6L6GC to CV428 conversion

[dougietamson]

A few months back I "MacGyvered" a pair of NOS CV428 5B/254M valves using the loctal base/top cap and an octal plug, just for testing.

Now that I'm happy with the performance in the amp it was time to ditch the bodge job converter for something more stable/safe.

The amp in question uses PCB octal sockets so isn't suitable for a simple base swap and rewire.

I've gone down the route of mounting the loctal sockets on standoffs and threading the wires from it through to the solder points on the pcb of the original octal sockets. The conversion is easily reversed.

I used heat proof wire salvaged from the HT supply from a microwave oven for the top cap leads. All the socket to socket wires are double insulated with heat shrink. The amp can take 4 x EL34 or 4 x 6L6GC, I'm testing with just the 2 installed (60W instead of 120W) and running double the load at the speaker (switch set to 4? but with an 8? speaker)

Photo attached.

Doug

[Radio Wrangler]

That sounds like half the load in terms of current taken.... double the resistance applied to the anodes. Definitions of words....

David

[paulsherwin]

Nice build anyway.

[dougietamson]

Originally Posted by Radio Wrangler

That sounds like half the load in terms of current taken.... double the resistance applied to the anodes. Definitions of words....

David

Yes, I was a bit unclear about the change I made of running a pair of output valves in push pull instead of the pair of 2 per side it was designed for.

I should have said I needed to double the reflected load from the OT secondary back to the pri of the OT (it's 2k?pri:4/8/16? taps) as the pair of 6L6GC would rather 'see' 4k?.

Though I may still be confused!

[Aub]

Very nice Job.

Aub

[dougietamson]

Here's the view from the PCB, pics attached with before and after wiring the loctal base.
I drilled an 8mm diameter hole in the PCB using the base index hole as a guide, the PCB is dual layer, all traces in plain view.

The wires were kept short.

Wiring I used:

H: Loctal pins 1 & 8 -> Octal pins 2 & 7
G1: Loctal pin 5 -> Octal pin 5
G2: Loctal pin 3 -> Octal pin 4
K: Loctal pin 7 -> Octal pin 8
A: Loctal TC -> Octal pin 3

[dougietamson]

I've added a circular black plastic strip between the local base and original octal base, pic with Chinese 'coke bottles' fitted for comparison.
You can buy 3 NOS Made in England CV428 valves for the cost of one of those coke bottle 6L6GC valves!

[dougietamson]

One thing I did note that doesn't seem correct is the heater voltage supply for the power valves and also shared with the PI (the preamp section has dedicated DC supply).

One issue when pulling 2 output valves from a set of 4 is the heater voltage increase due to lower current draw so I checked it.

I found that the heater voltage was low at 6.0v while running 2 x CV428 and 2 x 6L6GC and dropped to 5.8v when running 4 valves.

With the 4 output valves pulled it read 6.15v.

Today the mains reads 241v

I bought the amp with 4 x EL34 output valves fitted, the seller said he bought it new in 2010 with the EL34's fitted as stock, 6L6GC is a factory option and there is a switch at the bias adjust pot to select for the 2 option2.

EL34's would draw approx 50% more current so dropping the heater supply even lower.

Any issues with the lower voltage?

[Radio Wrangler]

The 5B/254M is a repackaged 807 in essence. A quite capable RF valve to start with. However, in the smaller package, the internal connections are shorter, pushing up its frequency capability a bit more. So you need to be careful about spurious RF oscillation when using them.

David

[dougietamson]

Originally Posted by Radio Wrangler

The 5B/254M is a repackaged 807 in essence. A quite capable RF valve to start with. However, in the smaller package, the internal connections are shorter, pushing up its frequency capability a bit more. So you need to be careful about spurious RF oscillation when using them.

David

Here's the schematic, 10K grid stoppers on each output valve.

[Richardgr]

This was my solution. £1.50 wafer loctal sockets from Billingtons, and £1 bakelite bases from China. Soldered longer pieces of solid core insulated wire to the socket, then bent them to match the octal pins, stripped the ends, and pushed the wafer onto the loctal base. It was the same diameter, so looks quite neat I think. A little bead of loctite finished the job.

I made a batch, but missed a trick because I could have used the same pinout for 7C5 and 5B/255M, but used an alternate pinout for 5B/255M instead (it has multiple cathode connections, I think it was).

Those look to be very nice loctal sockets, by the way! Where did you get hold of them?

[dougietamson]

Originally Posted by Richardgr

This was my solution. £1.50 wafer loctal sockets from Billingtons, and £1 bakelite bases from China. Soldered longer pieces of solid core insulated wire to the socket, then bent them to match the octal pins, stripped the ends, and pushed the wafer onto the loctal base. It was the same diameter, so looks quite neat I think. A little bead of loctite finished the job.

I made a batch, but missed a trick because I could have used the same pinout for 7C5 and 5B/255M, but used an alternate pinout for 5B/255M instead (it has multiple cathode connections, I think it was).

Those look to be very nice loctal sockets, by the way! Where did you get hold of them?

The bases are by McMurdo, chassis pulls from ebay.

I also fabri-cobbled a couple of adapters for pair of 7C5 ( = 6V6GT on loctal base).

I've started the 3rd adapter, this time with a stainless steel barrel and 4cm x 3mm screw.

Keep kicking myself for not buying a mini lathe when one came up locally! Would be nice to turn some dedicated standoffs.

[Richardgr]

Originally Posted by Radio Wrangler

... However, in the smaller package, the internal connections are shorter, pushing up its frequency capability a bit more. So you need to be careful about spurious RF oscillation when using them.

If the internal connections are shorter, shouldn't it be more immune to RF oscillation at normal audio frequencies?

[G6Tanuki]

Originally Posted by Richardgr

If the internal connections are shorter, shouldn't it be more immune to RF oscillation at normal audio frequencies?

The shorter leads mean lower losses at RF, meaning potentially increased gain at RF!

Your amp intended to amplify 30Hz to 30KHz will have ambitions to become a push-pull power-oscillator at 10MHz or similar, and unless you include control-grid screen-grid and anode-stoppers [either resistive or the likes of ferret beads or a few dozen turns of wire wound over the body of a resistor on the anode connectors] your amplifier can achieve its ambitions when you're not looking.

Just because it's stable in static conditions isn't a guarantee that it won't spring into oscillation at some point when the valves are being driven!

[dougietamson]

One of my CV428 had a top cap incident and went to air, here is a close up of the internals minus the glass, I could take some measurements and do a further disassembly with close ups if anyone is interested.

I'll scope the output and add some ferrite beads.

[Radio Wrangler]

Originally Posted by Richardgr

If the internal connections are shorter, shouldn't it be more immune to RF oscillation at normal audio frequencies?

The effects of capacitance to ground are felt more directly by the valve's electrodes and can be seen as RF decoupling. A small amount of inductance from the effective RF ground to the electrodes proper can be gyrated by feedback capacitance into a negative, real resistance component and it is this which powers any stray resonances into oscillating.

This sometimes gets used deliberately in some circuits "The electron-coupled oscillator" There is feedback in it, but it's hidden in stray coupling capacitances and the gain of the valve.

The shorter connections and the multiple cathode connections push up the frequency of potential oscillation and you need to apply stoppers good for that increased frequency.

These same things happen with transistors. The worst risk is an emitter follower after a filter. Lowpass filters likely end with a shunt C to ground. If whatever it drives looks like low-Z to ground at RF, then you're in for a fun time.

Because the feedback loop is almost all internal to the device, the loop can't be made any shorter and so these techniques can build oscillators right up to the highest frequency the device is capable of.

David

[dougietamson]

Originally Posted by G6Tanuki

The shorter leads mean lower losses at RF, meaning potentially increased gain at RF!

Your amp intended to amplify 30Hz to 30KHz will have ambitions to become a push-pull power-oscillator at 10MHz or similar, and unless you include control-grid screen-grid and anode-stoppers [either resistive or the likes of ferret beads or a few dozen turns of wire wound over the body of a resistor on the anode connectors] your amplifier can achieve its ambitions when you're not looking.

Just because it's stable in static conditions isn't a guarantee that it won't spring into oscillation at some point when the valves are being driven!

Are the CV428 valves still being used by enthusiasts in transmitters? They are currently low cost and plentyful but I'd rather not use them if they are still being used out there.

[Richardgr]

I think the consumers of all of the more esoteric NOS tubes will run out before long before the supply does. A lot of these loctal 807 equivalents were stashed for military purposes, so supply still comes along. When the 6.3V variants start to get a bit thin on the ground, then the 19V ones are three to be used too.

What would you think the annual supply to end users would be for CV428? 1000 max I would think. I am sure there are a few 10's of thousands yet to find a purpose in life.

[Ed_Dinning]

A thought for those of you thinking of using some of these trypes with strange bases and "odd" heater voltages . The heater power for most of these valves (V*A) is constant across the range, so a higher or lower voltage , transformed to the valve required takes very little additional heater power. This iss easiest achieved with a very small auto transformer that can easily be hidden under the chassis,

I've produced a few of these for various applications over the years

Ed

[G6Tanuki]

Originally Posted by dougietamson

Are the CV428 valves still being used by enthusiasts in transmitters? They are currently low cost and plentyful but I'd rather not use them if they are still being used out there.

Yes they're still being used by us ham-types, for HF transmitters/modulators/linear-amplifiers - but I wouldn't worry about 'using them up' - the world is unlikely to run out of them in our lifetime.

The Octal-based 6146 and its beefier descendant the 6146B [with 40 Watts anode dissipation] is somewhat more favoured by us RF types.