PMsix61 Limiter DIY

[alexc]

And by way of 'washup' review, I would like to do it again but with :

- full tube power supply
- sowter audio traffos
- elma switches
- select 8 tubes from 30

The negatives of this build (so far - I'm still working on these) have been

  i) low freq roll off a bit higher than the usual 20Hz due to the UTC gapped input traffo
ii) bit more distortion under deep GR than I think could be achieved with better suited traffos and tube balance
iii) bit more distortion at fastest time constants than could be achieved with better suited traffos and tube balance


Things that surprised me were :

- the relatively underwhelming performance of the utc input traffo in this admittedly non-optimised circuit
  in terms of noise susceptibility, even when shielded, and also in terms of low freq roll off

- the high degree of balance attainable in GR tubes with a large investment of measurement time and suitable sample size

- certain aspects of the gain staging in the control amp  ie. the whole class B operation and general twitchiness
  I don't understand some of the design decisions - this is just a reflection of my lack of understanding

- that the Edcor 8K:600  works as well as it does in this crude attempt at the venerable master that is the FC 660


Things that came out well :

- the amount of gain achieved with relatively low noise floor penalty in a relatively complex many-tube build
  Around 7dB of noise floor worsening for 22dB of added gain (20-20K SPL style of measurement)

- the amount of GR achieved without adding much at all in the way of instability and noise
  ie. the noise floor improved in line with GR, as expected and hoped for, but not surely guaranteed

- virtually flat frequency response from 55Hz to 38KHz and maintaining a good part of this under significant GR    and significant input/output levels

- perfect behaviour of Analag/Silent Arts PSU module even with a fair bit of jiggery-pokery and duress
  ie. not even one issue

- the basic timbre, or sound, of the thing is satisfying in a slick but primal sort of way
  ie. ggrrrrrrrhhhh

 
Things I want to understand more :

- control amp design choices

- performance tradeoffs and issues arising from different signal amp transformer loads
  ie. optimising match of  transformer to signal amp tubes at min GR vs max GR

- transient response of signal amp + control amp integration 
  ie. a control theory perspective - damping, overshoots, settling times and all that

------

Anyway - unless I figure out something else of interest with the PMsix61, 'that's all folks'

I truly feel that I have made an improvement on the diy enigma that is the original Poorman 660 Limiter.
Never have so many ..  held such disparate views .. on so much .. over so little

The PMsix61, while not all that I think it could be, has proven to sound very pleasing - hifi, capable of low distortion at significant GR, damn fast and sounding very weighty and impressive.

It's worthy of some of my best instruments and I'm very, very happy to have had the opportunity to build this.

I'd never have done this without the collective knowledge and goodwill sharing of GroupDIY

I now know what Jimi meant when he asked 'have you ever been experienced?'  I know I have.

The challenge going forward is to discover a new funding model that allows me to move to the next level.
I need to spend more time on the markets  8)



Regards
Alec

 

[skal1]

Hi Alec,

I'am still grinding away on my fairchild version , pure headaches , anyway what is the analog clamp ,did you use this instead of the bias tx ,have you got a circuit for the analog clamp.

cheers

Skal

 

[alexc]

The 'analag clamp' is from Analag's famous Poorman 670.

It's a small signal pnp transistor in the GR tube cathode circuit that is biased to provide a constant -ve voltage regardless of the current flowing thru the paralleled cathodes on each phase of the GR tubes.

I blended it with the FC670 style of GR tubes' adjustable cathode bias circuit (as opposed to the FC660 method which is different again).

It certainly works, both in my Analag Poorman 670 build as well as my own PMsix61 build.

The idea is that it provides a steady GR tube bias arrangement despite strongly changing GR current flowing.
ie. less second-order effects with respect to cathode voltages and resultant bias shifting (smearing)

ie. it clamps the cathode voltage for each phase rather than having it more around somewhat due to the currents slamming about.

As to the effectiveness, I haven't done a direct compare in my PMsix61 with/without  the 'analag clamp'.

I certainly measured very repeatable and stable biasing - very important no doubt. Not least when doing the detailed measurements for tube matching which is a must for any serious parallel tube GR section.

I did come to the position that it does make 'thump' free operation easier to achieve.

Probably can do just fine without but just another example of stretching my build really

cheers

 

[skal1]

8) i was thinking ,could you also use it to provide the neg voltage  for the rectifier at the control amp output  and the bias neg bias for the finals, may use two .

cheers

skal