Junk box low watt amp build

[dougietamson]

After going off-topic on another thread (sorry Tim) I thought I'd start a new one to keep on the straight and narrow

My plan is to build a low watt amp with middle son who is returning form Uni for the xmas/new year holidays.

He'll use it in his student flat share for a guitar so with that in mind if I was his one of neighbours I'd appreciate nothing too loud.

This will be a bits n bobs build to use up some of my "junk", this includes small power transformers (tho I may end up showing him how to wind our own).

A while back I bought a nice home made valve storage box on an auction site with a bid of £2. The box holds 60 B9A valves, also included was a bunch of old used low emission TV service valves with 5x PY800 1x ECL80 6x PCL83 5x EF91 2x EB91 10x PCF80 3x 6F12 4x PL83 and 4x EF184.

I tested the PCL83s in my Pye Black Box, all ok and I used the ELC80 successfully in a transformer-less spring reverb driver.

I'll assume that these valves were on the way out for TV service but may be ok for audio service.

I enquired on this forum if anyone knew if I could use the EF184/EF80 valves in a push pull audio amp with the goal of 1 watt or less, forum user "Valvepower" suggested the following as a good place to start:

http://www.r-type.org/articles/art-149.htm

The main issue might be matching the push pull output section to an output transformer. I have a working OT designed for the ECLL800 valve (triode phase splitter and 2xEL95 in one bottle) that I could use as a starting point, it has a 50:1 ratio for each side, trial and error might get us there.

I'll do a lot of prep work before he gets home, I've already started the project on my dev chassis. The EF184 valves are all ok giving 200x amplification and 80x triode strapped, I have a couple of old 1956 vintage EF80 valves to try out and I could throw in some used EF86 valves into the project too, I think I have 6 of these gathering dust.

All help/ideas welcome.

Photo attached of one of the EF184 valves in pentode mode with a 0.2Vpp sine wave.
I've installed 4 B9A sockets, 2 of them are the rubber mounted ones. A small power transformer giving a rectified/filtered 270vdc and 1.5A 6v filament supply.

We'll build a small cab too from scrap wood.

Doug

[G6Tanuki]

I'd suggest using the PL83s for the push-pull stage; they're a bit more robust for power-purposes than the EF80/EF183-types which are designed as small-signal RF amps.

Looking at them, I reckon a couple of PL83 in push-pull should give you 5 watts easily.

Only problem would be getting enough 'oomph' to power the heaters using your existing transformer.

[dougietamson]

Originally Posted by G6Tanuki

I'd suggest using the PL83s for the push-pull stage; they're a bit more robust for power-purposes than the EF80/EF183-types which are designed as small-signal RF amps.

Looking at them, I reckon a couple of PL83 in push-pull should give you 5 watts easily.

Only problem would be getting enough 'oomph' to power the heaters using your existing transformer.

If all goes well and he gets the amp building bug then a 5 watt amp will be our summer or xmas project next year, then we'll do a push pull 6V6/EL84 15/18 watt Fender Deluxe Reverb/Marshall clone or as he's a Beatles fan, an AC30 clone. These will include finger jointed tolex covered pine cabs and vintage AlNiCo speakers and nice grill cloth...

1 watt output should allow scope for some output valve overdrive distortion without waking the dead.

Today I'll work on the phase inverter, the Practical Wireless article is attached below.

Doug

[dougietamson]

PI stage built now, attached scope pics of normal and ch 2 inverted, wee bit of a level imbalance but it's only for guitar and often imparts a sonic "character". Maybe will cancel out depending on the level of the output valves...


Doug.

[dougietamson]

All connected up.

I made a couple of changes.

I removed the tone pot and just connected the 2nF capacitor that was being used for the tone control between the input and gnd at the input jack, I'll just use the guitar tone pot.

I moved the 1M volume control pot to after the 0.1uF coupling capacitor, pot tag 1, the wiper tag 2 goes to the grid of the triode strapped phase inverter and tag 3 to gnd. It's now more like a master volume control and I'll also use the vol pot on the guitar.

I put a trim pot to adjust the balance of the phase inverter to tweak levels/guitar amp distortion.

I hooked it up to one of the 12" speakers in my 120W quad EL34 amp cab.

How does it sound... damn good! and surprisingly loud, I cranked it up to 3/4 volume and the wife was down from the upstairs home office to complain. I'll wait till she goes to zumba class then let it rip at full volume.

I also tried the EF80 valves in the output stage, not as much volume but sounded just as good.

I'll try to make a decent recording using my Sure SM57 mic and Focusrite firewire interface hooked up to my macbook pro.

Photo attached of the dev build.

Doug.

[dougietamson]

Wee festive tune played on my junk box amp...

Not sure if this will work, here is my 1st attempt at a recording of the junk box amp, attached as a zip of and mp3.

Straight into the amp -> sure mic -> Focusrite FW Audio -> bit of reverb added in Logic pro.

Let me know if it works.

Doug.

[Herald1360]

Certainly works! Sounds clear with no obvious nasties going on.

Just to be clear, is it the EF91 circuit as per PW or a variation on that theme with other bottles?

[dougietamson]

Originally Posted by Herald1360

Certainly works! Sounds clear with no obvious nasties going on.

Just to be clear, is it the EF91 circuit as per PW or a variation on that theme with other bottles?

The amp as-built so far, is based on the PW article sticking quite close to the components values where possible.
I don't have any EF91 valves so I'm trying to shoehorn my EF184 and EF80 valves in, there may be other signal pentodes that work also.

The only issue I foresaw was the push-pull output section, the pre-amp and triode strapped PI should work on most signal pentodes. I also tried the preamp pentode in triode strapped mode, works but with less gain, may suit a line level input for use in a small audio amp.
1 watt power out is plenty, for the OT I went with the highest ratio one I had to hand (borrowed from an ECLL800 based amp).

Now it's making sound I'll wind a copy of the OT, adding extra taps to get something like an RS universal OT.

Here is an 'as-built' schematic I've been working on this morning.

Please feel free to find any typos etc.

Doug.

[Herald1360]

There's a bit of a rash of unwanted spots where suppressor grid-cathode lines cross over control grid lines!


V3 seems to have escaped though.

[buggies]

Originally Posted by dougietamson

Please feel free to find any typos etc.
Doug.

Well you did ask...
Too many dots in the circuit.
First EF184 anode load is shorted and both EF184 g1/cathode shorts.

[paulsherwin]

Once you get the circuit sorted out, it should work well enough at the modest volume levels you want. The EF80 in particular was a very tough valve and was used commercially in all sorts of unlikely applications, including class A audio output stages in tellies.

I've come across homebrew designs using several ECC82 valves paralleled in push pull, though these aren't the cheap throwaway TV valves they once were.

[dougietamson]

Thanks guys, amazing what a fresh pair of eyes can find!

I'm on v2 now, I'll also add in some grid stoppers to V1 and V2, I already had a ferrite bead on V1 g1.

Today I was playing through it with V1 triode strapped, I like the sound, may make it switchable...

Here's a sound clip of triode mode

Doug.

[dougietamson]

I've modified the schematics and added a triode strapped mode one too.

Doug

[Robert Gribnau]

I advise you to add a grid leak resistor (1M) at the input. This will prevent the control grid from floating under all circumstances.

C1 is unusual for a guitar amplifier but I didn't calculate what high roll-off it creates.

Edit: Can you post the schematic of your spring reverb driver with ECL80? I want to build something with the ECL80 and a spring reverb driver looks like a nice project.

[Radio Wrangler]

You've got five EF91s and a push-pull output transformer.

Hmm

That transformer would tend to eliminate thoughts of a single-ended amp, and besides the distortions you'd get out of a small push pull amp would be more like those of a big push-pull brute, just scaled down in power.

How about doing push-pull with a parallel pair of EF91 on each side?

Do some switching so that you can have triode or pentode mode

Do some switching so that the parallelled pairs can be cut down to singletons for lower power use. This way the outputpower is reduced, and equally so is the onset of distortion.

EF91s are still as common as muck and cheap. The snobby audio crowd haven't discovered them, or they decided they are non-U.

You could have a little amp which they could play about with.

David

[dougietamson]

Originally Posted by Robert Gribnau

I advise you to add a grid leak resistor (1M) at the input. This will prevent the control grid from floating under all circumstances.

C1 is unusual for a guitar amplifier but I didn't calculate what high roll-off it creates.

Edit: Can you post the schematic of your spring reverb driver with ECL80? I want to build something with the ECL80 and a spring reverb driver looks like a nice project.

Hi Robert, thanks for the input, now that the amp is making sound it's time to optimise it, your suggestions are spot on. The grid leak is a good idea, I'll put it across the 1/4" jack, the value of C1 can be solved by removing it.

I've just had a look at my ECC83/6V6 5 Watt amp with added valve driven reverb, my mistake, it has an ECL82 not ECL80.
For the reverb in an existing amp like my 5 watter, a single bottle transformer-less reverb driver works as follows:
Instrument -> high µ triode V1a -> signal tap -> dwell(level) pot -> V2 pentode -> spring -> V2 triode -> mix -> high µ triode V1b -> V3 output pentode.
For a stand alone reverb unit you'll always need 2 valves unless you have an external buffered effects send/return.
The ECL80 has been used in audio amps but it has shared cathode between the triode and pentode. At minimum we could use its pentode side only and add a dual triode for a 2 bottle stand alone unit. We'll still need an extra bottle (or solid state device) for the recovery side if we use the built-in triode to drive the pentode. So a 1 bottle 1 op-amp device is possible still getting the "tube drive sound"

Attached is a pic of my 5Watt plus single tube reverb, the extra B9A is wired to the octal socket of the 6V6 to take an EL84 instead (remove the 6V6 1st!!!!) Cab/chassis is a repurposed Marshall solid state amp.

Doug

[dougietamson]

Originally Posted by Radio Wrangler

You've got five EF91s and a push-pull output transformer.

Hmm

That transformer would tend to eliminate thoughts of a single-ended amp, and besides the distortions you'd get out of a small push pull amp would be more like those of a big push-pull brute, just scaled down in power.

How about doing push-pull with a parallel pair of EF91 on each side?

Do some switching so that you can have triode or pentode mode

Do some switching so that the parallelled pairs can be cut down to singletons for lower power use. This way the outputpower is reduced, and equally so is the onset of distortion.

EF91s are still as common as muck and cheap. The snobby audio crowd haven't discovered them, or they decided they are non-U.

You could have a little amp which they could play about with.

David

Hi David, how's it going just across the water in Fife?

I have 5xEF91 but only 3 B7G sockets.
One of the B7G sockets is pencilled in for a small SE amp/1950's bakelite radio cab with a nice Goodmans speaker/£18 ASDA DAB-Bluetooth contraption to go in the kitchen to replace my old 1980's Sony FM/AM/Cassette, FM (BBC Radio 2) reception in my part of Edinburgh is atrocious.
I have some EL95's and suitable OT waiting or could go down the push-pull route with the EF91's, wind an OT for them so just need another socket...might be a bit more interesting.

The OT I've been testing in the Frankenstein's monster mash-up p-p EF80/EF184 amp build (at the development stage as of now) will soon be returned to it's donor (a stereo ECLL800 based 10+10Watt amp/Tannoy bookshelf speakers).

I plan to wind a sort of universal PP OT (nothing too fancy, just some extra secondaries to reflect various loads through a CT primary).

The paralleled PP output stage had crossed my mind, for this amp, I want to keep the noise my son makes through it to <1Watt.

As a guitar player I must admit to liking the sound the poor wee tubes are making when pushed to 3/4 full volume, nice warm output tube distortion at a level that doesn't cause hearing loss.

One thing I have noticed when running the EF184 valves in the pp output stage is 50Hz hum which I don't get when using the EF80's so a bit of work to do there. I don't see any 50/100Hz modulation on the 1KHz sine wave out of the preamp or PI stages. I did see a bit when I built built the preamp stage until I added a couple of 100R resistors to ground to create an artificial CT for the filaments.

I'll knock-up some dummy loads (4, 6, 8, 12, 16 ohms) and probe the output and check the hum level and also estimate the actual power output. Will also find out how much current the valves are drawing and note all the voltages at the usual places and maybe revisit the PSU.

Doug

[Radio Wrangler]

It sounds like you're having fun. There are so many ways you can do this, so leaving it open for some future mucking around would seem sensible.

David

[dougietamson]

The amp I borrowed the OT from was designed for EL95's in push-pull, Philips data sheet says Raa as 10KOhms in Class AB, the CT to anode coil has a ratio of 50:1, anode to anode 100:1, so for a 4 Ohm speaker does that reflect back a 10K load?

Anyway, when using 2 EF80's in the output stage of the new amp build, a 0.2Vpp sine wave input, Vpp readings taken from the oscilloscope screen and converted to power in W for some dummy loads:

Dummy load(ohms) : Vpp. : Power(Watts)
4 : 3 : 0.28
8 : 5 : 0.39
10 : 7 : 0.61
16 : 10 : 0.78
20 : 14 : 1.06
26 : 14 : 0.94

Looks like a 20 ohm load is the sweet spot for 1Watt output though there was more visible distortion each time the load was increased, still, a bonus for a guitar amp where output stage distortion at low volume is a nice to have, I'll try a 16 Ohm speaker load.

For the hum when the EF184 valves are used in the pp output stage, I was getting a constant 0.1Vpp at 10 div at 1ms/div sine wave 100Hz by my mental arithmetic).
The preamp signal is clean, no measurable hum.
I found a potential ground loop so I made sure all grounds went to a single point and went round all solder joint to get them shiny, down to 2mVpp.

Doug

[G6Tanuki]

It's always good to experiment - take the theoretical equations/calculations as a reference-point but don't ever think it's a sin to deviate from them!

Valve power-amplifier matching is always a bit of a fudge: what gives the highest fidelity and what gives the best power-transfer-into-the-load are not always the same.

Most guitarists consider the whole guitar->pickup->amp->speaker path as part of their 'instrument' so seeking perfectly-linear, distortion-free amplification takes a back-row if they can coerce/overdrive an amp/speaker combo in order to get the particular sound they want.

I'm sure Eric Clapton, "Snowy" White, Mike Oldfield or Carlos Santana didn't get hung-up about issues of impedance-matching in their amplifiers.