There is a problem measuring voltage drop across primary for plate current read. Copper's resistance has pretty strong temperature dependence, so winding resistance increases when transformer warms up.
Lowthers are more sensitive speakers than Altecs, so no surprise in sound difference between 2A3 and 300B.
Following your progress with much interest.
Lowthers are more sensitive speakers than Altecs, so no surprise in sound difference between 2A3 and 300B.
Following your progress with much interest.
sser2,
You are correct!
Thanks!
DCR changes when the transformer gets warm.
The reason for problems is to keep us busy solving them.
Solution:
Drill 3 holes to mount 3 tip sockets into the chassis.
Have 3 tip sockets for easy access no need to open the amplifier bottom cover; connect tip sockets to plate # 1, center tap, and plate # 2.
Run the amplifier for 1 Hour.
Turn the amplifier off, set DMM to DCV, measure the center tap voltage to ground, when B+ has discharged . . .
Check that DCV is zero, from plate # 1 to center tap, and from plate # 2 to center tap.
Then set the DMM to Ohms, measure plate # 1 to center tap, and measure plate # 2 to center tap.
(the primary windings do not cool very quickly, it took one hour to get warm, so now you know the Hot primary DCRs).
"Necessity is the mother of invention"
If you do not want to mount 3 tip sockets . . .
With a cold amplifier, remove the bottom cover, get a power supply set to 60mA current limit, connect it from plate # 1 to plate # 2, and warm up the primary for 1 hour.
Turn the power supply voltage down slowly to zero, and remove the power supply connections.
Measure DCR of plate # 1 to center tap, and measure DCR of plate # 2 to center tap.
That is a "hot" idea.
Quality does not happen on its own, it takes time and effort.
$0.03 . . .
Accounting for Inflation
You are correct!
Thanks!
DCR changes when the transformer gets warm.
The reason for problems is to keep us busy solving them.
Solution:
Drill 3 holes to mount 3 tip sockets into the chassis.
Have 3 tip sockets for easy access no need to open the amplifier bottom cover; connect tip sockets to plate # 1, center tap, and plate # 2.
Run the amplifier for 1 Hour.
Turn the amplifier off, set DMM to DCV, measure the center tap voltage to ground, when B+ has discharged . . .
Check that DCV is zero, from plate # 1 to center tap, and from plate # 2 to center tap.
Then set the DMM to Ohms, measure plate # 1 to center tap, and measure plate # 2 to center tap.
(the primary windings do not cool very quickly, it took one hour to get warm, so now you know the Hot primary DCRs).
"Necessity is the mother of invention"
If you do not want to mount 3 tip sockets . . .
With a cold amplifier, remove the bottom cover, get a power supply set to 60mA current limit, connect it from plate # 1 to plate # 2, and warm up the primary for 1 hour.
Turn the power supply voltage down slowly to zero, and remove the power supply connections.
Measure DCR of plate # 1 to center tap, and measure DCR of plate # 2 to center tap.
That is a "hot" idea.
Quality does not happen on its own, it takes time and effort.
$0.03 . . .
Accounting for Inflation
Last edited:
6A3, euro21, and sser2: I've been thinking about all your suggestions and virtual center tap idea, in which two resistors from each filament pin meet at a (say) 10 ohm resistor that goes to ground, and then measuring the voltage drop across that resistor to get the plate current.
Maybe my drawing in #548 is unclear, but the 2A3 filament transformer is in fact a 2.5Vct model, and the transformer center tap goes to "ground" (at callout K2). Couldn't I just put a 10R (or any small value) resistor in series from the transformer center tap to the upper rail "ground" (or I guess you could call it the negative leg of the upper rail)? Or do the two legs from the filament pins have to be resistive, rather than inductive?
Maybe my drawing in #548 is unclear, but the 2A3 filament transformer is in fact a 2.5Vct model, and the transformer center tap goes to "ground" (at callout K2). Couldn't I just put a 10R (or any small value) resistor in series from the transformer center tap to the upper rail "ground" (or I guess you could call it the negative leg of the upper rail)? Or do the two legs from the filament pins have to be resistive, rather than inductive?
Yes.
Fixed Bias to the grid.
Single ended and 10 Ohms from the filament center tap to ground, works great!
For a Stereo single ended, you have to have 2 separate filament transformers, and 2 separate 10 Ohm resistors, so you can measure each individual 2A3 current.
Individual Self Bias:
For Self Bias, just put the 10 Ohm resistor from the bottom of the self bias resistor, and the other end of the 10 Ohm resistor to ground.
Fixed Bias to the grid.
Single ended and 10 Ohms from the filament center tap to ground, works great!
For a Stereo single ended, you have to have 2 separate filament transformers, and 2 separate 10 Ohm resistors, so you can measure each individual 2A3 current.
Individual Self Bias:
For Self Bias, just put the 10 Ohm resistor from the bottom of the self bias resistor, and the other end of the 10 Ohm resistor to ground.
The reason I use tip sockets on the top of the chassis of my amplifiers, is for Safety.
I often want to read high voltage parts of the circuits.
But I do not want to have to remove the bottom cover, turn the amp upside down, keep the tubes off the table, and then use my meter probes on the points of interest.
The danger is that if one of the probes slipped off the point, the probe will go further into the amplifier, stop, and then my hand may follow the probe.
If that causes my hand to contact a high voltage point . . . OUCH!
Safety First!
Prevent the "Surviving Spouse Syndrome"
Tip sockets work well.
I often want to read high voltage parts of the circuits.
But I do not want to have to remove the bottom cover, turn the amp upside down, keep the tubes off the table, and then use my meter probes on the points of interest.
The danger is that if one of the probes slipped off the point, the probe will go further into the amplifier, stop, and then my hand may follow the probe.
If that causes my hand to contact a high voltage point . . . OUCH!
Safety First!
Prevent the "Surviving Spouse Syndrome"
Tip sockets work well.
Member
Joined 2009
Paid Member
a bundle of low cost clip leads may help, with crocodile clips on both ends you can put them on points of interest and then you probe the other ends of the leads which you can place somewhere safer and more accessible. For example, clip the other ends onto a thin strip of wood or something non conducting.
You can also buy a couple extra meters and have them clipped into the amp before power up. After power up you don't need to go probing around because the meters are already in place.
You can also buy a couple extra meters and have them clipped into the amp before power up. After power up you don't need to go probing around because the meters are already in place.
I installed test tip sockets on the first amp I ever built back in 2015 -- the Tubelab SE with 300B outputs -- because that was fixed (grid) biased and required a lot of adjustment. With a couple of cheap multimeters from Harbor Freight, it worked great, allowing me to make adjustments on the fly while the amp was running (it had a number of potentiometers that you could turn), until the meters inevitably crapped out after 10 uses or so. I went through more of those than I care to count until finally buying a good Klein model.
I still have a bunch of test tip sockets, which would be good to add to the breadboard. However, I don't really feel the need because of the way everything is laid out and the conducting nodes are very accessible. It's actually quite easy to take measurements on the live amp. Clipped test leads would be useful, too. And maybe I should purchase a few potentiometers in my next Mouser order, so that I can make "live" adjustments without having to shut it down, wire up new resistors, turn it back on, etc., etc. The ability to make live adjustments was actually pretty nice with the Tubelab SE.
Just thinking ahead, if I end up doing a final build with one of these grid-biased output stages, then maybe I'll put current meters on the chassis faceplate so as to monitor the output tubes. And maybe even potentiometers for the sake of adjustment, so long as it wouldn't degrade the sound quality.
Anyway, work will be nuts until after tomorrow, when I'll try a measurement test on a resistor off the cathode.
I still have a bunch of test tip sockets, which would be good to add to the breadboard. However, I don't really feel the need because of the way everything is laid out and the conducting nodes are very accessible. It's actually quite easy to take measurements on the live amp. Clipped test leads would be useful, too. And maybe I should purchase a few potentiometers in my next Mouser order, so that I can make "live" adjustments without having to shut it down, wire up new resistors, turn it back on, etc., etc. The ability to make live adjustments was actually pretty nice with the Tubelab SE.
Just thinking ahead, if I end up doing a final build with one of these grid-biased output stages, then maybe I'll put current meters on the chassis faceplate so as to monitor the output tubes. And maybe even potentiometers for the sake of adjustment, so long as it wouldn't degrade the sound quality.
Anyway, work will be nuts until after tomorrow, when I'll try a measurement test on a resistor off the cathode.
Last edited:
My experience with clip leads . . .
Clip them on the amp. Clip the other end to the meter.
Turn the amp on.
Watch out! That clip is coming off where it is connected . . .
It "hops" over and shorts out a couple of wires or socket tabs, etc.
Amplifiers do not like shorting clip leads.
Now you know why I use tip sockets:
The prettyest amplifier that does not have tip sockets . . . can not produce music if the unintended 'shorting' clip lead burns something out.
$0.03
The above accounts for the inflated $0.02.
Clip them on the amp. Clip the other end to the meter.
Turn the amp on.
Watch out! That clip is coming off where it is connected . . .
It "hops" over and shorts out a couple of wires or socket tabs, etc.
Amplifiers do not like shorting clip leads.
Now you know why I use tip sockets:
The prettyest amplifier that does not have tip sockets . . . can not produce music if the unintended 'shorting' clip lead burns something out.
$0.03
The above accounts for the inflated $0.02.
I made several hook clip-banana cables.
For HV (over few hundred Volts) I use kV certified testing probes.
For HV (over few hundred Volts) I use kV certified testing probes.
OK, update on the question I posted above in #557, about how to measure the plate current of a grid-biased output tube (2A3) that is also in the top rail of a stacked power supply. The question was essentially whether I should pay attention to the voltage difference between grid and cathode, or whether I could use the DCR and voltage drop across the OPT primary to calculate it using Ohm's Law.
euro21 suggested putting a 10R resistor from the 2A3 cathode to "ground" and measuring the voltage there. Since posting that question, I made a version of the same circuit using 45 output tubes, which required few changes to run correctly. The attached image shows the measured voltages in blue pen (sorry for the crowded drawing). What I found is that the voltage drop across the DCR is consistent with the voltage drop across the cathode resistor, so it seems the OPT is a reliable way to measure plate voltage, at least in this instance.
With only about -42V bias on the 45 tube, where I expected a -50V bias, I measured 33mA of plate current in one tube and 38mA of current in the other. These were consistent with a 6V and 5V drop across the OPT primary (DCR 154 ohm) respectively.
So, especially with direct-coupled circuits, where the network "below" the output tube has a complex effect on its bias, you can't just go by measured Vgk and the tube datasheet.
I need to bring the current down in the 38mA tube. But this is a very good sounding circuit, just like the 2A3 version.
euro21 suggested putting a 10R resistor from the 2A3 cathode to "ground" and measuring the voltage there. Since posting that question, I made a version of the same circuit using 45 output tubes, which required few changes to run correctly. The attached image shows the measured voltages in blue pen (sorry for the crowded drawing). What I found is that the voltage drop across the DCR is consistent with the voltage drop across the cathode resistor, so it seems the OPT is a reliable way to measure plate voltage, at least in this instance.
With only about -42V bias on the 45 tube, where I expected a -50V bias, I measured 33mA of plate current in one tube and 38mA of current in the other. These were consistent with a 6V and 5V drop across the OPT primary (DCR 154 ohm) respectively.
So, especially with direct-coupled circuits, where the network "below" the output tube has a complex effect on its bias, you can't just go by measured Vgk and the tube datasheet.
I need to bring the current down in the 38mA tube. But this is a very good sounding circuit, just like the 2A3 version.
Attachments
Is this the same 2A3 SE project that was started Aug of 2022 or is my memory faulty?
How many iterations has the project gone thru. Any of them built & tested?
Why does it take so long to get such a simple amplifier completed.
Just me, John, a curious olde fart that built too many amps, some sitting here on the shelf,
but unused. But working.👍
In that dwg, the grid is biased Class A, the cathode current is simply the voltage drop across
the cathode resister divided by the resistance of the resister. Anything else, it is not Class A.
How many iterations has the project gone thru. Any of them built & tested?
Why does it take so long to get such a simple amplifier completed.
Just me, John, a curious olde fart that built too many amps, some sitting here on the shelf,
but unused. But working.👍
In that dwg, the grid is biased Class A, the cathode current is simply the voltage drop across
the cathode resister divided by the resistance of the resister. Anything else, it is not Class A.
The cathode resistor was just a temporary installation to measure the plate current, which I did by measuring the voltage across it. The resistors were a true 10 ohms (according to my Klein DMM), and at 330mV (left channel) and 380mV (right channel), the plates therefore were at 330/10 = 33mA and 380/10 = 38mA respectively. This is a temporary detour with a 45 in the output stage, just to have something to play around with until the anode chokes arrive for the 10Y, at which point I'll put the 2A3s back in.
Yes, since August 2022 it's been an open-ended project to build (breadboard), measure, and audition different circuits, some well known, some of my own design (the most recent). Many have been built and reported on along the way. Sorry, you'd have to go back through the thread to find them all, but most of those builds are probably early on. There were also several hiatuses (hiati?), one lasting about a year, due to busy work schedules.
When I wrap up this project, possibly with the next iteration using a 10Y as driver, then I'll post all the schemos and results in one place, either by editing the first post (with moderators' permission) or in an omnibus post at the end. Though in the latter case, if people keep replying then it'll get lost. I suppose I could post them to my profile bio or something, where the info would be more stable. We'll see.
This probably isn't the most interesting thread, so thanks for reading.
Yes, since August 2022 it's been an open-ended project to build (breadboard), measure, and audition different circuits, some well known, some of my own design (the most recent). Many have been built and reported on along the way. Sorry, you'd have to go back through the thread to find them all, but most of those builds are probably early on. There were also several hiatuses (hiati?), one lasting about a year, due to busy work schedules.
When I wrap up this project, possibly with the next iteration using a 10Y as driver, then I'll post all the schemos and results in one place, either by editing the first post (with moderators' permission) or in an omnibus post at the end. Though in the latter case, if people keep replying then it'll get lost. I suppose I could post them to my profile bio or something, where the info would be more stable. We'll see.
This probably isn't the most interesting thread, so thanks for reading.
Last edited:
Nice to see progress.
With the biasing thing.. of course .. and depends on what you want vs whats ideal, which may or may not be the same thing.
I really like DC circuits, you can do so many things.
For example: Type 26 B+ : CCS series filter set at the plate current, decouple with a capacitor. Stack the 45/2A3. Removes the input tube (sample) variation from affecting output tube bias - and you get a really low ripple B+ as an aside.
Then, you have fixed Vg, Ip of the input tube, and the variable is plate voltage (per sample).
At the output tube, this is fully compensated for by the fixed (CCS) through the DCR (choke and or resistor) bias.. and if you have an adjustable B+ (easy) you're off to the races.
Its the input tube, variation in DC condition spec which affects the 2 stage DC system. Easy to move that to the plate circuit of the output tube.
With the biasing thing.. of course .. and depends on what you want vs whats ideal, which may or may not be the same thing.
I really like DC circuits, you can do so many things.
For example: Type 26 B+ : CCS series filter set at the plate current, decouple with a capacitor. Stack the 45/2A3. Removes the input tube (sample) variation from affecting output tube bias - and you get a really low ripple B+ as an aside.
Then, you have fixed Vg, Ip of the input tube, and the variable is plate voltage (per sample).
At the output tube, this is fully compensated for by the fixed (CCS) through the DCR (choke and or resistor) bias.. and if you have an adjustable B+ (easy) you're off to the races.
Its the input tube, variation in DC condition spec which affects the 2 stage DC system. Easy to move that to the plate circuit of the output tube.
Quite the contrary, I'm enjoying following your journey and it inspires me to do some tinkering and listening myself. Beats a 'single and best' approach imo👍
Member
Joined 2009
Paid Member