Regarding the resistors, in the schematic I see 47 ohm in series with the input and a 100 ohm series feedback resistor. I wouldn't consider that low value at all, they are far above the resistance of a moving-coil cartridge. Still, if you want to drastically reduce the feedback resistance, you basically have to redesign the head amp. The 47 ohm at the input can maybe be replaced with a lossy ferrite bead.
The bias point seems to depend on matching between D101 and Q101, which will complicate transistor replacements. If the VBE and/or hFE values are substantially different from the original Q101, the head amp may not bias properly.
Last edited:
I think maybe I would be better to leave the head amp alone, you see I am replacing the Sony MM phono stage with a discreet one and I am thinking if this does sound better than the original Sony IC stage then in turn it will improve the sound of both inputs as the head amp then runs into the MM stage
Gary
Gary
) )
Hi,
as Rohm manufactured those famous 2SB737/2SD786, they still should have some lownosoise devices in their portfolio.
How´s for example the 2SA1455K/2SC3722K in SC59 SMT casing?
jauu
Calvin
as Rohm manufactured those famous 2SB737/2SD786, they still should have some lownosoise devices in their portfolio.
How´s for example the 2SA1455K/2SC3722K in SC59 SMT casing?
jauu
Calvin
According to the datasheets:
2SA1455K: noise figure of 3 dB at 100 ohm source resistance, 1 kHz and optimal collector bias current (1.8 mA)
Not quite as good as 2SC2547, 2SC3324 or 2SC2240, but certainly not bad. However, I don't find the 2SA1455K on the Rohm website. Is it still in production?
2SC3722K:
minimum noise figure of 3 dB at 200 ohm and presumably 1 kHz, 1.4 mA, the frequency labels of the graphs appear to be mixed up
2SA1455K: noise figure of 3 dB at 100 ohm source resistance, 1 kHz and optimal collector bias current (1.8 mA)
Not quite as good as 2SC2547, 2SC3324 or 2SC2240, but certainly not bad. However, I don't find the 2SA1455K on the Rohm website. Is it still in production?
2SC3722K:
minimum noise figure of 3 dB at 200 ohm and presumably 1 kHz, 1.4 mA, the frequency labels of the graphs appear to be mixed up
Hi,
seems You´re right. Couldn´t find them either any more.
How´s about the 2SAR293P/2SCR293P (SC62 SMT casing) pair instead.
Spice model quotes Rb of <3Ohm.
jauu
Calvin
seems You´re right. Couldn´t find them either any more.
How´s about the 2SAR293P/2SCR293P (SC62 SMT casing) pair instead.
Spice model quotes Rb of <3Ohm.
jauu
Calvin
There are no noise graphs in their datasheets, but judging from the data that are given, they could be quite good.
That is, RB=2.8446 (2SCR293P) or RB=2.3056 (2SAR293P) is a very good value and it sounds like Rohm seriously tried to extract base resistance, rather than just plug in some number (unlike the libraries that were supplied with PSpice that had RB=10 for all transistors). Of course this is no guarantee that they extracted it correctly for noise modelling, they could have extracted a number that gives the right VBE at high base currents without modelling the base resistance's current dependence.
These transistors also have a high hFE, that is, a favourable ratio between collector and base shot noise. The hFE versus collector current graph shows no sign of hFE roll-off at low currents; at least not down to 1 mA, they didn't measure any lower. That's important because the non-ideal base current component that causes hFE roll-off usually has a large amount of 1/f noise associated with it.
You would have to do noise measurements to be sure, but they definitively seem interesting.
That is, RB=2.8446 (2SCR293P) or RB=2.3056 (2SAR293P) is a very good value and it sounds like Rohm seriously tried to extract base resistance, rather than just plug in some number (unlike the libraries that were supplied with PSpice that had RB=10 for all transistors). Of course this is no guarantee that they extracted it correctly for noise modelling, they could have extracted a number that gives the right VBE at high base currents without modelling the base resistance's current dependence.
These transistors also have a high hFE, that is, a favourable ratio between collector and base shot noise. The hFE versus collector current graph shows no sign of hFE roll-off at low currents; at least not down to 1 mA, they didn't measure any lower. That's important because the non-ideal base current component that causes hFE roll-off usually has a large amount of 1/f noise associated with it.
You would have to do noise measurements to be sure, but they definitively seem interesting.
Last edited:
the 2SC3324 claims to have a NF of 0.2dB at 10kOhm Source resistance, but dunno how to convert this figure to noise density. Considering the extra resistance brought by the Potentiometer, it may be a good choice for input stage. Hfe max 350-700. The THAT340s have too low an Hfe, but their Vbe is matched with in 500uV, so its intended for current mirror i think. But doesnt using a 340 for current mirrors a little bit too luxury? costs like 7$ each.
These look like medium power transistors that have low resistances for good behavior up to 1A of Ic. You get the low rbb for free. The very low Vsat is another good indicator that they will have good noise performance. I remember someone who found a TO220 power transistor that he could coax .28nV noise out of in a MC pre.
The new edition of "The Art of Electronics" has some really good tables of transistor noise performance. Zetex came out verywell.
- Status
- Not open for further replies.