I do not know the specific of your playback and if you would like to get it for yourself or to use it as a dealer. After ML2.1 I very much lost trust in the respect to what Lamm did but my recent experience suggested that ML3 is the force they need to be acknowledged. Cuz they play back head the rest top of the Lamm electronics, which was produced after ML2.1 and frankly speaking I was anticipated less interesting result. I do have some kind of very subtle feeling is that original vintage ML2 voice slightly more kind and gentle than ML3, was slightly less showy but in the same way more respectful to music. But they are very subtle feelings and they might be to a great degree derivatives of other factors that the system I was listening had. When I was listening I was more listening not what the installation did but rather what capacity is had if some other minor things were addressed. I literally did not detect that Lamm electronics was the weakest link and to my surprise performed much better then what I anticipated.

Another small things. I know that Vladimir daughter is trying to revise the company and set up new production somewhere in Florida. My feelings that amplification they going to produce need to be evaluated as a new company. Obviously they going to follow Vladimir schematics and his assembling planning but I would still not automatically assume that the result they will produce would be identical to what we remember from end of the 90s.

The biggest mystery of Lamm ML3 

If anybody familiar with my attitude towards Lamm ML3  what's my disagreement that that amplifier of this level, should be permitted to have a coupling capacitor between 2nd and 3rd stage. Lamm, uses his standard schematic for ML2 with feedback applyed to cathode 12x7. The gm70 is direct heated that can easily enter grid current and work with a positive bias, pushing significantly more power then it's meant to but the most important do not go in very hard clipping like indirectly heated triods do. Here is where Lamm had a monster driver stage with  8 plates in parallel capable to produce immense current. In my view it is a humongous overbuild, the gm70 doesn't generate this huge current, but Lamm knew that his amplifier will be driving speakers for sensitivities around 90s and he made this overbuilt in my view completely reasonable, for this money he can throw a few more tubes in parallel. 

Here is where the main problem was for this amplifier in my view. And the picture here 

https://images.app.goo.gl/gKwKhUp6KtkTAUYJ8

You can clearly see that between second stage and GM 70 there are coupling capacitors. This is absolutely horrible idea because those capacitors works perfectly fine when they are polarized, means gm70 working with the negative bias in class A1. Now the gm70 need to throw more power, that output stage enter A2 operation mode and the output to running with positive bias, while Lamm's super powerful second stage buffer pushing current like no tomorrow. At that moment the coupling capacitor need to change polarity and this is absolutely the stupid idea that that should have a tremendous ramification to sound. 

The playback I was exposed I had top of the line karma load speakers, I do not know what sensitivity it was, I presume low 90s. The system had complementary subwoofers that I would like to discard for multiple reasons. This new provision of karma had both of low frequency sections ported it means in the bottom they should drop impedance and would demand a lot of current from amplification. This is exactly where's the topology of ml3 should shine. I was trying to crank the volume higher and try to detect the event when the coupling capacitor between second and third stage would makes difference. Of course it is not my system and everything is new but that is one of the primary interest of my visit was to detect when ml3, and it was my first listening of this amplifier, would auditably demonstrate it's main problem. However, this all my intention to detect auditable signature of this event I was not able to. The amp for sliding into high volume and low frequency events very very gracefully. So, I see three conclusions from this.

1. I am a full of shit who did not detect Sonic signature of capacitor recharging or impact was much much more settled that I anticipated. Another option would be that that apology off system subwoofers were not what I would like to see for system of this caliber and it might mask out the result. 

2. Lamm indeed uses direct coupling between his second and third stage and that capacitors add the picture are in fact three capacitors connected to parallel between first and second stage. 

3. Lamm was much more kinky that I thought and introduce illogic for biasing coupling capacitors between second and third stage and prevent it from changing polarity. 

I did not see ML3 schematic and I do not know what in reality is going on there but in this point I would hold my critical attitude about it. I really would like to hear this amplifier with a good, complicated tone enable low frequency section.

Poul, lamm ml2, the original not the division. It was from my point of view the best amplifier out there for a low frequency. When you have enough sensitivity in low frequency and you're able to idle ml2 output stage you was able to get base which is absolutely unknown for majority of audio people. Trust me, in that time I pretty much know all amplification available. The problem with ml2 is that it is 18 what only and if you need to suck out of it a little bit more that it instantaneously clip by voltage or by carrent, depends how you load output stage. ML3 is completely different concept. Is that output stage of that amplifier is specifically designed enter A2 mode and if you even drive it harder it's enters very very gentle soft clipping. the ML2 this indirectly heated trion purely topologically cannot work in a2 mode. 

No, the most interesting question is to learn if ml2 in its own best operation parameters deeply in A1 mode produces better or worse bass then ml3 in class A1 or A2. To answer this question you need to have a way back installation with no compromise low frequency section and so far I did not only personally experience it but I also looking in the vote other ml3 owners do, I did not feel that is there low frequency sections able to make Justice for either ML2 or Ml2. This is why I publicly stated that if somebody kills the day to be able to stress ML3 against truly serious low frequencies section then please invite me and I would like to hear it