I'm not sure if it's really oscillating here. Normally, the strange situation remains stable when switching the thing on. It reads 0V on upper, and 10V on lower tree, for example. Switching it off an on leads e.g. to both having 0V and so on.
very strange, though.
Rüdiger
very strange, though.
Rüdiger
Jocko, my comments about noise performance apply specifically to Onvinyl's first circuit. In this specific circuit, any voltage noise source connected in series with the input JFET's has much more influence than an equal voltage noise source connected in series with the bipolar cascading and current source transistors. Of course that doesn't mean that gate stopper resistors are necessarily unacceptable in other circuits, or that base stopper resistors are necessarily acceptable in other circuits.
I know that....
But you are going to try to convince me that a small resistor.....10-50 ohms...... is going to raise the noise floor THAT much??
No way do I buy that. Maybe JC has an opinion on that.
Jocko
But you are going to try to convince me that a small resistor.....10-50 ohms...... is going to raise the noise floor THAT much??
No way do I buy that. Maybe JC has an opinion on that.
Jocko
Of course it will significantly increase your noise. Why bother to use large area fets and put them in parallel, if you are going to use an input resistor with the same noise that you just removed? In this case, you have to work without an input resistor, at least anything more than 10 ohms.
My circuits run open loop, and they have the equivalent noise of a 10 ohm resistor. Your results are closer to a 50 ohm resistor, maybe much worse if you use any resistor at the input or as local feedback.
My circuits run open loop, and they have the equivalent noise of a 10 ohm resistor. Your results are closer to a 50 ohm resistor, maybe much worse if you use any resistor at the input or as local feedback.
Onvinyl,
There are 100 ohm resistors connected to the bases, but they are decoupled with relatively large capacitors. If the wire inductance from the bases to the capacitors is sufficient to cause oscillations, the 100 ohm resistors are not going to help, because they cannot spoil the Q of this inductance. If the bipolars are oscillating due to wire parasitics, you need a resistor or a high-loss ferrite bead as close as possible to the base without any decoupling between the resistor and the base.
By the way, if you try ferrite beads, be sure to take a lossy type intended for interference suppression, not a high-Q type intended for making high-quality RF inductors.
There are 100 ohm resistors connected to the bases, but they are decoupled with relatively large capacitors. If the wire inductance from the bases to the capacitors is sufficient to cause oscillations, the 100 ohm resistors are not going to help, because they cannot spoil the Q of this inductance. If the bipolars are oscillating due to wire parasitics, you need a resistor or a high-loss ferrite bead as close as possible to the base without any decoupling between the resistor and the base.
By the way, if you try ferrite beads, be sure to take a lossy type intended for interference suppression, not a high-Q type intended for making high-quality RF inductors.
I have small 'rings' lying a´round, that fit easily on devices' wire. Are you talking about these?
Rüdiger
Dunno if it came along already. Concerning the noise issue maybe this is of help:
http://www.klaus-boening.de/html/lab_page.html#selection
Cheers 😉
http://www.klaus-boening.de/html/lab_page.html#selection
Cheers 😉
Ferrites are made of more than one "mix", and the frequency range that they are designed for is optimised by selecting which "mix" is used.
There are 2 predominate types:
One is useful in the 20-100 MHz range. These are the most common. The other is used in the 5-10 MHz or so range.
Jocko.
There are 2 predominate types:
One is useful in the 20-100 MHz range. These are the most common. The other is used in the 5-10 MHz or so range.
Jocko.
"I have small 'rings' lying a´round, that fit easily on devices' wire. Are you talking about these?
Rüdiger"
They are probably OK.
Rüdiger"
They are probably OK.
By the way, if one of the JFET gates is not connected properly and does not have a path to a defined DC voltage, its voltage might start floating in all directions due to leakage currents. For example, suppose that there is no signal source connected to the input and that the bottom side of the upper 270 ohm bias resistor is accidently not connected to ground. In that case, almost anything can happen to the bias voltages.
(I once had this problem with a professional electret capacitor microphone. It always worked fine when I measured it, but not when someone actually wanted to use it. The reason was simply that a wire going to one pin of the 10Gohm bias resistor of the input FET was disconnected, but I always inadvertently discharged the microphone with my skin when I opened it to measure, bringing the FET gate at a usable bias voltage.)
(I once had this problem with a professional electret capacitor microphone. It always worked fine when I measured it, but not when someone actually wanted to use it. The reason was simply that a wire going to one pin of the 10Gohm bias resistor of the input FET was disconnected, but I always inadvertently discharged the microphone with my skin when I opened it to measure, bringing the FET gate at a usable bias voltage.)
Yes, I'm ready to find a nasty thing like that here. I will check for all such issues and a few more, not taking anything for granted and see where it gets me...
thanks,
Rüdiger
May be a little off topic, but I would like to point out that the drain and source of some of the Japanese-made general-purpose N-channel JFETs can be interchanged in a circuit without ill effect, or for that matter, much difference in measured performance or sound quality. I am currently holding a Toshiba 2SK30 in my hand, and the interchangeability of the source and drain pins is definitely true for this critter.
The reason is that for low-frequency general-purpose N-JFETs, the Japanese manufacturer often built the semiconductor die symmetrically relative to the gate.
This observation does not apply for high-frequency N-JFETs, very high Gm N-JFETs, and P-JFETs. In these latter cases, a symmetrical die construction is seldom used, due to the need to optimize the performance of the FET for its more specialized purposes.
Knowing this trick can make your life easier when you are designing a pcb layout. But do test this out on a breadboard prior to finalising the layout. No sense in taking unnecessary risks!
hth, jonathan carr
The reason is that for low-frequency general-purpose N-JFETs, the Japanese manufacturer often built the semiconductor die symmetrically relative to the gate.
This observation does not apply for high-frequency N-JFETs, very high Gm N-JFETs, and P-JFETs. In these latter cases, a symmetrical die construction is seldom used, due to the need to optimize the performance of the FET for its more specialized purposes.
Knowing this trick can make your life easier when you are designing a pcb layout. But do test this out on a breadboard prior to finalising the layout. No sense in taking unnecessary risks!
hth, jonathan carr
very interesing
Thanks Jonathan.
This is very interesting and I wasn't aware of it. It is interesting that a certain schematic of a certain very low noise phono preamp (anyone who has seen it will know which and don't ask. don't tell to avoid really annoying a very respected member of the forum) using the K147 and J72 shows the drain and source swapped.
In this reverse engineered schematic, the drain and source for two paralleled K147s are show as connected, that is, the drain and source have been swapped for one of the pair. The same is done for the pair of J72s. I assume this would be undesirable to connect asymmetrical die jfets in the manner.
Not having seen the PCB I can only assume this connection might been made to ease PCB layout or to couple the jfets together along the flat face surfaces for thermal coupling.
Are the J72 and K147 not asymmetrical? Which high transcondutance Japanese jfets are? I also notice the die for the second source version of the K147 appears to be symmetrical. I have also
been told by another usually reliable source that the drain and source can be swapped for the K170.
Thanks Jonathan.
This is very interesting and I wasn't aware of it. It is interesting that a certain schematic of a certain very low noise phono preamp (anyone who has seen it will know which and don't ask. don't tell to avoid really annoying a very respected member of the forum) using the K147 and J72 shows the drain and source swapped.
In this reverse engineered schematic, the drain and source for two paralleled K147s are show as connected, that is, the drain and source have been swapped for one of the pair. The same is done for the pair of J72s. I assume this would be undesirable to connect asymmetrical die jfets in the manner.
Not having seen the PCB I can only assume this connection might been made to ease PCB layout or to couple the jfets together along the flat face surfaces for thermal coupling.
Are the J72 and K147 not asymmetrical? Which high transcondutance Japanese jfets are? I also notice the die for the second source version of the K147 appears to be symmetrical. I have also
been told by another usually reliable source that the drain and source can be swapped for the K170.
Attachments
Source down the drain
Thanks Jonathan,
I remember the discussion in the some Kaneda articles from the early 80's where they discussed swapping drain and source. I had no real insight into why it was possible. The used FET was the 2SK30 (AGR?). Thanks.
Thanks Jonathan,
I remember the discussion in the some Kaneda articles from the early 80's where they discussed swapping drain and source. I had no real insight into why it was possible. The used FET was the 2SK30 (AGR?). Thanks.
Re: very interesting
Hi Fred,
It is in Horowitz page 120 in my edition.
But your friend Jocko knows it.
Now you are wondering how I know Jocko knows it?
Never mind.
Fred Dieckmann said:
Hi Fred,
It is in Horowitz page 120 in my edition.
But your friend Jocko knows it.
Now you are wondering how I know Jocko knows it?
Never mind.
Before some ask.......
I am of aware of asymmetrical jfet design
but was only curious if the particular devices I mentioned were designed in this way. I wonder if I should cut one open to look at it? If understand the referemce below, DS reversal will work for some asymmetric jfet designs but not as well as in the prefered direction.
http://www.play-hookey.com/semiconductors/junction_fet.html
I am of aware of asymmetrical jfet design
but was only curious if the particular devices I mentioned were designed in this way. I wonder if I should cut one open to look at it? If understand the referemce below, DS reversal will work for some asymmetric jfet designs but not as well as in the prefered direction.
http://www.play-hookey.com/semiconductors/junction_fet.html
Before some ask.......
I am of aware of asymmetrical jfet design but was only curious if the particular devices I mentioned were designed in this way. I wonder if I should cut one open to look at it? If understand the referemce below, DS reversal will work for some asymmetric jfet designs but not as well as in the prefered direction.
http://www.play-hookey.com/semiconductors/junction_fet.html
I am of aware of asymmetrical jfet design but was only curious if the particular devices I mentioned were designed in this way. I wonder if I should cut one open to look at it? If understand the referemce below, DS reversal will work for some asymmetric jfet designs but not as well as in the prefered direction.
http://www.play-hookey.com/semiconductors/junction_fet.html
- Status
- Not open for further replies.