The Schematic Thread (Amps, DACs, PSUs...)

[100VoltTube]

This is a thread for people to post their (or others') schematics. The idea is to have a thread full of ideas for people to build, improve on, and/or draw inspiration from.

Some rules:

1) In a post with a schematic, please include a description of the operation of the circuit.
2) Technical discussion is more than welcome, but keep it respectful, and keep in mind different views about harmonic distortion, feedback, tubes vs solid state, etc...
3) Try to keep all circuit descriptions and discussion newbie-friendly. If you are using an odd component or topology that wouldn't be understood by all potential viewers, explain it, or post a link to a site that explains it.
4) If you're posting a half-baked design, a new topology, or something that's not fully fleshed out with all the necessary components, tell everyone.
5) Schematics aren't limited to headphone-related designs. Speaker amps, preamps, PSUs, DACs, radios, and so forth, are all welcome.

I look forward to seeing some cool stuff

 

[100VoltTube]

Here's an output transformer coupled amp that can put out about 10 watts into a transformer with a 16kΩ primary:

The output tube is a BDT (beam deflection tube) Kevin Gilmore gives a good explanation of their operation in this thread. It's directly coupled to a long tail pair (tubecad's article here). If you have a balanced source, the grid of the triode marked U3 could be used as the input for the other phase of signal. The zener diode is to provide DC offset for the BDT's cathode so that the deflectors (the little < shaped things) are at about the same DC value as the cathode. It could be substituted for a resistor. resistor R1 is used to bias the control grid, setting the current through the tube. For a grid voltage of -0.5v, it should be about 18 ohms. In this implementation of the tube, -0.5v on the grid is ideal. At 0v, the dissipation would exceed maximum (3W per plate), and at -1V, the output current swing would be decreased, decreasing output power. The plate voltage of the BDT should be 200V at idle. The 300v B+ is just an example. It would allow for 100v on the input triodes' plates at idle, which is enough for most tubes.

For the two input triodes, most medium/high mu triodes will work just fine. Pentodes would work too. Just make sure that the output of the long tail pair can swing 50v plate to plate.

If you have any questions or ideas for improvement, fire away.

I have a few more designs lined up, and I'd love to see some of other people's designs.

 

[DutchGFX]

This should be fun

Once I start designing a bit more (just finished learning intro to MOSFETS in school) I'll certainly post. I'll draw up some simple PSU stuff when I get a chance

 

[WaffleIron]

Seems like a great thread idea. Keeping my eyes peeled.

 

[DutchGFX]

Figured I'd throw something up here just to keep/get it rolling. I'll try to explain using my limited and often incorrect knowledge...

Here's a filament supply without using the LM317, cause we all know that's cheating! (jk of course, but designing with limitations is fun!)

It's pretty much a CCS fed Zener, and that becomes the reference for a 2 transistor regulator. The regulator has a pass transistor so you can use good old MJE340's for the actual regulator portion of the circuit. It's assuming I can find a 5.5V zener, but any zener around that would be OK, you would get slightly less/more than 6.4V

Here's how it works:

Input
Simple transformer/bridge rectifier with a fairly sizeable filter cap. At 16V, a 10,000uF cap isn't a problem at all.

Pre-Regulator/Reference Voltage
The 2 PNP transistors form a constant current source. VEB of Q4 is roughly constant. This sets a roughly constant voltage drop across R2. But we can do even better than this. Q5 is on because of R4, which, if we assume Q5 is off, then we get a contradiction, so Q5 must be on. If Q5 is on, then Q4 must be on due to the voltage drop in R2. If Q4 is too much on, meaning the voltage across R2 wants to be more than .65V, then the voltage across R4 would rise, and that would decrease the current in Q5. Conversely, if the voltage across R2 is too small, then the voltage across R4 decreases, and the current through Q5 increases. This how the CCS maintains a roughly constant current. The current is fed to a zener diode, which, along with the filter cap, generates a pretty stable reference voltage.

Regulator
The output voltage, as seen again by the BE junction, is the zener voltage plus VBE. If the output tries to rise, R5 causes the base voltage on the top transistor to fall, reducing the output voltage/current. The same counter-action occurs if the output sags, hence regulation.

Potential Improvements
The reference voltage could be coupled to another pass transistor, and then that voltage could become the power supply for the second stage, which could be turned into a constant current source. This would allow a potentiometer to set the output voltage. This could then be coupled to the load by, you guessed it, another pass transistor. I'll play around later, but for now, here's a spice graph. Looks pretty decent

It outputs 3.2A with less than .1A ripple, that's not bad...


Let me know what y'all think/if I messed something up!