Whose afraid of Solid State
Konnichiwa,
Of course, if we are not mortally afraid of a little itty bitty bit of sand or two we can upgrade the above circuit to a hybrid one which gives us around 60db of gain and an input noise about equivalent to 100 Ohm (or 0.2uV flat, which is around 68db S/N ratio referenced to 0.5mV, unweighted)....
It would look like this:
The J-Fet cascode can be switched to provide around 44db gain (MM, High Output MC) or 64db (depends a little on the J-Fet), again the anode load is also the resistor that determines the first timeconstant (3180uS).
In the output stage a Supertex depletion high voltage FET (or if you like an IXYS IXCP 10M45, I like it better) acts as simple current source, the lower bits remain the same as does the "semi-fixed bias" biasing of the output stage, which eliminates these pesky sound killing cathode bypass electrolytic capacitors.
Anyway, have fun with these, if you like, convert your current preamp, it should not be too hard....
Ciao/Sayonara
Konnichiwa,
Of course, if we are not mortally afraid of a little itty bitty bit of sand or two we can upgrade the above circuit to a hybrid one which gives us around 60db of gain and an input noise about equivalent to 100 Ohm (or 0.2uV flat, which is around 68db S/N ratio referenced to 0.5mV, unweighted)....
It would look like this:
An externally hosted image should be here but it was not working when we last tested it.
The J-Fet cascode can be switched to provide around 44db gain (MM, High Output MC) or 64db (depends a little on the J-Fet), again the anode load is also the resistor that determines the first timeconstant (3180uS).
In the output stage a Supertex depletion high voltage FET (or if you like an IXYS IXCP 10M45, I like it better) acts as simple current source, the lower bits remain the same as does the "semi-fixed bias" biasing of the output stage, which eliminates these pesky sound killing cathode bypass electrolytic capacitors.
Anyway, have fun with these, if you like, convert your current preamp, it should not be too hard....
Ciao/Sayonara
Re: Whose afraid of Solid State
Thinking the same thing 😀
Is the network with the 68ohm resistor near the "0V" mark to obtain negative fixed bias for the second stage? It does so by lifting of some volts the whole first stage and the all other things connected to ground, isn't it?
If so, where do I have to connect the "0V" and the GND points? I assume that the GND point at the input goes to the input RCA ground (which are insulated from the chassis), and the rightmost GND goes to the star ground. What about the 0V point?
sorry for stupid questions but I'm not practical with this technique to obtain negative bias.
Kuei Yang Wang said:
Thinking the same thing 😀
Is the network with the 68ohm resistor near the "0V" mark to obtain negative fixed bias for the second stage? It does so by lifting of some volts the whole first stage and the all other things connected to ground, isn't it?
If so, where do I have to connect the "0V" and the GND points? I assume that the GND point at the input goes to the input RCA ground (which are insulated from the chassis), and the rightmost GND goes to the star ground. What about the 0V point?
sorry for stupid questions but I'm not practical with this technique to obtain negative bias.
Re: Re: Whose afraid of Solid State
Hi,
Yes, you basically stick it into the powersupply somewhere where there is clean power already (after regulation for example, but before the final RC decoupling networks per stage. You need to adjust the resistor that with all the current flowing in the actual preamplifier (all stages powered from this supply) it drops the voltage you want to bias your amplifier with.
The "audio ground" as shown is after this resistor. It is an old trick used in old audio gear in the days before big electrolytic capacitors became cheap. Often the choke was wound with sufficient DCR to drop the required voltage to bias the output stage and at the same time to filter the main HT with small value oil capacitors and was placed in the negative line of the supply.
To kill all the noise and get a nice clean DC for the gridbias they added just like I did a simple RC filter, again a low value oil capacitor did great there and they did not have anything else to use.... 😉
I find that many of the old design techniques can be resurrected in our century, this time to avoid having to use the same large value electrolytic capacitors the old designers lacked. We do have them, but just do not like their sound... 😉
You could also use silicon diodes connected in series untill they drop the desired voltage, if too much voltage is dropped a suitable resistor to (audio) ground will allow you to restore the desired voltage, I'd still leave the RC filter in place.
I find this a neat and simple trick. You could also a seperate negative bias supply, but the few volt we need to drope we can easily spare in the main supply, which needs to be clean and free from noise anyway, so why not "borrow" a little of it?
Ciao/Sayonara
Hi,
Giaime said:
Yes, you basically stick it into the powersupply somewhere where there is clean power already (after regulation for example, but before the final RC decoupling networks per stage. You need to adjust the resistor that with all the current flowing in the actual preamplifier (all stages powered from this supply) it drops the voltage you want to bias your amplifier with.
The "audio ground" as shown is after this resistor. It is an old trick used in old audio gear in the days before big electrolytic capacitors became cheap. Often the choke was wound with sufficient DCR to drop the required voltage to bias the output stage and at the same time to filter the main HT with small value oil capacitors and was placed in the negative line of the supply.
To kill all the noise and get a nice clean DC for the gridbias they added just like I did a simple RC filter, again a low value oil capacitor did great there and they did not have anything else to use.... 😉
I find that many of the old design techniques can be resurrected in our century, this time to avoid having to use the same large value electrolytic capacitors the old designers lacked. We do have them, but just do not like their sound... 😉
You could also use silicon diodes connected in series untill they drop the desired voltage, if too much voltage is dropped a suitable resistor to (audio) ground will allow you to restore the desired voltage, I'd still leave the RC filter in place.
I find this a neat and simple trick. You could also a seperate negative bias supply, but the few volt we need to drope we can easily spare in the main supply, which needs to be clean and free from noise anyway, so why not "borrow" a little of it?
Ciao/Sayonara
Thank you very much.
I would like to understand better your RIAA approach: you say the anode load resistor (33k) determines the first 3180uS timeconstant, so I need a 96nF capacitor from the anode to ground to form this timecostant, am I right?
Then I need the 75uS one, and that is 33nF // 2k4, right?
Then the 318uS is 33nF + 99nF and 2k4, or am I wrong?
I have to learn using spice
I would like to understand better your RIAA approach: you say the anode load resistor (33k) determines the first 3180uS timeconstant, so I need a 96nF capacitor from the anode to ground to form this timecostant, am I right?
Then I need the 75uS one, and that is 33nF // 2k4, right?
Then the 318uS is 33nF + 99nF and 2k4, or am I wrong?
I have to learn using spice
Konnichiwa,
Not really mine, it is the same as in the old RCA Phonostage. The Key advantage is that it works well using just 4pcs of identical capacitors.
The 75uS as you observe is the RC parallel combo, but there is some effect from the second capacitor which reduced the HF attenuation so the actual value of the timeconstant of R&C only is 79.2uS. It becomes 75uS once we account for the second capacitor.
Again, the 318uS is indeed 33nF + 99nF * 2K4 and thus in actuality is 316.8uS.
And the 3180uS are 33K//Ra//Rl which is high but not infinitly high * 99nF, giving pretty accuratly 3180uS if you account for the 1M Input resistance of the second stage.
As there are many interactive factors in the whole thing there is no simple textbook calculation, you need to account for all the major parasitic's in the circuit, hence the values of real RIAA Circuits that are flat actually differ often a little fromwhat calculation would lead us to expect.
Ciao/Sayonara
Giaime said:
Not really mine, it is the same as in the old RCA Phonostage. The Key advantage is that it works well using just 4pcs of identical capacitors.
The 75uS as you observe is the RC parallel combo, but there is some effect from the second capacitor which reduced the HF attenuation so the actual value of the timeconstant of R&C only is 79.2uS. It becomes 75uS once we account for the second capacitor.
Again, the 318uS is indeed 33nF + 99nF * 2K4 and thus in actuality is 316.8uS.
And the 3180uS are 33K//Ra//Rl which is high but not infinitly high * 99nF, giving pretty accuratly 3180uS if you account for the 1M Input resistance of the second stage.
As there are many interactive factors in the whole thing there is no simple textbook calculation, you need to account for all the major parasitic's in the circuit, hence the values of real RIAA Circuits that are flat actually differ often a little fromwhat calculation would lead us to expect.
Ciao/Sayonara
Its getting pretty hard to get your hands on the 2SK170 or 2SK369's. Except for sourcing them via Borbely or VacuumState, is there other sources?
There's an outfit in the Bay Area that has started making them. They've posted here in the past several months, so a little searching should turn up the name and contact info.
Konnichiwa,
Well, first, Borbeley and VS will match for you, making their services much cheaper for the DIY'er in the end.
Past that, I tend to buy mine from the local electronics "odds & ends" shop (Cricklewood Electronics) and schuro.de also stocs them. A bit of googeling and froogeling givs more options.
Right now 2SK170 & 369 tend to be common as much and cheap as chips. I remember the days when you could say this about the 2SK147 (50pence each!!!! try getting chips for that at McDonalds).
Sayonara
JoshK said:
Well, first, Borbeley and VS will match for you, making their services much cheaper for the DIY'er in the end.
Past that, I tend to buy mine from the local electronics "odds & ends" shop (Cricklewood Electronics) and schuro.de also stocs them. A bit of googeling and froogeling givs more options.
Right now 2SK170 & 369 tend to be common as much and cheap as chips. I remember the days when you could say this about the 2SK147 (50pence each!!!! try getting chips for that at McDonalds).
Sayonara
They are Linear Systems, the data sheet is here: http://www.linearsystems.com/datasheets/LSK389.pdf
Have a poke around their site.
-Chris
Have a poke around their site.
-Chris
I am not saying their services aren't worth it, I was just concerned about ability to continue to be able to source them for the foreseeable future in case you had to replace one. I did google for the 147, guess I didn't try the 170/369.
It has been said many times on the forum that Toshiba no longer makes 2SK170, but I recently checked the Toshiba website, and I couldn't find anything there indicating that they are obsolete or about to be. They are also still available from my local supplier, which has stopped selling certain other Toshiba devices due to delivery problems.
- Status
- Not open for further replies.