At your risk. Take all precautions to protect your hearing.
An amazing sound. How could it not be. It is just 2 transistors from production to reproduction in real time.
Signal path, setup:
Various sound sources - Sennheiser MD21 HN - initially Canford inline XLR to 1/4in input transformer but then the usual German XLR inline transformer with lower inductance - Pleiades K117 operated with 1.3V - Pleiades 2N3053 operated with 1.3V - Sennheiser HD580
Note: Both transformers are mentioned on previou posts. Inductance values were given at 1.2KHz?. Canford is 413mH but it is 1.7H at 97Hz. The German transformer is 232mH at 1.2KHz but 470mH at 97Hz. The German transformer is very similar to the GSM1 e + p which is XLR to XLR rather than XLR to 1/4in jack.
The setup sounds so real that it is hard to discriminate whether a sound comes to the ears from the real acoustic source or through the open back headphones. The volume control at the input of the 2N3053 has to be changed so as to realize whether the so real sound is indeed coming from the headphones.
Various acoustic source were used, for example:
Male singing voice at 1-12in.
Ticking clock form many meters away, different rooms, till just next to mic.
Ride or crush cymbals gently taped by finger.
Approaching from meters to close the MD21 HN mic to the Philips AD5046N high Z (800Ω?) full range loudspeaker driven by a Pleiades 2N3053 power amp supplied by 12V. The full range speaker was playing Just Hold Me - Maria Mena. Exceptional sound quality.
The excellent mic was used at 90degrees for grazing incidence, no HF boost, to 0 degrees for full HF boost due to the pressure doubling effect.
The setup is very low noise.
The sound is as smooth and detailed as possible. Great bass. For example deep bass can be heard when approaching the full range cabinet from the back. Possibly capturing the soudfield created by the vibration of the back wood of the speaker enclosure too.
Some bass heaviness at voice or treble exaggeration is apparent when live singing as expected. This must be Fletcher Munson, voice effort curves, pressure doubling effect at play. A lot of the bass heaviness disappeared by using the German input transformer of less primary inductance than the Canford.
So most of this experiment was done by German XLR to 1/4in inline transformer.
Further bass heaviness or treble heaviness may be gone if a Pleiades V6 electron tube front end is added with the electron tube at about 20-30μA anode current. This rather small value may give the bass and treble decrease needed for a full big and natural sounding vocal sound.
It is hard to realise that many of these Pleiades experiments were really hearing aid experiments.
A mic is amplified in real time to feed headphones.
In general in pop music at reproduction a voice of higher acoustic intensity than the original acoustic intensity is usually the case.
For correct frequency response from producer's to listener's brain a deliberate non flat, usually gentle low and high cut response is required.
If this is done with the minimum electronic components possible and thereby the simplest possible signal path operating in pure class A (electrons flowing all the time), realism is fantastic.
See the 2 next posts on how each of the 2 stages is made.
While this was being tested the 7 pin Nuvistor 2DV4 holders came from eBay. The 2DV4 is to be operated by 1.3V at the heaters and anode. Is this a good indication that the 3rd stage would be a 2DV4 so that all 3 stages be powered by just a 1.3V ordinary battery? What would be the role of 2DV4 apart from making the sound correct to listener's brain? Perhaps the front end? Or 2nd amplifying stage? Or output stage? If it is 1st stage would a hi Z to hi Z coupling transformer work best. For example an Ampex bridging transformer. Or an inductor coupling, auto transformer etc.
Further reading:
Flat frequency response from producer's to listener's brain, Sound Picture recording and reproducing characteristics - D. P. Lowe, K. F. Morgan - JSMPE
No comments:
Post a Comment