Designing/building a preamp for a Stage Rhodes

Started by sookwinder, September 16, 2017, 06:03:55 AM

Previous topic - Next topic


OK... as some of you may know, a couple of months back I bought a restored Fender Rhodes Mk1 electric piano.  It was the "Stage" version, meaning effectively it is a giant passive electric guitar with 73 single coil pickups, unlike the "suitcase" version that has its own preamp, power amp and speaker cabinet.

We all know that like an electric guitar the stage version of the Fender Rhodes needs to be plugged into a guitar amplifier that has a preamp that can raise the signal level, in a similar way they do for electric guitars. Modern keyboard amplifiers assume the signal from the keyboard is already at LINE OUT level and hence are basically no good for use with the "Stage" version, other than with the addition of a preamp or booster pedal unit.

This thread is the adventure that took about 7 - 8 months to complete, designing/building a preamp unit for the Fender Rhodes Stage EP.

Those who have read any of my build thread over at know that I write like I talk... a lot.... so grab yourself a cuppa tea, a coffee, a beer, maybe even a couple of sandwiches ... whatever  ... and delve into preamp design work and all the highs and lows that went with it.  And for those who are not trainspotters for amp designs, just cut to the chase and page down to see the photos.

Last December when I first ordered the EP, I knew that I needed some sort of preamp unit for the EP as I was not going to mic it up when I used it for recording. I had already built an amp that I have been using when I wanted to record my Pianet T, but it also had issues that many others had found in as much as overdrive, too much overdrive (distortion) can occur.... In short the attributes that a valve amp is well known for when used with a guitar, often are not those desired when using an EP. The designs that utilise 12AX7 valves in many preamps for electric guitar amps are just too hot.

So initially I discovered that a company in German,  Reußenzehn, has a "preamp box" available for use with a Stage EP.  The box had tremolo and reverb functions and also depending upon volume controls, different levels of overdrive could be achieved. The blurb on the website said it was designed to simulate the Suitcase model preamp.  It can do mono or stereo tremolo output.

As my intention was to use this EP when recording in my home studio, mic'ing up a full size valve amp is somewhat prohibitive, so this Reußenzehn unit seemed like the perfect solution. 

The unit utilises a 12AX7 and a 12AT7. I fitted NOS 1960s valves and I was happy with the sound. But what immediately threw me off and concerned me was that there is a "clicking" sound functioning when the tremolo section of the box is engaged. The click is quite prominent, especially with held notes/chords.  The clicking is coherent with the cycling of the tremolo.

Here is a trace of the resultant tremolo using the Reußenzehn unit, when compared to a tremolo that uses cathode bias varying design.

Large Fender amps that have tremolo circuits that use an a light dependant resistor (LDR) in the opto-isolator (AKA as  the 'roach' ) have a "click" when the tremolo is used.  But the "click" was either very quiet or fixed by an "infield fix" back in the day.

When I contacted the manufacturer of the Reußenzehn unit I received the following reply. (these are the salient points from the reply with all the hype removed):
QuoteSo the click of the Vibrato / tremolo is part of the sound. Like in most of all original Tube Tremolo the tube make an ignition and alternating two high voltage gas lamps, the clipping of the electrons does that floating alternating sound. You can hear that in each original Fender Twin Reverb Amp when the vibrato is on.

... And the click is very lower while playing, because the system works while playing with lower Z.

So in short it seems that the Ressenzehn's "Suitcase Preamp" is a preamp that attempts to reproduces the sound of a Fender Twin Reverb tremolo and not that of an actual suitcase preamp.  In fact it attempts to reproduces the sound on the SF Twin including the annoying click.

Now let me say that in a live situation the Ressenzehn's  "Suitcase Preamp" would work perfectly OK ...It would sound great ...  no one would even notice the clicking, not the audience, not the players.   But I do not gig ... all I do is record in  my home studio and I notice the clicking immediately.  So the Ressenzehn's  "Suitcase Preamp" now sits on my shelf collecting dust never to be used by me again... anyone want to make me a an offer for it ?

So by late January 2017 I decided to design/build myself my own all valve preamp unit for the Rhodes. It was not going to be a copy of the Peterson, Janus or Jordan preamp designs used in the suitcase versions, but rather a preamp based upon the various Fender valve amps that have been used successfully by players over the years.

It would have input and master volume controls, Treble/Middle/Bass controls, Tremolo  (only mono – I dislike stereo tremolo with a passion) and also a selectable overdrive function.

When I have asked various Rhodes players which amp they use (or used) with their Stage Rhodes (as well as reading on various newsgroups) the type of amp mentioned was usually a Fender Silver Faced Twin Reverb, a Fender Tweed Bassman or a Fender Princeton Reverb for small gigs.  Now anyone who knows Fender amps and their design attributes would immediately observe that these three amps are vastly different from each other, so different that I have never seen these three amps grouped together in any discussion about amps, other than the fact that Leo designed them.

If you think about it for a while ... the (only) common factor grouping these amps together is that depending upon how they are used, each amp has a reasonable amount of clean headroom before any level of overdrive kicks in.

The Princeton Reverb is a very clean amp with about 12 - 15 watts and only starts to break up with the volume at around 8.5 - 10.  Perfect for a small gig space. 

The SF Twin Reverb certainly can achieve overdrive distortion in levels that will make your ears bleed, but given that it is such a powerful amp, the levels needed for the Rhodes mean that the valves are a long way from really producing a gritty sound when a Rhodes is used with it.

Lastly the Tweed Bassman, an amp associated with a bunch of blues guitar players and also Neil Young's amp of choice.  An amp known for its overdrive and distortion.  But also like the Twin Reverb, the Tweed Bassman is extremely loud, so the levels needed for a Rhodes are achievable without sounding like Neil Young.

The Princeton Reverb uses a 12AX7 input preamp valve running at about 120V, the SF Twin reverb also a 12AX7 at about 210V and the Tweed Bassman runs a different preamp valve, a 12AY7, at about 150V. All very different input conditions.

So whatever design I come up with, one of the attributes must allow for a decent amount of clean head room.

I also like, in the right situation, a decent amount of warm overdrive with my Rhodes sound ... not death metal OD, but nice warm OD. So a decision of where in the circuit the OD control should be placed as well as the method of producing the OD itself needed to be developed.

Lastly the circuit needed to be able to convert the signal (after yet to be defined gain stages) to a voltage and impedance level that will be compatible with LINE OUT specs.

I didn't think this was that difficult to come up with, but I did know that every new design always has the ability to kick you in the arse when you are in the middle of designing/building it.  [this project would be no different as it turned out].

Circuit Topography
The 1960s "Black Faced" Fender amps and the 1970s "Silver Faced" Fender amps have effectively the same preamp/input topography.  This is:
- input into
- First triode of the first preamp valve (12AX7)
- Fender BF tone stack, known these days at the FMV tone stack
- Second triode of the first preamp valve. (12AX7)

In the larger amps such as the twin reverb a third gain stage occurs  (First half of another 12AX7).

After discussing this project with a friend who also designs valve amps, I decided to try the circuit for creating OD outlined in chapter 6:15 of Merlin "The Valve Wizard" Blencowe's book "Designing Tube Preamps for Guitar and Bass, 2nd edition".   This involves a gain stage followed by a DC coupled Cathode Follower, where the size of the preceding (to the CF) grid stopper resistor can achieve different levels of OD. If this grid stopper was a variable resistor the amp would have a OD control.  Mmmm now we are getting somewhere!  Mind you even now as I write this up, I still do not fully understand DC coupling in valve amps ... maybe I never will.

I knew that by the time the signal gets past the 3rd gain stage the voltage levels can be in the ballpark range of 150 – 250V. So the Master Volume also needed to be able to dramatically reduce the final output voltage to a level that can be used as LINEOUT.  And finally a circuit was required to enable the output impedance of this PREAMP unit to match those required when plugging direction into a desk, a modern keyboard amp or a Digital Audio Workstation. Some reading of my amps books & on line searches lead me to discover that a "Cathode Follower" is the best way to achieve impedance matching and no loss of the high frequencies for the final stage of this unit.

This is how the topography would look like [Note at this stage of the design I was not even considering the tremolo part of the circuit, just amplifying, having an OD circuit, and a master volume with impedance matching was enough to consider initially]:

Even before I built up the unit or finalised the design, I checked some of my other amps and determined (for a 500mV input) what the rough voltage levels after each of the gain stages, to give me a ball park idea of what I was dealing with.

The final cathode follower, initial thoughts were to use a 12AX7, but ultimately I went with a 12AU7 valve, has effectively unity gain and is purely in the circuit to allow matching of impedances.   I was thinking that the Master Volume would be some sort of voltage divider where I would wash away 90% of the signal, leaving 10 - 14V at the output. The input gain stages and tone stage was already defined via the Fender BF/SF design. All I had to think about (at this stage) was the OD part of the circuit and how could I vary a grid stopper resistor from 5k up to even values of 1M.

Given the voltage levels at this point in the circuit (hundreds of volts), a normal 1/4watt potentiometer would not be sufficient, so after much searching I came up with the idea of using a ceramic 1 pole/6 position rotary switch, which I sourced quite cheaply via ALIEXPRESS from China.  It was perfect, robust, could handle 4 watts and was small enough to fit into the amp chassis I was intending to use.

I have learnt over the years to never rush any amp (electrical) design/build, as mistakes can be made and electrical shocks can occur.  I have been shocked by 240V wall voltage 3 times, now ... but none since I realised it is better to take one's time and live.

A rough outline of the schematic was starting to take shape, with the preamp sections on the left essentially "borrowing" the standard large Fender BF or SF amp, including the addition of a "Bright Switch" and Mids control.

At this point I am compressing 5 or 6 weeks' worth of emails and discussions with friends and colleagues at work, where the ideas for the master volume and the final cathode follower developed.  I was now at a stage where I could breadboard this design, and for the first time (and I must say THE VERY LAST TIME) I decided to use pre-made turret boards for the first build.  Did I mention I now hate turret boards?

One of the early incarnations. I used a Vibro Champ sized chassis and had planned to use the same type of power transformer that I I had used on the "Reverb Unit" that I had built 18 months back, a Mercury Magnetics PT that developed for new or replacements into Fender style stand-alone Reverb Units.

The filtering circuit in the "dog house" looks like this.  Given the price of electrolytic capacitors is relatively cheap compared to 40 or 50 years ago, my design used enough caps to ensure a quiet power supply. Also given that I was using a solid state rectifier and not a valve rectifier, there was no limitations on the amount of capacitance that the rectifier can "see".  The section of the photo I have highlighted in yellow I will discuss later.

By this stage I had (stupidly) decided that I knew what I was doing and this "bread boarding" exercise could be also the final version, so I had added in the tremolo section of the circuit.  The tremolo circuit was based on a simple design my fellow amp building friend had used in his amps over the past couple of years: The layout was

12AX7 triode 1  Gain stage 1 – valve 1
FMV Tone stack
12AX7 triode 2  Gain stage 2 – valve 1
12AX7 triode 1  Gain stage 3 – valve 2
12AX7 triode 2  OD  stage2 – valve 2
12DW7 triode 1  Tremolo – valve 3
12DW7 triode 2  Cathode Follower (for line out) – valve 3
The 12DW7 is a hybrid 9 pin valve that combines one 12AX7 triode with a 12AU7 triode.  I could use the 12AX7 half of the 12DW7 for the tremolo circuit and the second half of the 12DW7 (a 12AU7 triode) for the cathode follower.  Perfect.  When one has limited real-estate on amp design, these hybrid valves come in really handy.
Late 60s Pianet N - Late 70s Pianet T - Ensoniq ESQ1 - Hammond XK2 - Wurli 206A converted to a 200A - 1973 Rhodes Stage 73 - Roland RD150 - Vintage Vibe 64 EP - s**t load of guitars, basses & amps



But this is also where I fucked up big time.... and basically wasted almost 2 months of building, redesigning, starting over yet again ... this is the dark period when building a new design...  no one else can help you, they can give advice, they can suggest solutions, but in the end it comes down to checking and rechecking everything from the schematic, to the layout drawing (did I create my layout drawing exactly how the schematic had the circuit?), to the actual amp and even as I discovered , even the way the valves I was using were manufactured.

I have told myself over the years, after barely passing university electronics, never assume anything when dealing with electronics and circuits. But I never seem to remember my own advice.

The 12DW7 has two different triodes within the glass envelope, unlike a 12AX7 or 12AT7 or 12AU7 9 pin valves, were both triodes are the same.  If one wires up backwards any of these type valves it doesn't really matter, each triode is the same. But with a 12DW7, each triode is doing a completely different function in the circuit so you must wire the 12DW7 correctly.

From the spec sheet for a 12DW7 I know that the first triode is the 12AX7 type triode and the second triode is the 12AU7 type triode. Pretty simple, not hard to get the wiring correct I hear you all chuckle. 

The standard 9 pin miniature type valve uses pins 1,2,3 for one triode and pins 6,7,8 for the other triode.  Now any sane person, any logical person, anyone when asked would say Pins 1,2,3 would be the "first" (#1) triode and pins 6,7,8 would be for the second (#2) triode,  And that is how I wired up the 12DW7 valve into my circuit.

But I was wrong that is not how the two triodes are specified . Pins 1,2,3 are for the #2 triode and pins 6,7,8 are for the #1 triode. I would love to know which idiot back in the 1940s or 1950s thought it would be a good idea to designate it this way... totally illogical.

When I finally decided to plug the amp in and turn it one, while I was able to get a signal through, which sort of sounded OK, there was no tremolo at all and the OD circuit would have been perfect for Ozzi Osborne, but not for a Rhodes player. Plus even without the OD circuit engaged, if I turned the volume past about 6 it was just "crunch city"

mmmmm... damn!

So I had 3 issues:
- Signal is OK, up till 6, but then not usable past that point
- Specific OD circuit appears to have very little variation no matter where I set the controls
- who stole the tremolo?

At least the pilot lamp was turning on and the PT/rectifier/heater circuits were performing as they should and the measured DC voltages all roughly lined up where I thought they should be.

Mind you I did have another stuff up courtesy of the turret board used for the filtering circuit in the dog-house.  If you look at the photo above where I high lighted a section of the circuit in yellow, notice that unused hole... it is joined under the black surface coating with the turret directly above it. Four months later I can now see that quite clearly, back then I did not notice it in the photo or even when looking at the turret board itself.

When I mounted the power supply filter circuit into the dog house I decided to use that spare hole as a mounting point, rather than drilling a new hole.  Seems logical. Well it is until one turns on the power, ... I had created a short circuit to ground and suddenly the PT is getting  very very hot. I eventually tracked down this stuff up and it started to confirm my hatred for turret boards.  Give me eyelet boards (just as Leo used) any day of the week.

For five or six weeks I was in almost in continual email conversation with my friend James about possible reason for the design not working as I thought it should. My lunchtimes at work were almost solely dedicated to discussing the same issues with my manager who is an electronics engineer.

Even though I use my third bedroom as my workshop, my lounge room had this on the coffee table, scope, sig-gen, alligator clip leads, probes and electronic components for many weeks:

By this stage the restored Rhodes that I had commissioned back in December 16 arrived (albeit 2 months after the promised arrival date).  The pressure was on to get this circuit correct and finish the project.

After yet another evening that ended up with ...

... I came to the realisation that I needed to start again, dump the turret board (you can barely fit 2 or 3 wires into the holes, you are limited to the layout on a grid   aaaahhhhhhhhhhhh) and design the circuit again using the old faithful eyelet board process.

I had decided to use the variable cathode biasing type tremolo circuit that is used on the Fender Vibro Champ ... I had used this circuit before on eight previous builds (of various designs) ... it is a simple circuit, working on the cathode of the second or third gain stage and produces a lovely sounding tremolo.

At this stage I still had not realised that I had the 12DW7 wired up incorrectly, but the 12AX7 triode was still operating sort of OK as a Cathode Follower, rather than the correct 12AU7 triode in the 12DW7.

For the VC tremolo I needed to use a complete 12AX7, so an additional hole in the chassis needed to be drilled.  I had also surmised that no matter how I set up the first valve (a 12AX7), I could not get enough clean headroom even without engaging the OD part of the circuit.  It was just too hot having 3 gain stages. The unit lost any semblance of clean output once the volume went past 4.

So I decided to change the first valve from a 12AX7 to a 12AY7.  This is the same type of preamp valve used in the Tweed Bassmen and has about half the gain of a 12AX7. It also has a smoother gradient of OD as the signal increases than the 12AX7...  all good points.

The layout was

12AY7 triode 1  Gain stage 1 – valve 1
FMV Tone stack
12AY7 triode 2  Gain stage 2 – valve 1
12AX7 triode 1  Gain stage 3 – valve 2
12AX7 triode 2  OD  stage2 – valve 2
12AX7 triodes 1 & 2  Tremolo – valve 3
12DW7 triode 1 Empty unused triode valve 4
12DW7 triode 2  Cathode Follower (for line out) – valve 4

What's the old saying ... pride come before a fall ...  two more problems arose ... and it is times like these that just stopping and doing nothing for a couple of days is good advice.  When dealing with electricity (house voltages) never rush things, never just do an extra how before midnight of work ...  always stop before you get too tired physically and mentally.

The first issue was something so stupid I am embarrassed to show it ... but if it helps others then that is good.

If you look at the photo below of the underside of the populated eyelet board showing the connecting wires that are underneath, I still have not trimmed the legs of the components.  Problem is that I forgot to trim them before I wired the board into the chassis. And notice how some of the extraneous wires/legs are touching other components. Oh s**t !

So I wired everything into the chassis, power it up and nothing seems to be as it should be (I wonder why?) 

I grabbed the DVM and started to measure voltages throughout the circuit and at one stage while measuring the voltage on one of the pins of the 12AX7 used for the tremolo part of the circuit I slipped and shorted something out.

Pooooooooffff .. bang  ...magic smoke escaping from somewhere ...

To cut a longish story short, I stuffed up the power transformer.  Never done that before.  It only puts out about 2/3 of the DC voltage it should now.  Normally someone in this situation would have to go and buy another PT, but because I purchase my PTs from the US and a significant portion of the cost is shipping them to Australia, I always purchase a second (spare) PT.  Never had to use the spare PT before...

Another day/evening lost

The eyelet board was removed (yet again), wires trimmed and along with the spare PT everything was re-wired back into the chassis.  Another replacement PT was  ordered from the USA.

Powered up the unit again.

The change to the 12AY7 worked a treat ... now OD cause by running the volume control at higher levels only kicked in at about 7 to 7.5. But the tremolo circuit, the VC style tremolo circuit that I had just changed to would not work. Swapped valves in the trem circuit, measured voltages, swore a lot, but even though I had successfully used the VC style tremolo many times before and it is a circuit I know inside out (I even understand how it actually works), I could not get the damn thing to oscillate. No matter how many times I measured voltages, compared the schematic to the layout to the board itself .. nothing, zip

So I decided to watch an episode of the latest series of THE BLACK LIST and came back to the project a week or so later.  I had a 6 day business trip to the Manchester/Liverpool UK, the same week that the bombing occurred.  Things like that make you realise what is important in life and not to get too worried about a preamp that doesn't want to work.

Upon returning to Oz, a week or so later, I still did not have any idea why the tremolo circuit would not work, but my time away from the project was useful.  It still amazes me how one's brain can be working on a problem even though you have set it aside for the moment.

I had finally realised that I had been wiring up the 12DW7 incorrectly.
Amazing what information one can learn from the 1960s valve spec sheets when one actually ready everything rather than just skimming through it.

This explained why the first version of the tremolo circuit did not oscillate.  The 12AU7 section of the 12DW7 did not give out enough voltage swing to even come close to oscillation.  If I had have used the 12AX7 half of the 12DW7 it would have worked.  So for the moment I was going to leave the tremolo circuit alone... the VC style trem circuit had beaten me and eventually I would go back to the original circuit I have wired in, all those months ago.  You live and learn.

I also had some revelations concerning the DC coupled OD circuit.
After spending some time looking at the scans from the scope of the OD created by the DC coupled gain stage and the cathode follower pair, it was apparent that if I was into northern European style death metal then this type of OD would be perfect. 

But no matter how I adjusted the set up of the grid stopper resistor voltage located between the two triodes that made up the OD circuit I could not achieve anything near a smooth or creamy OD sound. Effectively it went from clean to flat top OD and nothing in between. All that it would produce was the sound of a Kenworth truck crashing into a steel girded bridge.

I needed to final another circuit that would allow me to adjust the levels of OD.

Just like ground hog day, out came the circuit board yet again...

And a modified version was put back in.  This time with the original trem circuit design, that only required one half of a 12AX7 and a new variable OD circuit.

This new variable OD circuit uses the method of varying the cathode resistor on the cathode of the third gain stage to achieve different levels of and different types of OD.

In general the size of the cathode resistor for a single 12AX7 triode is 1.5k.  If this cathode resistor is can be varied from the normal 1.5k down to say 680 ohms, the value will run "warm" and then to "hot' and produce grid current clipping.  This is what that "creamy" OD is. Take it too far and it gets ugly, but control it at the right level relative to the input signal and you get a dynamic creamy three dimensional OD that every guitarist loves and it will make the Rhodes sound awesome... even adds a bit of a Wurlitzer style grit to the sound.

On the other hand if you run the triode in a cold state one can achieve cut off clipping style OD.  In this situation rather than the cathode resistor being 1.5k, let's make it 5k, or anything in between. This produces cool or cold biasing of the valve and the OD is more two dimensional, less dynamic, but a style of OD  that I prefer with the Rhodes when you are just wanting the sound to be thicker (slightly more creamier or "hairy").

There is also another dimension to this OD set up, that being the level of signal input into the valve, obviously controlled by the volume control at the start of the circuit.  Have the volume really high, then the 3rd gain stage OD circuit described above just adds to the existing OD. Have the volume in the middle, then the 3rd gain stage OD circuit can add a delicate level of thickness to the sound or a lot more or even none, depending upon where you set the OD control.  (that being a 5k linear pot).

This is the schematic for the circuit.  Very simple and, again, the idea was from gleaned from various sections of Merlin "The Valve Wizard" Blencowe's book "Designing Tube Preamps for Guitar and Bass, 2nd edition"

Late 60s Pianet N - Late 70s Pianet T - Ensoniq ESQ1 - Hammond XK2 - Wurli 206A converted to a 200A - 1973 Rhodes Stage 73 - Roland RD150 - Vintage Vibe 64 EP - s**t load of guitars, basses & amps



So I had my OD circuit working ... maybe the light at the end of the dark tunnel I had been in for some months, wasn't a train coming in the other direction...

Next I went back to the tremolo circuit it and amazingly it worked.  Everything was coming together.  This trem circuit uses a red LED as the method of biasing the triode. (rather than the usual 3.3k res parallel with a 25uF capacitor).  Every time the trem circuit oscillates the LED will flash off and on, so you get a visual reference as well.  A number of amp manufacturers use this feature so you can see if the trem is on or not on the front of the amp control panel.

While I was experimenting with this circuit I did discover something about this trem circuit and the use of the red LED to bias the valve.  I noted that when the trem speed was raised, the pulses seem to blur. I do not mean that they were so fast I could not perceive them, but rather there seem to be a loss of the high point of each pulse (that was noted aurally not visually on a scope). Yet when I replaced the red LED with 3.3k//25uF set up, this blurring at high trem speeds seemed to no longer occur. In the end my ears preferred the set up with the cathode using the 3.3k//25uF rather than the red LED.

When the design and build is almost finished it is almost an anticlimax because you know that there is just a few more things to do (take photos of the completed circuit, clean up any lead dressing that is messy, finish of the artwork for the schematic and layout drawings, mount it in the cabinet).

But I did have a couple of ideas that I wanted to pursue.

The first being the on/off foot controls for the tremolo circuit.  Up until now I had been just lifting the cathode which immediately would stop the tremolo.  Or start it again, once the cathode was back in the circuit.  One of the problems with valve tremolo circuits is that, especially at low speeds, it takes a couple of seconds for the tremolo to kick in. Lifting/dropping the cathode was an instantaneous switching option, but it is dangerous for the valve and eventually it poisons the valve. (how long that takes, I have no idea).  But after talking with Merlin, he suggested an add on to my circuit that allowed me to safely keep using the lifting/lowering of the cathode as my tremolo switching position.

The second idea I wanted to look into was also to do with the tremolo circuit. 
This would be to add a switch that allowed me to lower the range of speeds that the tremolo functioned at.

I grabbed a toggle switch from my stash of trusty NOS "eastern block cold war" Russian military spec switches (available cheaply from a number of ebay stores) ...  these switches are "real men" switches... when they switch, they feel and sound like a switch should... definite, solid, dependable.... not like much of the crap one can buy these days.  Also grabbed a 47nF capacitor to be added in parallel with the other caps of the tremolo circuit.   

If I want to, I can now have a tremolo speed of less than 1Hz.

The last thing to do was change the valves from my "working B team" valves (anything I had lying around), to the "A team".
1962 GE 12AY7
1958 RCA 12AX7
1955 RCA 12AU7

By this stage I had replaced the final valve (a 12DW7) with a 12AU7, as I only required one of the 12AU7 triodes for the final cathode follower.

The circuit now was back to three valves:
12AY7 triode 1  Gain stage 1 – valve 1
FMV Tone stack
12AY7 triode 2  Gain stage 2 – valve 1
12AX7 triode 1  Gain stage 3 with variable OD – valve 2
12AX7 triode 2  Tremolo circuit – valve 2
12AU7 triode 1 Empty/unused triode valve 3
12AU7 triode 2  Cathode Follower (for line out) – valve 3

While all of the above had been occurs over the past 6 months I had ordered from MOJOTONE a cabinet for this preamp unit that matched the colouring of my Rhodes.  It arrived some months ago, but I was still awaiting the faceplates that I had ordered from my supplier in Thailand.   Jay had helped me with the artwork for the faceplates and I assumed all would be ok.  Never assume anything I keep telling myself but always forgetting.

When I went to fit the faceplates to the chassis, it dawned on me that they had been manufactured incorrectly.  There should not be an outer white edge/band.  The outer edge should be black. The faceplates did not fit. 

I contacted my supplier in Thailand and managed to convince him that he needed to have the plates manufactured again.  The next day I was flying to Germany for a 10 day business trip and hopefully by the time I got back the correctly manufactured faceplates would be waiting for me. And they were.

So before we get to the final photos some points of note:
- Even though it was a nightmare during some of the development stages, in the end the design I came up with achieves exactly what I wanted. A preamp unit  that allows me to plug my Rhodes directly into a modern desk or amp or DAW.
- The unit sounds like a valve guitar amp, again what I wanted. It sounds like I have plugged it into a SF Twin Reverb or a Princeton reverb.
I was not attempting to mimic the sound of the solid state preamps used in the suitcase versions of the Rhodes EP.  There are organisations out there who have done that with lesser or greater levels of success and if you want that sound then they are available.
- I can plug in an electric guitar and achieve a "valve" amp sound without blood spurting from my ears. From chimey clean to Hendrix like OD.
- I can also plug in my Wurlitzer EP or modern day Roland EP and "warm up" their sounds as well.
- My Pianet T also sounds awesome when run through this unit.
- While the "Bright" control adds sparkle to a guitar plugged in, it doesn't seem to add much to the Rhodes sound, which seems to gel with the notion that the frequency range of the Rhodes is quite a bit lower than for the harmonics in an electric guitar.

Thanks to James and Jay (from for the invaluable help with this project.


Late 60s Pianet N - Late 70s Pianet T - Ensoniq ESQ1 - Hammond XK2 - Wurli 206A converted to a 200A - 1973 Rhodes Stage 73 - Roland RD150 - Vintage Vibe 64 EP - s**t load of guitars, basses & amps


Beautiful writeup.  Beautiful preamp.

I share your frustration with engineering design phases - good idea, first attempt, utter screwup, embarrassing realization, fill the trashcan with valuable components, rebuild, refine, tuxedo dressing, and final reveal.  (Then carefully hide the financial details and pretend it was accomplished with frugality and efficiency.)

Awesome work.  Thanks for sharing.


The Real MC

I'm an EE and I read this with great interest.  Piano is my primary instrument; I also play guitar and bass so I have an assortment of good amps and speakers, which permits me to experiment using the rhodes with different combinations of amps and speakers.

First off: good call on using the 12AY7 for the input stage.  Rhodes pianos work best with the lower gm triodes, just like bass amps.  12AX7s just get too distorted for the dynamic range that the Rhodes can do.

Note that the pickups in rhodes pianos are extremely sensitive to loading, much more than guitar pickups.  Enough that it will alter their tone drastically.  You used the 1Mohm input impedance which is a good start.  1Mohm is the minimum that should be used.  DI boxes can alter the tone of the Rhodes, even those with 1Mohm impedance.  The one I prefer is the Countryman Type 10 which has 10Mohm impedance.

With that in mind, note that the pickups have changed over the production era of rhodes pianos.  Be mindful of this as you are voicing the amp.  Also voice the amp using the RCA jack right off the harp assembly not just the passive stage electronics.  I had mentioned the loading of the pickups, and the passive electronics of the stage piano happen to do exactly that.  I used to own two rhodes pianos, nine years apart in age.  The older one from 1967 had very good tone, with nice round midrange and bell tone; the newer piano from 1976 would not duplicate this tone even with the same wiring configuration, it was dull and muddy no matter what I did.  I kept the older piano.

I own a silverface Dual Showman Reverb from 1976, which is the head version to the Twin Reverb exact same chassis, circuit, and transformers.  The one I have is so clean that it still has the original RCA tubes (a little old lady only played it on Sundays).  I tested the tremolo and the clicking does not exist on my unit.

My other favorite amp for Rhodes is the old "croc-skin" Selmer Zodiac Twin Thirty.  These are more prevalent in Europe but I landed one in a store in the US, they didn't know what it was.  This one has the original G12 alnico speakers and Partridge OT (Twin Thirty is the only model with these speakers).  I like to bridge the channels and use the Selectortone buttons to enhance the sound.  The Twin Thirty is a very unique preamp as the input 12AX7 has a meek plate voltage and doesn't overdrive like a Marshall, but it uses a pentode EF86 for the tone recovery stage - a configuration I have yet to see in another amp.  The tone stack is also rather unique.  This amp really loves Strats but it works very well for Rhodes too.

Here is the subtle detail that few people emulate in the Rhodes tremolo - trapezoidal modulation.  The waveshape of the signal modulating the tremolo should not be sine or triangle, but trapezoidal.  This is what the original Suitcase preamp did.  This is a signature sound of Rhodes and is very effective in stereo.  Guitar players may not like stereo tremolo, but it is an essential sound for Rhodes players.  The Twin Reverb tremolo doesn't sound the same.

You can probably convert the modulation signal to trapezoidal by using a pair of diode clippers.

Finally - Guitar tone aficionados know that speakers make a big difference in the sound of their guitar.  This is also true of the Rhodes.  Fender put some real dud speakers in the suitcase pianos, and the sound can be vastly improved by changing out speakers.  My two favorites for Rhodes are the Celestion G12 alnicos and the 75w or 65w speakers (why should guitar players have all the fun?).  For recording I like to use speaker emulators, my two favorites being the old Groove Tube Speaker Emulator (which emulates the G12) and the Red Box (which emulates 4x12 cabinet with 75w speakers).


I'd love to hear some sound samples, if you're up for posting any.



Late to the party but I have to give a kiddos. So well done. You've got to have clips by now, right? ;D


Well done. Beautiful case. Is there anyone that can make such preamp cases commercially?