With or without input cap, with or without a floating power supply, with bipolars or jfets, I believe a power on/off current pulse through the cartridge is unavoidable. That would be very hard to simulate or otherwise evaluate, since the magnitude and duration of the pulse would depend on circuit and device asymmetries, power supply sequencing, parasitics, etc...
I wasn’t referring to on/off current pulses but to permanent DC current when using a gnd referenced supply because of a hfe mismatch between NPN and PNP.
Hans
P.s. It will be quite easy to simulate this permanent DC current
Last edited:
Maybe you can be so kind to explain why to your opinion a Duraglit etc. with floating supply is rendered useless when not using an input cap.
Hans
The only commercial version of Newton/Weinberg/Archimedes is the circuit in my orginal document, which was part of a 2x50W integrated amplifier.
IIRC, only 1000 were made so quite rare. But this little 50W amplifier now (2012) has a cult following in certain circles more than 20 yrs on … almost certainly due to its MC amplifier stage.
We never had any problems with direct coupling to the cartridge but only ran 0.5mA cos that simplified the LN requirements on the power rails.
Today, with small 1mF caps and especially Bonsai's LN PSU solutions, higher currents might be possible but I would caution anyone doing this to pay serious attention to earthing & shielding and be prepared for some trial & error .. especially if you are using a ferrous enclosure.
Current through the cartridge won't be so much from hfe mismatching but the resistor chains from each power rail.
So ... No we don't need an input capacitor on pure battery Duraglit versions. We certainly don't need an input cap on the +/- earthed supply version at 0.5mA and probably not on the higher current versions.
Due to their construction, MC cartridges are much less affected by magnetising effects of coil current than MM
Last edited:
Not when but if an input cap is not required. Why would one get through the complexities and constraints of a floating supply (low voltage, or batteries, capacitive coupled hum, etc...) if one can use a regular ground referred supply with a conventional bipolar gain stage?
Or let me put it this way: the only reason for the floating supply in Leach, Duraglit (same thing, really), #375, Hawksford LFD, etc... is to avoid the cartridge DC current. Otherwise, if a little DC current through the cartridge is acceptable, or if an input cap is acceptable, there’s no point in using a floating power supply.
Last edited:
I was not referring your point, and definitely not talking about simulating a DC current through the cartridge. Perhaps a more careful reading process would help.
Richard, yes, blue is a fraction of the output signal, red is the distortion which is almost H2.
The circuit is a duraglit with 2 pairs at 10mA total, input is 500µV 20kHz.
Agree.
I may be the only one to have misinterpreted this because IMO it can be read in two ways, but yes I fully agree when reading the way you meant.
Hans
As I said, the commercial Newton/Weinberg/Archimedes with earth referenced rails didn't have an input cap.
But my reason for sticking with the floating battery supply for the very lowest noise was the difficulty of getting superlative LN earth referenced rails without making them more complex then a complete 2x50W integrated amp.
In particular, the earthing & layout wasn't trivial. In Jurassic days, I considered myself an earthing guru though I'm still learning. I pontificate on the subject in Yahoo MicBuilders under LNprimer.
______________________________________
syn08, now that you've actually made & used #375 (floating SolarCell/LED power) and numerous HPS versions (supadupa earthed referenced LN power) ... which power supply approach would you recommend for practical ease of achieving near Duraglit levels of noise & THD for the DIYer? Please include getting hum & buzz below these levels. Also cost .. though this doesn't have to be a priority.
Your opinion is the most important cos you've actually built real-life versions of each.
Feel free to add various caveats as you see fit.
Last edited:
BTW, this, w/o servo, is sorta how Great Guru Baxandall did the QUAD 44 & 34 preamps. Part of the RIAA network 'replaced' the '2x50R' feedback resistors
As I said, the commercial Newton/Weinberg/Archimedes with earth referenced rails didn't have an input cap.
But my reason for sticking with the floating battery supply for the very lowest noise was the difficulty of getting superlative LN earth referenced rails without making them more complex then a complete 2x50W integrated amp.
In particular, the earthing & layout wasn't trivial. In Jurassic days, I considered myself an earthing guru though I'm still learning. I pontificate on the subject in Yahoo MicBuilders under LNprimer.
______________________________________
syn08, now that you've actually made & used #375 (floating SolarCell/LED power) and numerous HPS versions (supadupa earthed referenced LN power) ... which power supply approach would you recommend for practical ease of achieving near Duraglit levels of noise & THD for the DIYer? Please include getting hum & buzz below these levels. Also cost .. though this doesn't have to be a priority.
Your opinion is the most important cos you've actually built real-life versions of each.
Don’t fool yourself, the power supply noise is low not because is floating, but because it’s batteries, solar cells, or acid elements (like in the Hawksford LFD). A ground referred power supply build with the same type of cells would have equally low noise and having as a plus a lower hum, a problem due to capacitive coupling in the floating power supply. Try to make a floating power supply out of the same parts as a ground referred power supply, it will certainly be worse, at least in the hum performance.
Since you are into simplicity before anything else, with fitting in a Duraglit can as a performance metric, I don’t have a good power supply solution for you. Stay with a battery, you are safe with that.
