It depends on the circuit. The book value (10R) is probably based on the most basic Vbe multiplier (which is what you have when your tools are only paper and pencil), where it really depends on how you choose Rbc/Rbe ratio. In real circuit, Vbe multiplier can be complicated by addition of diodes and paralleled R (not to mention caps). But the good thing about using simulator is that you don't need to calculate, and in the end you may just be able to get rid of this unnecessary resistor.
If you're looking to use something like the Naim regulator, it doesn't cost anything to invert the LTP and VAS so their references aren't the dirty power rails.
This schematic also adds a simple CCS for the LTP, but most of the performance gain over the original is due to inverting the LTP and VAS:
Here are the attenuation curves (original in green, above schematic in blue):
Cheers,
Jeff.
PS: idea is not mine; PigletsDad on PFM suggested it to me.
Attachments
What was the VAS current? What was Rbc? Rbe? Junction temperature?
Even using simulator you need to be skillful to get to such precise values. Doing it by ears?? It's like choosing passive speaker crossover values using ears only.
If you're looking to use something like the Naim regulator, it doesn't cost anything to invert the LTP and VAS so their references aren't the dirty power rails.
This schematic also adds a simple CCS for the LTP, but most of the performance gain over the original is due to inverting the LTP and VAS:
Here are the attenuation curves (original in green, above schematic in blue):
Cheers,
Jeff.
PS: idea is not mine; PigletsDad on PFM suggested it to me.
interesting, thank you
The graphical comparison is important
but in the end what is more important in a final power amplifier?
- a nice scheme but stuffed with a cap or servant to check the offset
- a scheme of type 140 but revised in the quality of a capacitor (the input cap is eliminated)
- or a 140 model but fully stabilized in the input, driver and end VCC voltages?
- a nice scheme but stuffed with a cap or servant to check the offset
- a scheme of type 140 but revised in the quality of a capacitor (the input cap is eliminated)
- or a 140 model but fully stabilized in the input, driver and end VCC voltages?
My NAP clone vas current is 20mA. P3A is 5mA. You can optionally use higher Rbc/Rbe ratio. At such high current (and Asian countries), the temp can be higher. With such conditions you will get higher than 10R.
The references for the ltps are identical, the dirty rails, and it is okay. The vas could be better with your circuit. But i'm not convinced. For me the regulator is not about as much as regulating the supply voltage. How about other performances such as output impedance?
Clearly, as I could, I would stabilize this scheme, but it is not clonable
Why do you want to clone that amplifier? Yes, the input JFET has no equivalent, but there are many ways to do the first stage with different part. The second stage is strange as the JFET Vds is very high. Doesn't have to be fet tho. Harder part imo is the complementary fet drivers. Nothing equals the J313 Toshiba parts which is hard to find. What amp is that? Spectral?
I would like to clone it because it sounds very good.
there are no caps on the signal and it is coupled in DC with a very stable offset
The problem is precisely 2n 5566 to which I can not find a substitute
Tried to replace with a dual transistor, but I lose a lot.
The emitter follower with a j fet P-channel , Toshiba J 109, is very good for the amp
The scheme is Spectral revised by Norma Audio
.
if you have suggestions for the first stage, I am grateful
the original drivers were Hitachi 2SK 216 and complement 2SJ 79, no more products, but can be replaced.
My problem is the Dual j fet input and input stage
Last edited:
Sloppy wording on my part. The LTP is inverted only because that's the easiest way to drive the inverted VAS.
Output impedance is also improved, again more-so in the upper frequencies. At 1KHz it goes from 1.5mOhms to 0.2mOhms; at 10KHz 10mOhms to 0.6mOhms.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.