The DN2535, 2540 being a Depletion MOSFET, even though it has a built in protection diode, high static charge on the soldering iron tip or floating voltage in ungrounded soldering equipment can damage the device. I suggest that you short all three pins with bare, thin guage wire, while soldering/desoldering.
One more of my questions that a more knowledgeable DIYer would be able to answer for themselves ...
The input capacitor which i have queried about in the past probably too many times.
Just noticed you have the capacitor at 1uF in the latest FET schematic. It has usually been 10uF which you have said was a nominal value.
If I may ask: what value are you using?
Now to see if I can get this current source to work. Having new ones I was able to compare to the blown ones using the diode function and there is no doubt something is wrong with the ones I assumed were destroyed.
Wish I could figure out how to keep DN2535s alive past the initial installation, if there is a need to change something.
Obviously this complaint has nothing to do with the circuit - I have read of others finding these to be fragile at much higher voltages than we are using but then we are using much higher voltages than the typical solid state circuit.
Sorry to go on about that.
As always, with thanks, Rahul
The input capacitor which i have queried about in the past probably too many times.
Just noticed you have the capacitor at 1uF in the latest FET schematic. It has usually been 10uF which you have said was a nominal value.
If I may ask: what value are you using?
Now to see if I can get this current source to work. Having new ones I was able to compare to the blown ones using the diode function and there is no doubt something is wrong with the ones I assumed were destroyed.
Wish I could figure out how to keep DN2535s alive past the initial installation, if there is a need to change something.
Obviously this complaint has nothing to do with the circuit - I have read of others finding these to be fragile at much higher voltages than we are using but then we are using much higher voltages than the typical solid state circuit.
Sorry to go on about that.
As always, with thanks, Rahul
I am using 1 uF at the moment. The input and out caps form high pass filters with the resistor to ground that follows. So if you have 47k from the gate of the bottom jfet to ground, then you can calculate where the response will be 3 db down using this calculator:
https://www.digikey.com/en/resource...sion-calculator-low-pass-and-high-pass-filter
With 1 uF I get 3.3 hz, plenty good enough for me. Similarly, if your TDV has a 10k resistor from gate to ground, then 10 uF will give a 1.5 hz corner (-3 db point). If you drop the output cap to 1 uF, corner shifts to 15 hz. Roll off begins at a higher frequency than the corner. You can use that calculator to figure out how big a cap you need.
If you look at Nelson’s cap-coupled amps, they have 10,000 uF on the output to the speaker. Assuming an 8 ohm load, that calculator will tell you the -3db point. Now, for the TDV amps driving your tweeters, you can use a smaller higher quality cap!
https://www.digikey.com/en/resource...sion-calculator-low-pass-and-high-pass-filter
With 1 uF I get 3.3 hz, plenty good enough for me. Similarly, if your TDV has a 10k resistor from gate to ground, then 10 uF will give a 1.5 hz corner (-3 db point). If you drop the output cap to 1 uF, corner shifts to 15 hz. Roll off begins at a higher frequency than the corner. You can use that calculator to figure out how big a cap you need.
If you look at Nelson’s cap-coupled amps, they have 10,000 uF on the output to the speaker. Assuming an 8 ohm load, that calculator will tell you the -3db point. Now, for the TDV amps driving your tweeters, you can use a smaller higher quality cap!
How is the new buffer singing? Is it permanently in-circuit? So many questions!
Bought a case for the new build!
Bought a case for the new build!
I’m working on it. Still testing the sound of the preamp buffer. I’m close and should be able to close the design this weekend. If the sound doesn’t match up with the no buffer case, I’m ready to drop it.
I’ve only just discovered this thread, incidentally while building a MoFo amp. I did intend to use FE2022 driver stage, but after using it as a gain stage in my active crossover, it didn’t do as well as I had hoped.
I’m excited by this project but notice many of the active devices in V3.1 are now no longer manufactured, so I’m holding fire on building something.
Looking forward to seeing the next phase of development!
I’m excited by this project but notice many of the active devices in V3.1 are now no longer manufactured, so I’m holding fire on building something.
Looking forward to seeing the next phase of development!
There are the kinked ones: J175-D26Z
The board footprint might need to be adjusted accordingly, I suppose.
I tried to unkink them once, my OCD killed them.
Not even sure parts are finalized.
I finally figured out my problem. But first a confession.
While playing around with some copper mesh I accidentally shorted something (with the stage I have been trying to get the J113 to replace the second DN2535). I thought: since it was broken this would be a good place to try the J113/DN2535 combo again.
Never could get the J113 to work so I went back to two DN2535 and I could not get that to work either. I thought I was destroying these devices - so I kept replacing them and nothing changed.
I had almost completely rebuilt the thing - I did incorporate the new small film cap in place of the large electrolytic in the "regulator" while I was at it.
The only thing I had not touched was the zener diode for the regulator mosfet - I clipped the leads and everything worked as it should - voltages were easily attained. Damnit. I will never know if the zener sacrificed itself to save the rest of the circuit but that is certainly possible. I am a dumbass in so many ways ...
So if something goes funny with your SCG look there first!
Now I suspect I will be able to get the J113 to work.
The more I listen with the choke input raw supply the more I am sure this is worthwhile. Using a 5 H HAMMOND choke with a 57 Ohms resistance - with a 90 volts transformer I get about 88 volts to the line stage - 82 volts of zener for about 79 volts to the current source.
Funny thing is my gain is about half - who knows why - will replace the J111 and J175 this afternoon and see if that clears that up. With all of the jiggling around with the board I could easily have damaged one of those if not both. Luckily they are so cheap.
Even with having to crank up the gain to this stage the thing sounds good and is very quiet.
One thing after another. Now to update the regulators in the other three,
While playing around with some copper mesh I accidentally shorted something (with the stage I have been trying to get the J113 to replace the second DN2535). I thought: since it was broken this would be a good place to try the J113/DN2535 combo again.
Never could get the J113 to work so I went back to two DN2535 and I could not get that to work either. I thought I was destroying these devices - so I kept replacing them and nothing changed.
I had almost completely rebuilt the thing - I did incorporate the new small film cap in place of the large electrolytic in the "regulator" while I was at it.
The only thing I had not touched was the zener diode for the regulator mosfet - I clipped the leads and everything worked as it should - voltages were easily attained. Damnit. I will never know if the zener sacrificed itself to save the rest of the circuit but that is certainly possible. I am a dumbass in so many ways ...
So if something goes funny with your SCG look there first!
Now I suspect I will be able to get the J113 to work.
The more I listen with the choke input raw supply the more I am sure this is worthwhile. Using a 5 H HAMMOND choke with a 57 Ohms resistance - with a 90 volts transformer I get about 88 volts to the line stage - 82 volts of zener for about 79 volts to the current source.
Funny thing is my gain is about half - who knows why - will replace the J111 and J175 this afternoon and see if that clears that up. With all of the jiggling around with the board I could easily have damaged one of those if not both. Luckily they are so cheap.
Even with having to crank up the gain to this stage the thing sounds good and is very quiet.
One thing after another. Now to update the regulators in the other three,
Good to hear that you found the problem! I’ve done that many times where it looks like the problem is here but it is actually there—you are not alone.
Sound testing is complete. With buffer the bass is a bit more controlled and tight. Without buffer the sound is more open, more transparent, and just more enjoyable for me in my system. I’ve decided to keep the buffer but put a toggle so you can take the output directly from the gain stage and bypass the buffer. Then you have a choice and can decide what works best in your system.
Sound testing is complete. With buffer the bass is a bit more controlled and tight. Without buffer the sound is more open, more transparent, and just more enjoyable for me in my system. I’ve decided to keep the buffer but put a toggle so you can take the output directly from the gain stage and bypass the buffer. Then you have a choice and can decide what works best in your system.
At the risk of being needier than your children - what do you think of this?
Sunday I get the thing working - voltages as they should be - current slightly high at 17.5 but the gain is well below where it should be.
Yesterday I replaced the J175 (reversed) and the J111 (as specified in the schematic) and my voltages are much too high and the current has risen to almost double. Nothing else was changed.
I tried another set of J175 and J111 and got the same results.
All of the other three boards attain the correct voltages and currents with the same values of resistor (I do have a trimpot in series with a 15K resistor for the voltage divider - but I bet the trim pot is not that different for each of them) I substituted a 9.8K resistor with the pot at zero and the voltage does not lower anywhere near enough.
Can I assume that if Q101 sees over 60 volts and 27 mA it is likely dead almost immediately?
Any idea what has gone wrong? Does it sound like I have really damaged the CCS this time? (How I cannot imagine)
At your leisure. Thanks, Rahul
Sunday I get the thing working - voltages as they should be - current slightly high at 17.5 but the gain is well below where it should be.
Yesterday I replaced the J175 (reversed) and the J111 (as specified in the schematic) and my voltages are much too high and the current has risen to almost double. Nothing else was changed.
I tried another set of J175 and J111 and got the same results.
All of the other three boards attain the correct voltages and currents with the same values of resistor (I do have a trimpot in series with a 15K resistor for the voltage divider - but I bet the trim pot is not that different for each of them) I substituted a 9.8K resistor with the pot at zero and the voltage does not lower anywhere near enough.
Can I assume that if Q101 sees over 60 volts and 27 mA it is likely dead almost immediately?
Any idea what has gone wrong? Does it sound like I have really damaged the CCS this time? (How I cannot imagine)
At your leisure. Thanks, Rahul
Hi Rick,
The current is controlled by the CCS, so it should not change. If a J75-J111 pair is giving different drain voltage, see if it can be adjusted using the pot connected to the gate of J175.
Start with the CCS. Measure current through the CCS resistor. Is that coming out right? If not, adjust resistor.
If that’s working, then look at the voltage on the gate of j175. Should be in the 7-10V region to get about 40V drain of J111. That’s pretty much it.
I am doing the device Idss and Vp spread for J111 and J113 and should have some results to share before the end of the week.
The current is controlled by the CCS, so it should not change. If a J75-J111 pair is giving different drain voltage, see if it can be adjusted using the pot connected to the gate of J175.
Start with the CCS. Measure current through the CCS resistor. Is that coming out right? If not, adjust resistor.
If that’s working, then look at the voltage on the gate of j175. Should be in the 7-10V region to get about 40V drain of J111. That’s pretty much it.
I am doing the device Idss and Vp spread for J111 and J113 and should have some results to share before the end of the week.
I have three other boards that do exactly as they should, as you have specified.
Why this one is doing this to me is a mystery.
Last night replaced the 2 DN2535s and installed the DN2535/J113 combo. I figured there must be something wrong to get such high current with the 95R resistor so I thought I would give the combo another try. To be safe I used 130R to set the current and it comes out about 6 mV - I had used 95 with the 2 DN2535s in the other three boards. I was not sure if the J113 would give the same output as a DN2535 in the position and in this case it does not.
Ran out of time and did not substitute smaller value resistors - but I will tonight.
Did nothing with the voltage divider - with 9K76 in place of the trimpot it was around 6 volts where it should be 40 volts.. Of course, I am concerned the J111 and J175 could be damaged by the high current. I will replace these before proceeding. I have the above resistor in series with the trimpot - I have 5K I can add to the 9K76
I assume too low a value of current will affect the voltage at J111? But the question is moot since I am unsure about the J111 and J175.
This afternoon I will continue.
THANKS for your help and patience, Rahul
Why this one is doing this to me is a mystery.
Last night replaced the 2 DN2535s and installed the DN2535/J113 combo. I figured there must be something wrong to get such high current with the 95R resistor so I thought I would give the combo another try. To be safe I used 130R to set the current and it comes out about 6 mV - I had used 95 with the 2 DN2535s in the other three boards. I was not sure if the J113 would give the same output as a DN2535 in the position and in this case it does not.
Ran out of time and did not substitute smaller value resistors - but I will tonight.
Did nothing with the voltage divider - with 9K76 in place of the trimpot it was around 6 volts where it should be 40 volts.. Of course, I am concerned the J111 and J175 could be damaged by the high current. I will replace these before proceeding. I have the above resistor in series with the trimpot - I have 5K I can add to the 9K76
I assume too low a value of current will affect the voltage at J111? But the question is moot since I am unsure about the J111 and J175.
This afternoon I will continue.
THANKS for your help and patience, Rahul
Finally got the CCS to work with the J113.
The resistor values required are so different from the two DN2535 version it took me lots of creeping up on the needed values.
Did not get a chance to listen to it - all I know is I get 40 volts after the CCS and 14.5 mV.
To get the 14.5 mV I am using a resistor just above 8R (in parallel 20/20/43). I would have liked to get to 15 mV but this value of resistor seemed so low I did not want to take any chances. To set the voltage at 40 volts requires just over 10K.
To KNOW the value for the voltage divider you have to unsolder the 47K and I did not want to worry with that last night. I will check just before installing it in its place.
In comparison - the three boards using the two DN2535s are using 95 Ohms to set the current and somewhere above 15K to set the voltage at 40.
Whether sound comes out of the thing will have to wait until tonight. Will be interesting to hear if there is any difference. Certainly not ideal with just one of the boards set up this way.
One of those things I had to get to work. At my level of inexperience this was educational. I always remind myself that education never comes cheaply.
Rahul, do these values sound right to you?
The resistor values required are so different from the two DN2535 version it took me lots of creeping up on the needed values.
Did not get a chance to listen to it - all I know is I get 40 volts after the CCS and 14.5 mV.
To get the 14.5 mV I am using a resistor just above 8R (in parallel 20/20/43). I would have liked to get to 15 mV but this value of resistor seemed so low I did not want to take any chances. To set the voltage at 40 volts requires just over 10K.
To KNOW the value for the voltage divider you have to unsolder the 47K and I did not want to worry with that last night. I will check just before installing it in its place.
In comparison - the three boards using the two DN2535s are using 95 Ohms to set the current and somewhere above 15K to set the voltage at 40.
Whether sound comes out of the thing will have to wait until tonight. Will be interesting to hear if there is any difference. Certainly not ideal with just one of the boards set up this way.
One of those things I had to get to work. At my level of inexperience this was educational. I always remind myself that education never comes cheaply.
Rahul, do these values sound right to you?
Yes, they do. Glad you got it working. On to the listening. My bet is that the difference will be marginal at best compared to two dn2535s. There are many ways to do a CCS and the two dn2535 version provides the same level of performance.