TGM Amplifier ?

[Bigun]

Nice plot !

I see the roll off in the solid line and the 90deg phase shift in the dotted line. If I understand properly the bad news comes at 180 degrees and this doesn't happen until 5MHz where I imagine the gain of the amp is well below unity. Why would I pick this frequency, for example, why not roll off earlier ?

I think there is something important about where the HF feedback comes from. I've only read about the claimed purpose as better stability against nasty speaker loads but I'm wondering (don't laugh) if there's another factor because the VAS won't have any cross-over distortion. Hopefully the output doesn't either, but what remnants might exist there would then enter the NFB loop ?

I was reading this just now: http://www.dself.dsl.pipex.com/ampins/dipa/dipa.htm and it mentions the importance of picking the feedback point to avoid picking up cross-over currents. It also discusses the trade-off between LTP balance and 2nd harmonic. These things are starting to look more familiar now. It also introduced me to 'blameless' (I like the name!). I was curious to see a diode across C2 (in the last schematic) - what function does this serve ?

many thanks...

[ostripper]

Why would I pick this frequency, for example, why not roll off earlier ?

That might degrade slew rate and "slow the amp down"
at or below 20k (audio).

I think there is something important about where the HF feedback comes from. I've only read about the claimed purpose as better stability against nasty speaker loads but I'm wondering (don't laugh) if there's another factor because the VAS won't have any cross-over distortion. Hopefully the output doesn't either, but what remnants might exist there would then enter the NFB loop ?

That is very good reasoning, BG ! I have seen small "artifacts"
of Xover distortion enter my perfectly balanced LTP's. GNFB
"corrects" them and I end up with my super low THD .. BUT -
what if this introduces higher order harmonics into
the LPT. In The aska way , only the pure class a VAS is
NF'ed back to the input stage. Here you might have slightly
higher high freq. THD , but no Xover"garbage" from the OPS
gets into the loop.

feedback point to avoid picking up cross-over currents

This is the "physical" part of design , try to make your takeoffs
and at the exact midpoint of your
output trace , hopefully Xover (current) transients will cancel
out. Same with your ground returns to the central grounding
point .(attached) (ps.. that board is a blameless or a aska
or any thing in between).

I was curious to see a diode across C2

and the diode.. that is if you don't want to use a non-polarized
cap , so no "pop" if it gets reversed biased. I use
63v non-polarized so it is a non- issue with my designs.
OS

[Bigun]

TGM 1 - base version

Maybe to put a stake in the sand I'll record the 'final' schematic (subject to some resistor values being adjusted) for the base version of the amp which doesn't include the improvements I've been looking at.

[Bigun]

I'm probably spending too much time with the circuits instead of building because now I'm not sure I will be satisfied with version 1 given the knowledge gained. By borrowing from the FrugalAmp LTP, the Blamess protection diodes and many others the version 2 is simply more interesting - providing that the LTP 'balance' control still allows me to inject a bit of low order harmonic 'colour' into the results should I wish.

I also have to remember to scale back to lower supply rails for the final implementation as I don't need 50+W for the HT (but would like the design to be future-capable of 50+W).



I have to sort out a power supply and other things before I think about a pcb plus MJL21193 has kindly provided one of his obsoleted Patchwork Amps that will allow me to make some interesting comparisons between design elements.

I have noticed a common topology with pcb's so far so there must be a lot of proven 'wisdom' to leverage from. What are the pro's and con's of making up a simple home-pcb compare to struggling with proto-board for making up first prototypes ?

[ostripper]

What are the pro's and con's of making up a simple home-pcb compare to struggling with proto-board for making up first prototypes

I'm on "round" 3 &4 (pairs of amps), with the PCB you might
"get it right" the first time and keep them. Worst case you can
"regurgitate" the parts for your next one.

If you use MJL's HCL/peroxide etchant, a 4 X 6 costs 4$
and a 6 X6 costs 6$ (full price- PCB/ one sheet design/+etchant).
You might luck out and be satisfied with your first creation.

PS . R43 would be the one to affect "balance" "P3" will mostly
just offset the DC component in your output.
OS

[Bigun]

That's pretty inexpensive, less than a Pint of the good stuff. I'll definitely want to have a play with that.

[Bigun]

p.s. am I missing any decoupling caps - such as across C2 or from signal ground to chasis earth, across C4 maybe ?

[Bigun]

Just for fun I am thinking also about JFETs. I don't see them so often used in DIY amps in this forum but they are popular elsewhere.

Here's the TGM JFET version...

[Bigun]

Making progress - now have a box and am compiling a BOM for Digikey for the basic version of the TGM

[Lumba Ogir]

Bigun,
it starts to look like an amplifier. I would remove C14. Nice box.