The Long and Winding Road

Part the Third

“You certainly usually find something, if you look, but it is not always quite the something you were after.”
J.R.R. Tolkien, The Hobbit

In this the penultimate part of the journey the path forks and I take a look at some experiments with using a valve preamp and improvements to the original FET preamp.

Opening the Valves

Eric Thacker, better known as Ecca, had demonstrated on the Shadows Music forum that Piet’s FET and op-amp preamp design could be replaced with a valve (tube) preamp that was available from a number of eBay sellers in Hong Kong and China. (Here, for example).

IMG_7055_5843_edited-1

The valve preamp and Eccamatic pcb mounted in an enclosure.

The circuit was originally designed to be used as a Hi-Fi stereo preamp using two 6N3 dual triode valves, one valve for each stereo channel. For the eTap2HW module we only require a mono audio pathway so the left channel of the stereo pair was used as the input preamp feeding the echo module and the right channel became the recovery amp, taking the echo module output and acting as a buffer and output level control.

Eric had produced a PCB for his ‘Eccamatic’ design which allowed the echo module to be plugged in, provided a step-up power supply to generate the HT voltage and the LT heater supply for the valves and had connections for the three pots that adjust the echo settings.

IMG_0513_edited-1

The bits for Eric Thacker’s Eccamatic pcb.

Having already designed the automation circuitry it was easy to adapt the design so that the Arduino Uno would feed the pot inputs on the Eccamatic board, reading the required settings from pots mounted on the front panel.

The front panel also had input and output pots, controlling the signal levels.

I have described this valve EchoTapper build in a series of earlier posts and if you’d like more details just click on the Valve entry in the Categories list on the right of this page.

When the new unit had been completed, the sound was quite pleasing with less hiss and noise than the FET preamp but rather ‘sterile’ as might have been expected from it’s origins as a Hi-Fi design. As I am sure you will know, guitar amplifiers and the early Meazzi echo units have circuits that are far from Hi-Fi for a very good reason (at least in the case of the typical valve-based guitar amp) – the almost perfectly flat frequency response of a Hi-Fi design and it’s deliberate lack of harmonic distortion (achieved largely through the use of negative feedback) is really not what guitarists want or indeed need to get the sound they desire. The original Meazzi preamp is hardly a classic of valve design but somehow has that magic sound that we have come to love.

Piet’s FET preamp design had made use of the ‘Fetzer Valve’ circuit which uses an FET to emulate the sound of a triode valve so it does create harmonically rich sound when the input level control is set correctly to match the guitar pickup level to ‘push’ the Fetzer ‘valve’ into generating the pleasant sounding harmonics we all like.

Steve Mitchell soon had his thinking cap on and he came up with a number of modifications to the valve circuitry that, according to his simulations of the circuit which he had made using the TINA SPICE modeller, should tailor the frequency response better to the typical guitar pickup signal and also generate those desirable harmonics to give that rich sound we all love.

Steve and I came up with a three stage process that culminated in a final design that sounded great.

Stage 1: Remove the built-in stereo volume control pot and increase the input impedance

Stage 1 was essential to allow the input and output level controls to be separated and to provide a better match to the typical guitar pickup’s high impedance.

Stage 2: Remove the negative feedback from the preamp and add Meazzi-like tone shaping

Stage 2 was regarded as highly recommended to obtain a better tone and more gain which would encourage the desirable valve harmonics and tailor the frequency response which was very flat and resulted in a ‘lifeless’ sound, thus improving the sound for guitar purposes.

Stage 3: Add a separate Gain pot and three-way tone switching

The Stage 1 and Stage 2 modifications had resulted in a really good sounding preamp but Steve came up with Stage 3 which would take the whole thing to another level. Whilst not regarded as essential, these mods did make the unit even more flexible. There would now be three ‘level’ controls, the new Gain pot controlling the drive to the second valve stage in the left channel of the amp. The Volume pot controls the signal to the echo module and the Master pot controls the level from the output socket.

Modified Vavle Preamp Flowchart

Flowchart by Steve Mitchell (SCM)

Modified Valve Preamp Freq Responset

Predicted frequency response of the modified circuit

IMG_2570_edited-1

Front Panel of the final Valve-based Echotapper

The 3-way tone switch provided ‘normal’, ‘vintage’ and ‘warm’ options by tailoring the bass roll-off. On the unit in the picture (built for Mario Voltolini in Italy) I also added, at his request, a Cutting Edge Filter (CEF) which could be switched in and out if required.

Biting the Bulletins

In parallel with the improvements to the valve preamp, Steve was also working on a series of modifications to Piet’s FET preamp design.

One of the problems with any FET is that individual devices, even from the same batch, can have parameters that vary over quite a wide range and this makes it difficult to design a circuit that will allow the FET to operate in the optimum part of its characteristics.

Steve and Rolf Holmberg collaborated and came up with a spreadsheet that would calculate the best resistor values to bias the FET into the optimum operating point. This would be a bit complex for the average user to get to grips with so Steve and I came up with a simpler ‘flowchart’ approach. This would allow the best operating point for the FET to be found using a simple voltage measurement and a little trial an error with a few preferred resistor values.

Steve put together a series of ‘bulletins’ that were published on Piet’s EchoTapper blog which also included simple modifications to improve positive signal headroom, improving the FET gain and reducing white noise (hiss). These mostly involved removing or changing some resistor values and adding some extra capacitors.

Piet has since modified his original preamp PCB design to accommodate these improvements and this is documented on his blog.

What Next?

Now I had two excellent sounding echo units but I had an urge to see if it would be possible to squeeze all this technology into a smaller package, one that would fit on a pedal board or sit neatly on the floor at my feet like a typical guitar effects pedal.

Thus began the germ of an idea that would result in the Blue Nebula. It had to be compact, so that ruled out the use of valves and it needed to be much easier to build, with minimal off-board wiring so that, hopefully, anyone with sufficient experience would be able to put one together from a kit or buy a completed built-to-order unit at reasonable cost.

We had to go from this …

IMG_2803_edited-1

My first automated eTap2hw

 

to this …

IMG_3927

The final Blue Nebula Design

 

To be continued.

Optimizing & Modding

I finally got around to optimizing my FET-based eTap2HW, the one with automation. (I also have a manual eTap2hw that I already optimized a while ago now).

What’s this optimizing I hear you ask? It’s a well known fact that FETs are notorious for having widely varying electrical characteristics even when they come from the same batch and this makes it difficult to design a circuit that will work well with any FET you plug into it. Aside from buying a big batch of FETs and testing them to find the most suitable ones, it’s usually necessary to resort to trying different biasing resistors and measuring the source and drain voltages, and keep repeating this until the FET is operating in the desirable part of it’s characteristics. By doing this we are arranging that the FET will behave very much like a triode valve and thus give us the sweetness of tone that every guitarist seeks 🙂

This could all get very tedious but thankfully Steve Mitchell came up with a fairly straightforward ‘flow chart’ approach that will get your FETs biased nicely with the minimum of trial and error. Steve has kindly produced a set of ‘Bulletins’ that explain what’s required. You can download these over on Piet Verbruggen’s Echotapper blog. The ones required for optimizing your FETs are Bulletins 1 and 2 and Bulletin 5 is worth following as well as it improves the gain of the FETs and reduces the white noise or hiss that some people find problematic, especially for recording.

While I had the unit opened up I decided to make a few modifications I’d been thinking about doing for a while.

1. First up, I decided to cut a small opening in the end panel directly opposite the USB connector on the Arduino so I could easily connect the unit to my PC whenever I wanted to update the firmware and when I’m using the Librarian software to backup my patches and arranging them into a set list. Before this I had to unscrew the end panel to get access to the USB connector which was inconvenient.

The new USB 'easy access port'

The new USB ‘easy access port’

2. My second modification replaced the ‘Wet’ output mono 1/4″ jack socket with a stereo jack socket that will accept the patch change foot pedal I described in an earlier post. I never found a use for a ‘Wet only’ output signal so this jack socket was never used and it was easy to replace it with a stereo TRS 1/4″ socket. The sleeve (S) connects to Ground, the tip (T) goes to the ‘Up’ switch and the ring (R) goes to the ‘Down’ switch.

The footswitch socket labelled FS.

The footswitch socket labelled FS.

3. My original tactile push buttons were beginning to play up. Frankly I wasn’t surprised as I had always thought this type of little ‘clickety-click’ switch wouldn’t be very durable and so it proved. I decided to remove the strip-board sub-assembly that held all five switches and replace the Edit and Mem switches with panel mounted miniature momentary push buttons.

IMG_6875_5505

The old buttons on their strip-board sub-assembly.

I was able to retain the ribbon cable and connector and make up a suitable matching connector for the new switches and patch change wiring.

As I was going to be adding a rotary encoder I no longer needed the Up, Down and Select buttons so the holes for the up and down were plugged with a couple of little rubber grommets that I happened to have lying around and the rotary encoder went in the  hole where the Select button used to be, after enlarging it with a reamer.

The new switches and rotary encoder. Colour coding will help me remember which switch is which!

The new switches and rotary encoder. Colour coding will help me remember which switch is which! Note the neat little blanking plugs.

4. I’ve described how to add a rotary encoder to the eTap in a previous post and I followed that when adding the new encoder to this project. The Sparkfun breakout board makes it easier to wire the encoder. I used an RGB encoder even though I don’t use the built-in LEDs as this type has a mounting bush to fit it to the front panel; other cheaper variants can’t be easily panel-mounted. I really recommend the rotary encoder as an alternative to punching the up and down buttons like a mad thing when choosing patches or editing  the patch name!

LCD/Buttons board and rotary encoder

Buttons and rotary encoder on its breakout board.

And here’s the EchoTapper Vintage Echo Unit with all it’s mods and optimizations completed.

Catalinbread ECHOREC pedal shown for size comparison.

Catalinbread ECHOREC pedal shown for size comparison.

I just wish there was room left for a bypass switch 😉IMG_2803_edited-1

And here it is on my new ‘Shadows’ pedalboard 🙂IMG_2816_edited-1

Frequency Response Tests Part 2

These tests examine the effect of the Cutting Edge Filter (CEF) which is a steep high pass filter originally designed by Charlie Hall to remove any muddiness in the guitar tone by filtering out the unwanted low end frequencies.

Test 1

CEF: OFF, Presence: MIN.

Bode-#1---CEF-OFF,-Pres-Min

Test 2

CEF: ON, Presence: MIN

Bode-#1---CEF-ON,-Pres-Min

The gain now drops off steeply below about 400 Hz; for example at 200 Hz the gain goes from about -7.5 dB with the filter off to about -22 dB when it’s turned on.

But what’s with the rise again around 80 Hz? It’s probably not significant as the gain there is still about 8 dB down on the response from 1000 Hz up.

Please feel free to leave comments.

Frequency Response Tests Part 1

Bode plots were performed on the FET eTap2HW unit to study the effect of the Presence control and the Cutting Edge Filter (CEF).

The input signal was a 250mV peak swept frequency sine wave. Sweep range 50 – 5000Hz. Echo Model 7, all controls zero, Level = 6

Test 1

CEF: OFF; Presence: Min;

Some crazy phase shift stuff going on!

Some crazy phase shift stuff going on!

Test 2

CEF: OFF; Presence: Max;

Bode-#1---CEF-OFF,-Pres-Max

Note the presence peak at around 2.24 kHz

Signal Analysis of the FET eTap2HW Part 1

Using a recently acquired National Instruments myDAQ and the Bode Plotter, Dynamic Signal Analyzer and Function Generator Virtual Instruments (VIs) included with it, I’ve been making some detailed measurements of my FET eTap2HW unit.

This is the circuit designed by Piet Verbruggen with the optimizations suggested by Steve Mitchell. It is also fitted with a variable presence control (another Steve M suggestion) and a Cutting Edge Filter (CEF) from the design by Charlie Hall.

Test run #1

Input Signal: 1 kHz sine wave, 400 mV pk-pk

Presence = minimum; CEF = OFF, Echo Model 7 (Reverb + Trem), all Controls = 0, Level = 4

Note there are few harmonics and THD is low

Note that there are few harmonics and THD is low (0.25%) with this low input Level setting, as expected.

Test run #2

Input Signal: 1 kHz sine wave, 400 mV pk-pk

Presence = minimum; CEF = OFF, Echo Model 7 (Reverb + Trem), all Controls = 0, Level = 8

A higher input level is designed to hit the modified ‘Fetzer valve’ circuit’s ‘sweet spot’ and produce desirable valve-like harmonics. This plot shows the harmonics are indeed present now.

The higher input level has 'turned on' the harmonics. THD now about 1.3%

The higher input level has ‘turned on’ the harmonics. THD now about 1.3%

eTap2HW Poll Results

Thanks to everyone who took the time to listen to the samples from my three eTap2HW-based EchoTapper variants and cast a vote in my poll on which sound they preferred.

There was a very clear preference for Track ‘C’ which polled 55% of the votes, in second place was Track ‘B’ which polled 30% and coming well up the rear was Track ‘A’ with a mere 15%.

Now it’s time to reveal the details of the ‘competitors’…

  • Track ‘A’ was the unit with the valve preamp (15%)
  • Track ‘B’ was the unit with the unmodified FET preamp (30%)
  • Track ‘C’ was the unit with a modified FET preamp with optimized bias (55%)

Which really makes me wonder about all the fuss and hullabaloo on some of the Shadows forums regarding how ‘good’ the valve preamp-based units sound. My personal opinion is that the valve preamp sounds too ‘mellow’ with not enough bite to get the sound of the early Shadows recordings. It does sound nice but comes nowhere near ‘that sound’ but maybe with some modifications to the circuit it could get closer – we shall see.

Incidentally, I also prefer the sound of the modified FET preamp, it gives more gain and punch though Piet’s original FET design is also excellent and the results of individual units can vary depending on the wide manufacturing spread of the specs for individual FETs.

The original sound samples are still available here if you’d like to audition them (again).