Thanks, Andynor. It took me while to get to this. My wife just had a hip replaced today.
Here's what I measured/calculated:
R38 2.28 V /150r 15.2 mV stable
R27 61.7 mv /10r 6.17 mV stable
R28 0.1 mV /10r 0.1 mV stable
R29 50–90 mv /10k 005–009 mV very unstable
R31 2.315 V /150r 15.4 mV stable
One thing is not like the others, but what does it mean?
Here's what I measured/calculated:
R38 2.28 V /150r 15.2 mV stable
R27 61.7 mv /10r 6.17 mV stable
R28 0.1 mV /10r 0.1 mV stable
R29 50–90 mv /10k 005–009 mV very unstable
R31 2.315 V /150r 15.4 mV stable
One thing is not like the others, but what does it mean?
With no current through R28, I would suspect there's a problem with the jfets in that leg, Q21 and/or Q24. Those little guys can be difficult to install properly....overheated when soldered, contamination beneath part (flux), bad solder joint, etc. I talked about some of that here... If I recall properly, Wayne has this setup to have ~2mA through each leg of this input stage when everything is working as designed.
Last edited:
Thanks, Andynor. Here goes:
R38 ------ 61.7 mV -------- /10r = ---------6.17 mV ---------- stable
R27 ------- 2.28V -------- /150r = ---------15.2 mV ---------- stable
R28 ------ 0.1 mV --------- /10r = --------- 0.01 mV ---------- stable
R29 -----50–90 mV ------- /10k = -----0.005–0.009 mV ---- unstable
R31 ------ 2.315V --------- /150r = ---------15.4 mV ---------- stable
A couple of these are obviously suspect. what action is recommended?
R38 ------ 61.7 mV -------- /10r = ---------6.17 mV ---------- stable
R27 ------- 2.28V -------- /150r = ---------15.2 mV ---------- stable
R28 ------ 0.1 mV --------- /10r = --------- 0.01 mV ---------- stable
R29 -----50–90 mV ------- /10k = -----0.005–0.009 mV ---- unstable
R31 ------ 2.315V --------- /150r = ---------15.4 mV ---------- stable
A couple of these are obviously suspect. what action is recommended?
Sorry for the duplicate post. I didn't notice it had jumped to the next page. Also, that third column should be mA, not mV.
Will replace Q21 and Q24. What about R29 (R17 on circuit diagram)? That was drifting continuously and shows even less current than R28. Would that be Q21 (Q2 on diagram)?
Will replace Q21 and Q24. What about R29 (R17 on circuit diagram)? That was drifting continuously and shows even less current than R28. Would that be Q21 (Q2 on diagram)?
Very very nice, Buck! Just follow Williams advice, and I’ll chime in laterz if I have anything to add (unlikely). Have to make dinner for five pax first.
Hope wifeys ops went ok!!
Hope wifeys ops went ok!!
Agree. What happens across R17 is amongst other things a result of what happens between the JFETs. So as long as the voltage drops @ the 10R’s is unstable or low, just start with the JFETs 🙂
Is there a simple way to increase gain in this design, like changing a resistor value, etc. paired with my Pearl 2, even with a higher gain tweak -R14, if I recall- in it I wish sometimes for more gain
thanks!
thanks!
Yes. Just reduce feedback. Plenty of posts about that a few years back. You can choose between adjusting two resistors. I’d choose the one outside the RC network.
Target R16/17
https://www.diyaudio.com/community/threads/waynes-ba-2018-linestage.329240/page-123
Target R16/17
https://www.diyaudio.com/community/threads/waynes-ba-2018-linestage.329240/page-123
OK, here's the report.
Lifted Q21 off the board to reveal a shiny gold patch of circuit pad underneath one leg. Obviously this was a problem
Replaced Q21 and Q24. Man, those little buggers are a pain. I have got to get some soldering paste when I do this again.
Across R28, same as before: 0.1 mV for 0.01 mA; across R29, an entirely different picture, 4.21 V, for 0.4 mA, very stable.
Was able to adjust offset down to around 2 mA.
Music playing! Thanks everyone for your invaluable assistance and encouragement.
Lifted Q21 off the board to reveal a shiny gold patch of circuit pad underneath one leg. Obviously this was a problem
Replaced Q21 and Q24. Man, those little buggers are a pain. I have got to get some soldering paste when I do this again.
Across R28, same as before: 0.1 mV for 0.01 mA; across R29, an entirely different picture, 4.21 V, for 0.4 mA, very stable.
Was able to adjust offset down to around 2 mA.
Music playing! Thanks everyone for your invaluable assistance and encouragement.
I was so thrilled, but now a new problem.
Hooked up the other half of the board to check offset and function before installing in chassis. Was able to set offset nice and low, but as soon as I attached the output to my test amp, offset went to 16.89V and when attempting to play music, speaker visibly pumped in and out when I adjusted volume. Tried this a couple of times and cannot budge the offset off that number with or without amp attached. Another maybe helpful clue: the bias LED glows at maybe half the brightness of the other half of the board.
This half was playing music just fine before, and I haven't touched anything on it as I was fixing the "bad" half. Suggestions?
Hooked up the other half of the board to check offset and function before installing in chassis. Was able to set offset nice and low, but as soon as I attached the output to my test amp, offset went to 16.89V and when attempting to play music, speaker visibly pumped in and out when I adjusted volume. Tried this a couple of times and cannot budge the offset off that number with or without amp attached. Another maybe helpful clue: the bias LED glows at maybe half the brightness of the other half of the board.
This half was playing music just fine before, and I haven't touched anything on it as I was fixing the "bad" half. Suggestions?
On hand-soldered printed circuit boards using SMD parts, the vast majority of faults in my experience have been: opens.
Open circuit where you wanted short circuit. Component lead not in electrical contact with PCB trace.
Luckily this can be tested before power-up, if you have a circuit schematic including pin numbers, and a continuity test feature on your digital multimeter. If leg2 of component X is connected to leg 3 of component Y on the schematic, put your continuity test probes right on (leg2 of X) and (leg3 of Y). Right on the component leads themselves. If you get continuity then you know that both (leg2 of X) and (leg3 of Y) correctly make contact with the PCB copper traces. So you can put a check-mark on those pins on your schematic diagram. Proceed with the others.
If the board is a mix of thru-hole parts and SMD parts, you can probably get away with assuming that the thru hole parts are soldered correctly. Thus you can probably get away with testing only the leads of the SMD parts, for unwanted opens.
Open circuit where you wanted short circuit. Component lead not in electrical contact with PCB trace.
Luckily this can be tested before power-up, if you have a circuit schematic including pin numbers, and a continuity test feature on your digital multimeter. If leg2 of component X is connected to leg 3 of component Y on the schematic, put your continuity test probes right on (leg2 of X) and (leg3 of Y). Right on the component leads themselves. If you get continuity then you know that both (leg2 of X) and (leg3 of Y) correctly make contact with the PCB copper traces. So you can put a check-mark on those pins on your schematic diagram. Proceed with the others.
If the board is a mix of thru-hole parts and SMD parts, you can probably get away with assuming that the thru hole parts are soldered correctly. Thus you can probably get away with testing only the leads of the SMD parts, for unwanted opens.
A tempting and not unlikely hypothesis is that half (lower half) if that board is inactive for some reason. Try measurements again, and focus on JFETs first and then the bipolar transistors. PS: remember to very the correct transistors are in the correct locations and orientated correctly. Easy to mess up.
For example, measuring the lower half across the closest resistor to the bias LED, should perhaps show something.
Then, you could also put your fingertip on top of all bipolar transistors, and identify which ones are possibly cold(er) than the others. Then it will be easier to search in the correct places.
But, given the extreme offset, you sure that voltage is not the same as you regulated PSU is giving? If so, just tighten your power wires and give it a go 🙂
For example, measuring the lower half across the closest resistor to the bias LED, should perhaps show something.
Then, you could also put your fingertip on top of all bipolar transistors, and identify which ones are possibly cold(er) than the others. Then it will be easier to search in the correct places.
But, given the extreme offset, you sure that voltage is not the same as you regulated PSU is giving? If so, just tighten your power wires and give it a go 🙂
Eureka.
I went through all the suggestions: continuity, voltage/current measurements, PS connections. Nothing amiss.
Then, per Andynor's suggestion, I started putting my fingers on the FETs to see if any were warmer. When I touched Q1, the bias LED lit up at its full brightness. Turns out the middle leg was not connected to the board. Hit it with the soldering iron and voila: music can play. Showed continuity because I was pressing down on the leg with the probe.
I'll send some pix when I get it all together. Lots of fun things in a semi-custom case with handmade hardware.
Thanks again!
I went through all the suggestions: continuity, voltage/current measurements, PS connections. Nothing amiss.
Then, per Andynor's suggestion, I started putting my fingers on the FETs to see if any were warmer. When I touched Q1, the bias LED lit up at its full brightness. Turns out the middle leg was not connected to the board. Hit it with the soldering iron and voila: music can play. Showed continuity because I was pressing down on the leg with the probe.
I'll send some pix when I get it all together. Lots of fun things in a semi-custom case with handmade hardware.
Thanks again!
The next project coming out shortly is laid out with SMT or through hole part options.