Also have a 16.5, served me well since the 90's, but Ultrasonic is worth the $$$ and effort. The other thing that I swear by is Last Record Persevative. We use this at work where records get used and abused, some have been used for over 30 years! Not only does it seem to extend the life of LPs but I have had many records have their noise floor dramatically improved.
Have the PSU board soldered up. Started doing the parts check for the signal boards. I looked at the SMD components. My eyes hurt.
Working with the new Hakko soldering station is interesting. So used to my little 30W Weller I'm having to relearn. The SMD practice kit is going to be worth the money.
Working with the new Hakko soldering station is interesting. So used to my little 30W Weller I'm having to relearn. The SMD practice kit is going to be worth the money.
Could this be solved by lowering the value of R1? That is what I did because my 2SK170 BL grade also draw more current and pulled the voltages down. Lowering the value of R1 works with regard to the voltages, but I have no idea if it introduces another potential problem.
Lowering R1 will reduce the OL gain of the first stage, I think. ~6dB in Mr Wellerman's case in message #2,320.
Will this hurt overall performance? If so, reduce loop gain as a fix?
I too have a matched octet of 2sk170bl's to which I am nostalgically attached (acquired for original Pearl). I juggled R1, and R2 and R3 in simulations in order to get some Vce greater than saturation across Q5.
On the other hand, @wayne reported in message #610 that he had 9.1V across R1 when using 2sk170. Can the cascode work in this case?
2sk209 looking better all the time. ;-)
Will this hurt overall performance? If so, reduce loop gain as a fix?
I too have a matched octet of 2sk170bl's to which I am nostalgically attached (acquired for original Pearl). I juggled R1, and R2 and R3 in simulations in order to get some Vce greater than saturation across Q5.
On the other hand, @wayne reported in message #610 that he had 9.1V across R1 when using 2sk170. Can the cascode work in this case?
2sk209 looking better all the time. ;-)
The values around the cascode could be modified. The drain load change would decrease the gain but the transconductance of the BF862 is higher compensating. The 2SK209 does the job really well and at less than 20 cents for a genuine Toshiba part is hard to beat. They are just really consistent.
The 1/F noise is not really much different in the finished circuit as other noise dominates.
I have a roll of BF862 for some reason but haven't tried them yet.
The 1/F noise is not really much different in the finished circuit as other noise dominates.
I have a roll of BF862 for some reason but haven't tried them yet.
So, if I'm right, looking at the data sheets 2SK209 Yfs with Id of 2 mA (typical for the P3) is around 12. 2SK170 Yfs with Id of 3.6 mA (my P3) is around 25. Double the gain due to higher Yfs, 6 dB loss of gain due to drain load according to @Auricle. Net result would then be the same gain?
As noise is particularly important at the input stage, would there be a difference between halving the value of R1 (as I currently have done) or doubling the value of the source resistors and keeping R1 at 750R?
The thing with the 2SK209 is that I currently have nothing else to order at Mouser, and that would make ordering them quite expensive. If I can get the same result with my 2SK170's and changing some resistor values, I prefer that. How many 2SK209's would it take to get 2 tightly matching quartets (or ideally an octet)? 20? 50? 100?
As noise is particularly important at the input stage, would there be a difference between halving the value of R1 (as I currently have done) or doubling the value of the source resistors and keeping R1 at 750R?
The thing with the 2SK209 is that I currently have nothing else to order at Mouser, and that would make ordering them quite expensive. If I can get the same result with my 2SK170's and changing some resistor values, I prefer that. How many 2SK209's would it take to get 2 tightly matching quartets (or ideally an octet)? 20? 50? 100?
I matched a few quads/octects of 2sk209s, if I remember correctly the first three parts I tested off the reel were relatively tight matches (less than 0.1mA), then after testing 6 or 7 parts in the Idss jumped by 2mA, so it seems luck of the draw. To make a long story short with a little luck you might be able to get two quads with 20 parts, but with 50 you are more likely to and will probably get tighter matching. it was tedious using tweezers to pick and place those SOT-23s correctly into my fixture.