advertisement


Arcam Class G amps - Class A

PhilDick

PhilDick

pfm Member
I asked a question about this in the 'all amps sound the same' thread but it was a bit off topic. However I would like to hear from our techies about this.

I have just acquired an Arcam a49 and to me it sounds fabulous.

Now Arcam claim that this amp and others run in class A up to 50 watts. however, looking further into this they state:

"The A49’s output stage includes a proprietary error correction circuit that modulates the modest standing currents in the output stage and ensures a near-constant output impedance for peak currents of up to about +/- 4 amps, corresponding to well over 50W in to 8 ohms. The A49 thus behaves exactly like a classical class A amplifier up to this power level in terms of performance but without the heat penalty"

How much of this is actually true - does it really behave like a Class A amp? it doesn't get hot!

Arcam's doc:
https://www.arcam.co.uk/ugc/tor/p49/Technical Notes/The_technologies_behind_the_A49_101014_A.pdf
 
They absolutely do not claim it runs in class A for the first 50W and neither it nor most similar amps do anything of the sort!

I'm not going to answer the OP's question any further I'm afraid as a proper explanation would get VERY technical and probably the only one's who would understand it would be the other engineers and skilled DIYers on pfm who probably already understand it...

I will say that IMO only hot running climate destroying class A is the real thing as there are a multitude of advantages to class A beyond merely having no crossover distortion and most of the "clever" "cool class A" ideas address only this point. Error correction is a different thing.
 
I only know classes A to D.

Is Class G an actual thing or is it a made up name for their particular system of error correction?
 
Even though Arkless thinks I'm too thick to understand ;) I know that Class G amps are ones with 2 or more power rails. When demand dictates the higher power rails switch in giving more power output.
 
It is not class A operation! Class G is a way of saving money and weight in high power amplifiers, yes with two or more voltage rails. Class H is very similar and they can be considered "two sides of the same coin". Although the first use of it I can think of was in a hi fi amp, Hitachi "Dynaharmony" series, 1978, it is usually used in PA amplifiers and can be considered a negative in hi fi amps though if done well it needn't make things noticeably worse.
 
its a totally bullshit claim. "Class A alike operation through error correction upto 50 watts would be factual" but then we have to assume class A alike has meaning.

There's no possible means for that amp to dissipate the heat generated by 50 watts of class A operation. Theyre likely using feedforward error correction and a sliding/switching bias and rail voltage scheme that preceeds the signal power requirements, but it aint the same thing.
 
PhilDick said:
Yeah that's what I thought. How do they get away with it? They are still claiming the same on one of the current range:

https://www.arcam.co.uk/product,hda,power-amplifiers,pa240.htm
Click to expand...

They state idle power from mains as only 90W so it ain't class A to 50W as they lie.

An interesting possibility with class G though (and a design has been published by Doug Self) is that because the extra voltage rails are only switched in when needed their voltage does not combine with the static bias current of a real class A section powered by the lower rail voltages.

Why is this so interesting?

An amplifier with voltage rails high enough for lets say 200WPC is incredibly inefficient when set to give the first Watt or two in class A! To give only 1W in class A it would need to get much hotter and use much more power than a 15WPC pure class A amplifier that runs out of steam at its rated 15W. Much? Prob X 3 ish!

However... with class G it is perfectly feasible to have a genuinely class A amp of say 15WPC powered by the lower set of voltage rails and for the second higher voltage rails to switch in at "15.01W" and then take it up to 200W output. Up until that "15.01" switching point the 15W class A amp would function as a perfectly normal proper class A amp and not even "know" it is part of a bigger amp of 200W. Assuming you are not using more than 15W then it runs at the expected heat and power consumption for a 15W class A amp.. cos that's exactly what it is!
 
It's error correction. No mystery and not class A. Class A vastly improves almost every aspect of an amps measured performance other than of course efficiency. Whilst nothing to do with class A, error correction can improve some aspects of the measured performance in terms of THD etc. If we were to arbitrarily say there were 10 aspects of performance to be taken into account in judging output stage performance then I expect that the marketing department has caught wind of say 3 or 4 of these being improved by a similar degree to class A operation and has let their imagination run away with itself!

Beware that the LM4702 IC at the heart of the A49 has been made obsolete fairly recently. I don't know what stocks Arcam has bought up as spares but it may well be difficult to obtain one to enact a repair in the near future.
 
So not class A. I did learn that Charlie Brennan is/was the md at Arcam. He of one box simple server device fame.
 
What is it that rega does? They say it is a class A emulated output stage. I know their amps get hot.

Below is description from Elicit-R

"The output amplifier used in the Elicit-R was born after extensive research by our engineers to develop a low source Class A driver stage, based around a complementary pair of Darlington output transistors, forming an emulated Class A driver stage."
 
You must log in or register to reply here.


advertisement


Back
Top