Got the multi-checker. Voltage reg LM336z-2.5 is no longer. What happens when a voltage reg is no longer a voltage reg? Have readings for the 2sa970 bjt-pnp that I'll compare to a new one. Also have measurements for the p and n mosfets. Don't these work in tandem and shouldn't they have similar measurements? Will also check the measurements against the new parts. Worried one or both of these could be fakes.
I think we have the same little tester - mine says a brand new LM317 is a damaged part. Hard to say if your 336s are still good. At least they're cheap to replace. I just got a box of stuff from Digikey and threw in extra LM334s since they seem pretty useful with small signal tubes.
Readings & specs for transistors are beyond what little SS knowledge I have. I'm hoping to learn a bit by building the Pass clone kit soon, but not sure I really understand what's going on with that circuit yet.
John H - do you remember anything about Mosfets from your college days?
Is there anything on this board I didn't fry??? Those little e202 (D1a/D1b) current reg diodes are $2 each. Top row pulled from board, bottom row new part. D3 tests bad, 2sj79 does not test like the new one, and D1b/D1a current reg diodes test as a cap.
All this from not installing the tube???
Aha, check this out. BOM calls for "J508 or CR200 2.4mA constant current diode". According to Newark, cr200 is Vishay "CR200 Zener Single Diode, 300 mW, TO-18, 10 %, 2 Pins" but I do not see that J508 is a zener. I subbed e202 for these hard to find diodes to keep from chancing ebay fakes and I don't think e202 is a zener.
Datasheets attached.
Whas up wit dis?
Just to be sure I have this straight; a zener diode can conduct in both directions, a regular diode cannot.
No where else have I found J508 listed as a zener and the board manu says he stuffs D1a/D1b with e-352&e-152 (<specs can be found on the previously attached e202 diode pdf.) to total 5ma. I subbed it with two e-202 (total 4ma) but doubt this was THE major problem.
So ... back to the biggie ... WTF happened?
Of course Mouser is out of stock, even the smd version.
Now you are getting into stuff I sure can't answer. I sort of understand zeners, but didn't know there were constant current diodes. I'm sure somebody here could explain that circuit, but once you add SS devices more complex than a zener to the schematic, I need somebody to walk me through it.
Add a 22pf mica cap to the DOA list. Since that cap is rated 500vdc I guess it's not the voltage but the current that killed it? Mouser didn't have the recommended size diodes in stock so had to buy extra and will have to 'stack' them for the correct value.
Are you sure all those parts were destroyed? C2 is across a 10K resistor, seems strange that it would be taken out. D1 diodes are available at Mouser 1N5306 is a 2.4ma current regulating diode, and should be the equivalent, but not cheap. The E202 that you listed should be close enough. I would look at the power supply and all grounding first, after that I would look to a possible oscillation, considering that you said that there were no tubes installed. Can you look at this on a scope? Did you use a dim bulb tester when you fired this up?
Nope, not sure at all, but I am sure of one thing, must find out what I did wrong before doing it again. Highly suspect it is my wiring and question some of the ground points.
Had to revisit the add for the handy dandy $20 multi-tester since it came with zero documentation. Lower limit is 25pf, so yeah, C2 (@22pf) may be good. Also too small for DATS to measure. @14, those 1n5306 are certainly not cheap.
No scope. Should have enough scraps laying around to build something like this dim buld tester diyaudioprojects.com/Forum/viewtopic.php?t=1051
after that, what do I look for?
Looking at the preamp schem ...
+audio input goes to pin2gridB, - audio input is the shield or just ground since this is unbalanced, -input for pin7gridA goes to ground ...
"the non-inverting input terminals are +INP and SGND. When using this input, -INP need to be grounded. This is a must otherwise feedback cannot be accomplished and the amplifier gain is too high to be used".
To the right of the schem, it shows the separate heater circuit with pin9shield to sgnd/pgnd1 (there is no sgnd/pgnd1 on board, only sgnd/pgnd which I assume is the same), pin9shield trace is not connected to anything. The heater uses a separate power supply and separate power switch so it can warm up before powering up the board, and it appears to work with tube in place.
Does pin9shield get grounded? Why or why not?
I'm not a tube guy, but according to the drawing pin 9 gets grounded, and you are correct in that this design has a common power and signal ground. Do you have real transistors? I don't believe that they have been available for years. If you are getting the ss active devices from ebay, there is about a 99% chance that they are counterfeit junk.
Was not aware of the NLA pieces when ordering this board or I wouldn't have ordered it, and yes, the validity of the ebay parts are concerning. Am I wrong to think 2sj79 and 2sk216, being a p and n mos, should be sort of matched? The limited functionality of the handy dandy multi-tester shows a huge difference in gate to source, gate cap, drain to source resistance, and forward voltage of protecting diode between the two different trans, and with my very limited knowledge, it's impossible for me to convert and compare these measurements to those of the oem datasheets.
The next step appears to be; where do I source known good transistors?
I just did some digging, and it seems that some of the fake Chinese FET's have the pins in the wrong order. The originals are gsd and the fakes are gds. That would certainly cause smoke, and be easy to miss, if not tested and checked for.
You beat me to it!
Just ran across that after wasting most of a day digging around for a reputable supplier. Checked it and the source and drain pins of the 2sk216 are reversed. Should be 1(gate), 2(source), 3(drain) but it tests 1(gate), 2(drain), 3(source). Wondering if the thing will work as a proper 2sk216 if the d/s are jumpered back into s/d position?
The pinout of 2sj79, if that's what it is, tests correctly.
This could be the entire problem???
These darn mosfets are proving to be very elusive. May have to order them from Ireland! Not going to be cheap.
I would guess that is what caused the smoke. You can make an adapter with a header and perf board. Properly oriented it should function, how well or how long is anyone's guess. The counterfeits may be a close match relabeled or the could have an undersized die, so a proper heatsink might be necessary.
Before you fire this thing back up, center the dc offset potentiometer, and be aware that because we don't know exactly what FET's you have it's possible that it will sef oscillate, so if you get any heat shut it dow.
While not able to nail down a defining observation of real vs fake, it sounds like the fakes are close enough to work but not correctly. I'll set the pot to mid now. Shipping from Ireland is going to take a while, plenty of time to build the dbt.
I know that you are invested in this project, but it might make sense to just buy a different set of boards that don't use unobtainium parts, at least from a financial perspective.
As you alluded to, I am already vested in this project. As many a diy project, it might be cheaper to just buy something other than build it. I'll have a spare pair to build another set or sale the extra trans, and chalk the rest up as a $ learning experience.
... and yes, now I know better. Next time, I'll be sure parts are available and affordable before buying.
I was originally looking for a board based on the Bravo headphone amp but instead a preamp. If you run across one, let me know.
The handy dandy ebay multitester says both old and new e-202 current regulating diodes https://mouser.com/datasheet/2/362/P22-23-CRD-1729293.pdf are 27pf capacitors. Using a multimeter on diode setting says .924v forward and .65 reverse bias, with the new e-202 measuring close to those numbers.
The handy dandy ebay multitester gives uf=713mv, c=0pf, Ir=23ma and the multimeter says .62v forward and oL reverse bias, for the used 1n4148 gen purpose diode https://mouser.com/datasheet/2/427/1n4148p-1766760.pdf and close to those numbers for the new ones.
Since both new and used measure nearly identical, I'm thinking the e-202 may not have fried. Why does the e-202 not measure like the other diodes? Is there another easy way to use a multimeter to test the e-202 and what do I look for?
Received new parts from Ceitron in Dayton, OH. He also sells on ebay. The ebay 2sk216 are no telling what, but definitely not a 2sk216. Even if the pinout had been correct, there was no way this trans was ever going to work.
The ebay 2sj79 is closer to the oem, but the pinout is still wrong. If the pinout was correct, maybe it would have worked? The one from the board must have been fried as the measurements do not match the other new ebay parts.
New parts in the board. Time for smoke test later today.
How to differentiate fake 2sj79 and 2sk216 from the real thing before buying and wasting six weeks waiting on fake parts?
You got me.
They are some visible differences but not much. First thing noticed is the thick part of the legs is much longer on the real ones. The back side of the heatsink is shaped more like a bell with a small tab vs a more square looking backside and upper portion. Although you cannot see it in the pic, the insignia on the lower left corner of the real 2sk216 is larger. The letter inside of the bellybutton of the 2sj79 is a upper case T, the real one is blank. The letter inside the bellybutton of the fake 2sk216 is T while the real one is blank.
Even if they're not the genuine part, maybe they'll be close enough to work in your circuit. I have fingers crossed!! If you get sparks & smoke, there's still a lot you can do with the 6DJ8s. I can probably put together a little care package to build something cool.
Need to drop the current on the 6.5dc heater circuit, datasheet says 365ma heater current for a 6dj8. The dc/dc convertor drops the 24v, from the smps, down to 6.5v with max 1a output, but it measures 1.8a after the reg. Can I just add a resistor to the circuit till the correct current is reached and should this resistor go before the voltage reg or after, or does it matter? Without the tube heaters connected, and a 10ohm resistor after the voltage reg; 6.5v and .62a. Am I correct in adding both tube heater requirements together for a total of ~700ma?
Comments
Still less expensive than a college degree.
I'm following because, DIY.
Also making payments on that and it is waaaaaaaay more expensive.
As you can tell, it wasn't for an EE degree.
You also have a tube/mosfet project. When finished, I'm anxious to hear your opinion.
Just so you don't fell like you're the only one - I had a little smoke on my tube project last night. I'll post the details in my thread.
Got the multi-checker. Voltage reg LM336z-2.5 is no longer. What happens when a voltage reg is no longer a voltage reg? Have readings for the 2sa970 bjt-pnp that I'll compare to a new one. Also have measurements for the p and n mosfets. Don't these work in tandem and shouldn't they have similar measurements? Will also check the measurements against the new parts. Worried one or both of these could be fakes.
I think we have the same little tester - mine says a brand new LM317 is a damaged part. Hard to say if your 336s are still good. At least they're cheap to replace. I just got a box of stuff from Digikey and threw in extra LM334s since they seem pretty useful with small signal tubes.
Readings & specs for transistors are beyond what little SS knowledge I have. I'm hoping to learn a bit by building the Pass clone kit soon, but not sure I really understand what's going on with that circuit yet.
John H - do you remember anything about Mosfets from your college days?
Is there anything on this board I didn't fry??? Those little e202 (D1a/D1b) current reg diodes are $2 each. Top row pulled from board, bottom row new part. D3 tests bad, 2sj79 does not test like the new one, and D1b/D1a current reg diodes test as a cap.
All this from not installing the tube???
Wow - that is a lot of damage! Can you still get those mosfets from a reputable source?
So far, ebay, and who knows if they are legit.
Hard for me to believe all this damage is from not installing the tube. The plates only see 24vdc.
Aha, check this out. BOM calls for "J508 or CR200 2.4mA constant current diode". According to Newark, cr200 is Vishay "CR200 Zener Single Diode, 300 mW, TO-18, 10 %, 2 Pins" but I do not see that J508 is a zener. I subbed e202 for these hard to find diodes to keep from chancing ebay fakes and I don't think e202 is a zener.
Datasheets attached.
Whas up wit dis?
Holy $hit, Vishay cr200 $60 on ebay.
Just to be sure I have this straight; a zener diode can conduct in both directions, a regular diode cannot.
No where else have I found J508 listed as a zener and the board manu says he stuffs D1a/D1b with e-352&e-152 (<specs can be found on the previously attached e202 diode pdf.) to total 5ma. I subbed it with two e-202 (total 4ma) but doubt this was THE major problem.
So ... back to the biggie ... WTF happened?
Of course Mouser is out of stock, even the smd version.
Now you are getting into stuff I sure can't answer. I sort of understand zeners, but didn't know there were constant current diodes. I'm sure somebody here could explain that circuit, but once you add SS devices more complex than a zener to the schematic, I need somebody to walk me through it.
Add a 22pf mica cap to the DOA list. Since that cap is rated 500vdc I guess it's not the voltage but the current that killed it? Mouser didn't have the recommended size diodes in stock so had to buy extra and will have to 'stack' them for the correct value.
Are you sure all those parts were destroyed? C2 is across a 10K resistor, seems strange that it would be taken out. D1 diodes are available at Mouser 1N5306 is a 2.4ma current regulating diode, and should be the equivalent, but not cheap. The E202 that you listed should be close enough. I would look at the power supply and all grounding first, after that I would look to a possible oscillation, considering that you said that there were no tubes installed. Can you look at this on a scope? Did you use a dim bulb tester when you fired this up?
Nope, not sure at all, but I am sure of one thing, must find out what I did wrong before doing it again. Highly suspect it is my wiring and question some of the ground points.
Had to revisit the add for the handy dandy $20 multi-tester since it came with zero documentation. Lower limit is 25pf, so yeah, C2 (@22pf) may be good. Also too small for DATS to measure.
@14, those 1n5306 are certainly not cheap.
No scope. Should have enough scraps laying around to build something like this dim buld tester diyaudioprojects.com/Forum/viewtopic.php?t=1051
after that, what do I look for?
Looking at the preamp schem ...
+audio input goes to pin2gridB, - audio input is the shield or just ground since this is unbalanced, -input for pin7gridA goes to ground ...
"the non-inverting input terminals are +INP and SGND. When using this input, -INP need to be grounded. This is a must otherwise feedback cannot be accomplished and the amplifier gain is too high to be used".
To the right of the schem, it shows the separate heater circuit with pin9shield to sgnd/pgnd1 (there is no sgnd/pgnd1 on board, only sgnd/pgnd which I assume is the same), pin9shield trace is not connected to anything. The heater uses a separate power supply and separate power switch so it can warm up before powering up the board, and it appears to work with tube in place.
Does pin9shield get grounded? Why or why not?
Are we good up to this point?
I'm not a tube guy, but according to the drawing pin 9 gets grounded, and you are correct in that this design has a common power and signal ground. Do you have real transistors? I don't believe that they have been available for years. If you are getting the ss active devices from ebay, there is about a 99% chance that they are counterfeit junk.
Was not aware of the NLA pieces when ordering this board or I wouldn't have ordered it, and yes, the validity of the ebay parts are concerning. Am I wrong to think 2sj79 and 2sk216, being a p and n mos, should be sort of matched? The limited functionality of the handy dandy multi-tester shows a huge difference in gate to source, gate cap, drain to source resistance, and forward voltage of protecting diode between the two different trans, and with my very limited knowledge, it's impossible for me to convert and compare these measurements to those of the oem datasheets.
The next step appears to be; where do I source known good transistors?
I just did some digging, and it seems that some of the fake Chinese FET's have the pins in the wrong order. The originals are gsd and the fakes are gds. That would certainly cause smoke, and be easy to miss, if not tested and checked for.
You beat me to it!
Just ran across that after wasting most of a day digging around for a reputable supplier. Checked it and the source and drain pins of the 2sk216 are reversed. Should be 1(gate), 2(source), 3(drain) but it tests 1(gate), 2(drain), 3(source). Wondering if the thing will work as a proper 2sk216 if the d/s are jumpered back into s/d position?
The pinout of 2sj79, if that's what it is, tests correctly.
This could be the entire problem???
These darn mosfets are proving to be very elusive. May have to order them from Ireland! Not going to be cheap.
I would guess that is what caused the smoke. You can make an adapter with a header and perf board. Properly oriented it should function, how well or how long is anyone's guess. The counterfeits may be a close match relabeled or the could have an undersized die, so a proper heatsink might be necessary.
Before you fire this thing back up, center the dc offset potentiometer, and be aware that because we don't know exactly what FET's you have it's possible that it will sef oscillate, so if you get any heat shut it dow.
While not able to nail down a defining observation of real vs fake, it sounds like the fakes are close enough to work but not correctly. I'll set the pot to mid now. Shipping from Ireland is going to take a while, plenty of time to build the dbt.
$14 each for the 2sj79&2sk216. $5 for the 2sa970. Supposed to be Renesas. Anyone need some, I'll order a couple extra?
I know that you are invested in this project, but it might make sense to just buy a different set of boards that don't use unobtainium parts, at least from a financial perspective.
As you alluded to, I am already vested in this project. As many a diy project, it might be cheaper to just buy something other than build it. I'll have a spare pair to build another set or sale the extra trans, and chalk the rest up as a $ learning experience.
... and yes, now I know better. Next time, I'll be sure parts are available and affordable before buying.
I was originally looking for a board based on the Bravo headphone amp but instead a preamp. If you run across one, let me know.
Learning curve ....
The handy dandy ebay multitester says both old and new e-202 current regulating diodes https://mouser.com/datasheet/2/362/P22-23-CRD-1729293.pdf are 27pf capacitors. Using a multimeter on diode setting says .924v forward and .65 reverse bias, with the new e-202 measuring close to those numbers.
The handy dandy ebay multitester gives uf=713mv, c=0pf, Ir=23ma and the multimeter says .62v forward and oL reverse bias, for the used 1n4148 gen purpose diode https://mouser.com/datasheet/2/427/1n4148p-1766760.pdf and close to those numbers for the new ones.
Since both new and used measure nearly identical, I'm thinking the e-202 may not have fried. Why does the e-202 not measure like the other diodes? Is there another easy way to use a multimeter to test the e-202 and what do I look for?
Received new parts from Ceitron in Dayton, OH. He also sells on ebay. The ebay 2sk216 are no telling what, but definitely not a 2sk216. Even if the pinout had been correct, there was no way this trans was ever going to work.
The ebay 2sj79 is closer to the oem, but the pinout is still wrong. If the pinout was correct, maybe it would have worked? The one from the board must have been fried as the measurements do not match the other new ebay parts.
ebay 2sa970 looks like it may work.
New parts in the board. Time for smoke test later today.
How to differentiate fake 2sj79 and 2sk216 from the real thing before buying and wasting six weeks waiting on fake parts?
You got me.
They are some visible differences but not much. First thing noticed is the thick part of the legs is much longer on the real ones. The back side of the heatsink is shaped more like a bell with a small tab vs a more square looking backside and upper portion. Although you cannot see it in the pic, the insignia on the lower left corner of the real 2sk216 is larger. The letter inside of the bellybutton of the 2sj79 is a upper case T, the real one is blank. The letter inside the bellybutton of the fake 2sk216 is T while the real one is blank.
Left the pic full size so you can zoom in.
Even if they're not the genuine part, maybe they'll be close enough to work in your circuit. I have fingers crossed!! If you get sparks & smoke, there's still a lot you can do with the 6DJ8s. I can probably put together a little care package to build something cool.
Need to drop the current on the 6.5dc heater circuit, datasheet says 365ma heater current for a 6dj8. The dc/dc convertor drops the 24v, from the smps, down to 6.5v with max 1a output, but it measures 1.8a after the reg. Can I just add a resistor to the circuit till the correct current is reached and should this resistor go before the voltage reg or after, or does it matter? Without the tube heaters connected, and a 10ohm resistor after the voltage reg; 6.5v and .62a. Am I correct in adding both tube heater requirements together for a total of ~700ma?