Thanks for the work, effort, and ideas so far fellas! What I really need after you experts get all the theory and debate worked out is a simple A, B, C, 1, 2, 3 for hands-on changes in a real noob simple form. As-in do this, then this, etc.
My grand daughter told me today that her journalism teacher 101 says anything you discuss or write publicly has to be in about a 9th grade level comprehension level anymore so it could be easily read and done by just about anyone. I laughed and agreed and thought about my dilemma and this thread.
All Ive done so far hands-on is replace 4 'lytics, and have not replaced any of the two transistors (as I only have one). Cant replace the two 22uf 450+v caps with huge solen polypropylene as I dont think there is any way possible to physically fit them anywhere.
So... am thankful so far for your ideas so far with this cheap concertina engineering..
My grand daughter told me today that her journalism teacher 101 says anything you discuss or write publicly has to be in about a 9th grade level comprehension level anymore so it could be easily read and done by just about anyone. I laughed and agreed and thought about my dilemma and this thread.
All Ive done so far hands-on is replace 4 'lytics, and have not replaced any of the two transistors (as I only have one). Cant replace the two 22uf 450+v caps with huge solen polypropylene as I dont think there is any way possible to physically fit them anywhere.
So... am thankful so far for your ideas so far with this cheap concertina engineering..
Yaquin made a huge direct coupling voltage level mistake.
Serious enough to defeat proper function of NPN Concertina.
Diagram of Post#35 PNP fix is good, cheap, simple.
And requires no irreversible alterations to the board.
Actual fix of a real problem will make more difference
than throwing random cap upgrades at it.
MJE350 is dirt cheap and widely available everywhere.
MJE350 Transistor PNP 300V 0.5A TO-126 (SOT-32) | Parts-Express.com
I don't yet see clear enough picture of your plate and screen
connections to mark up where to swap those. But you can't
switch to PNP concertina without, else feedback will oscillate.
Snap me pictures of those transformer connections, so I can draw over.
Serious enough to defeat proper function of NPN Concertina.
Diagram of Post#35 PNP fix is good, cheap, simple.
And requires no irreversible alterations to the board.
Actual fix of a real problem will make more difference
than throwing random cap upgrades at it.
MJE350 is dirt cheap and widely available everywhere.
MJE350 Transistor PNP 300V 0.5A TO-126 (SOT-32) | Parts-Express.com
I don't yet see clear enough picture of your plate and screen
connections to mark up where to swap those. But you can't
switch to PNP concertina without, else feedback will oscillate.
Snap me pictures of those transformer connections, so I can draw over.
Attachments
Last edited:
I thought this amplifier was working as delivered. If Ken is right about the idle voltages from the 1st stage, then it should sound horrible I would think. Some voltage measurements at the splitter transistor's base and collector should be made during operation (in a no signal state, relative to ground) so we can be sure of what is going on there (caution, high voltage present). Also the circuit needs to be checked (Ohm-meter will do, while turned off) to see which splitter load resistors are connected to the splitter transistor's collector and emitter (and the resistor values), since there is an issue with the schematic already.
Last edited:
I suspect base to emitter leak is still driving one pentode, and the same leak is bringing
SRPP voltage down, but SRPP would not behave as constant current source anymore.
Assumes Yaquin faithfully built mistakes as drawn in Post#1 without undocumented fixes.
Just cause it makes noises don't mean "working" as intended. You can read a lot of
good intentions into that topology. Yet simply blew it when direct coupling to NPN
Concertina at voltage wouln't have worked even with an old fashioned Triode.
Look at the 1K and 2K cathode resistors in the SRPP stack. If both are same current,
and both triodes the same Mu. Then 2K must bias the bottom triode to drop twice as
much voltage as top. Again we got huge current dump into NPN base when voltage
approaches B+/2, the SRPP will never actually reach its own voltage of equilibrium.
I'm constantly having to stop boneheads from pushing things out the door
that only halfway sort of work. Some regulators with the enable pin floating,
can we lets at least ship the half that work?!? NoOOO! fixit... At every successful
company there must exist at least one clone of me, else nothing would ever work.
Granted without me, a lot more things would ship on time...
SRPP voltage down, but SRPP would not behave as constant current source anymore.
Assumes Yaquin faithfully built mistakes as drawn in Post#1 without undocumented fixes.
Just cause it makes noises don't mean "working" as intended. You can read a lot of
good intentions into that topology. Yet simply blew it when direct coupling to NPN
Concertina at voltage wouln't have worked even with an old fashioned Triode.
Look at the 1K and 2K cathode resistors in the SRPP stack. If both are same current,
and both triodes the same Mu. Then 2K must bias the bottom triode to drop twice as
much voltage as top. Again we got huge current dump into NPN base when voltage
approaches B+/2, the SRPP will never actually reach its own voltage of equilibrium.
I'm constantly having to stop boneheads from pushing things out the door
that only halfway sort of work. Some regulators with the enable pin floating,
can we lets at least ship the half that work?!? NoOOO! fixit... At every successful
company there must exist at least one clone of me, else nothing would ever work.
Granted without me, a lot more things would ship on time...
Last edited:
If Yaquin swapped locations of 1K and 2K cathode resistors, that also might work.
Yeah, measure first to see if they might already have done something like that...
Don't change to PNP if the base voltage measures less than 150 to ground.
Thats the pin on the far right side of ECB.
I still think changing to PNP at 200V better than NPN at 100V by swapping the
cathode resistor locations... Choose one fix or the other, you can't do both.
Yeah, measure first to see if they might already have done something like that...
Don't change to PNP if the base voltage measures less than 150 to ground.
Thats the pin on the far right side of ECB.
I still think changing to PNP at 200V better than NPN at 100V by swapping the
cathode resistor locations... Choose one fix or the other, you can't do both.
Last edited:
If one was to get all theoretical and hypo-theatrical about it...
Here's simulation of proper bias point for maximum headroom.
Centered a few volts down from B+/2, on account of abusing
the PNP's VBE to double duty as a constant current source
for the 12AX7's plate...
Not perfectly stiff as a current source, the slope is only 190K.
But hey, what do you want from a 75cent transistor?
OK, I did sabotage the global loop in favor of a local one that
stabilizes the DC operating points. Not to be taken seriously
as a proposed fix for the Yaqin. I drew this thinking of driving
6L6 into A2 in some future project of my own.
To run the sim, you need LTSpice from Linear Technologies:
http://www.linear.com/designtools/software/ltspice.jsp
Afer downlaod, rename "PSoup1.asc.txt" to just "PSoup1.asc"
And dmtriodep.inc from Duncan Amps:
http://www.duncanamps.com/spice/valves/
NH12AX7, simply means "No Heater"12AX7 in Duncanese.
Its not a special valve made only on Mars or anything...
MJE350 spice model I've already pasted for you into the sim.
Ripped from the big zipfile at OnSemi. Just so you know...
Here's simulation of proper bias point for maximum headroom.
Centered a few volts down from B+/2, on account of abusing
the PNP's VBE to double duty as a constant current source
for the 12AX7's plate...
Not perfectly stiff as a current source, the slope is only 190K.
But hey, what do you want from a 75cent transistor?
OK, I did sabotage the global loop in favor of a local one that
stabilizes the DC operating points. Not to be taken seriously
as a proposed fix for the Yaqin. I drew this thinking of driving
6L6 into A2 in some future project of my own.
To run the sim, you need LTSpice from Linear Technologies:
http://www.linear.com/designtools/software/ltspice.jsp
Afer downlaod, rename "PSoup1.asc.txt" to just "PSoup1.asc"
And dmtriodep.inc from Duncan Amps:
http://www.duncanamps.com/spice/valves/
NH12AX7, simply means "No Heater"12AX7 in Duncanese.
Its not a special valve made only on Mars or anything...
MJE350 spice model I've already pasted for you into the sim.
Ripped from the big zipfile at OnSemi. Just so you know...
Attachments
Last edited:
I also thought if triode saw feedback from both plate and cathode it might
help equalize impedances? Not sure I've accomplished that, but step in the
right direction... Gain would be 100 if Plate Mu were the only feedback, I've
knocked it down to 50 at the cathode. Might be worth further tweaking.
Differential gain is still 100, and thus equal to Mu if that figures equal Z's?
I dunno. I need to put various loads on it and see what happens.
------------------------
I should have said centered slightly below (3/4)B+. Mirrored slightly below
(1/2)B+. 14V headroom nearest B+ was sacrificed to gain use of free CCS.
3mA plate current must drop 14V uselessly across R3, an acceptable loss...
The PNP biases at 220V, plenty for proper operation of the triode.
If we were driving an NPN concertina, no free CCS. And you have to bias
the base at (1/4)B+ = 75V. I believe this low voltage not best for triode.
Cathode feedback would be posistive, so Z equalization not an option.
help equalize impedances? Not sure I've accomplished that, but step in the
right direction... Gain would be 100 if Plate Mu were the only feedback, I've
knocked it down to 50 at the cathode. Might be worth further tweaking.
Differential gain is still 100, and thus equal to Mu if that figures equal Z's?
I dunno. I need to put various loads on it and see what happens.
------------------------
I should have said centered slightly below (3/4)B+. Mirrored slightly below
(1/2)B+. 14V headroom nearest B+ was sacrificed to gain use of free CCS.
3mA plate current must drop 14V uselessly across R3, an acceptable loss...
The PNP biases at 220V, plenty for proper operation of the triode.
If we were driving an NPN concertina, no free CCS. And you have to bias
the base at (1/4)B+ = 75V. I believe this low voltage not best for triode.
Cathode feedback would be posistive, so Z equalization not an option.
Last edited:
Maybe I got P Concertina Z match first shot by accident or luck?
1/Mu plate feedback + 1/100 cathode feedback by Schade = 1/50.
Equal feedbacks are what make impedances, not so much R2 = R3.
I may not have any need to tweak, if already perfect...
An N soloution can't see what happens on the collector side.
So impedance of NPN B+ side is very high except for R3 itself.
No match to emitter impedance, reduced by Mu feedback, in
parallel with R2 leak to ground.
1/Mu plate feedback + 1/100 cathode feedback by Schade = 1/50.
Equal feedbacks are what make impedances, not so much R2 = R3.
I may not have any need to tweak, if already perfect...
An N soloution can't see what happens on the collector side.
So impedance of NPN B+ side is very high except for R3 itself.
No match to emitter impedance, reduced by Mu feedback, in
parallel with R2 leak to ground.
Last edited:
Cool! Sim for yourself confirmation of Z match:
Replace R5 or R6 with 10K and watch identical
exaggerations of gain on the other 47K. There
is perfect symmetry of expected misbehavior.
This is way better balance than I had hoped!
Replace R5 or R6 with 10K and watch identical
exaggerations of gain on the other 47K. There
is perfect symmetry of expected misbehavior.
This is way better balance than I had hoped!
Last edited:
Ken- PM sent 2 weeks ago.
Stay with me fellas. Working on better pics, etc.
Finally got a brand new set of JJ's, not much improvement. As has been said, circuitry/design/mods count more than tubes most of the time. I did have a popping, bad EL84. New tubes almost broken in, but need to some of the ideas/mods Ken and other have come up wit so far.🙂
Stay with me fellas. Working on better pics, etc.
Finally got a brand new set of JJ's, not much improvement. As has been said, circuitry/design/mods count more than tubes most of the time. I did have a popping, bad EL84. New tubes almost broken in, but need to some of the ideas/mods Ken and other have come up wit so far.🙂
I have the same JJ's in my Fisher X100, fine choice.
But you still need to measure the original splitter bias voltage to determine
if the problem seen in schematic is real or imaginary, or already patched in
some other way not documented in the drawing. Measure at B or E of the
transistor, speakers MUST be connected for safety, but no music playing.
Fortunately your cathode bias appears self adjusting. I doubt you are gonna
burn up the JJ's just rolling them blindly. Still you might wanna measure the
point marked "A" in the original drawing to see if the new tubes are drastically
different from the old set.
Some EL84 and 7189 have difference of pins set aside for internal use only.
If those points are abused as free lugs to solder other things, a change of
tube could be disastrous. You have already plugged and nothing burned up,
so its a little late to worry over how the extra pins of the socket might or
not be wired to things they shouldn't be.
But you still need to measure the original splitter bias voltage to determine
if the problem seen in schematic is real or imaginary, or already patched in
some other way not documented in the drawing. Measure at B or E of the
transistor, speakers MUST be connected for safety, but no music playing.
Fortunately your cathode bias appears self adjusting. I doubt you are gonna
burn up the JJ's just rolling them blindly. Still you might wanna measure the
point marked "A" in the original drawing to see if the new tubes are drastically
different from the old set.
Some EL84 and 7189 have difference of pins set aside for internal use only.
If those points are abused as free lugs to solder other things, a change of
tube could be disastrous. You have already plugged and nothing burned up,
so its a little late to worry over how the extra pins of the socket might or
not be wired to things they shouldn't be.
Last edited:
You don't need to post "PM sent", I got PM notifications turned on.
I don't answer any tech questions by PM, only contact info when
I might occasionally trade for parts or books. My PM is 97% full...
Yes, "A" appears two places in the drawing. I'm assuming they are
connected together for purpose of "elevating" DC offset of heaters.
The top SRPP triode might be destroyed if cathode voltage is much
higher than the heater. So the heater DC is allowed to rise halfway
between the idle voltages of the two SRPP cathodes. To minimize
the average heater to cathode voltage stress...
I suppose this protects the EL84 heaters from Bias Voltage stress
too, though this is not really a big issue.
It don't matter which "A" you measure (to ground if not otherwise
specified) cause all "A" are the same wire, same trace, same net.
ECB on the silkscreen is a pinout, not a NET of the schematic.
The E's of the two transistors do not connect together anywhere.
Though they should measure about the same idle voltage anyway.
I don't answer any tech questions by PM, only contact info when
I might occasionally trade for parts or books. My PM is 97% full...
Yes, "A" appears two places in the drawing. I'm assuming they are
connected together for purpose of "elevating" DC offset of heaters.
The top SRPP triode might be destroyed if cathode voltage is much
higher than the heater. So the heater DC is allowed to rise halfway
between the idle voltages of the two SRPP cathodes. To minimize
the average heater to cathode voltage stress...
I suppose this protects the EL84 heaters from Bias Voltage stress
too, though this is not really a big issue.
It don't matter which "A" you measure (to ground if not otherwise
specified) cause all "A" are the same wire, same trace, same net.
ECB on the silkscreen is a pinout, not a NET of the schematic.
The E's of the two transistors do not connect together anywhere.
Though they should measure about the same idle voltage anyway.
Last edited:
And when you measure, it will be with one hand in your pocket.
Never with other hand holding the chassis. Use something else
besides yourself to make sure it won't tip over before you apply
any power.
The idea here is that when you eventually electrocute yourself,
at least it won't be across your chest. Have someone near with
a wooden broomstick to pry you away and/or yank the plug out
of the wall is good backup. Have charged cell phone nearby.
Use gator clip for the ground probe or both, that way you don't
have two hands in there at once.
Never with other hand holding the chassis. Use something else
besides yourself to make sure it won't tip over before you apply
any power.
The idea here is that when you eventually electrocute yourself,
at least it won't be across your chest. Have someone near with
a wooden broomstick to pry you away and/or yank the plug out
of the wall is good backup. Have charged cell phone nearby.
Use gator clip for the ground probe or both, that way you don't
have two hands in there at once.
Last edited:
I have one of these Yaqin MC-84L amps on its way. When it gets here, I'll take those vital voltage measurements on the amp itself as supplied and will post them here. I'll see if there's any differences in actual amp to the schematic too....
I would prefer a fully tube signal-path and was thinking of something along the lines Wavebourne mentioned earlier,--MOSFET on top of the gain-stage triode--which as I'm going to add a tube hidden down in the chassis, I'll strap the AX7 in parallel DC coupled to a 6N16B-V medium Mu dual triode, sorta like a SN7 in some ways, a tiny wire-ended part and use one triode half for each channel.....
I'm quite impressed by Kenpeter's hybrid Mu/gain/phase-splitter posted earlier in this thread though--Might give that a try first, as its probably the simplest to implement without chopping the thing about too much.....
Allegedly--It'll be here Friday....
I would prefer a fully tube signal-path and was thinking of something along the lines Wavebourne mentioned earlier,--MOSFET on top of the gain-stage triode--which as I'm going to add a tube hidden down in the chassis, I'll strap the AX7 in parallel DC coupled to a 6N16B-V medium Mu dual triode, sorta like a SN7 in some ways, a tiny wire-ended part and use one triode half for each channel.....
I'm quite impressed by Kenpeter's hybrid Mu/gain/phase-splitter posted earlier in this thread though--Might give that a try first, as its probably the simplest to implement without chopping the thing about too much.....
Allegedly--It'll be here Friday....
Yaqin MC-84L is an el84 amp... push-pull el84 can be very good so long at the iron is not wound by a monkey on methaqualone.
I would stick with EL84 push-pull. Everything else should be re-designed, as previously noted. I would suggest looking at Mr. Morgan Jones' book.
The pictures of this amp on the internet suggest there is a little bit of space on hand.
Well--Here's the Voltages of the gain-stage and phase-splitter as is in the amp--
Transistor,
E = 157V
B = 158V
C = 181V
Collector resistor = 39K
Emitter resistor = 47K
Top triode Anode = +B at 330V
Top Grid = 157V
Top Cathode = 158V
Bottom Triode-
Bottom plate, 157V
Bottom Grid, = Didnt measure it.....
Bottom Cathode, = 1.1V
The sound is OK ish--but sorta flat/hollow sounding Already started re-designing as already noted above, One channel complete and testing......
Transistor,
E = 157V
B = 158V
C = 181V
Collector resistor = 39K
Emitter resistor = 47K
Top triode Anode = +B at 330V
Top Grid = 157V
Top Cathode = 158V
Bottom Triode-
Bottom plate, 157V
Bottom Grid, = Didnt measure it.....
Bottom Cathode, = 1.1V
The sound is OK ish--but sorta flat/hollow sounding Already started re-designing as already noted above, One channel complete and testing......
If you make the cathode resistor of the upper triode greater the output voltage of the SRPP drops.No need to switch to PNP.
With a Hfe of more the 20 it make no sense to have different resistors at the collector and emitter.
The output ~voltage can do for driving EL84, a bit short for EL34.
Then you have to change the SRPP tube or lowering the current in the 12AX7.
The 2M2 resistance prevents getting the full 340V on the transistor at startup.
Mona
With a Hfe of more the 20 it make no sense to have different resistors at the collector and emitter.
The output ~voltage can do for driving EL84, a bit short for EL34.
Then you have to change the SRPP tube or lowering the current in the 12AX7.
The 2M2 resistance prevents getting the full 340V on the transistor at startup.
Mona
Attachments
- Status
- Not open for further replies.