@jenimitso, it's an example for the circuit and the capacitor, no real world amplifier. I don't want subsonic response, right, but I also don't want about +3dB at 100Hz. But this are just my two cent.
...and completely off topic.
I have some new ideas and will simulate and test some of the inverter stages. I will also think about the double mu-follower. Actually I'm not sure about the high gain it will provide, but this is off topic as well.
Thanks again for the great ideas and discussions!
But please allow me to clarify, that I don't have the simulation models Calvin used and tried to do it quick and dirty to support us here. Calvins circuit using my FETs don't provide a good riaa correction and therefore I select a different one.
So you are right, that if you select other components, it will not be that dramatical. But even at 3u3 you see a slightly increase of gain with my lower riaa impedance. If this is now an issue or maybe a feature... that's up to you.
And finally, I don't know the reason why Calvin made his example circuit like this. But I believe it’s a great idea and good starting point for own designs. Actually I don’t use the 68u elco at the moment. 😉
That's a good input, thanks. Actually I use a cascode at the input and will think about this as well. 👍
I do not understand why we should use transistors to make an inverter.
Almost any opamp would do a perfect job.
Almost any opamp would do a perfect job.
Now I'm surprised... have I missed something? With the cascode it's having a good response... with 68u. Again, wrong Fets, but the impact of the capacitor is bigger than with the quick and dirty example, is't it?
Why, @Calvin used a cascode as well and I like to compare the influence of the capacitor, nothing else? Why makes it again no sense for you? I mean, you said that it's necessary to leave the circuit as it is to compare and if I don't use a cascode to compare the influence of the cap... it's a different circuit. 🤔🤔🤔
Last edited:
This was what you recommends. I don't like to be the smart-***. I thy to understand and learn from you guys ways more familiar with audio circuits than I am.
I'm joking @jenimitso ! Let me have a deeper look and better understanding first. I don't know today what are the "weird parameters" that will make this big capacitor necessary. The influence of the 2Meg7 resistor at the upper bias circuit is also remarkable.
Any idea how to do that? 🤔 I guess with feedback this should not be an issue, but without?
Hi,
since You have a high gain first stage the Miller capacitance will be high also.
The lownoise JFETs already come typically with increased capacitance values.
The overall input capacitance of the circuit can then easily reach a range where it influences on the amplitude response of the cartridge.
AT MM-pickups for example like low capacitances as load. Shure and Ortofon seem less sensitive in this regard.
If Your cable from pickup to amp-input already has ~100pF of capacitance You won´t like to add another couple of hundreds of pF to it.
With MC pickups the capacitance doesn´t play a role anymore (but then you need to parallel the input JFETs to get enough gain).
Also the cascode reduces distortion due to the almost constant Vds the lower JFET works under.
A good cascoding JFET for the SK170 are the 4391 which generate ~4V for the 170 at typical drain currents.
jauuu
Calvin
@EUVL
long time no see pal .... happy new year to You also
since You have a high gain first stage the Miller capacitance will be high also.
The lownoise JFETs already come typically with increased capacitance values.
The overall input capacitance of the circuit can then easily reach a range where it influences on the amplitude response of the cartridge.
AT MM-pickups for example like low capacitances as load. Shure and Ortofon seem less sensitive in this regard.
If Your cable from pickup to amp-input already has ~100pF of capacitance You won´t like to add another couple of hundreds of pF to it.
With MC pickups the capacitance doesn´t play a role anymore (but then you need to parallel the input JFETs to get enough gain).
Also the cascode reduces distortion due to the almost constant Vds the lower JFET works under.
A good cascoding JFET for the SK170 are the 4391 which generate ~4V for the 170 at typical drain currents.
jauuu
Calvin
@EUVL
long time no see pal .... happy new year to You also
Try mmbf5486 as lower JFET, and add a source resistor, say ~1k.
The top JFET can be mmbf5484.
The followers can be 2SK2145GR.
Will get you about 20x gain per stage.
(According to Spice)
Patrick
PS you need a JFET with higher Yos for low mu.
Hi guys. What will be the right upper Fet for JFE2140? In the datasheet they use the same Fet (same package) but I'm a little curious about Vgs match, because I read a higher Vgs for the upper Fet will be better.
Assuming you want to design a diff amp, I'd use something that guarantees Vds for the mains FETs will be well-matched, notably when Vds is low (near the knee). Thus, another JFE2140 biasing each main FET's drain at least 2V above its source. Biasing of the cascode FETs is easy because of the high gate impedance. Cascoding sort of steals supply voltage, though, because both the main FET and the cascode FET want to have enough Vds at all times.
Thanks @KSTR . As I already said, I'm thinking in many directions and beside the two-stage mu-follower with reduced gain, I'm thinking also about a mu-follower with higher gain but the cascode FETs connected to the source resistor... like below. I’m sorry, it may be confusing, but I’m still in the phase of looking in many possible directions before I decide what’s the right way to go.
There’s the link I mentioned. It's an example and although it's for a current source, you see the relevance of the right selection.
https://www.preamp.org/static/cascoded-jfet-current-source/index.html
I learned from @jenimitso that a cascode may be not necessary for a phono input stage, but to evaluate that, I like to make a comparison and are wondering if there are special requirements for the upper FET. During this research, I found the link attached.
I like to go with JFE2140 because I know them. On the bench and in simulation the cascode is working good using them as upper and lower FET. But, and this is important, I don’t have the experience to decide if others may work ways better. By changing the upper FET with J112 in my simulation, there is a huge gain difference.
Why do not use non-linearity compensation instead of GNFB?
circuit I've prepared
has following parameters in LTSpice:
k=-1.000 V/V
Partial/Total Harmonic Distortion: 0.005046%
Bandwidth (-1dB) 0- 2.5MHz) with absolute stability.
I've prepared series of posts about non-linearity compensation without NFB in audio amplifiers in Polish DIY forum:
Saint Graal - linear amplifiers without NFB
circuit I've prepared
has following parameters in LTSpice:
k=-1.000 V/V
Partial/Total Harmonic Distortion: 0.005046%
Bandwidth (-1dB) 0- 2.5MHz) with absolute stability.
I've prepared series of posts about non-linearity compensation without NFB in audio amplifiers in Polish DIY forum:
Saint Graal - linear amplifiers without NFB