No.danielwritesbac said:
Let's concentrate on the output stage to see why: Two sets of output transistors are connected to two different rail voltages in class G. Since all of the output devices of the LM1875 are tied to one voltage, and since we can't really muck with the guts of the chip, there is no way to generate class G since we are "stuck" with only one rail voltage.
You can do Class G and H with the LM1875 (or any chipamp) but you have to build the rail voltage controls external. I have done both with LM3886. The drawback is a tick up in THD when going into different modes but I never did extensive listening test to see if the difference was audible. There are circuits in some ST chipamps showing how to do it, although may either not be the best or too complicated or something else. I have never used them or looked at them in great depth.
The best option is to use something like a LM4702 or one of it's cousins (LME49810/811/830) so that you can keep the rail voltage changes only on the output devices and the rest of the amplifier is fixed. This should give better performance but I haven't gotten to it yet. Someday. . .
-SL
The best option is to use something like a LM4702 or one of it's cousins (LME49810/811/830) so that you can keep the rail voltage changes only on the output devices and the rest of the amplifier is fixed. This should give better performance but I haven't gotten to it yet. Someday. . .
-SL
My (admittedly limited) understanding of class G amps is that a separate output device is used for each rail voltage. What you are describing is more of what I thought class H was, IE variable rail voltages modulated by output voltage requirements. As such, it would follow that a class H chipamp would indeed be possible, but a class G amp wouldn't. If, however, you define class G as just stepwise adjustment of rail voltages (as opposed to continuous in the case of class H) then your description does make sense.SpittinLLama said:
I guess I am getting tied up in the technicalities, and like I said, I am definitely not the ultimate expert on these definitions.
PS Either way, I agree with only varying the output stage voltage of an LM4702 or related amp would be the best chip based solution. Not really sure how to do that, but I am guessing a variable regulator would be involved somehow. . . .
I'm relatively new to this, but I imagine it would be possible to design a PSU whose supply voltage varied according to the output voltage of the amp.
Doing this using a voltage regulator (I think probably easy to implement) seems fairly pointless though as you'd just be burning off all that extra heat in the regulator instead of in the chip
Doing this using a voltage regulator (I think probably easy to implement) seems fairly pointless though as you'd just be burning off all that extra heat in the regulator instead of in the chip
dfdye, Class G & H is not about how to implement but what happens at the supply rails. Class G is stepped voltages, 2 or 3 tier type of thing. The idea, obviously, being that for most signals under 'normal' listening conditions are small enough to use lower supply rails and waste less power. When the signal demands more the rails jump to a higher level. If you go to infinite number of rails then you hit Class H. There is plenty of confusion about which is G and H so no point in arguring that just that one is stepped rails and limited in number and the other is output tracking or envelop or infinite steps. Stepped rails can be done with a FET acting as a switch and some diodes. Output tracking is not done with a redulator as all the power saved in the output stage is wasted in the regulator. Instead some sort of SMPS is needed to gain efficiency in the system not just the output stage. With chipamps, you have to do a hybrid because there is some minimum operating voltage. So you set solid DC rails at say +/-20V and then envelop tracking for signals that need more. The point of doing either Class G or H is all about efficiency but it comes at a cost in THD and sonic performance, cost, etc.
-SL
-SL
Class G
I am presently working on a SM tracking power supply to track output to provide 12 to 40VDC to a set 8 parallel bridged LM3886's , as to run class A at low power and keep dissipation down at high levels. The whole idea is to provide a 300W Class variable chip amp in a 5 pound 12"*12" case.
I am presently working on a SM tracking power supply to track output to provide 12 to 40VDC to a set 8 parallel bridged LM3886's , as to run class A at low power and keep dissipation down at high levels. The whole idea is to provide a 300W Class variable chip amp in a 5 pound 12"*12" case.
You guys would never believe the reason that I wanted to explore Class G with LM1875.
Its not just for the power boost, although I'd be happy with that.
Its not just for the energy savings possible with Class-G, although I'd like a smaller electric bill and/or less heat output for summertime.
The reason is actually a beneficial errata, in that the small amount of distortion in LM1875 has an effect on even vs odd harmonics, similar to a BTL tripath or a good tube amplifier.
Somewhere, and I totally forgot where, this is documented with some sort of graphs and scans. As it is quite obvious to the ear, well that's good enough for my use of the $2 chip, which, does indeed, involve an ear.
So, LM1875 represents an opportunity to make a solid state Class-G amplifier with some aspects of a luxury-class hifi sound.
And that's why.
Its not just for the power boost, although I'd be happy with that.
Its not just for the energy savings possible with Class-G, although I'd like a smaller electric bill and/or less heat output for summertime.
The reason is actually a beneficial errata, in that the small amount of distortion in LM1875 has an effect on even vs odd harmonics, similar to a BTL tripath or a good tube amplifier.
Somewhere, and I totally forgot where, this is documented with some sort of graphs and scans. As it is quite obvious to the ear, well that's good enough for my use of the $2 chip, which, does indeed, involve an ear.
So, LM1875 represents an opportunity to make a solid state Class-G amplifier with some aspects of a luxury-class hifi sound.
And that's why.
banana said:
Although the TDA7294 is able for high fidelity, its not the nearly-automatic experience of LM1875.
I'm proposing utilizing the LM1875 as a "$2 hifi effects box," much like the "effects" of a vacuum tube pre. Yes, I'm proposing a cheat/shortcut, in hopes of a lot less "sound tweaking" of the finished product.
The LM1875 starts off with a good high fidelity baseline, which is measurable--not a confusing botique deal.
Do you know what I'm saying?
Its that the LM1875 is one of the rare cases where science and hifi have met. That's an awesome thing to make louder.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.