Electronics Service & Repairs in Portland, Oregon
Here’s a look at a design project I’m working on. This is a driver circuit for a push-pull amplifier designed around the 6KV8 tube. The 6KV8 tube is a high-gain triode with good linearity and high transconductance coupled with a frame-grid pentode also having a high transconductance. Tubes like this were designed for television applications, not audio. Here the triode sections of the two 6KV8s are connected as an active-loaded gain stage, which then feeds the two pentode sections acting together as a phase inverter and driver section. The pentodes use IXYS programmable current sources to help linearize their operation. The circuit includes local power supply filtering and the negative bias supply. There are two options for feedback loops. One loop goes to the cathode of the lower pentode, and makes a standard feedback loop. The other is plate-to-plate loops from the output stage via connections LF1 and LF2. These are both subject to some experimentation when the prototype PCBs arrive and are stuffed.
This is the art of the PCB itself. This is where my project is different from what you’ll usually see in PCBs for vacuum tube audio. Other than the tube sockets, all components are surface mount. This is the more modern assembly method for electronics. The possible disadvantages are how the PCB will handle waste heat, and the limits on component selection. You can see in red the large copper area used as a heat sink for the current source components, which if effective will eliminate the use of a standard aluminum heat sink.
More to come when the prototype is ready to test!
In the traditional “Darling” style, it gets .75 watts per channel! The 6C8G is a really nice tube, unfortunately it gets overlooked.
Here it is, finally! Oh, and I hope you like green! This is a complete rebuild of my very first amplifier. Using everything I’ve learned since that time, I began this project with the goal of creating an artistic and functional version of that design. The tubes, output transformers, and a few internal parts remain from the original build. A new power transformer was used, as well as the new chassis, sockets, and other internal parts. Being mostly the same circuit and parts, it sounds exactly the same.
The schematic shows what few changes were made:
My next project has been cut, painted, and mechanically assembled. This is going to be a small single-ended amplifier with speaker and headphone output similar to the 1626 “Darling” amplifier below. The 89Y output tube is a military grade “type 89 – triple-grid amplifier” with a micanol base and graphite coated glass. Type 89 is a precursor to Type 41, 6K6/G/GT, 7B5, and 6AR5. It has a cap for control-grid connection and separate screen-grid connection pin at the base. Here it will be connected as a triode and driven by the 6C8G. The 6C8G is a tube I haven’t used before which has good linear curves and looks good in the amplifier. It contains two triodes, one of which has a grid connection at the top cap. If both stages prove linear enough and the output impedance isn’t too high, no loop feedback is planned. The schematic I’ve worked up for it is still on paper and not installed underneath the chassis yet. The output transformers aren’t cheapies this time! These are Transcendar 8K:8Ω from Gerry at Transcendar Audio Transformers.
More to come!
This amplifier was built on-the-cheap. I used salvaged transformers purchased well below online prices and 807 tubes which are still reasonably priced. 7 watts per channel. I won an impromptu local “shootout” with this. Recently added speaker/headphone switch and headphone jack. Now it’s personal!
How it was done: Single-ended pentode mode, with primary-side feedback, and fixed bias operation.
A look at the construction:
Because I could. This piece of wood uses 18FY6 and 17C5 or 19FX5 tubes to achieve an astounding 1.5 watts downhill with a tailwind. How to do this without, say, wasting money on it? That’s today’s contest: Build the best amp for under $20. I think the most expensive part was the pine board from Michael’s. But good pine improves the sound so much… Ok, back to reality. The power supply is two big 17V transformers back to back. The filaments of the tubes are connected to the 17VAC windings, and the 120VAC on the reversed second transformer is rectified and filtered for the 165VDC plate supply. Everything else was precision calculated on a napkin, except the coupling capacitors. They were randomly chosen and math on the napkin made sure they were just good enough to work. The input tubes are in those shielded sockets which are soldered to the ends of the ground buss “T”.
Does is work? Yeah, it actually isn’t bad. No hum or noise, but obviously not loud. Play it gently! Here’s the art:
Yes, the 150uF and 33uF are just connected in parallel. 33uF was not enough, so in the spirit of the design I just pasted that 150uF on there, but placed it closer to the stages it was filtering for. That was the last of the audible hum. Output transformers are some $5 items from AES. Please, DON’T BUILD THIS unless you can do it FOR $20. It just wouldn’t be right.
Here’s a long time workhorse in my setup. Four line-level inputs are on the selector switch, and active/passive mode is selected by the toggle switch in the center. Can easily be rewired to accept 7N7 or 7AF7, so that more types become usable. Collecting tubes is more fun when you can use some!
The Schematic. This preamp has its two 14N7 or 14AF7 tubes connected as CCS (Constant Current Source) loaded single gain stages. The idea here is to be able to use one stage of gain without feedback. Using the CCS load makes the stage linear enough that it does not require feedback. This does nothing for hum & noise of course, so the power supply is critical and thus somewhat complex. To read more about CCS loading, check out section 11-7 of Vacuum Tube Amplifiers by Valley & Wallman:
Built for my Mother-In-Law as a Christmas gift! This little amplifier has a headphone jack and volume control up front, speaker terminals in the back, and a power switch with blue LEDs. The color, a light textured metallic blue, is a new color for me and with that blue power switch just looks awesome! She was so impressed with my wife’s darling amplifier (below), that this one was intentionally modeled after it. If I can find a matching pair of 6K6G tubes (you know, the big bottles), those would look really nice in this!
The schematic. This is a very similar design to the 6Y6SE I’ve already posted, which was also a made-to-fit-the-parts design. The 12K:8 output transformer has a primary tap but isn’t a distributed load type, just multiple impedances. Connecting the 6K6 this way was just like triode connection, and no loop feedback was needed. It has limited power output but is very nice through headphones, with no hum or noise.
This is a design I’ve built five times! I don’t own one myself, as I had other desires. This particular one I built for my father, and he wanted it set up so any rectifier tube, including a solid-state plugin, could be used. Again we have the standard 17″ x 10″ x 3″ hammertone-painted steel chassis, this time with Hammond 1645 output transformers and a surplus power transformer. My father insisted on painting the transformers I would have kept as black. The solid-state rectifier is the painted cylinder in the middle. The completed unit puts out 15 watts per channel.
This, unlike the single-ended 6B4G amp, is not a simple affair. This is a three stage circuit, using one section of the input 6SL7 as a paraphase phase inverter. The second stage is the driver, using the 6SN7, to achieve the high voltage swing the 6B4Gs require. The output stage is very straightforward, except for the connection of the cathode-bias circuitry. In order to set up each channel’s output tubes, two 6.3 volt center-tapped windings are required. The three stages are enclosed in a negative feedback loop, without any step circuits to prevent oscillation! This is possible because the 6B4G has a high input capacitance, acting as it’s own step circuit or filter. The circuit can still be unstable unless a sufficiently wideband output transformer is used. This design is a close copy of a design in the Acrosound catalog, and in this case, the Hammond transformers have a somewhat lower bandwidth. This made it necessary to carefully tune the feedback loop for stability. The end result is a solid performer with decent power, sounding unlike push-pull pentode designs. It does sound louder than the watts indicate. Anyone building this circuit should spare no expense on the output transformers.