Simplistic NJFET RIAA

Member
Joined 2003
Paid Member
Andrew. 211mA-60mA = 151mA or 140mA if you allow for control current. I was looking for an easy way to measure it vs. calculating it.

Salas, Let me make sure my understanding is correct. No assumptions here. Q1(irf9610) is the CCS fet. Q4(irfp9140) sets the voltage. If so, what is interesting is that as I ramp up the Vin, the current will stay <70mA until it hits the regulation voltage. At that point it jumps to 215mA. If I understand the circuit, Q1 is full on and Q4 is off till it has enough voltage headroom to regulate.
 
diyAudio Chief Moderator
Joined 2002
Paid Member
The voltage is set by the Vref which is the translation of current through a resistor by a Jfet CCS. The main CCS references to Vout in this, hence it won't open properly until you get the expected voltage conditions. If it will turn on instantly without a variac it should pull instantly too.
 
Member
Joined 2003
Paid Member
I put .1uF mkp on Vin. Big change but not a solution. Frequency jumped to 14mHz and is a cleaner sawtooth. See before and after in attached.

I have to pull board off of heat sink to implement change on gate resistor. (will do however if proper next step)
I also measured Vref across LED as 1.92V
 

Attachments

  • DSC02605.JPG
    DSC02605.JPG
    148.2 KB · Views: 263
  • DSC02606.JPG
    DSC02606.JPG
    144.9 KB · Views: 247
Last edited:
diyAudio Chief Moderator
Joined 2002
Paid Member
Reinforces my suspicions that the resonance resides in the CCS. Make that 0.1uF a 47uF local decoupler, so we see where it goes, maybe some nH from layout resonates with something. Lets see about the gate stopper and possibly the base stopper values around the Mosfet and its control element next. Is the input voltage pre filtered BTW?
 
diyAudio Chief Moderator
Joined 2002
Paid Member
80mV is too much, its oscillation. Lets see, can we move the compensation from the error amp nearer to the CFP? 9140 has 140pF Crss when 9240 has 81pF. We can be having a resonance there that can be read across the reg. I have done that once when I had to use a more capacitive shunt Mosfet because I was out of 9240.
 

Attachments

  • Comp.gif
    Comp.gif
    2.4 KB · Views: 235
diyAudio Chief Moderator
Joined 2002
Paid Member
Something is not right for sure. So the 0.1uF is good enough value across the input, I use it too sometimes. Look at my post about moving and changing the compensation scheme.

P.S. To get it right, you never saw a straight line on the scope, yes? It was a sonar ping there on all knob settings?
 
Member
Joined 2003
Paid Member
With it regulating, I never saw a straight line. I only see the sonar ping with the lytic on the output. Without the lytic on Vout it is a steady state oscillation. I should clarify the cap on Vin makes it SS oscillation closest to sine wave.

Darn. Just blew the semi's again. I accidentally unplugged Vin. Won't start-up now.
 
Member
Joined 2003
Paid Member
With it plugged in, I watched the output voltage slowly climb until it hit regulation voltage. The LED lit up and away it went working again. Will set voltage fine, but still has HF oscillation just like before.


The error amp is a BC546B rated for 65Vceo and 85Vcbo
 
Last edited:
diyAudio Chief Moderator
Joined 2002
Paid Member
Its not normal, its struggling with some oscillation. If when touching with a probe the 9610's gate it will go on fast, its sure sign the oscillation upsets it. Try my post #4830. Its a thing that has a chance over revising the layout which will involve a lot of guess work. It does not change performance.