Regarding reflowing and rinsing, that comment was for starbender. But I guess his problem has little to do with soldering work.
I have been thinking about improvising. But it wont look as good! Then again, the lid will be on.
Maybe I will run the improvisation through you. I will check out the PCB traces later and get back to you.
I have been thinking about improvising. But it wont look as good! Then again, the lid will be on.
Maybe I will run the improvisation through you. I will check out the PCB traces later and get back to you.
silly me - pcb is stereo 
tripple check positioning and orientation of all parts, soldering joints
doubt in everything - resistor values, swapped places of semiconductors etc
if nothing pops up as suspicious, replace outputs ........ then input JFets
let's blame Gremlins if any of them is Dodo
tripple check positioning and orientation of all parts, soldering joints
doubt in everything - resistor values, swapped places of semiconductors etc
if nothing pops up as suspicious, replace outputs ........ then input JFets
let's blame Gremlins if any of them is Dodo
After all the fun discussion re: the input stage bias, I decided to re-check mine after it has been in service for a while. During the same process, I reduced the output stage bias to a "more sane" level to reduce the heat off the sinks. Note, this is while the IPS is attached to the complimentary BA-2 output stage. So, a BA-3 amplifier.
Some observations.
1) We all know this, but the IPS bias is definitely both temperature dependent and temperature sensitive. When I reduced the output stage bias, and thus the internal temps of the chassis, the overall bias for the input stage dropped. Offset drifted a bit.
2) Several months ago, I had made some adjustments to P3 to adjust H2 / H3 balance using REW. However, I had not taken offset or bias measurements when doing so. After checking my measurements, P3 definitely affects both the relative bias of each device and the offset. Maybe I should have known this prior, but I did not.
So, for my own fun... after reducing the output stage bias and allowing the temp to become stable; I rebiased / nulled the offset for the front end. I also made an attempt to re-center P3 just by # of turns.
3) I can use P1 and P2 to get the bias generally up to where I wanted it at about 0V95 fully warm. However, very slight tweaks to either P1 or P2 could cause 10s to 100s of mV change in the offset, and it never quite seemed to stabilize.
4) Minor (within a turn or two) tweaks to P3 from center allowed both channels of offset to be very stable below 5mV.
So, I used P1 / P2 to get close to the bias points on both channels and null the offset as close as practical. I then used P3 to null the offset. I am sure P3 is not "centered" nor is the # of turns away from center the same on either channel. I am going to try and take a few measurements using REW and see how closely the channels are matched. It only took a few turns from "center" for each side. Also, the adjustment using P3 was far less sensitive, and it was far more stable.
For others that have used P3 to intentionally alter the H2/H3 balance - Did you adjust P3 and then attempt to null the offset after the P3 adjustment using P1 and P2? OR Did you have a standard "center" starting place and simply adjust P3 on its own? Just curious.
Still learning about how all this ties together.
Some observations.
1) We all know this, but the IPS bias is definitely both temperature dependent and temperature sensitive. When I reduced the output stage bias, and thus the internal temps of the chassis, the overall bias for the input stage dropped. Offset drifted a bit.
2) Several months ago, I had made some adjustments to P3 to adjust H2 / H3 balance using REW. However, I had not taken offset or bias measurements when doing so. After checking my measurements, P3 definitely affects both the relative bias of each device and the offset. Maybe I should have known this prior, but I did not.
So, for my own fun... after reducing the output stage bias and allowing the temp to become stable; I rebiased / nulled the offset for the front end. I also made an attempt to re-center P3 just by # of turns.
3) I can use P1 and P2 to get the bias generally up to where I wanted it at about 0V95 fully warm. However, very slight tweaks to either P1 or P2 could cause 10s to 100s of mV change in the offset, and it never quite seemed to stabilize.
4) Minor (within a turn or two) tweaks to P3 from center allowed both channels of offset to be very stable below 5mV.
So, I used P1 / P2 to get close to the bias points on both channels and null the offset as close as practical. I then used P3 to null the offset. I am sure P3 is not "centered" nor is the # of turns away from center the same on either channel. I am going to try and take a few measurements using REW and see how closely the channels are matched. It only took a few turns from "center" for each side. Also, the adjustment using P3 was far less sensitive, and it was far more stable.
For others that have used P3 to intentionally alter the H2/H3 balance - Did you adjust P3 and then attempt to null the offset after the P3 adjustment using P1 and P2? OR Did you have a standard "center" starting place and simply adjust P3 on its own? Just curious.
Still learning about how all this ties together.
Eureka!
I will do the same.
But seems your investigative might be in Nelsons article all along. Not sure if applicable in your specific case, but nevertheless:
«If you get lower values, you may want to consider higher values for P1 and P2, say 1 Kohm. They (read: JFETs) are best matched, but P3 can be adjusted to compensate for some mismatch.»
P3 is connected to the (gate?) of the JFETs. So considering there is no feedback loop, it makes sense that the offset can be affected when adjusting P3.
Exactly how this happens, if my assumption is correct, Nelson or Zen Mod might have the explanation for?
Cheers,
Andreas
I will do the same.
But seems your investigative might be in Nelsons article all along. Not sure if applicable in your specific case, but nevertheless:
«If you get lower values, you may want to consider higher values for P1 and P2, say 1 Kohm. They (read: JFETs) are best matched, but P3 can be adjusted to compensate for some mismatch.»
P3 is connected to the (gate?) of the JFETs. So considering there is no feedback loop, it makes sense that the offset can be affected when adjusting P3.
Exactly how this happens, if my assumption is correct, Nelson or Zen Mod might have the explanation for?
Cheers,
Andreas
Last edited:
Btw: my bias boards work. In dialogue with Zen Mod, I have concluded that one of my N channel MOSFETs is not matched. 60%-ish more IQ draw to one of them, leaving the other two rather in the cold, so to speak. An error from the store in other words. Hoping for a partial refund. Have ordered a new set of FETs. It is a bummer, since the amp was finished.
Well well, those who wait for something good.
But the guys do a great job for and at the store, and **** happens, a wise man said. For future projects, I may have to make a test jig for FETs myself. And buy a lab PSU. And lots of other things. When I am rich maybe? Yes...
Well well, those who wait for something good.
But the guys do a great job for and at the store, and **** happens, a wise man said. For future projects, I may have to make a test jig for FETs myself. And buy a lab PSU. And lots of other things. When I am rich maybe? Yes...
Double checked, P3 is connected to JFET source. I’m no physicist. But as far as I can tell, adjusting P3 off center, increases or decreases the bias balance between the two JFETs, but probably minutely. As far as my limited understanding goes, this is adding or subtracting the source resistance the JFET «sees», and in consequence affecting DC offset in the end. Amateurishly explanation, but might explain the effect you describe.
Regards,
Andreas
Regards,
Andreas
Last edited:
That is a shame. I am very happy you figured it out.
I appreciate your explanation. Anyone that makes an attempt is A+ in my book. I am still a complete novice to electronics / circuits. Where I got a bit lost is how the JFETs Q1 and Q2 affect the bias / offset from Q3 / Q4. The balance across Q1 / Q2 is affected by P3. It is connected to the source of both as you say. I think I understand that part. As an example, I think about how Papa uses two values of source resistance on the outputs of the F6 to intentionally affect the H2/H3 profile. I can't ever say that I truly, fully understand how it works, but I see the similarities in principle.
It is the drain (Q1/Q2) to gate through R8/R9 (Q3/Q4) tie that still has me thinking. I am sure it is simple, but my brain is not picking it up. I am trying to learn LTSpice, but it is a VERY slow process for me.
Note - re: Post #466. I have no idea how I missed that in Nelson's article. I read it too many times perhaps.
Edited to correct my misreading of the schematic (again). Still not 100% on the ball with this circuit.
Last edited:
I do not understand it per se. But for now, I think of it like this: altering bias on one JFET, affects the symmetry of bias between the devices. This, in effect, affects the offset. Since the circuit has no feedback, what happens with the JFETs, affects the MOSFETs. Is this the truth? I dunno... I will ask Zen Mod to have a look at the US-NO dream team of wannabe physicists thoughts and derivations =D
if you have THD Spectra measuring apparatus (decent soundcard and few progs) , then fiddling with P3 is inevitable and progress leading
if not, carefully set it (P3) to mid position (measuring with ohmmeter, not counting turns) and forget it, till moment of having all of above, and idea how to use it
of course, same as P1-P2 play is - both influencing little mosfet's Iq and output node voltage potential, when you start fiddling with P3, it'll be in group of iterative fiddling with other two trimpots ....... simply because P3 is changing DC balance , thus Iq and output node voltage level
same story as with F5
if not, carefully set it (P3) to mid position (measuring with ohmmeter, not counting turns) and forget it, till moment of having all of above, and idea how to use it
of course, same as P1-P2 play is - both influencing little mosfet's Iq and output node voltage potential, when you start fiddling with P3, it'll be in group of iterative fiddling with other two trimpots ....... simply because P3 is changing DC balance , thus Iq and output node voltage level
same story as with F5
After all the fun discussion re: the input stage bias, I decided to re-check mine after it has been in service for a while. During the same process, I reduced the output stage bias to a "more sane" level to reduce the heat off the sinks. Note, this is while the IPS is attached to the complimentary BA-2 output stage. So, a BA-3 amplifier.
Some observations.
1) We all know this, but the IPS bias is definitely both temperature dependent and temperature sensitive. When I reduced the output stage bias, and thus the internal temps of the chassis, the overall bias for the input stage dropped. Offset drifted a bit.
2) Several months ago, I had made some adjustments to P3 to adjust H2 / H3 balance using REW. However, I had not taken offset or bias measurements when doing so. After checking my measurements, P3 definitely affects both the relative bias of each device and the offset. Maybe I should have known this prior, but I did not.
So, for my own fun... after reducing the output stage bias and allowing the temp to become stable; I rebiased / nulled the offset for the front end. I also made an attempt to re-center P3 just by # of turns.
3) I can use P1 and P2 to get the bias generally up to where I wanted it at about 0V95 fully warm. However, very slight tweaks to either P1 or P2 could cause 10s to 100s of mV change in the offset, and it never quite seemed to stabilize.
4) Minor (within a turn or two) tweaks to P3 from center allowed both channels of offset to be very stable below 5mV.
So, I used P1 / P2 to get close to the bias points on both channels and null the offset as close as practical. I then used P3 to null the offset. I am sure P3 is not "centered" nor is the # of turns away from center the same on either channel. I am going to try and take a few measurements using REW and see how closely the channels are matched. It only took a few turns from "center" for each side. Also, the adjustment using P3 was far less sensitive, and it was far more stable.
For others that have used P3 to intentionally alter the H2/H3 balance - Did you adjust P3 and then attempt to null the offset after the P3 adjustment using P1 and P2? OR Did you have a standard "center" starting place and simply adjust P3 on its own? Just curious.
Still learning about how all this ties together.
Watching this closely! I'll be right behind you with the REW thing. Before the 2i2 units and REW had leaked down to mere minor builders like myself, there were posts from people kind of winging it with P3.
While my understanding is that with REW measurements the idea is to get the H2/H3 thing as you like it with both channels spectra is as close to the same as possible.
In my scientific wild *** view, winging P3 is not really going to get you there, but those that did try seemed to describe their process as playing with P3, then touching up the offset. I was lost, as my BA3 units seemed that no matter which of P3, P2, or P1 you twiddled, it affected the other settings.
This both confused me and also caused me to look into this REW/2i2 method. Otherwise, with no ability to know what results of P3 adjustment would be after touching up would be, it didn't make sense for me to proceed.
Reading 6l6 and zenmod talking about positive and negative H2 showed me I was over my head, so my P3 pots are still at the mid point setting.
Now, this can be inexpensively accomplished, but here comes the learning curve.
Oh, when you adjusted P3 for offset purposes, did it matter which direction you turned it?
Carry on brave soldier.
Russellc
Last edited:
never adjust P3 for offset purposes
you have P1 and P2 there - starting from minimal resistance of both - iteratively playing with these - turn one to increase Iq, then turn other to reach desired voltage level at output node ( say 0mV)......... then again, up to full desired Iq
only when you are ready to play with THD spectra .......... :
first rec THD SPectra situation with P3 at mid;
rotate P3 in one direction for 1 turn; re-set Iq and DC Offset (P1 and P2, again) back to target values
re-measure THD SPectra; think are you going in desired direction or not ........ should you proceed in same direction or you want to go back
it is handy to document each step ( screenshots & paper-pen) for each step, prevent you of being lost ........
which is, frankly, not so bad - I'm lost most of the time
you have P1 and P2 there - starting from minimal resistance of both - iteratively playing with these - turn one to increase Iq, then turn other to reach desired voltage level at output node ( say 0mV)......... then again, up to full desired Iq
only when you are ready to play with THD spectra .......... :
first rec THD SPectra situation with P3 at mid;
rotate P3 in one direction for 1 turn; re-set Iq and DC Offset (P1 and P2, again) back to target values
re-measure THD SPectra; think are you going in desired direction or not ........ should you proceed in same direction or you want to go back
it is handy to document each step ( screenshots & paper-pen) for each step, prevent you of being lost ........
which is, frankly, not so bad - I'm lost most of the time
never adjust P3 for offset purposes
<SNIP>
it is handy to document each step ( screenshots & paper-pen) for each step, prevent you of being lost ........
which is, frankly, not so bad - I'm lost most of the time
Incredibly helpful, ZM. Thank you!
That is how I did it the first time around. I'll get it back to "stock" middle and adjust with REW similar to past measurements. I got H2 just about even with H3 and left it alone.
Definitely can't wing it. Previously, I had did the proper bias / offset adjustment only with P1 and P2. Then, I used REW and adjusted only P3 to get the H2/H3 balance set. At that time, I didn't know how to check relative phase of H2. I want to try that as soon as I learn DIANA.
Yes, it definitely mattered. Now that I know it's a no-no, I'm going back to the mid-point until I am ready to take a chunk of time and re-set it with REW again.
I understand and have seen how movement of any of the three pots affects settings of the other. I didn't mean to adjust offset via P3, but was and still am trying to figure out why it would stabilize offset measurements.
All I have seen is, move any one and affect the other 2. Well, at least with P1 and P2, I haven't measured P3 as of yet but did accidentally turn P3 once while measuring, I didn't notice right off, as the multimeter's reading was changing.
Just would like to know why nudging P3 would have this effect.(the effect of stabilizing, not the multimeter moving.)
Russellc
All I have seen is, move any one and affect the other 2. Well, at least with P1 and P2, I haven't measured P3 as of yet but did accidentally turn P3 once while measuring, I didn't notice right off, as the multimeter's reading was changing.
Just would like to know why nudging P3 would have this effect.(the effect of stabilizing, not the multimeter moving.)
Russellc
Last edited: