I do not doubt that there are many way to make them. I only know of the three you have published. It had become a challenge to get the FET to work.
The fact that you published the schematic made it seem like something worth trying. So now if someone else wants to use them they have an idea where to begin.
The question lingers: is there a best way to make a CCS and would the best approach make a sonic difference?
I doubt I will be able to hear the difference, either.
With the addition of the FETs to the main circuit it does seem complimentary to use them in the CCS is my justification for the time spent (and wasted since I did so many dumb things).
THANKS, Rahul
The fact that you published the schematic made it seem like something worth trying. So now if someone else wants to use them they have an idea where to begin.
The question lingers: is there a best way to make a CCS and would the best approach make a sonic difference?
I doubt I will be able to hear the difference, either.
With the addition of the FETs to the main circuit it does seem complimentary to use them in the CCS is my justification for the time spent (and wasted since I did so many dumb things).
THANKS, Rahul
For these types of adjustments/calculations, learning enough about LTSpice has been a huge time saver for me.
The thing does work and work well.
Can't tell if what I am hearing is due to replacing the electrolytic at C14 on the original schematic with 0.22 uF WIMA PP or from the J113 but there is something better happening. I will do the regulator mod to the rest of the boards before I begin converting to J113. Need to order some resistors to do that, first.
ElArte - what does LTSpice predict the resistor values to be if you have it handy?
Can't tell if what I am hearing is due to replacing the electrolytic at C14 on the original schematic with 0.22 uF WIMA PP or from the J113 but there is something better happening. I will do the regulator mod to the rest of the boards before I begin converting to J113. Need to order some resistors to do that, first.
ElArte - what does LTSpice predict the resistor values to be if you have it handy?
I didn’t do the conversion to JFETs, as I decided to wait for the new board.
Sorry, if I make a model later, I’ll share it here.
I'm going to work on an upgrade board also, so you can keep most of the current PCB, the power supply, input/output caps, and mount a small new PCB on the top that carries the signal circuitry with the new FETs.
That’s very nice of you, Rahul. But I tinkered so much with my board, and a couple of times way too much current went through. 2 of the limiting resistors acted as fuses in those instances, but I can’t imagine it did anything good to the parts downstream from that. So, I’ll definitely build from scratch. 😆
JFET Sample Spread
Here are some data on parameter spread for J113 and J175. I found an interesting relationship between JFET pinch-off voltage (Vp), the voltage at which the JFET stops conducting current, and the Idss, the current that flows through the JFET when gate and source are at the same voltage (Vgs = 0).
It seems that for each JFET type, there is a linear relationship between Vp and Idss. This means you can make just one measurement and get the other from an equation. If you are testing hundreds of JFETs, this is very useful. Of course, for this design, it probably doesn't save much time, but interesting nevertheless.
The J111 JFET has a pretty high Idss figure, >70 mA, so it heats up pretty quick and doesn't leave enough time to settle. For Vp, most of the J111 samples were between 7.75V and 8.00V. So, that makes it easy. We'll be operating it at around 10 mA, and because we have the ability to adjust the drain voltage, pretty much any J111 will do.
Here's the summary:
1. J113s for CCS: Use the smallest Vp samples in your set. Smallest Vp means highest current and gives most latitude for adjust the CCS.
2. J111s for gain: No specific selection needed.
3. J175 as input follower: Use highest Vp samples in your set.
Couple more notes:
1. For the buffer, I am switching to J111 for the signal device. I like its sound better than J113 in that position.
2. For the CCS, I might yet ditch the J113 if it proves too much trouble. Or, we can also use J111s and not use J113s at all in the design. Fewer parts to order. We can even select J113s or J111s at a specific CCS resistor, say, 47 ohms and 10 mA. But I need to test that. It would make the design simpler.
I want to get this over the line as soon as possible and am finalizing the PCBs.
Here are some data on parameter spread for J113 and J175. I found an interesting relationship between JFET pinch-off voltage (Vp), the voltage at which the JFET stops conducting current, and the Idss, the current that flows through the JFET when gate and source are at the same voltage (Vgs = 0).
It seems that for each JFET type, there is a linear relationship between Vp and Idss. This means you can make just one measurement and get the other from an equation. If you are testing hundreds of JFETs, this is very useful. Of course, for this design, it probably doesn't save much time, but interesting nevertheless.
The J111 JFET has a pretty high Idss figure, >70 mA, so it heats up pretty quick and doesn't leave enough time to settle. For Vp, most of the J111 samples were between 7.75V and 8.00V. So, that makes it easy. We'll be operating it at around 10 mA, and because we have the ability to adjust the drain voltage, pretty much any J111 will do.
Here's the summary:
1. J113s for CCS: Use the smallest Vp samples in your set. Smallest Vp means highest current and gives most latitude for adjust the CCS.
2. J111s for gain: No specific selection needed.
3. J175 as input follower: Use highest Vp samples in your set.
Couple more notes:
1. For the buffer, I am switching to J111 for the signal device. I like its sound better than J113 in that position.
2. For the CCS, I might yet ditch the J113 if it proves too much trouble. Or, we can also use J111s and not use J113s at all in the design. Fewer parts to order. We can even select J113s or J111s at a specific CCS resistor, say, 47 ohms and 10 mA. But I need to test that. It would make the design simpler.
I want to get this over the line as soon as possible and am finalizing the PCBs.
Thanks for that, ra7.
Converting my other boards to the J113/DN2535 has been pretty easy.
I will try the J111.
Ordered some 20R metal foil trimmers from Michael Percy - using discreet resistors would require lots of low value resistors if you wanted them all to be similar. And that gets ungainly if you want to keep a tight layout. These trimmers are expensive but about a third of what they cost at other suppliers.
For the obsessive, only? This thing is in the signal path so I think this is a worthwhile extravagance.
I think the circuit is worth it.
Converting my other boards to the J113/DN2535 has been pretty easy.
I will try the J111.
Ordered some 20R metal foil trimmers from Michael Percy - using discreet resistors would require lots of low value resistors if you wanted them all to be similar. And that gets ungainly if you want to keep a tight layout. These trimmers are expensive but about a third of what they cost at other suppliers.
For the obsessive, only? This thing is in the signal path so I think this is a worthwhile extravagance.
I think the circuit is worth it.
I had learned there can be gain differences - I assume due to J175.
One of the "tweeter" boards is either 4 dB more sensitive or the other is 4 dB less.
Digital EQ makes it easy to work around this though I prefer all of the boards to be the same.
I have one of those PEAK DCA75 devices on the way along with lots of FETs.
The next step in building good stuff is knowing what you are using. That device does so many neat things I could not resist getting one instead of building lots of clunky stuff that would not work as well.
Very excited about the FET selection recommendations.
As always. THANKS, Rahul
One of the "tweeter" boards is either 4 dB more sensitive or the other is 4 dB less.
Digital EQ makes it easy to work around this though I prefer all of the boards to be the same.
I have one of those PEAK DCA75 devices on the way along with lots of FETs.
The next step in building good stuff is knowing what you are using. That device does so many neat things I could not resist getting one instead of building lots of clunky stuff that would not work as well.
Very excited about the FET selection recommendations.
As always. THANKS, Rahul
Tested 103 J175s and have attached the spreadsheet to show the distribution of the group I just got from MOUSER. I had three from a previous purchase. They came on a paper strip which makes it easy to label them. They were measured in the order they came off the roll. Thought it might give folks an idea of how many to buy.
Using the PEAK meter. I am assuming Vgs(off) is the same thing.
I am wondering if the highest Vp J175 has any effect on gain? I know one of my SCGs seems to have a lower gain than the others. That has not always been the case.
Are my highest value J175s the way to go? Can they be too high?
Using the PEAK meter. I am assuming Vgs(off) is the same thing.
I am wondering if the highest Vp J175 has any effect on gain? I know one of my SCGs seems to have a lower gain than the others. That has not always been the case.
Are my highest value J175s the way to go? Can they be too high?
Thanks Rick! I’ll check the attachment and get back to you.
Yes, J175 can affect gain. It is not the highest transconductance part and it can lose a bit of gain in follower mode.
Yes, J175 can affect gain. It is not the highest transconductance part and it can lose a bit of gain in follower mode.
Going through 100 J113s - got through 75 before I had my fill for the day.
Even thought this PEAK meter is very handy I have to keep it near the computer - no table to work on - lots of bending over. I can feel it in my hamstrings this morning. Audio exercise.
Ranging from 2.0 to 1.55 - I know you said low is best but wonder if there is a too low?
When I get them all done (J111s come next) I can insert those of varying values and see/hear what happens, nonetheless I would like to hear what you think are ideal values.
Thanks, Rahul
Even thought this PEAK meter is very handy I have to keep it near the computer - no table to work on - lots of bending over. I can feel it in my hamstrings this morning. Audio exercise.
Ranging from 2.0 to 1.55 - I know you said low is best but wonder if there is a too low?
When I get them all done (J111s come next) I can insert those of varying values and see/hear what happens, nonetheless I would like to hear what you think are ideal values.
Thanks, Rahul
WOW, @rickmcinnis !
I never thought there would be so much variance on a single ammo pack.
I am now wondering if I should get a DCA75.
I never thought there would be so much variance on a single ammo pack.
I am now wondering if I should get a DCA75.
Hope you are well fellas. Been busy last week but getting back to this.
The results with the J111 are far more consistent. I am selecting for transconductance with these.
J175s are more like the J113. I will post what I have found later.
I love the meter. If we are going to work with FETs I guess we need one?
For two hundred dollars you can test any semiconductor. I do not regret the purchase.
Not knowing anything about measuring FETs (or any other device) I have wondered if the voltages used might be too low?
J175s are more like the J113. I will post what I have found later.
I love the meter. If we are going to work with FETs I guess we need one?
For two hundred dollars you can test any semiconductor. I do not regret the purchase.
Not knowing anything about measuring FETs (or any other device) I have wondered if the voltages used might be too low?
What I have found -
Rahul's advice about the lowest Vp for the J113 in the current source is if you would like to use higher values of a current set resistor - similar to what you were using with the two DN2535 one.
With my 20R trimmer as my "limiter" I have found 1.67 to 1.68 to get me to 15 mA. Somewhere close to 2/3rds of the way to 20.
I continue to think there is a sonic advantage to the J113/DN2535 combo.
Rahul's advice about the lowest Vp for the J113 in the current source is if you would like to use higher values of a current set resistor - similar to what you were using with the two DN2535 one.
With my 20R trimmer as my "limiter" I have found 1.67 to 1.68 to get me to 15 mA. Somewhere close to 2/3rds of the way to 20.
I continue to think there is a sonic advantage to the J113/DN2535 combo.
Have had a buzz and an elevated gain with one of my line stages.
I have replaced EVERYTHING on the board but the Schade resistors and last night as a final check I measured the resistance of the 10K resistor (actually two 20K in parallel) and find one of them is damaged - no reading or inconclusive reading.
Last thing I did before going to bed so I was not able to investigate properly. Will do that this afternoon.
Hope that is the problem though it has given the impetus for lots of exploration so it has been for the good.
IF that is the case a gain of 20 does not bring about the gain one would expect but it does (surmising) put a strain on the power supply. When I look at the output through my speakers with REW RTA low level noise spikes from 400 to 20K Hz.
What is interesting is the sound is good - nothing obviously bad as far as sound quality - other than the annoying as HELL buzzing.
I was starting to think the SIT had gone funny.
Another interesting observation - especially to folks like me with a proclivity for destroying semiconductors - if you give the J111 and the J175 too much voltage (had not turned down the trimpot for the bias before turning on with a "new" device installed) the device will still measure as it did on the PEAK meter. Does this make sense? One thing I thought the meter would be good for is verification that a part had been ruined.
I have replaced EVERYTHING on the board but the Schade resistors and last night as a final check I measured the resistance of the 10K resistor (actually two 20K in parallel) and find one of them is damaged - no reading or inconclusive reading.
Last thing I did before going to bed so I was not able to investigate properly. Will do that this afternoon.
Hope that is the problem though it has given the impetus for lots of exploration so it has been for the good.
IF that is the case a gain of 20 does not bring about the gain one would expect but it does (surmising) put a strain on the power supply. When I look at the output through my speakers with REW RTA low level noise spikes from 400 to 20K Hz.
What is interesting is the sound is good - nothing obviously bad as far as sound quality - other than the annoying as HELL buzzing.
I was starting to think the SIT had gone funny.
Another interesting observation - especially to folks like me with a proclivity for destroying semiconductors - if you give the J111 and the J175 too much voltage (had not turned down the trimpot for the bias before turning on with a "new" device installed) the device will still measure as it did on the PEAK meter. Does this make sense? One thing I thought the meter would be good for is verification that a part had been ruined.