I did some sims some time ago but, nobody wants to here that bla bla do they. Keep in mind it's only as good as the models and sim. My 2sks are about 11mA in the model and I only have a IRFP240/9240 with who knows what parameters.
The front end has very little gain as I remember and that makes sense. The output I think was most of the gain of 300. I did not do much higher level stuff, but, now that Patrick suggests this looks like a 741, I should go look at OLG and freq responce etc. and see what we have here...
UUUUURGH!
The front end has very little gain as I remember and that makes sense. The output I think was most of the gain of 300. I did not do much higher level stuff, but, now that Patrick suggests this looks like a 741, I should go look at OLG and freq responce etc. and see what we have here...
UUUUURGH!
If you lower the frontend current to 6mA, you will double the frontend gain (with the same 2nd stage bias).
You need to know how to change your Spice model to do this.
Contact Melon Head for details. I am sure he is very willing to help.
And yes, look at OLG & fre resp, like an opamp.
Bob Cordell's book also have a very good section on this.
Patrick
You need to know how to change your Spice model to do this.
Contact Melon Head for details. I am sure he is very willing to help.
And yes, look at OLG & fre resp, like an opamp.
Bob Cordell's book also have a very good section on this.
Patrick
Last edited:
Early in the eighties I was the front end guy for Motorola OpAmps. I think I can dig it out. I should look into Bob's new book though. Looks like a good investment.
I can probably figure out the Spice thing enough for a ballpark idea where we're goin with this. Something about those claims of MHz response made me think I didn't really need to think about that much, if I tried to stay faithfull to the original design? But, OpAmps do that too, under some conditions. It was a little surprising to see a Pass design with so much gain and F.B.?
Melon Head, are you listening? (Never stops making me laugh, when I see that. I'm talking to a Melon Head!). I need a little time though. I'm free tonight but, I really wanted to play with some heatsink testing tonight...
I can probably figure out the Spice thing enough for a ballpark idea where we're goin with this. Something about those claims of MHz response made me think I didn't really need to think about that much, if I tried to stay faithfull to the original design? But, OpAmps do that too, under some conditions. It was a little surprising to see a Pass design with so much gain and F.B.?
Melon Head, are you listening? (Never stops making me laugh, when I see that. I'm talking to a Melon Head!). I need a little time though. I'm free tonight but, I really wanted to play with some heatsink testing tonight...
My Spice models for the 2SK170 and 2SJ74 have about 11mA of Idss and who knows what transconductance. It's fine for getting ideas worked out but, in the F5 case, they are the wrong devices(V). We are trying to tweek the model to get a 7ma Idss(?) part like the F5 uses.
I might get a chance to play with the sim tonight? Depends on what the S.O. wants to do. But, I really want to finish setting up my Heat Sink testing ideas first. Once things are cooking, I can play Pspice...
Any ideas about Lambda, Vgsth, and anything else I'm not thinking about now?
What about a model for the 1530 and 201?
Any ideas about Lambda, Vgsth, and anything else I'm not thinking about now?
What about a model for the 1530 and 201?
post65 is quoting model values for a simulator.
This post is talking about measurements and 6mA in circuit.
What are you finding? What are you simulating? What are you measuring?
I was interested in the characteristics of gain and freq resp from the std F5. And, what I would end up with by changing the JFET Rdrains, output MOSFET(s), Bias, Rload, etc., etc..
First problem I know about is the model parameters of the 2SK170 and 2SJ74 in my Pspice. They are nowhere near a 6-7mA device. In order to be more accurate with the sim, I need a more realistic model.
First problem I know about is the model parameters of the 2SK170 and 2SJ74 in my Pspice. They are nowhere near a 6-7mA device. In order to be more accurate with the sim, I need a more realistic model.
I am measuring Id vs vgs, Idss, and Id when the part is placed in the circuit.
Patrick said to flg if he lowers the current running through the jfets eg 6mA then he can increase the the drain resistor an hence increase the open loop gain.
I was merely telling him what to adjust to get 6mA. I am not endorsing anything here
Patrick said to flg if he lowers the current running through the jfets eg 6mA then he can increase the the drain resistor an hence increase the open loop gain.
I was merely telling him what to adjust to get 6mA. I am not endorsing anything here
The original circuits from NP runs at about 6mA at the frontend.
I merely pointed out that if your spice model is at 11mA Idss, it will run in circuit at about 10mA due to 10R source resistor. You lose OL gain and this gives different overall results. And Dave kindly offered you tips how to change it back to 6mA.
Patrick
I merely pointed out that if your spice model is at 11mA Idss, it will run in circuit at about 10mA due to 10R source resistor. You lose OL gain and this gives different overall results. And Dave kindly offered you tips how to change it back to 6mA.
Patrick
I have been able to spend a little time simmimng with Ppsice. Keep in mind YMMV
I used Melon Head's suggested modification to the JFET model and it seems to be working more like we want. I made a test circuit that has about 6.5mA Iq in the front end with a volatge gain of about 34. The JFET drain resistors adjusted to about 740 ohms. I tried to keep passive components and everything faithull to the F5. Completeing the circuit a pair of un-modified IRFP240 and 9240s with 1.3A bias.
The stock F5 shows OLG of about 333 with an 8 ohm load and OL BW is -3db @ 63kHz. Closed loop gain measures about 5.88 and has a -3db of 3MHz!
I used Melon Head's suggested modification to the JFET model and it seems to be working more like we want. I made a test circuit that has about 6.5mA Iq in the front end with a volatge gain of about 34. The JFET drain resistors adjusted to about 740 ohms. I tried to keep passive components and everything faithull to the F5. Completeing the circuit a pair of un-modified IRFP240 and 9240s with 1.3A bias.
The stock F5 shows OLG of about 333 with an 8 ohm load and OL BW is -3db @ 63kHz. Closed loop gain measures about 5.88 and has a -3db of 3MHz!
Very hard to tell. Appears to not have one at least to Unitty gain of about 25MHz
I assume you have a plot of OLG vs Freq. After the first corner freq the OLG has -6dB slope. After the second, it has -12dB. So you cannot miss. If you do not see a change of slope in OLG all the way up to 0dB (at which point OLG drops to 1 from 333), then the amp is unconditionally stable even at unity gain. But I doubt that. You could also look at the phase and see where you change from 0° to 90° to 180°.
Patrick
Patrick
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.