Single-Ended Stereo Console Rebuild
This amplifier began life, sometime in the 50's, as the receiver inside an RCA console. A console stereo was an all-in-one unit somthing like a boombox, except that it was the size of a chest of drawers and styled to look like one. Panels in the nondescript piece of furniture would slide back to reveal the tuner controls, and the record player. The speakers were cleverly conceiled behind wooden latticework. The electronics were buried deep inside, and were never meant to be seen except to change tubes. It often happens that people want to keep the furniture and get rid of the electronics inside, so you can pick up the amplifier part for not very much money.
The Amplifier Circuit
RCA employs here a low-power, single-ended circuit using 6V6GT output tubes. The 6V6GT is a beam tetrode, and good for 5.5 watts single-ended. There are two 6CG7 dual triodes, one for the input, the other as the driver. The rectifier is the popular 6CA4.
That's the revised circuit, current version. There are some oddities, and a few changes :
- The power supply was originally more complex because it had to have rails for the tuner section as well as the phono preamp. The 6CA4 is rated for 50 µF, so the 66 µF now used on the first stage is pushing it. It was originally 35 µF. I tried 33 µF but the sound was too thin. Since the current is 30% lower than the rated maximum, 20% more capacitance is probably ok... The transformer unfortunately does run rather hot as a result of the high conduction angle. (Mental note: next amp will use a pi-filter...) Elsewhere filter capacitance values are comparable.
- Cathode bypass resistors were originally 1 uF on the input 6CG7 and 20 uF on the 6V6s. This sets the bass cutoff frequency at 150 Hz. This would have been done either to filter off turntable rumble, or curb cabinet resonances, or simply because the rest of the chain produced no useable output at low frequencies, so there was no reason to invest in larger caps. In any event I have found that the output transformers are good to much lower frequencies, and I presently use 50 µF on the front and 220 µF on the back with good results.
- Where the circuit shows a 100 kohm volume control, there was originally a tone control network. No loss! A balance control was between the input and driver tubes, on the right channel, replacing the 220 and 270 kohm resistors there now. The idea was to adjust the balance by adjusting the volume of one channel holding the other fixed. Gak! Bye-bye! I get a bit more gain by replacing the two resistor voltage divider with a single 470 kohm grid resistor, and also improve bandwidth, so this is how I've been running it recently.
- I kept the shared bias circuitry which was put on the input and output tubes by RCA for no other reason than economy. The advantage is smaller parts count and garenteed equal bias settings between both channels. In theory channel separation suffers, but in one of those quirks of deep electronics, it turns out that while common (mono) signal is attenuated by degenerative feedback, stereo signal is actually injected into the opposite channel in inverse phase at the same time as it is only slightly attenuated in the original chanel. This actually enhances stereo bass! Whatever. By setting the bypass caps properly, one can tune the bass to arrive at the appropriate balance between rhythmic control and outright extension.
- The transformers were not identical in that the wire colour scheme was different. Originally one was wired in inverse phase. This was done so that the 'third channel' speaker could be wired accross the positive terminal of both transformers. Think of it like a primitive center channel, which would only output the mono signal. I rewired them so that they appeared to be in phase, and of course connected them for true stereo.
- The output transformers don't look like much, but since they have to pass less than 45 mA of current through the primary, there is no reason they have to be especially large. I have measured their 3 dB down point at over 50 kHz. I haven't measured them since I've changed the bypass capacitors, but the amp puts out significant signal at 50 Hz, about the point where my speakers roll off.
- One unusual circuit element is the series resistor and capacitor in parallel with the output tube. It has been suggested it was included to curb high frequency resonances of the output transformer.
- Lastly, the feedback network is wholly resistive, and excludes the input stage. Note this particular arrangement keeps a resistor across the speaker terminals. This prevents the amp from being damaged in the event that it is run without speakers attached, and also keeps the part count low. It might even be a really good arrangement, with no capacitor to add distortion at this critical point.
Regarding the Rebuild
- Once satisfied that everything was working well, and properly, I got a nice 16 gauge stainless steel chassis made up by a local steel fabrication shop. It was a lot of work to finish, but I'm glad I made the effort, and I'd do it again. Looks great. For the rebuild, I kept only the transformers. All passive components and hardware was new.
- For this, my first project, I learnt many valuable lessons:
- 16 ga. stainless is hard! You need a serious punch at your disposal, and any further work is best done by people with professional tools.
- The punch size needed for an 8 pin socket, 1 1/16", is not a common one!
- Point to point layout is an art, and the men at RCA could do it very well. Lean all you can from the original and draw up the schematic!!! before taking it apart.
- As a corallary to that, assume that everything in the circuit has a vital function until you know exactly what that function is.
- Tube rectifiers have a rated capacitance. Avoid the temptation to stick 100's of uF on the first stage of the filter. Hum is not an issue with this amp at the listed values. Photoflash caps and tube rectifiers don't mix.
- I didn't use any premium parts initially, though the RCA jacks and binding posts are good quality gold plate. Resistors are metal oxide, except the cathode resistor on the output tube, which is wire-wound. Capacitors are electrolytic in the power supply and cathode bypass, Solen polypropylenes for the interstage coupling. Tube sockets are ceramic but not gold plated. Internal wiring is 24AWG solid core copper, FEP insulation. The volume control is an ALPS carbon.
Then, the Tweaks
On the subject of tubes, the RCA clear top is the best sounding 6CG7, but even within a set of these, individual tubes sound different from each other. I haven't messed about with the recifiers much, though the Mullard-branded-something-else 6CA4 is the best, followed by the 6 blade RCA 6CA4. There is a cheap 4 blade version which sounds really poor. On the subject of the 6V6, I like the RCA grey glass 6V6GT best of the ones I have tried. The Marconi 5871 has lots of power but is a little rolled off, though the Marconi 6V6GT is identical to the RCA. The Sylvania 6V6GT is a B league substitute, I found it harsh and cold. One of the nice things about a 6V6 based amp is the variety of cheap output tubes kicking around, though the better ones are highly prized by musicians for their classic guitar amps.
I would upgrade the internal wiring, except that the point to point job is so efficient that there is only a few centimeters per channel that is actual wire and not component lead.
While changing the wire-wound cathode resistor back to the original carbon comp didn't change anything, messing about with the values of the cathode bypass capacitors has a large effect on the sound. I presently had a Rubycon Black Gate 220uF/25V FX type on the output cathode, and a regular Black Gate 50uF/16V on the input. I found they were worth the expense. Coupling capacitors are Multicap PPMFX, and I have replaced the power supply input capacitance from two 33 µF electrolytics to one 33 µF electrolytic and one 24 µF Solen polypropylene. Improvements here were smaller but still worthwhile.