Saturday, 19 October 2013

More CBA mods

Found the treeware diagrams

The T'Watt VIP has the following tone stack, but bear in mind it has two extra stages of gain in the pre-amp to make up for the loss in the stack.

The 860pF cap shown at the input to the circuit can be left out. It was a throwback from an earlier tone circuit that didn't decouple the DC between the pre and power stages.. With this tone stack the decoupling between pre and power stages happens with the three caps in the circuit. This drawing was done from the actual amp layout a while ago. I've since physically removed that cap.

This tone stack is the one I've been happy with for over a year now in the current T'Watt. It does everything I want it to.

The original T'Watt had no tone stack at all, and didn't suffer because of it. If I wanted less trebles, I simply wound down the tone on the guitar. ;-)

However, I started experimenting.

With a cathode follower in the circuit to drive the tone stack...

...I found that this bass and treble circuit gave me the best results.

And in that circuit, the 1uF cap is there to decouple the DC between pre and power stages. You can fiddle with that value, just don't leave it out. I tried a version with a middle control, but it was too fussy. The controls were just too interactive. This one gave good results for the lowest component count.

I would strongly recommend downloading and playing with Duncan's TSC (tone stack calculator) to get some ideas and to "see" what the response curves look like when you tweak the values.

So, CBA mods...

Apologies. It's been a busy week in the office, and this project had to go on the back burner. The other reason is that the physical build of this one-off for Peter is now complete. I'm just waiting for him to call in (all the way from Australia) to collect it next Sunday (27 October 2013) so, in my head, the pressure is off.

However, my other awesome friend, Bob 1B King, has started down his own path to valve amp addiction. Bob's amp is already going to be a little different.

One of the mods Bob's looking at is purely cosmetic (but fun) and he's adding an EM80 "magic eye" power indicator. I have these on my WTF Evo X and the T'Watt VIP, so here's the circuit.

The EM80 is fed from one of the later B+ nodes to get the voltage down below 250 vdc. On Evo and T'Watt, I added an extra node just to isolate the EM80 from the pre-amp circuit a little more. Not needed, but I had the components and the physical space. The supply is fed direct to pin 9 and through a 470K resistor to pin 7.

Pin 2 goes to signal ground.

Pin 1 is the signal feed. The signal is "read" from one of the anodes of the output valve (either pin 1 or pin 6 of the 6N1P). The 0.1uF 400 volt cap blocks the DC from the output valve's anode. The 390K to 470K resistor and the 220K preset are used to adjust the deflection of the magic eye beam. I found that a 470K fixed resistor was just fine in my own amps, but you may want to tweak the values if you're getting too much or too little deflection for the preset to cope with. The pair of diodes just rectify the signal to get a DC level for the EM80 to work with. The 0.1uF cap marked "Hold" in the diagram determines the rise and hold time of the display. A bigger value gives a slower reaction speed. Feel free to tweak.

BE CAREFUL! There is a temptation to tweak the deflection preset with the amp powered up and under load. Remember the warnings about LETHAL voltages and currents. DON'T do this unless you are ABSOLUTELY sure about what you are doing! It's far better to be patient and take your time to power off the amp, tweak, power on and test than to electrocute yourself. I cannot be held responsible for you poking your finger/screwdriver/butter knife into a live DC circuit!

It's fun and gives a wonderful moving light display, but I wouldn't say you could calibrate it to any degree of accuracy for recording and what not.

The whole circuit can be built directly on to the EM80 valve base. It's fiddly, but it can be done. Just keep an eye on the orientation of the two 1N4148 diodes (that's where I usually mess it up first time round).

To come - the dreaded tone stack - and the need to add another valve!

Saturday, 12 October 2013

Where to get parts

Please bear in mind that this is solely down to me hunting around on various forums. I don't endorse (with one exception), nor do I have any connection with, any of these suppliers, I'm just posting some links. And I'll try to indicate the relevant country.

And this is NOT an exhaustive or definitive list by any stretch of the imagination. If you have a favourite supplier for valve amp parts, use the comment box to let me know and I'll add the link to this blog. :-)

Hoffman Amplifiers - US

Apex JR - US and Canada

Digikey - US (but has global delivery option)

Mouser - Global

Ampmaker - UK (but will ship globally)
Barry is the one exception to my "don't endorse" statement. I buy 99% of my parts from him!

That'll have to do for the moment. I've a living room carpet that isn't going to vacuum itself.

Thursday, 10 October 2013

Okay, last lap for the actual build. But there's more to come in this blog. Over the next few installments I want to take a brief look at some of theory behind what makes the CBA tick.

But for now, I'm going to post tonight's little diversion.

I want to give Peter something that is pretty much self-contained so it was always my plan to add a speaker cabinet to the package. And this would also be built into a cigar box.

So, we have a speaker from a car stereo system, simply because it's rated at 30 watts RMS and is small enough to fit in the box I've got to hand. And you can see the box and socket in the pic too.

Not only do I prefer my amps to be old school and use retro technology, I'm the same about my wristwatches. And this spring-powered beast didn't deserve to be magnetised while working on the speaker so it had to come off.

So we find a centre point by drawing across the diagonals.

Then measure the mounting diameter of the speaker. There is a cutting template in the box, but I always like to check anyway.

Out with the trusty old pair of compasses.

And then I hit it with a large hole saw. No pics of that sadly.

Speaker mounted and socket soldered into place.

Time for testing (with the lights off so you can see the glow).

And with the lights on so you can get a clearer view.

It sounds better than I expected, but it's more an exercise in completeness. The amp sounds stunning through a big cab, so I'm guessing it won't be too long after Peter gets back to Australia before he sources something more appropriate. Oh, and tonight's "ah bollocks" moment... I've fitted the speaker grille upside down. There's lettering that reads "Mutant X" on one of those curved sweeps and that kind of gives it away. And it was a sod of a job getting the nuts on the bolts too! Still, I've got until 27 October to correct that little error though.

Tuesday, 8 October 2013

And Bob has come through with the schematics! He's even included a parts list. All that's missing from the parts list are the transformers, the power sockets, valve sockets and valves. They're all things that are pretty freeform and depend on what you can source locally.

For builders in the UK, the transformers and valve bases can be bought from Ampmaker. I'll do a bit of research for other countries and post up some suggested sources here.

But before we go any further, here's the disclaimer!

Building and modifying amplifiers is DANGEROUS (like most fun things). Valve amplifiers invariably contain both very high voltages and high currents, capable of killing you. Never work on a live amplifier!

This blog is here to share information about a particular amplifier project. It is up to you to exercise caution and common sense at all times to avoid electrocuting yourself, and make your amplifiers safe to use. I am not responsible for your negligence. Building valve amplifiers is also more addictive than nicotine; you have been warned.

Now, here's a link to a PDF that has the schematics and suggested parts list.

And the pics below are what you get in the PDF. If you'd rather not download the PDF, clicking on any of the pics will open them full size.

Would you like to see my etchings?

I thought I ought to make a special, one-off nameplate for Peter's CBA. So the copper sulphate and power supply came out this evening.

It's difficult to photograph, but I'm really pleased with the way this has turned out!

I've glued it in place for the moment, but tomorrow I'll be hunting through my watchmaking bits to see whether I've got four screws small enough to just add the finishing touches.

I'm now even deeper in Bob's debt!

He's taken my hand-scribbled circuit drawings and he's started turning them into a very professional set of schematics for the CBA. They'll be making an appearance here soon.

Watch this space.

Sunday, 6 October 2013

Remember yesterday's "ah bollocks" moment with the dropper resistors fouling on the input socket?

While I was in that floaty, pre-sleep stage last night, I "saw" the solution. And when I woke up this morning, after the second coffee, I gave it a go.

Those big 5 watt 1k resistors I'd used were overkill. So I replaced them with these 2 watt 1k resistors and slightly altered the layout so that there's now loads of clearance round the input socket. Problem solved. On with the hum fixing.

The heaters in CBA are running AC. And AC heater wiring should be kept electrically balanced! Which is exactly what I hadn't done up to this point. For testing, I was more interested in just making sure the valves warmed up.

With a centre-tapped heater transformer, this balancing can be done by simply connecting the tap to ground somewhere. It doesn't usually matter where, and it's normal practice to connect it to the chassis at the closest point to the transformer.

The transformer I've used doesn't have a centre-tap though, so I had to create an artificial one. There are a couple of ways to do this. I could have just added a pair of 100 ohm resistors, one to each side of the winding, with their ends connected together then down to ground. It's worth remembering that this would add a little to the current draw from the transformer too. With a pair of 100 ohm resistors, this would be another 32mA. Still well within what the transformer can supply. The valves are drawing 900mA and the transformer is good for around 1000mA.

I chose to go the "humdinger" route though. That's just a small preset potentiometer soldered across the heater pins on the pre-amp valve with its wiper connected to the chassis ground.

The advantage of the humdinger is that it can be adjusted to dial out the hum by achieving the right electrical balance. Each of a pair of fixed resistors are never usually quite the same value, whereas by tweaking the preset it's possible to get pretty close to perfect balance.

On to the juggling act. I think I'll let the pictures do the talking. It was tight, but not impossible to get the whole thing to come together.

And it was at this point that I remembered I still had to add a speaker socket!

The next few pics are especially for Bob "1B" King. Bob's awesome! He sent me a fabulous shirt all the way from Florida, and in payment I started this blog.

Bob was asking about the RC filtering in the HT supply. Before I screwed the lid down on CBA, I connected my oscilloscope to the different nodes of the supply.

This is straight from the first smoothing capacitor after the bridge rectifier. The scale is 1 volt per centimetre, so we're seeing a ripple of around 3 volts peak to peak sitting on top of the 350 volt supply.

In the next picture, we're still on the same scale and the probe is on the first RC node, the B+1 in my original sketch. We're now looking at just over 1 volt peak to peak ripple.

Next, the B+2 node. Scale now is 0.2 volts per centimetre, and we're seeing probably around .05 volts (50mV) peak to peak.

And B+3 is on the 20mV per centimetre scale, so we seeing around 4mV peak to peak.

Apologies for the slightly blurred image quality. They were taken without flash to bring out the traces and I couldn't be bothered running downstairs for a tripod to grab these pics.

So... it's screwed together and fully working! And I'm delighted with the way it sounds just through a nondescript speaker sat in free air on the bench! It definitely has a growly, bluesy sound. And I think it looks pretty neat too!

Next job is to build a speaker cab to go with it.

Watch this space. :-)

Saturday, 5 October 2013

So what have we here? Yep. It's wired up and ready for testing!

Today's "ah bollocks" moment came when I tried to dry assemble the chassis to the box with all the parts in place. The input socket snags on the dropper resistors that I had to relocate! There are a couple of options. I can move the resistors to a completely separate board mounted in the bottom of the box, or I can move the input socket and fit an on/off switch in the hole it leaves. Or I can put the whole thing in a bigger box...

Anyhoo. First job was to make sure everything was electrically safe. I laid the amp out in a line on the bench. The output transformer is floating free at the moment. I didn't fancy sticking it in the coffee tin until I knew it was all working.

First power up is always done without the valves in place. And the first check I did was to make sure I'd got 6.3VAC across pins 4 and 5 of each of the valve bases. The next check was to make sure I'd got HT in all the places it should be...

Yep. That's 378 volts of DC. More than enough to seriously ruin your day if you touch it! That's why I keep harping on a bit about safety.

Once the valves are in place, that will drop to the 320 volts or so that I'm expecting to see on the output transformer. It will be a little lower on the pre-amp stages.

Power off. Then wait! Those decoupling caps in the power supply take a while to discharge! There are ways of discharging them more quickly, but I was prepared to let it take its time. I could fit some discharge resistors across the caps to let them come down more quickly, but space is a bit tight already. What I have (somewhere) in my toolbox is a lead with a couple of crocodile clips and a 20K 5 watt resistor soldered into it. This lets the capacitors bleed down gently but quicker than leaving the amp standing.

During the power-on test, nothing caught fire or even got hot, so it was time to drop the valves in place.

Power back on, with the volumes turned all the way down! And it lives! The valve heaters glow nicely - and there's an awful hum from the speaker! But I've got sound. The next thing is to connect a guitar up and see whether it will actually amplify an input signal. And it does!

The hum is down to a couple or three things. I need to add a couple of resistors to create an artificial centre-tap and reference the heater circuit to ground. I need to connect the chassis ground to the signal ground. And I need to shield that output transformer inside its coffee tin.

But I did manage to record a fragment of video on my camera. I think the CBA sounds quite growly and well suited to a bluesy CBG...

CBA clip

Next jobs (but not tonight) are to make those additions and alterations I mentioned, then start on the speaker cab. Tonight is eating chocolate and watching a movie night.
Okay. Taking another break, this time for some lunch. I haven't got quite as far as I wanted, but I'm happy with progress so far. Some of the slow-down was due to redesigning the layout on the fly. I'll explain why with the pictures.

First component of the pre-amp goes in place. It's the 68k resistor soldered to pin 2. This is what we call the grid stopper. Its main function is to act as a filter to prevent radio interference and we try to solder it as close as humanly possible to the valve base.

You'll also notice in the picture that the solder tag on the top left of the tag board has been squashed down. And that's to allow the rather large decoupling capacitor some room to sit close to the board.

The decoupling capacitors form part of the power supply. Their job is to work in conjunction with the power supply dropper resistors to form a high-pass filter, helping to stop hum in the amplifier. In my early amps I used to build a separate power board with a series of "nodes" - a resistor and capacitor filter - then run flying leads to each part of the circuit that needed power. Through trial and error, I discovered that this wasn't always the most stable way of doing things. Far better from a stability point of view to have the capacitors as close to the part of the circuit they're feeding as possible.

By "unstable" I mean problems with unwanted oscillation and that low frequency thumping sound called "motorboating".

So now we get to the red LED connected to the first cathode on pin 3. This is to provide the bias voltage for the stage.

In order for a valve amp to work correctly it needs the grid to be sat at a more negative voltage potential than the cathode. I'm oversimplifying massively here for the sake of a quick and dirty explanation by the way. One way of setting the grid more negative than the cathode is by sticking an LED between the cathode and ground, effectively making the cathode more positive. We can achieve the same effect with a diode or a resistor or a resistor in parallel with a capacitor (which has an effect on gain at different frequencies).

As I said, that's a quick and dirty explanation. I chose LED biasing because I've used it in most of my amps and I've found I like the sound and the reliability. Plus it reduces the component count in this amp by two!

You might also have noticed in the pic above that I haven't soldered the LED or capacitor into the circuit. And that's because I had a feeling I was going to have to do some juggling of the layout. 

I needed to use the capacitor ground connection to run the shield for the input lead...

I also added the ground cable for the capacitor at that point. This will loop back to the ground point of the first smoothing capacitor in the power supply. And again this is done to keep hum to a minimum by adopting what's known as a "star ground scheme" for the amp. All the signal and power grounds are taken back to a single point in the amp to avoid ground loops.

So now it can all be soldered down to the board and the excess leads trimmed.

Next up we have another minor layout problem. The anode resistor for the first stage of the pre-amp will be sat at around 250 volts. And it's also sat very close to the head of the mounting bolt that is a direct connection to the chassis.

Which is why the resistor has been given some insulating sleeves and is standing quite high from the board. The grey cable you see running under the board is the feed to the power amp stage that will be taken from the centre pin of the master volume pot.

Okay. More parts going into place. The black cable is the feed and return along with the grounded screen for the gain control. This sits between the first and second stages of the pre-amp.

Decoupling capacitor for the second pre-amp stage and the 100k anode resistor in place. Also in this pic you can make out the two interstage coupling capacitors (the little orange discs). Their job is to block the high DC voltages from the anodes while allowing the AC signal voltages to pass. However, they also have a filtering effect. By adjusting the values of these capacitors we can shape the voice of the amplifier. When the tweaking starts, it's these capacitors that will get looked at first.

And there it is, almost complete and ready to test. I still have to wire in the control pots and output transformer then connect up the power to the board. Then testing can start. If you look, there are two large blue resistors on the right of the tag board. These are the dropper resistors in the power supply that I was talking about earlier, the ones that form part of the supply filtering. In my original layout, these were supposed to be sat alongside the decoupling capacitors. When it came to putting the parts together, I realised they were just too big! So that's why the resistors are sat way over to the right, feeding their respective decoupling capacitors through the red flying leads. Stability shouldn't be an issue though. The important factor is where the capacitors sit rather than the resistors.