Greetings friends,
I've been playing with this Angela Super Single-Ended project amp for a while now. I believe the OG article on the circuit was posted in the Angela Instruments Catalog in 1995 and sadly the website where I found the schematic and description is no longer online. So I'm turning to you. Here's the schematic:
Now of course I didn't build exactly that: I only put in one rectifier tube and changed the reservoir cap from 40uF to 16uF. And I used a PT I had on hand, Hammond 276X, with an output of 640v vs the 660v of the spec'd PT. So the b+ is lower than what you'd find in the OG circuit. Here's the schematic as built, with voltages. According to the RobRob calculator I'm running the power tubes at 100%
And here's a gutshot
Now the amp is stable and there's a small hum that's present with the Vol knob all the way off, and I get to maybe 10% on the volume before I get horrible squealing/squelching. I've switched the leads from the OPT to the output jack and it got worse, so I'm pretty sure the OPT is in phase correctly. This problem doesn't happen when I use the variac to run the amp at 90% and it doesn't happen with the light bulb current limiter, either. So is my voltage just too high? Should I put resistors btw the PT and the rectifier to drop my b+ ? any other strategies?
thanks!
I've been playing with this Angela Super Single-Ended project amp for a while now. I believe the OG article on the circuit was posted in the Angela Instruments Catalog in 1995 and sadly the website where I found the schematic and description is no longer online. So I'm turning to you. Here's the schematic:
Now of course I didn't build exactly that: I only put in one rectifier tube and changed the reservoir cap from 40uF to 16uF. And I used a PT I had on hand, Hammond 276X, with an output of 640v vs the 660v of the spec'd PT. So the b+ is lower than what you'd find in the OG circuit. Here's the schematic as built, with voltages. According to the RobRob calculator I'm running the power tubes at 100%
And here's a gutshot
Now the amp is stable and there's a small hum that's present with the Vol knob all the way off, and I get to maybe 10% on the volume before I get horrible squealing/squelching. I've switched the leads from the OPT to the output jack and it got worse, so I'm pretty sure the OPT is in phase correctly. This problem doesn't happen when I use the variac to run the amp at 90% and it doesn't happen with the light bulb current limiter, either. So is my voltage just too high? Should I put resistors btw the PT and the rectifier to drop my b+ ? any other strategies?
thanks!
yes, it squeals with no input signal, or with a cable plugged in. I have a 1/4 to 1/8 TRS adapter and plug in my phone. when playing music I get clean sound til the knob is at 1, than I get onset of distortion and the squeal just above that.
The marked-up schematic is as accurate as I know to make it.
unfortunately I am not 'scope savvy, but I've been hoping to learn.
unfortunately I am not 'scope savvy, but I've been hoping to learn.
Your NFB mod, as shown in the schematics, does next to nothing. So I'm scratching my head why you observed different behaviour when swapping the OT leads.
Best regards!
Best regards!
yes it did not cause any change
swapping white and black leads on the output jack or putting the 22k NFB resistor in there, got the squeal all the same
swapping white and black leads on the output jack or putting the 22k NFB resistor in there, got the squeal all the same
It depends where exactly the wires were reversed ..
But the source of disturbance is somewhere else.
But the source of disturbance is somewhere else.
Ok, that's quite the contrary of what you wrote in #1:
Best regards!
Anyway, this isn't your issue's cause. I suspect parasitic capacitive coupling between the input and the power stage.
Best regards!
there's a passage in the Morgan Jones book on parasitic oscillation and he seems to think grid stoppers are the solution...is this what you mean? please forgive my ignorance on the subject. As shown, the power tubes have 1.5k grid stoppers.
I'm just now remembering that the input jacks are configured as in the 5F1, allow me a few minutes to correct the schematic
I'm just now remembering that the input jacks are configured as in the 5F1, allow me a few minutes to correct the schematic
corrected schematic. The 68K resistors are soldered directly to the input jacks and a length of shielded cable runs to the socket. Should there be grid stoppers on the grid pins of the 12AX7?
Yes, for DIY constructions grid stoppers should be used everywhere. Commercial products have folk that come along after the design and remove parts to save a few pennies, but for DIY there's no reason to skimp. But my first three points of attack would be G2 stoppers on the output valves (100R or so is fine), a Zobel across the OPT secondary (10R and 0u1F in series) and some decoupling in the B+ feed between the first two stages. Normally two stages with a common supply impedance is negative feedback, but that's only solidly true for flat amplifiers. Another RC filter section can't hurt (1K Ohm and some 10's of uF).
The 6.3VAC filament supply needs to be grounded or referenced to some small (10s of VDC) voltage. It's usual to make a center tap with two 100R resistors in series across the 6.3VAC and to connect the center point to the output valves' cathodes.
The remaining issue with parasitics is the grounding scheme. They're all awful, but some are less awful than others. What have you chosen to implement?
All good fortune,
Chris
edit: your feedback resistor is shorted out by the cathode bypass capacitor.
The 6.3VAC filament supply needs to be grounded or referenced to some small (10s of VDC) voltage. It's usual to make a center tap with two 100R resistors in series across the 6.3VAC and to connect the center point to the output valves' cathodes.
The remaining issue with parasitics is the grounding scheme. They're all awful, but some are less awful than others. What have you chosen to implement?
All good fortune,
Chris
edit: your feedback resistor is shorted out by the cathode bypass capacitor.
Last edited:
Possibly the transformer windings need to be reversed on the output. Otherwise you have no feedback but feedforward... just saying.
Thank you, Chris, for the comprehensive reply. I'll try to reply to each point in turn, with photos where helpful.
I had no idea that screen grids benefitted from grid stoppers, but on my earlier single-tube builds, I guess I was effectively using one by using the resistor as the connector between node and pin. Right now there's solid copper wire connecting the pins of the paralleled output tubes, and a 1.2k resistor connects the copper wire to the B node:
Shouldn't be too much trouble to put 100R resistors in there. I am concerned the screen voltage is too high, even the 1.2k resistor isn't doing enough.
you can also see the 100R resistor pair that references the heaters to ground on the left power socket. I had the center tap connected to the primary ground node but read somewhere that was a good way to fry the PT in case of tube failure, so I snipped the CT and installed the 100R pair.
A Zobel network? that's new to me. what would be the suitable voltage rating for that 0.1uF cap? assuming Film type. I have a few on hand, no prob there. ok to mount at the output jack?
Are you suggesting a 4th RC stage in the power supply, creating a D node to power the first stage of the preamp tube? makes sense. I have more 16uF 475v on hand.
as far as grounding, before installing the turret board I linked the turrets that would be connected to Ground side of components, and link that to the reservoir cap. K bypass caps have flying leads to their switches, there's a few more flying leads for the (-) side of the output jack and input jacks, No Loops. Maybe I should pull the turret board and clean up the paths of those flying leads...
will
I had no idea that screen grids benefitted from grid stoppers, but on my earlier single-tube builds, I guess I was effectively using one by using the resistor as the connector between node and pin. Right now there's solid copper wire connecting the pins of the paralleled output tubes, and a 1.2k resistor connects the copper wire to the B node:
Shouldn't be too much trouble to put 100R resistors in there. I am concerned the screen voltage is too high, even the 1.2k resistor isn't doing enough.
you can also see the 100R resistor pair that references the heaters to ground on the left power socket. I had the center tap connected to the primary ground node but read somewhere that was a good way to fry the PT in case of tube failure, so I snipped the CT and installed the 100R pair.
A Zobel network? that's new to me. what would be the suitable voltage rating for that 0.1uF cap? assuming Film type. I have a few on hand, no prob there. ok to mount at the output jack?
Are you suggesting a 4th RC stage in the power supply, creating a D node to power the first stage of the preamp tube? makes sense. I have more 16uF 475v on hand.
as far as grounding, before installing the turret board I linked the turrets that would be connected to Ground side of components, and link that to the reservoir cap. K bypass caps have flying leads to their switches, there's a few more flying leads for the (-) side of the output jack and input jacks, No Loops. Maybe I should pull the turret board and clean up the paths of those flying leads...
will
I'm not sure if additional grid stoppers will cure the symptom. Squealing, as he describes it, is a phenomenon in the AF range, while grid stoppers prevent RF oscillations. There are grid stoppers yet present at the most relevant location, the power tube grids. And I've learned from experience that low gm small signal triodes, such as 12AX7's, don't benefit from them.
I suspect the issue's cause is in the build arrangement here.
Best regards!
I suspect the issue's cause is in the build arrangement here.
Best regards!
Did you try reversing the output transformer cables?
The output of the transformer is a low impedant source and can give trouble even with a big capacitance as a load.
The output of the transformer is a low impedant source and can give trouble even with a big capacitance as a load.