Since I don't know yet whether and which problems may be associated with the bootstrapping buffered from the emitters, I'm a bit hesitant to have this as the only option. By adding two Zener diodes as options (D7 and D8 in the schematic attached), I could "downgrade" to bootstrapping from the tail node. Two CCS and two Zeners basically in parallel are kind of redundant, but I have them on the PCB anyway.
Since the CCS show good matching, I could also connect the Zener anodes to the emitter, thus having bootstrapping from the emitters without buffers.
What do you think?
Would using two CCS and two Zeners fed to each emitter be a good or bad idea? The benefit would be bootstrapping from the emitters without buffers.
Since the CCS show good matching, I could also connect the Zener anodes to the emitter, thus having bootstrapping from the emitters without buffers.
What do you think?
Would using two CCS and two Zeners fed to each emitter be a good or bad idea? The benefit would be bootstrapping from the emitters without buffers.
Sorry for the messy schematics I posted.
I wonder how the magic floating voltage (battery) of the boostrapped cascode shown in textbooks could be generated in reality.
Wouldn't a CCS and Zener be the most obvious solution?
I wonder how the magic floating voltage (battery) of the boostrapped cascode shown in textbooks could be generated in reality.
Wouldn't a CCS and Zener be the most obvious solution?
Elegant solution since this is self biasing. The gate of the JFET could also be tied to the emitter.
Vds of JFETs is rather limited so the usefulness of the JFET cascode is limited to low voltage applications.
I just have Kolinummi's book open on page 127.
The circuit with the battery cancels the cascode transistors base current, but this works only with the battery generating the offset. The cancellation effect is gone using a buffer. And the problem is nonexistent with a JFET.
Vds of JFETs is rather limited so the usefulness of the JFET cascode is limited to low voltage applications.
I just have Kolinummi's book open on page 127.
The circuit with the battery cancels the cascode transistors base current, but this works only with the battery generating the offset. The cancellation effect is gone using a buffer. And the problem is nonexistent with a JFET.
I found some photos I saved from the internet to my phone.
Unfortunately without context.
Seems like the idea with the Zeners and CCS is not too far off.
Both examples have capacitors in parallel either to the Zeners or resistors generating the offset. Looks like the capacitors are important.
Unfortunately without context.
Seems like the idea with the Zeners and CCS is not too far off.
Both examples have capacitors in parallel either to the Zeners or resistors generating the offset. Looks like the capacitors are important.
Attachments
Those are from Hawksford's paper on the bootstrap cascode technique if I'm not mistaken.
Edit: Yes, see here
Edit: Yes, see here
Last edited:
I had a closer look at the mechanism of the emitter resistors added to the cascodes. What they do is spoiling the positive feedback of the cascode arrangement, but this also means that they may spoil the performance if the value is too high. A value of 100R seems far too high. Maybe 10R is enough to avoid instability without ruining the performance. I read that those resistors are good from multiple sources, like here Nelson Pass: https://www.diyaudio.com/community/threads/cascodes-the-truth-is-out-there.63187/post-7047078.
I illustrated the effect of the cascode emitter resistors stepped 1R, 10R and 100R:
@Lee Knatta
Yes, 100 Ohm is quite large and mitigates the benefit of having cascodes, I am glad you found out the negative consequence of these resistors when being too large, but I understood also your general notion of using additional resistors for damping purpose.
Another thing while typing here, you mentioned the bootstrapping of the cascode using a battery in post #42, I don't know what Kolinummi have written in his book, but I think the battery symbol is some times in theoretical discourse used to just symbolize a needed fixed potential between two nodes without going into what the practical circuit would look like.
And a practical circuit could be as shown in your post by using something based around a zener diode, a capacitor in parallel would be a good addition as zeners have a bit high impedance and are somewhat noisy too.
Yes, 100 Ohm is quite large and mitigates the benefit of having cascodes, I am glad you found out the negative consequence of these resistors when being too large, but I understood also your general notion of using additional resistors for damping purpose.
Another thing while typing here, you mentioned the bootstrapping of the cascode using a battery in post #42, I don't know what Kolinummi have written in his book, but I think the battery symbol is some times in theoretical discourse used to just symbolize a needed fixed potential between two nodes without going into what the practical circuit would look like.
And a practical circuit could be as shown in your post by using something based around a zener diode, a capacitor in parallel would be a good addition as zeners have a bit high impedance and are somewhat noisy too.
Yes, the battery symbol is used exactly for that: Kolinummi's schematics use the battery symbol all the time to indicate there needs to be a voltage offset. It is up to the reader to figure out how to practically implement this voltage offset then. The book is excellent BTW, but some ideas may be difficult to implement in reality because often he does not explain what kind of hacks one might need to add in order to get something really stable in reality.
Apart from shunting noise, which is only effective if the impedance of the reference (Zener, LED, resistor) is high, I believe the parallel capacitor is aiding stability since the reference would be bypassed at high frequency reducing positive feedback.
Apart from shunting noise, which is only effective if the impedance of the reference (Zener, LED, resistor) is high, I believe the parallel capacitor is aiding stability since the reference would be bypassed at high frequency reducing positive feedback.
I wonder if bootstrapping concept can be applied to cascoded VAS and if it is of any benefit. Comparing to Hawskford enhanced cascode?
