The Class - H Amplifier

Jump to Latest
andrewlebon

andrewlebon

R.I.P.
Joined 2008
dear sir
greetings thanks for your reply i make amplifiers only as a hobby no sale
just to learn and gain knowledge please can you tell me some books on h class amplifiers
to get an insight how h class works
1 where do you get your parts from as h class parts arevery hard to get.
2 do you have a shop in your city as i visit your city often wish i could see the great
workhorse i admire your knowledge
thanking you
andrew lebon
 
kASD

kASD

Member
Joined 2004
andrewlebon said:
dear sir
greetings thanks for your reply i make amplifiers only as a hobby no sale
just to learn and gain knowledge please can you tell me some books on h class amplifiers
to get an insight how h class works
1 where do you get your parts from as h class parts arevery hard to get.
2 do you have a shop in your city as i visit your city often wish i could see the great
workhorse i admire your knowledge
thanking you
andrew lebon
Click to expand...

Hi Andrew,

Class-H is not covered by any text book yet. You can learn a great deal by experimentation and also by glancing on various schematics. Parts availability is a problem in India, we source our parts from Delhi for samples and sometimes from abroad.

regards,
Kanwar
 
event horizon

event horizon

Member
Joined 2008
Hi Kanwar, i have a quick question for you if you don't mind :)

You mention the switching glitches on the output when rail switching with a hard switch like class H & that these were visible on an oscilloscope. What kind of voltage are we talking about here just out of interest?

I guess it must be small & high frequency, enough to upset some people obviously :rolleyes:

Thanks for sharing your circuits & your insight.

Bests, Mark.
 
event horizon

event horizon

Member
Joined 2008
Hi Kanwar, yes i understand that totally ;) What i'm trying to gather is what kind of peak voltage these spikes might be, no doubt the speed of the switching to the higher rail will have an effect.

I'm assuming that the faster the switching the higher the peak voltage of the spikes. As i haven't actually messed around with this arrangement yet & knowing you have, i thought you might be able to possibly give a ballpark figure of what you yourself experienced in reality :)

Are we looking at a few tens of millivolts, or hundreds of millivolts, or greater?

Bests, Mark.
 
kASD

kASD

Member
Joined 2004
event horizon said:
Hi Kanwar, yes i understand that totally ;) What i'm trying to gather is what kind of peak voltage these spikes might be, no doubt the speed of the switching to the higher rail will have an effect.

I'm assuming that the faster the switching the higher the peak voltage of the spikes. As i haven't actually messed around with this arrangement yet & knowing you have, i thought you might be able to possibly give a ballpark figure of what you yourself experienced in reality :)

Are we looking at a few tens of millivolts, or hundreds of millivolts, or greater?

Bests, Mark.
Click to expand...

The max voltage amplitude of spikes i have seen is less than 2 volts
 
event horizon

event horizon

Member
Joined 2008
wg_ski said:
Try class G (non-switching) with a Schottky commutator. It takes more output transistors, but nothing's free.
Click to expand...
Agreed, that would be an ideal solution.

There is or should i say could be problems with this method though ;) Have you considered that the instant the upper rail class G transistors turn on that they are going to be passing the full current draw of whatever the speakers demand? I looked into it & depending on the rail voltages you are looking at a hell of a lot of powerful transistors. Admittedly there dissipation will reduce as the voltage increases (above a certain point) but for my money it's just not economical.

If i was rich i'd go that route, as it is i'm not, so i'll use a cascoded output stage as i already have transistors to make this with the power output i desire. I figure that the cascode will eliminate any switching glitches on the output, or at worst case seriously reduce them. I can put up with some non linearity as the H amp will be driving a class A output stage :D

Might be the worlds first class AH amp lol.
 
W

wg_ski

Member
Joined 2007
event horizon said:
There is or should i say could be problems with this method though ;) Have you considered that the instant the upper rail class G transistors turn on that they are going to be passing the full current draw of whatever the speakers demand? I looked into it & depending on the rail voltages you are looking at a hell of a lot of powerful transistors. Admittedly there dissipation will reduce as the voltage increases (above a certain point) but for my money it's just not economical.
Click to expand...

You end up using the same size upper bank as you do in the lower bank. If you have 4 in parallel, you need 4 more on the upper tier for class G. C5200's aren't THAT expensive, and there's no reason to use anything with higher SOA because the max VCE on the upper bank is the difference between tiers. Even with reactive loads.

Of course, in The Real World, the bean counters won't even splurge for two miserable 50 cent predrivers (which would significantly improve performance of the RMX line), so I'm sure extra output pairs are out of the question there. But for a DIY effort, what's $10? Or $20 if you put 8 in parallel?
 
kASD

kASD

Member
Joined 2004
djk said:
A 1KW Crest typically uses four 150W pair on the lower rails and three pair on the upper rails.
Click to expand...

yeah, along with asymmetrical Tier voltages to further enhance the reactive load handling capability of output stage.


Still i will vote for Class-G if 2 tiers and decent efficiency is required along with good sound like mx2000 old qsc, if the job is high power PA go for 3 steps class-H straightforward like qsc ex4000.
 
You must log in or register to reply here.
Top Bottom