Hi Pete, thanks for the comments.
I have some ideas on how bump up the value of R2, but they won't be revealed until much later
. The preview is that in typical design R2 needs to match RF (the feedback resistor) hence why I did not want to make it too large, else RG will be large, and the amp will be noisy. On top of this, if the input stage diff pair are BJTs, if R2 is too large you get problems due to base current matching, etc. Hence, as you correctly point out, non-JFET amps are around the 20K mark. I know I did not mention this yet in the video, I'll talk about it when designing the gain network. Now, what if there was a way to break this need of R2=RF... keep tuned .
Question, is the output impedance of these tube pre-amps larger than 1K (I really don't know, not a tube person)? The danger with high output impedance sources is really C1. If the impedance (RS) is too large, the low pass filter formed by RS+R1 and C1 will low pass the source's signal and you don't want that. It is true that R2 will also divide the signal down, but you can always crank the volume up more... ok, we'll bump up R2 later, but will not change R1, maybe adjust C1.
Finally regarding the NP0, you are right, made a mistake... but NP0/C0G should also work. Not sure if you can find something large enough for C2, but definitely you can find one for C1.
Thanks again for the comments Pete.
I have some ideas on how bump up the value of R2, but they won't be revealed until much later
. The preview is that in typical design R2 needs to match RF (the feedback resistor) hence why I did not want to make it too large, else RG will be large, and the amp will be noisy. On top of this, if the input stage diff pair are BJTs, if R2 is too large you get problems due to base current matching, etc. Hence, as you correctly point out, non-JFET amps are around the 20K mark. I know I did not mention this yet in the video, I'll talk about it when designing the gain network. Now, what if there was a way to break this need of R2=RF... keep tuned .Question, is the output impedance of these tube pre-amps larger than 1K (I really don't know, not a tube person)? The danger with high output impedance sources is really C1. If the impedance (RS) is too large, the low pass filter formed by RS+R1 and C1 will low pass the source's signal and you don't want that. It is true that R2 will also divide the signal down, but you can always crank the volume up more
... ok, we'll bump up R2 later, but will not change R1, maybe adjust C1.Finally regarding the NP0, you are right, made a mistake... but NP0/C0G should also work. Not sure if you can find something large enough for C2, but definitely you can find one for C1.
Thanks again for the comments Pete.
Last edited:
You can find the specs and schematics here, there is no cathode follower (for economy)
at the output. Output Z is rated at 1K and loading should be 100K ohms or higher. This
pre was commonly used with the ST-70 tube power amp that had a 470K ohm input
impedance:
http://www.audioman.com/Dynaco/pas3/pas3.htm
Forgot to mention that the tone controls are at the output of the line pre and are probably
influenced by loading.
at the output. Output Z is rated at 1K and loading should be 100K ohms or higher. This
pre was commonly used with the ST-70 tube power amp that had a 470K ohm input
impedance:
http://www.audioman.com/Dynaco/pas3/pas3.htm
Forgot to mention that the tone controls are at the output of the line pre and are probably
influenced by loading.
I see your use of the simulation tool as an essential part of the design process. You will eventually have to cross that bridge at some point. I don't recommend leaving your content to public popularity vote; there is too much variance in the audience.
At worst, you could make a parallel series for LTSpice.
At worst, you could make a parallel series for LTSpice.
Hi Shaun, I actually come from the integrated circuit industry where we design the entire IC using simulation software (there is no other way). Once the design is complete, we send it to the fab, get it fab-ed and then evaluate.
I have designed about 15 IC's in this fashion and the plan is to follow a similar approach here, i.e. design the entire SW-VFA-01 in LTSpice, then build it, and finally evaluate it. I believe that LTSpice is powerful enough and the models good enough to follow this approach. We shall see.
Worst case, it will blow up in my face. Best case, we get a super-awesome amplifier. Guaranteed case, everyone learns something out of this.
I have designed about 15 IC's in this fashion and the plan is to follow a similar approach here, i.e. design the entire SW-VFA-01 in LTSpice, then build it, and finally evaluate it. I believe that LTSpice is powerful enough and the models good enough to follow this approach. We shall see.
Worst case, it will blow up in my face. Best case, we get a super-awesome amplifier. Guaranteed case, everyone learns something out of this.
Don't worry, I simulated many amps with LTspice and built some of them and all worked with very minor changes.
One example is this amp, 200W and 100W version built and in use.
200W MOSFET CFA amp
Damir
3.1. Audio Amplifier Design (SW-VFA-01): Fundamentals of negative feedback SW-VFA-01
Hi all, just to let you know that I just uploaded the next video of the series:
3.1. Audio Amplifier Design (SW-VFA-01): Fundamentals of negative feedback
SW-VFA-01 Video 3.1
Also will be launching parallel video series: LTSpice: How to. The idea in this one is to help you guys upgrade your game on LTSpice.
Enjoy,
Sandro (SW Audio)
Hi all, just to let you know that I just uploaded the next video of the series:
3.1. Audio Amplifier Design (SW-VFA-01): Fundamentals of negative feedback
SW-VFA-01 Video 3.1
Also will be launching parallel video series: LTSpice: How to. The idea in this one is to help you guys upgrade your game on LTSpice.
Enjoy,
Sandro (SW Audio)
Good luck and thank you !
I have question about ground loop and the resistor + diodes.
This decouples the "source ground" (input jack chassis ?) from the "amplifier ground" (the first gain stage ?) ?
I see R2 is grounded too, if I ground the input rca to chassis where should R2 be grounded ?
I learned that the first gainstage should be grounded to the input jack and nowhere else. Please help me understand.
Sorry to bother, I am still learning, and asking too much & reading too much the forum
I have question about ground loop and the resistor + diodes.
This decouples the "source ground" (input jack chassis ?) from the "amplifier ground" (the first gain stage ?) ?
I see R2 is grounded too, if I ground the input rca to chassis where should R2 be grounded ?
I learned that the first gainstage should be grounded to the input jack and nowhere else. Please help me understand.
Sorry to bother, I am still learning, and asking too much & reading too much the forum
Hi Federico,
Let me address your questions.
First and foremost, it is absolutely no bother. So ask away. I like people asking questions, it means they are paying attention.
The ground loop breaking / ground decoupling resistor do decouple the "source ground", i.e. the ground in the input jack, and the amplifier ground. The latter is the star ground of the amplifier.
So let's think this through again:
- The gain of the amplifier will be 1x at DC and ~35x after ~2Hz.
- Resistor R2 defines the DC ground reference for the amplifier, which may lead you to believe that the noise in this ground will be gained by 1x.
- If, the feedback network of the amplifier is referred to this same ground though, the ground noise becomes common mode and is rejected. At this point, we rely on the high CMRR of the amplifier to reject it.
- Referring to the local ground prevents having unnecessary current being returned to the source, but, the source's ground will probably be quieter since the currents being injected to that ground are much lower (no huge supply capacitor current or return currents from the speakers being injected to the ground).
Given all the facts above, I think I am changing my mind on this one, and probably the better trade-off is to return a bit more current to the source ground line and take advantage of the quieter ground. This way we put less pressure on the CMRR of the amplifier.
I am attaching a modified schematic.
Thanks for pointing this out and make me think it through again.
Best, Sandro
Let me address your questions.
First and foremost, it is absolutely no bother. So ask away. I like people asking questions, it means they are paying attention.
The ground loop breaking / ground decoupling resistor do decouple the "source ground", i.e. the ground in the input jack, and the amplifier ground. The latter is the star ground of the amplifier.
So let's think this through again:
- The gain of the amplifier will be 1x at DC and ~35x after ~2Hz.
- Resistor R2 defines the DC ground reference for the amplifier, which may lead you to believe that the noise in this ground will be gained by 1x.
- If, the feedback network of the amplifier is referred to this same ground though, the ground noise becomes common mode and is rejected. At this point, we rely on the high CMRR of the amplifier to reject it.
- Referring to the local ground prevents having unnecessary current being returned to the source, but, the source's ground will probably be quieter since the currents being injected to that ground are much lower (no huge supply capacitor current or return currents from the speakers being injected to the ground).
Given all the facts above, I think I am changing my mind on this one, and probably the better trade-off is to return a bit more current to the source ground line and take advantage of the quieter ground. This way we put less pressure on the CMRR of the amplifier.
I am attaching a modified schematic.
Thanks for pointing this out and make me think it through again.
Best, Sandro
Sandro
While on the topic, I wonder if you might perhaps take the opportunity to explore and discuss various grounding approaches:
1) From mid-80s DIY electronics magazines and electronics books of the time I learned about the practice of a star ground in order to avoid ground loops.
2) Later, still observing the star approach, I used the practice of thinking in terms of "signal ground and power ground" or "small signal ground vs large signal ground". Essentially, the two respective grounds would exist separately and then be linked by a low value resistor or 0R link at the star point. The resistor was a way of keeping the nets separate in the schematic when routing the PCB.
3) Nowadays, with very high speed electronics being commonplace, there seems to be a trend to use ground planes with no apparent electrical separation between small signal paths and power routes (I say apparent, because I have not explored this to any substantial extent).
There may be other variations on grounding practice, but those are what I know.
If not appropriate or withing the scope of the work you are kindly sharing, I apologise
While on the topic, I wonder if you might perhaps take the opportunity to explore and discuss various grounding approaches:
1) From mid-80s DIY electronics magazines and electronics books of the time I learned about the practice of a star ground in order to avoid ground loops.
2) Later, still observing the star approach, I used the practice of thinking in terms of "signal ground and power ground" or "small signal ground vs large signal ground". Essentially, the two respective grounds would exist separately and then be linked by a low value resistor or 0R link at the star point. The resistor was a way of keeping the nets separate in the schematic when routing the PCB.
3) Nowadays, with very high speed electronics being commonplace, there seems to be a trend to use ground planes with no apparent electrical separation between small signal paths and power routes (I say apparent, because I have not explored this to any substantial extent).
There may be other variations on grounding practice, but those are what I know.
If not appropriate or withing the scope of the work you are kindly sharing, I apologise
Shaun ;
That would be very good for the DIY community.
Because besides M.Blencowe chapter on grounding, there is very low amount of content on grounding for audio amplifier.
But I think he will because he said grounding will be addressing in future videos.
I also experimented with star grounding, the return currents going to a star point are different and the wire/resistor will inevitably create a potential difference at the star.
Which will inevitably create noise that is amplified ? So keep the star wires/plane thick I guess.
Also mister Sandro don't thank me, I should be the one being grateful.
I am looking forward to this series. I will probably "bother" you again, so I thank you!
That would be very good for the DIY community.
Because besides M.Blencowe chapter on grounding, there is very low amount of content on grounding for audio amplifier.
But I think he will because he said grounding will be addressing in future videos.
I also experimented with star grounding, the return currents going to a star point are different and the wire/resistor will inevitably create a potential difference at the star.
Which will inevitably create noise that is amplified ? So keep the star wires/plane thick I guess.
Also mister Sandro don't thank me, I should be the one being grateful.
I am looking forward to this series. I will probably "bother" you again, so I thank you!
Last edited: