA amps are still running in 40 years. I bet none of them without significant repair. Smps just ain't built for the long haul in consumer goods.

 

[sarumbear]

I called Sela and asked them if they had schedules for it but they said that in the 80's there was a fuc*'' g quarrel between their amp designer and the owners. The designer became so angry that he resigned immediately and with that he took with him all the drawings, schedules. Since then, Sela has never manufactured any, in - house, amplifiers.

They never got hold of him after he left Sela.

I’m sure the argument was about cost. After owners saw the cost of the design asked the designer if he is effing mad? He was supposed to build a consumer device not an industrial quality one, which is what I see. It’s an overkill, but nothing wrong with overkill if you can afford it.

Maybe the designer joined Krell?

 

[ZolaIII]

Old 2N's 2N3773 (15~25W) in a cascade, made some bords with those when I whose rather young. Nothing to understand there really.
Edit: reason for a such design is because those where really wide spread/cheap and available (almost anywhere). You can buy them easily enough even today.

 

[DanielT]

I’m sure the argument was about cost. After owners saw the cost of the design asked the designer if he is effing mad? He was supposed to build a consumer device not an industrial quality one, which is what I see. It’s an overkill, but nothing wrong with overkill if you can afford it.

Maybe the designer joined Krell?

Note, just so there is no misunderstanding, Sela is very good at what they do. Good service. Recommended.

It may not be the easiest thing to combine the commercial, what the market is willing to pay with something that a good designer wanted to do.

 

[DanielT]

Old 2N's 2N3773 (15~25W) in a cascade, made some bords with those when I whose rather young. Nothing to understand there really.
Edit: reason for a such design is because those where really wide spread/cheap and available (almost anywhere). You can buy them easily enough even today.

Do you know if 2N3773 is still used in PA / Hifi, or have newer transistors with better performance replaced them? Curious ..

 

[ZolaIII]

Do you know if 2N3773 is still used in PA / Hifi, or have newer transistors with better performance replaced them? Curious ..

That (N2) is not stellar performer (just average) but it whose quite a work house and widely used especially for a stage (active mixers/amplifiers and cetera) equipment. Biggest advantage they arguably have even today is how easy is to get to some of those that are still in working condition (patch, replace and carry on). There are far better solutions from the end of AB class era such as late STK higher end first gen hybrids for instance.

 

[DanielT]

That (N2) is not stellar performer (just average) but it whose quite a work house and widely used especially for a stage (active mixers/amplifiers and cetera) equipment. Biggest advantage they arguably have even today is how easy is to get to some of those that are still in working condition (patch, replace and carry on). There are far better solutions from the end of AB class era such as late STK higher end first gen hybrids for instance.

Thanks ZolaIII.. It was interesting information.

What do you think about this, my observations. Is there something in this:
Wondering if better transistors were created around the late 70s and early 80s, compared to how they were before. If I look at technical data on medium, even the slightly better models of amplifiers, from the beginning to a bit into the 70s, it is often stated 0.5% distortion. But somewhere around the 80s, the distortion levels seem to be falling. Now I do not know if it is a coincidence that I happened to see the amplifiers that show the mentioned level distortion or not. Maybe you measure in different ways at different eras, different measurements and accounting standards, that is?
Maybe due to something other than the transistors, reduced distortion levels?

Edit:
What you mention:
"There are far better solutions from the end of AB class era such as late STK higher end first gen hybrids for instance."
Did this start in the early 80's, the improvment, and did transistors (and hybrid) became even better, more refine in the coming decades?

These are beginner questions, but we are in the right section of the forum, so.

 

[ZolaIII]

In short engineering - testing - back to drowning board process trough whole development cycle.
Those late STK's had a spec's similar to today's top D class circuits (second gen D class hybrid). You for instance still have plane AB class transistor amps which perform admirably enough(better than cheap class D one's) especially considering the price (for instance Sanken ones in Yamaha Amp's and active speakers).
Even today we usually take up to 1% THD power rating for power amplifiers as acceptable.

 

[DanielT]

In short engineering - testing - back to drowning board process trough whole development cycle.
Those late STK's had a spec's similar to today's top D class circuits (second gen D class hybrid). You for instance have still have plane AB class transistor amps which perform admirably enough(better than cheap class D one's) especially considering the price (for instance Sanken ones in Yamaha Amp's and active speakers).
Even today we usually take up to 1% THD power rating for power amplifiers as acceptable.

Superb information ZolaIII.. Thanks.

Just when you posted, I created a new thread:

Historical expose. Development of better performing transistors for hifi.

Anyone know a little about it? Now I think from a historical perspective. For example, decade or year Sanken transistor Y which enabled better... Year / decade came hybrid Z. The origin of thread wire comes from this thread where ZolaIII writes about, among other things, the transistor 2N3773...

audiosciencereview.com

Edit:
STK you say...
Check attached photos. I bought that receiver for almost no money at all at the beginning of the summer. The owner / seller had intended to throw it away.

My mother has been using it now. Works just as it should.

People, I just say..Throw away fully functioning stuff. Wear and tear mentality ..

https://www.hifigoteborg.se/Luxor%205082%20R%20%20spec.htm

Edit:
Certainly early version of STK in the Luxor. But that does not matter. I do not think my mother will be sleepless if she finds out that she has an older version of STK in her receiver.

 

[ZolaIII]

Well in second half of 80's to early 90's lot of Japs (aka Denon, Onkyo, Martinez and so on) among many others used them (sadly usually early lower performance one's).

 

[audio2design]

Partially spreading thermal loading as noted SOA can come into play but even a single unit of 2N3773 is rather robust.

Second issue is maintaining high hfe. hfe drops quickly at high current. Parallel devices and you avoid that issue keeping the drive circuit simpler and/or distortion lower.

 

[Graaf]

Facts:
Power amplifier brand Sela:
Power, I guess 2x120 W (in ...? Ohm).
Transistors: 2N3773 (6 per channel)
Electrolyte capacitors in the power supply: 4700 uF, 63 V (times2, per channel, so 9400 uF, 126 V per channel),
Fuse, see picture
Class amplifier? No idea AB, maybe ..

I would guess it's a class A. The output stage seems to be using npn transistors only.
Class A produces much more heat than AB. So thermal management is the best candidate to answer your question, as mentioned in other replies.
Cheers

 

[DanielT]

I would guess it's a class A. The output stage seems to be using npn transistors only.
Class A produces much more heat than AB. So thermal management is the best candidate to answer your question, as mentioned in other replies.
Cheers

Thanks for input Graaf.

Clearly interesting if that is the case. But class A and PA, well.. Then I really understand if there was a quarrel between the designer and Sela.

It may have been manufactured by Sela (the PA company) but the idea was that the model would be sold to Hifi at home. Unfortunately, I know nothing about that.

In any case, it dont get particularly hot when I use it. Of course the volume on it, I need to turn up a bit to get normal listening volume..hmm.

Now I will definitely take it to my friend with the occciloscope so he can measure the effect where it is driven into the clipping.

Edit
OT. Class A by the way that class I associate with the past. Hopelessly low power single end tube amplifier. Krell class A transistor power amps that weighed lead. Nowadays, a well-designed class AB amplifier has such low distortion levels that you no longer need to fire for the crows.

Was it not Douglas Self (guru who wrote Audio Power Amplifier Design Handbook) who said that he would even rather take a class B amplifier than a class A, because of the generally low power in a class A, increased risk of driving into clipping, among other things?

 

[solderdude]

Points to class-AB. When it was class-A and would be idle for 10mins the heatsinks would become hot.
Some AB amps had idle currents as high as 100mA which also would generate some heat in idle.
In general idle currents are around 30-50mA per transistor. There are exceptions.

 

[tomtoo]

We need someone with more knowledge of amp topology.

How do you build a class ab with only one sort of transistors?