As far as I know, no, it is a function of the LTspice model formula rather than Bob's or any other specific model values.
I have a recollection that Self mentions this briefly because he couldn't successfully simulate some observations.
Says he discussed it with Baxandall and/or Cherry and they said you would need a quadratic term at least.
Will try to find the reference and also see if LTspice has had any updates.
Best wishes for the season
David
You receive the Cherry CE schematic I sent you?
This is a personal thing of course but for me a took literally a few years to get rid of an uneasy feeling that I was loafing and cheating by not going to work.
But I'm over that now, finally
Jan
The default Spice Gummel Poon model does not include a nonlinear Early effect. If the VAF value has any impact on distortions then it is indirect, through the Ccb contribution.
Thanks Andrew, good pick up. R121 has been selected so that the voltage dropped across it is equal to that across R107/R108 for best input stage balance. I've split the resistor to prevent the protection transistor from turning on too soon.
I've seen a lot of schematics with a resistor in the collector of the VAS EF to protect it when overdriven. But wouldn't this put a voltage at the collector, causing Cbc to modulate with the signal?
I have seen from 0r0 to 10k in the collector lead of the EF helper.
I have also seen that it should not be 0r0.
I have seen comment that this resistor affects output quality.
I use the minimum that protects the transistor. 1k to 2k usually does that. (with your high value emitter resistor you could use a lower value collector resistor, but base current into VAS needs to be estimated).
You can use the simulator and ask it what it predicts for the effect on signal quality.
Last edited:
For those who desire (a) quick power-up and (b) maximum PSRR at deepest bass frequencies, consider installing two components from collector to negative rail: a zener diode and a capacitor in parallel.
Hi Mark, what's not apparent from the schematic I posted at here is that the rails feeding the IPS are fed from regulated supplies, which is why I haven't implemented the usual measures such as filtering the CCS Vrefs either.
The power supply board's schematic is attached. I used to obtain the boosted and regulated rails by adding windings to the PT as described by Bob in his book, but I now find it far more convenient to use a voltage doubler circuit instead. Only two "extra" parts (C9, C10) vs. Bob's method - a no brainer IMHO to avoid fiddling around with magnet wire and kapton tape.
When I revisit this board I'll do away with the boosted +ve rail altogether and simply bootstrap the VAS CCS and feed the LTP tail from the +15V opamp rail.
The main rails under the control of an 8-pin PIC microcontroller (off board, not shown) which delays turn on to eliminate thump and disconnects power if DC is sensed on the speaker line.
The power supply board's schematic is attached. I used to obtain the boosted and regulated rails by adding windings to the PT as described by Bob in his book, but I now find it far more convenient to use a voltage doubler circuit instead. Only two "extra" parts (C9, C10) vs. Bob's method - a no brainer IMHO to avoid fiddling around with magnet wire and kapton tape.
When I revisit this board I'll do away with the boosted +ve rail altogether and simply bootstrap the VAS CCS and feed the LTP tail from the +15V opamp rail.
The main rails under the control of an 8-pin PIC microcontroller (off board, not shown) which delays turn on to eliminate thump and disconnects power if DC is sensed on the speaker line.
Attachments
It took me many years to adjust to the slow pace and unlimited time to do nothing..... ..... the pace and people and work-load goes to zero with nothing much to fill the hole. And, the average man on the street/neighbor isnt a good substitute, either. Thanking the Gods for our hobbies.
-Richard
AFAIK they both use a Modified Gummel Poon model for BJTs and are essentially equivalent in that respect.
But LTspice has made proprietary improvements to way FETs can be modeled so that's why I wrote that I would check if there were any updates, in case he had made similar improvements to the BJT models.
Best wishes
David
You bet, I was lucky to come in at a good time and do the retirement thing on my own terms. I'm sorry for those who did not have that opportunity but at least in some cases the severance packages are still very generous.
BTW Dick I noticed California's homestead protection is pretty meager, in MA it costs $35 to file for $1M protection of property from ALL litigation after you are 65. A few years ago they expanded it to include money held in certain trusts.
Last edited:
Micro-Cap 11.0.2.0
Models
Parker-Skellern Gaasfet Model
A fourth GAASFET model option was added.
New BSIM4 Model
The latest BSIM 4.7 model was added.
Updated JFET Model
Added the ALPHA, VK, ISR, N, NR parameters.P
M Multiplier
The M multiplier command line parameter was added to diodes, BJTs, and JFETs.
I do not know enough to compare them. They have been at a while with 11.0.2.0 version
Models
Parker-Skellern Gaasfet Model
A fourth GAASFET model option was added.
New BSIM4 Model
The latest BSIM 4.7 model was added.
Updated JFET Model
Added the ALPHA, VK, ISR, N, NR parameters.P
M Multiplier
The M multiplier command line parameter was added to diodes, BJTs, and JFETs.
I do not know enough to compare them. They have been at a while with 11.0.2.0 version
Since we are stimulated by life and electronics to stay alive.
May the magic smoke always carry the same fragrance.
It's interesting that the ink sticks used in Chinese/Japanese painting are scented with rare herbs and jasmine which you smell while preparing your ink.
Mr. C.
it seems several boyz are interested in your Hafler DH 200C , over at tweak thread
http://www.diyaudio.com/forums/solid-state/31131-hafler-dh-200-220-mods-26.html#post4924742
I think it would help them if you just post sch file , to save them staring at ootoobe video and drawing schmtc that way
Hi Zen Mod,
Thanks for the tip. I'll head over there and check it out.
Cheers,
Bob
If the current source is being used as the load for a single-ended VAS that is Miller-compensated, bear in mind that the impedance at the VAS output load is really dominated by the low impedance created by the shunt feedback of the Miller compensation. This low impedance persists to quite low frequencies in many designs. The key is that we want very low nonlinearity the impedance of things hangling off the Hi-z VAS node. Also, do not overlook the shunt impedances on this node created by the output stage pre-driver or driver transistors via their Ccb and Early effect, not to mention the amplifier load reflected back through the 2 or 3 emitter followers in a 2EF or 3EF output stage.
Cheers,
Bob