I was thinking the output mosfet's insulators could have been bad thus conducting the outputs drains (center pin) to ground. However, looking back at the F6 schematic, I don't know if my dummy test is that simple.
The way the offset works is you are essentially adjusting the bias of each mosfet individually. You adjust the first mosfet's offset to a specific current (in your case 520mv across the 0.47R source resistor) then you adjust the second mosfets current to match (higher mv reading because of the 0.56R resistor).
So the bias of one mosfet or both is/are wandering just slightly.
I don't recall the bias on that F6 that I had in the past hovering around at all even unloaded. However, +-6mv isn't generally a big deal. You can try what Zen Mod suggested and ground the input and add I suppose an 8R load resistor or something like that.
The way the offset works is you are essentially adjusting the bias of each mosfet individually. You adjust the first mosfet's offset to a specific current (in your case 520mv across the 0.47R source resistor) then you adjust the second mosfets current to match (higher mv reading because of the 0.56R resistor).
So the bias of one mosfet or both is/are wandering just slightly.
I don't recall the bias on that F6 that I had in the past hovering around at all even unloaded. However, +-6mv isn't generally a big deal. You can try what Zen Mod suggested and ground the input and add I suppose an 8R load resistor or something like that.
Back at it again today. I planned to take the advice of @Zen Mod and short the inputs and try taking a new DC offset reading. Then, once offset was double checked I was going to try feeding the amp with my iPad like @von Ah suggested, thinking maybe my trusty passive pre sat too long and the silver contacts tarnished. Fingers crossed.
To begin, I dug out my shorting RCAs that I have been saving for special occasion and plugged them in. Checked to make sure they were functioning with the DMM. Set up my meters on each channel to test DC offset and fired up the amp. With the plugs in place I still got wandering readings and vastly different reading on each channel. After 30 seconds or so while observing the readings on my DMMs, I heard the faintest click sound, looked at the enclosure and saw a very tiny puff of smoke coming from one channel. Quickly shut down to inspect. The zip tie on the Q3/Q4 jfets had melted and popped on one channel. That was the click and smoke. Jfets look ok, but that got me thinking, and remembering, that sometimes I am a real idiot. So I dig out the schematic and study. Ok Q3 is 2SK120 and Q4 is 2SJ74, makes sense I ordered the NP channel quad originally. Figuring these were probably cooked if they melted the zip tie I pried them apart to check they were in the correct location. Q3 is 2Sk120, check. Pry a little more and to peak at Q4 and it turns out it is also 2Sk120. Dumb. Of course next is to check the other channel because maybe I was sent all 2SK120, that must be it. No, both 2SJ74. Good grief.
Long story short, new jfets on order from the store. More to come in a few days.
To begin, I dug out my shorting RCAs that I have been saving for special occasion and plugged them in. Checked to make sure they were functioning with the DMM. Set up my meters on each channel to test DC offset and fired up the amp. With the plugs in place I still got wandering readings and vastly different reading on each channel. After 30 seconds or so while observing the readings on my DMMs, I heard the faintest click sound, looked at the enclosure and saw a very tiny puff of smoke coming from one channel. Quickly shut down to inspect. The zip tie on the Q3/Q4 jfets had melted and popped on one channel. That was the click and smoke. Jfets look ok, but that got me thinking, and remembering, that sometimes I am a real idiot. So I dig out the schematic and study. Ok Q3 is 2SK120 and Q4 is 2SJ74, makes sense I ordered the NP channel quad originally. Figuring these were probably cooked if they melted the zip tie I pried them apart to check they were in the correct location. Q3 is 2Sk120, check. Pry a little more and to peak at Q4 and it turns out it is also 2Sk120. Dumb. Of course next is to check the other channel because maybe I was sent all 2SK120, that must be it. No, both 2SJ74. Good grief.
Long story short, new jfets on order from the store. More to come in a few days.
@No 4 : While this is a setback and undoubtedly a disappointment, you might also consider this an opportunity to do some measurements. With the parts desoldered from the boards, I recommend you try to measure their Idss. (In attached drawing, you can replace the DC amp meter by a 100R resistor and measure the voltage across this resistor instead). (https://www.diyaudio.com/community/threads/2sk170-matching.141162/#post-1784617)
If you get anything near 0.6V to 1.2V for a particular jfet then it probably is still good.
If you get anything near 0.6V to 1.2V for a particular jfet then it probably is still good.
@Dennis Hui this is definitely frustrating, but also a good opportunity to learn. Unfortunately for me it seems the best lessons have been learned the hard way. Luckily the mistake does not appear to be catastrophic thus far. I appreciate the link in your post. I will definitely test the jfets I take out of the circuit. I was actually doing some light reading on testing jfets this morning. It will be interesting to see if they survived or not. During original startup and biasing, this amp has been running for a few hours all together.
Replacement jfets arrived, and have been installed in the CORRECT positions this time. Powered up the amp, and let it heat up. Bias set to 520 mV on both channels, and DC offset (now steady) set to zero. Let that cook for awhile to make sure there was no drift, so far so good. Connected amp to test rig, using the same passive pre as the first attempt. Power up, no thump, no noise, no hum just black background. Fed some black keys through an old blue Ray player to the f6 and all is good.
I moved the f6 into my main setup paired with Lii fast 8s in the SLOBs. I have had to adjust the gain a little with the DSP to accommodate the increase in power on the fast 8s from the ACAs that have been powering them until now. I don’t really want to dial the f6 down so I may have to find some more power for the bottom end. However, I am really enjoying what I have going now.
Thanks to those who assisted and thanks a million to @Nelson Pass
I moved the f6 into my main setup paired with Lii fast 8s in the SLOBs. I have had to adjust the gain a little with the DSP to accommodate the increase in power on the fast 8s from the ACAs that have been powering them until now. I don’t really want to dial the f6 down so I may have to find some more power for the bottom end. However, I am really enjoying what I have going now.
Thanks to those who assisted and thanks a million to @Nelson Pass
@No 4 actually it turns out a similar thing happened to me, that was likely the cause of my problems with the F6 build. I actually think the bags from diyaudio store had some of the wrong JFET type in them, but it's my fault I did not double check. I am now triple checking every single component, like I should have done in the 1st place. Hey does anyone have a line on where I can find a .56 ohm 3W resistor? Seems like Mouser is out, I've tried Arrow and a few others, and I'm not trusting myself to know what to buy. I have a ton of Panasonic .47 resistors from other Firstwatt projects/power supplies, is there enough of a difference to try and track down the .56ohm? Thank you!!
The .47 will work fine. However, it'll change the Distortion characteristics. The difference in source resistors is there to add some second harmonic distortion. I forget which phase but I would suspect negative phase. But that being said, try the .47. It's not like you can't play around with different values.
I would guess having both be the same value would null out the second harmonic distortion leaving you with third. Some people like that sound. I would also think that you would end up with lower Distortion overall.
I would guess having both be the same value would null out the second harmonic distortion leaving you with third. Some people like that sound. I would also think that you would end up with lower Distortion overall.
Good advice from Mikerodring27.
Instead of wirewound or metal oxide parts for source resistors, how about using Current Source resistors in that position?
https://www.mouser.com/ProductDetail/KOA-Speer/BPR58CR47J?qs=4NRufenSBVJ3ZS9Vxft0Wg==
https://www.mouser.com/ProductDetail/KOA-Speer/BPR58CR56K?qs=1He0yLMpldAuIe9wU3nkyw==
Any potential drawbacks?
Instead of wirewound or metal oxide parts for source resistors, how about using Current Source resistors in that position?
https://www.mouser.com/ProductDetail/KOA-Speer/BPR58CR47J?qs=4NRufenSBVJ3ZS9Vxft0Wg==
https://www.mouser.com/ProductDetail/KOA-Speer/BPR58CR56K?qs=1He0yLMpldAuIe9wU3nkyw==
Any potential drawbacks?
Those are current sense resistors, not current source resistors.
They work well enough in power supply applications, though I haven’t tried them in the output stage.
I used 1% precision thick film resistors in my F6 and was quite happy with those. Wrote about it fairly recently.
They work well enough in power supply applications, though I haven’t tried them in the output stage.
I used 1% precision thick film resistors in my F6 and was quite happy with those. Wrote about it fairly recently.