Skip to content

Back from Holiday 2023

THANK YOU to all my customers for your support! I didn't have a 'holiday' as such but simply had to make time to finish off my new premises. Now I feel like a holiday!!! 😀 This week though, I'll be jumping repairs and order-builds that have come in recently so rest assured that things are back to normal.

As many will know, I hate DIY mainly because I don't think I'm very good at it. Having said that, this final phase of the project has taught me a lot of patience. Being fussy, I've learnt when to draw the line. The property is a self-build house from the seventies and no adjacent walls are quite perpendicular and no opposing walls are exactly parallel. Kind of good for a recording studio but a pain to paint, decorate and build stuff in. Of course I also learnt a lot about paint, LOL.

as you can see in the picture above, I still have a couple of things to finish off like the recoat of the other side of the door that's behind me and most obviously, I need to mount a TV screen.

Sub Bass Enclosure Just Needs A Couple More Coats

I'm delighted that my sub-bass enclosure is almost finished. A couple more coats of paint and I can think about putting in the hi-fi. I'll be doing a feature on the build of this soon. Of course it's the front that most people will be interested in.

Furniture has been ordered and while the seating will be here in a few days, I'm not expecting the bespoke coffee table before the end of the year! 🙁


Dodgy diode in Vox AC15HW

I'm taking a few weeks off next month so that I can complete my new premises (yes, it's more than two years) and I'd like to get as many units back to customers as I can before 1st September. Things like this dodgy amp wiring seriously don't help.

This Vox AC15HW came in smelling bad and the customer saying that it had started smoking. It was evident that the power transformer had fried. While doing the swap-out, I noticed the little fella (diode) pictured below, sitting uncomfortably close to a fuse. This just looked really bad so I decided to put a jacket on it (bottom picture). Better safe than sorry, eh?

Dodgy diode in Vox AC15HW Fixed

I have a feeling that this amp had been repaired before so I'm reluctant to point the finger at Mr. Vox, on this occasion. 🙂


Another busy week at Plasma Music

So far August has been well, a bit crazy. On top of an insane amount of repairs poring in, I've been busy building stuff and next week I have a meeting about a complete studio revamp. What the hell has happened?

Pictured below are 'some' of this week's builds which will all be fully tested over the weekend ready to go out on Monday.

  • At the back in red is a Supernova modular switched-mode power supply for the Juno-106.
  • In front of that and also to the left are two of Guy Wilkinson's P0004 modular switched-mode power supplies for the Roland MKS-70.
  • In front of those is the dual-stacked Nebula balanced outputs jack-board for the Roland MKS-70.
  • And In front of the Nebula and also the small PCB to the right, is an Aurora modular switched-mode power supply for the Roland MKS-80.

Another busy week at Plasma Music

Not pictured because they're already boxed up, is an Eclipse bounce eliminator, a couple of my new CR2032 battery adapters and a PML-TX01 redesigned power transformer, all for the Marshall JMP-1.

Yamaha RX5 Repair

The Yamaha RX5 is one of my favourite drum machines and when I acquired an example that wasn't 100% working, A Yamaha RX5 repair definitely seemed on the cards at some point in the future.

Every once in a while, I receive a mysterious parcel, sometimes from as far away as Japan. It took me a while but I eventually sussed out that my ex-frontman / singer, Rob would occasionally get the urge to buy me something. Anyway, a couple of years ago, one of those parcels contained a Yamaha RX5 drum machine (yes, you read correctly) that indeed had come all the way from Japan.

Being a die-hard Sisters of Mercy fan and like many, wanting to get those big Doctor Avalanche drum sounds, it was quite obvious where Rob was coming from. He'd given the game away, LOL! 😀

While much appreciated, the RX5 wasn’t in great cosmetic condition and outputs 10, 11 and 12 were dead. This wasn't a pad problem as I was getting MIDI. It was defiantly a problem with the outputs. Despite the issues however, this RX5 came in really useful for testing my Nebula balanced outputs jack-board for the Roland MKS-70. I made a special lead that connected four outputs from the RX5 to the audio inputs on Nebula, where the outputs from the MKS-70’s voice-boards would go. It worked like a charm but to be honest, I would have preferred to have a Yamaha RX5 in my keyboard room.

So, as you do, every once in a while, I’d check out the Yamaha RX5 action at the usual on-line places. Well, a few weeks ago, a RX5 popped up. Cosmetically in great condition it did have a fault and so the auction started at 50 GBP. To cut a long story short, I ended snapping it up for 108 GBP. On top of that, collection was only twenty-five minutes away from where I live. GEEEERATE!!!!!!

Yamaha RX5 on ebay
My patience paid off and I finally found a nice condition Yamaha RX5 for a very reasonable price.

When I went to pick up the unit, the seller showed me the fault and indeed as described, it simply wouldn’t boot properly, with garbled characters on the display and randomly active and sometimes blinking LEDs.

Having sussed out that the fault on the RX5 that Rob bought me, was down to a couple of faulty op-amps on the jack-board and that the fault on the unit I’d just bought was on the main CPU-board, I did the obvious thing and just swapped out the main-board in my machine with the one I’d just bought. Okay, so now I have one pretty good condition, working Yamaha RX5. That can now go upstairs into the studio. Yippee!

I put the main-board from the unit I’d just bought into the case of my old unit. Of course, it still didn’t boot properly.

Before switching on loads of test gear, I always do a quick visual inspection. If the item in question will switch on, then I switch it on. On this occasion something was definitely not right as I noticed that an analogue-to-digital converter was getting rather warm. In fact, after half-an-hour or so, it was about 50°C while adjacent devices were only a little over ambient temperature.

Normally associated with the DX range, the Mitsubishi M58990P 28-pin ADC can be found in many early Yamaha digital instruments from that period. Now-a-days it’s not too easy to source but I managed to find one that was unused and for a decent price.

Faulty components removed from Yamaha RX5 repair
The Yamaha RX5 main-board. To the right are the components that I removed; a M58990P and two M5218L dual op-amps.

Unfortunately, IC 12 (the M58990P) wasn't what was causing the malfunction. This ADC simply converts the analogue reading from the Click and Data sliders. While it had to be replaced anyway (because it was cooking), I knew it wouldn't fix the problem as it doesn't have anything to do with the computer side of things. This RX5 still wasn't booting. Damn!

I had to dig a little deeper and soon discovered that the FIRQ counter to the processor was only running at about half of what it should be. A lot of back-tracking and things were becoming quite challenging. The counter clue was great and in conjunction with the symptoms, it certainly pointed me in the right direction. Several logic chip changes later and I eventually got to the heart of the problem.

Interesting; the RX5 I'd just bought, actually had two seemingly unrelated problems; the ADC and some duff logic. Hmm... The guy I bought it off, threw in a GliGli RX5 USB card interface and I'm wondering if that screwed things up somehow.

All in all, my Yamaha RX5 repair took about three hours. While I was at it, I also changed the battery. After putting everything back together, reinitialising the unit and restoring factory data, I switched it on and YES, life and.... SOUND!

Mitsubishi M5218LYes, of course ICs fail. That's a big reason that I do what I do! While accepting the ADC going down and even a couple of logic gates, I have to admit to being quite surprised at the failure of the two M5218Ls. Remember that these were technically out of my first RX5. These humble single-in-line dual op-amps are used extensively in equipment manufactured during this period and are amazingly reliable, with a pretty good specification for the time. In fact, I can't remember the last time I had to change one of these, let alone two in the same machine.

Being ex-Roland, you wouldn't be blamed for thinking that I should almost have a natural affinity towards machines like those in the TR range and even the R-8 but I have to confess that the solid, punchy feel of the Yamaha RX5, was more my cup of tea than the rather traditional, rounded, perhaps even 'jazzy' sounds associated with Roland, at the time.

Yes, I know the RX5 is only 12-bit at 25kHz but just put the specs away for a moment and LISTEN! 😀

A busy weekend at Plasma Music Limited - 30th July 2023

Had a busy weekend just gone building a load of stuff:

1 x Aurora replacement power supply for the Roland MKS-80
2 x Supernova replacement power supplies for the Roland Juno-106
1 x AT-JX-10 FSR- based replacement aftertouch sensor for the Roland JX-10
1 x AT-D-50 FSR-based replacement aftertouch sensor for the Roland D-50

Hmm... I guess I do look a little tired!

Busy weekend at Plasma Music (2023.07.30)

Despite a busy weekend building stuff for customers, I managed to find time to  renovate this now gorgeous Yamaha RX5 drum machine.

Yamaha RX5 rebuild

Yeah, I know what you're thinking... being ex-Roland I should instinctively favour classics like the TR-808, TR-909 and TR-707. To be honest though, I've never been a big fan of Roland drum sounds. The Yamaha RX5 on the other hand, well... I just love it.

CR2032 adapter installed in Marshall JMP-1

Desperately trying to remain modest, it seems I've developed a bit of reputation for the work I've done to keep the Marshall JMP-1 going for another thirty (or more) years. Indeed, I often receive e-mails requesting I look into various aspects of this legendary MIDI valve pre-amp. One request I receive a lot has led me to develop a CR2032 adapter for the Marshall JMP-1. This replaces the factory soldered battery, with a clip CR2032 holder allowing for the battery to easily be replaced.

CR2032 adapter for the Marshall JMP-1
CR2032 adapter for the Marshall JMP-1

While readily available several decades ago, CR2032 batteries with solder tags have dropped out of fashion and even if you can find one, there's a good chance that it won't be exactly the same as the one that fits in your gear. As I often say "the nice thing about standards is that there are so many to chose from".

Original battery in Marshall JMP-1
Original battery in Marshall JMP-1.

My little CR2032 adapter for the Marshall JMP-1 takes care of that, well at least in the JMP-1. 🙂 It's not rocket science but can make life a little easier.

Unfortunately the main PCB has to be removed to fit this adapter and that's not an easy job. It's a process which requires time, patience and of course, the right tools but which is detailed in the installation manual, available after purchase. In fact, the installation manual goes into so much detail, it's ended up being the most comprehensive manual I've written for something so incredibly small!

Marshall JMP-1 main board
This bad boy doesn't come out too easily but don't worry; my installation manual is quite detailed.

To prevent tilting during battery swap, there's a couple of self-adhesive rubber pads underneath of the adapter PCB which make it very secure.

CR2032 adapter installed in Marshall JMP-1
Fits like a glove and perfectly secure thanks to the rubber pads underneath the PCB.

IF YOUR JMP-1'S BATTERY ISN'T DEAD, THEN PLEASE DO REMEMBER ONE THING BEFORE REMOVING IT: If you have stored presets in your JMP-1, you MUST back them up prior to removing the battery, otherwise they'll be lost.

If you need to reinitialise your JMP-1, here's how:

  1. Switch off JMP-1 via the power button on the far right.
  2. Hold down the OD1 button and the CLEAN 1 button.
  3. While holding down these buttons, switch on the JMP-1.
  4. Wait a few seconds while the display flashes and then release the OD1 and CLEAN 1 buttons.

Marshall JMP-1 Factory Reset

Now then, you're JMP-1 might NOT reset! Yes, that's right. If your machine is locked, performing a factory reset will be useless unless you unlock your JMP-1 first.

To check the memory protect status of your JMP-1, simply try to save a patch. If the display shows 'St L', then your JMP-1 is locked and you will need to unlock it prior to performing a factory reset.

Unlocking is simple. Just follow this procedure:

  1. Try to save a patch.
  2. While 'St L' is displayed, press the <CHANNEL> button.
  3. The unit will unlock.
  4. You can now perform a factory reset as above.


While you're here, you may find it useful to know how to dump the entire memory of your JMP-1 to a sequencer or sysex package like MIDI-Ox or SEND-SX.

  1. Just connect the MIDI OUT from your JMP-1 to the MIDI IN of your sequencer or computer's MIDI interface.
  2. If using a computer, select that port in your sysex package.
  3. Now just press <Patch> and <Volume> simultaneously on your JMP-1.
  4. THE END!

Being able to convey feel to the sound after the keys are played, can transform a performance into something quite unique and almost magical. Indeed, Roland was definitely on to something when aftertouch began appearing on it’s synthesisers in the mid-eighties. Today, many classic synthesisers have aftertouch strips that either don’t work or are a shadow of what they used to be. Almost forty years later, my modern FSR-based replacement aftertouch sensor for the Roland D-50, has considerably more dynamic range and is infinitely more reliable than the second-generation transducers that Roland originally used in the D-50 and transcends the instrument into another dimension.

AT-D-50 FSR replacement aftertouch sensor for the Roland D-50
AT-D-50 FSR replacement aftertouch sensor for the Roland D-50.

Following on from my AT-JX-8P, AT-JX-10 and AT-AJ-2 projects and as promised, development of my AT-D-50 replacement aftertouch sensor for the Roland D-50 begins today! I'm so excited. 😀

Being basically refined versions of the type of carbon-track sensors that Roland originally used but delivering a lot more dynamic range as well as other benefits, FSRs are perfect for this kind of application. In fact, modern aftertouch sensors and even some drum pads use FSRs. The principle is the same but the resistive polymer material is of a modern composition and the manufacturing process makes FSRs considerably more robust and reliable than previous pressure sensors.

How FSRs work
Cross-section of a force sensitive resistor (FSR) illustrating how they work.

So, last night I picked up the instrument that's going to be the first to have my AT-D-50 replacement aftertouch sensor for the Roland D-50 installed. It's in lovely condition and really does deserve something special.

Donor for AT-D-50 replacement aftertouch sensor for the Roland D-50
Here's the guinea pig for my AT-D-50. An almost pristine example of this classic synth.

Before I opened up this D-50, I thought I'd reacquaint myself with the aftertouch buffer / amp circuit. Wow, a little more involved than pervious synthesisers and Roland started using SMD passives. That's gonna be fun!

Of course, another interesting issue is that if I remember correctly, the D-50 was the first Roland Synthesiser to have everything, including the keyboard chassis, secured to the inside of the top-case. Unscrew the bottom case and you're in and there's nothing bolted to it.

Roland D-50 with new upside-down assembly
Roland D-50 with the new 'upside-down' assembly. In  I guess in comparison to building half the synth on a piece of wood and then connected to it the other half which was built into metal / plastic, this approach saved a lot of manufacturing time in 1987.

This aspect of the project presented some challenges which made me consider that the AT-D-50 might be my first FSR-based replacement aftertouch system that would be supplied in kit form. Here's why...

Just about every keyboard that features aftertouch and that I've worked on, has a small notch cut into the keyboard side of each end-cheek, just underneath the key-line, which allows for a margin of error in the length of the sensor. I take full advantage of this notch as however hard I try, I cannot cater for extremes of assembly / construction tolerances in a machine that's almost forty years old.

While the D-50 and similarly constructed synthesisers, also have this notch, it's quite small and things can be tight.

Aftertouch sensor notch in D-50 case left
Note the notch cut into the inside of the left-hand end-cheek. There's one on the other side, too. They allow any excess aftertouch sensor to be neatly slipped out of the way. My sensors are a little wider than the original so I have to engineer things very precisely as I can't use this and I definitely don't expect customers to make this larger!

The idea of having to provide one of my aftertouch systems only in kit form didn't appeal to me at all. Many of my customers have a good degree of competence with electronics but the issues here were really mechanical. I didn't want to exert unnecessary pressure on those who already have a potentially challenging installation job ahead of them and so I had no option other that to persevere and find solutions to the problems.

Well, it took me a couple of weeks but I got there in the end. Once word of my replacement aftertouch sensor for the Roland D-50 got out, amazingly I had another D-50 come in which allowed me to confirm my measurements and installation procedure.

Similar to my other FSR-based aftertouch sensors, my AT-D-50 replacement aftertouch sensor for the Roland D-50 comprises two force sensitive resistors. The terminals of each sensor are passively connected on a little FSR Aftertouch Interface (FAI) PCB which is supplied with the AT-D-50.

FAI D-50 FSR aftertouch interface PCB
FSR Aftertouch Interface (FAI) PCB combines the outputs of each FSR and provides a simple way to select one of two sensitivity ranges.

The output from the FSR combination is quite different to the original carbon-track based system that Roland used in the eighties. FAI compensates for that difference but also offers something else.

FAI provides the facility to select one of two sensitivity ranges. Yes, that's right! What a cool little feature. 🙂 On top of that, the aftertouch control next to the volume control on the top of the D-50 still works as it did before.

With the aftertouch fader on the D-50 set to about 80%, a MIDI velocity of about 90 and an average pressure, responses were recorded into Cubase with FAI in Norm and Hi modes. As you can see, a MIDI aftertouch value of 127 is reached in both modes. It's just that you get there quicker in hi mode.

Don't forget that you still have control over how aftertouch interacts with each sound within the patch settings.

FAI D-50 MIDI response in norm and hi modes
FAI D-50 MIDI response in norm and hi modes

The 'upside-down' construction (as I call it) of the D-50, means that it's simply not practical to mount FAI to the bottom-case. With so little room on the inside of the top-case, FAI's four self-adhesive feet safely secure it to the inside of the D-50's left side-cheek. I've tried to make things as ergonomic as possible but of course the bottom-case has to be removed to access FAI. Positioning FAI on the inside left side-cheek however, does offer an excellent degree of access to FAI once you're in so switching aftertouch sensitivity ranges in the future, won't be too difficult.

FAI for the D-50 required a redesign of previous versions. Being side-mounted, I decided to use straight (vertical) Molex connectors, for example. It just makes life easier.

Also, up until FAI D-50, the facility to switch between two aftertouch sensitivity ranges was done by changing the value of the feedback resistor in the aftertouch buffer op-amp circuit. In FAI D-50 however, I decided to do the same job by switching the value of the series resistor between the FSRs and the op-amp.

Remember I mentioned earlier that the D-50 has a lot of passive SMDs? Well one big bonus of adjusting aftertouch gain by varying the series resistor to the buffer op-amp, is that no components on the bender-board need to be replaced. In fact, the bender-board doesn't even need to be removed. All you need to do is solder a two wires to the bender-board.

Roland D-50 bender-board
The bender-board in the D-50 is packed with SMDs. FAI D-50 uniquely delivers the option of switchable aftertouch sensitivity ranges, without you having to touch them!

You've developed an FSR-based aftertouch sensor for one synth (in fact, three) and you think it would be easy to knock up an FSR-based aftertouch sensor for another. Hmmm... not quite. Each synthesiser has it's own challenges and I'm so glad I persevered with this project. I'm delighted that AT-D-50 not just works brilliantly but is now a system which can be installed by anyone with a little patience and technical competence.

With an abundance of sensitivity and very high dynamic range, my replacement aftertouch sensor for the Roland D-50 truly gives this classic synthesiser, a whole new lease of life. It's like suddenly having velocity sensitivity! 😀 The added expression is just magical.

Modern FSRs are very reliable and it's not uncommon to see specifications quoting figures such as 'more than 10,000,000 actuations'. Being sealed units, FSRs are vulnerable to contaminants and oxidation problems like the old carbon-track sensors. It's no surprise then, that modern instruments that feature aftertouch, use FSRs. As previously mentioned, FSRs have a much greater dynamic range than their carbon-track counterparts which makes them ideal for modern drum pads.

I sometimes get asked to provide audio samples but I've backed off from doing that. Aftertouch isn't a sound, it's the result of a feel. You don't 'hear' aftertouch. You hear the effect of aftertouch. So posting an audio file with no reference to how the synthesiser was played, just seems a bit pointless to me. What I can tell you, is that the feedback I've had from users who have installed my FSR systems is humbling and inspiring. For a couple of hundred bucks, it's really worth a shot.

Roland D-50

AT-D-50 is now available to purchase here:

Being able to convey feel to the sound after the keys are played, can transform a performance into something quite unique and almost magical. Indeed, Roland was definitely on to something when aftertouch began appearing on it’s synthesisers in the mid-eighties. Today, many classic synthesisers have aftertouch strips that either don’t work or are a shadow of what they used to be. Almost forty years later, my Roland Alpha Juno 2 replacement aftertouch sensor which uses modern FSR technology, has considerably more dynamic range and is infinitely more reliable than the second-generation transducers that Roland originally used in the AJ-2 and transcends the instrument into another dimension.

Doner test bed for replacement aftertouch sensor for the Roland Alpha Juno 2

A couple of months ago, I received two Roland JX-8Ps with amongst other issues, failed aftertouch sensors. It was a great opportunity for me to try out something I've been meaning to do for a few years now. In no time, I'd developed a force sensitive resistor (FSR) based aftertouch system for the JX-8P. My friend Guy Wilkinson then lent me his JX-10 and 'AT-JX-10' appeared a couple of weeks later. After launching these products, something unexpected happened; I got contacted by a few AJ2 owners asking if I had considered developing a Roland Alpha Juno 2 replacement aftertouch sensor. Well, I hadn't but hey, here it is! 🙂

Roland launched the Alpha Juno 2 a year or so before I joined the company in '86 / '87 but like most music tech' at the time, instruments like the MKS-80, Juno-106, JX-8P and JX-10 were the ones that seemed to pop my eyes. As such, I never paid much attention to the Alpha Junos until many years after I'd left the company.

Roland Alpha Juno 2

Today, the Alpha Juno 2 is a much respected synthesiser which has also proved to be very reliable... apart from the aftertouch. It was kind of a no-brainer therefore, to sort this out, especially following on from the success I'd recently had with the JX-8P and JX-10 replacement aftertouch sensor projects. There was only one problem; I didn't have a Roland Alpha Juno 2. 🙁

The development costs of designing peripherals for vintage equipment can be considerable and I was reluctant in this case, to whack on a good sum for a second-hand AJ-2. Having said that, I very almost did. But then...

Jumping to the rescue, my friend Guy Wilkinson suggested he contact a friend of his, who's an Alpha Juno 2 owner and indeed Jonathan was only too happy for me to use one of his his synths as a test bed. Yes, Jonathan has two Alpha Juno 2s!

So now I had a Roland Alpha Juno 2 but...

You need to be careful when taking apart an Alpha Juno 2 because the front of the main-board isn't simply screwed to any part of the chassis. Instead, Roland decided to have it supported in slots underneath the keyboard chassis. Removing the keyboard chassis without knowing this and hence, taking precautions, could easily snap the main-board! 🙁

Slots in the keyboard chassis supports of the Roland Alpha Juno 2
Here's the underside of the AJ-2's keyboard chassis. Highlighted are slots that the front of the main-board slide into.

The original aftertouch strip looks kinda FSRish but it's not. Similar to the sensor found in other synthesisers at the time, it's simply a couple of plastic strips with carbon tracks sandwiched together.

Original aftertouch sensor in Roland Alpha Juno 2
The original aftertouch sensor looks pretty hi-tec for the mid-eighties.

The Alpha Juno 2 is a slim instrument and wherever you look, space is limited. Even the front of the keyboard chassis is narrow. My replacement aftertouch sensor for the Roland AJ2 is 20mm wide and I just got lucky that it fitted.

My Roland Alpha Juno 2 replacement aftertouch sensor only just fitted
Not much space for a replacement aftertouch sensor on the Roland AJ-2.

Being basically refined versions of the type of carbon-track sensors that Roland originally used but delivering a lot more dynamic range as well as other benefits, FSRs are perfect for this kind of application. In fact, modern aftertouch sensors and even some drum pads use FSRs. The principle is the same but the resistive polymer material is of a modern composition and the manufacturing process makes FSRs considerably more robust and reliable than previous pressure sensors.

How FSRs work
Cross-section of a force sensitive resistor (FSR) illustrating how they work.

Having done this twice already and very recently and being very familiar with this synthesiser, developing a replacement aftertouch sensor for the Roland Alpha Juno 2 mainly involved measuring up and fine-tuning my FSR Aftertouch Interface (FAI) PCB.

FSR aftertouch interface PCB for Roland Alpha Juno 2 replacement aftertouch sensor
My FSR Aftertouch Interface (FAI) PCB has a cool little secret on-board; switchable aftertouch sensitivity ranges!

What I mean by 'fine-tuning' is that FAI provides the facility to switch aftertouch sensitivity ranges (what a cool little feature). By simply moving a jumper on FAI, you can select between 'Norm' or 'Hi' sensitivity. Hence, I need to play around with some resistor values until things feel right in both modes.

Below you can clearly see the difference between the the two modes. In Norm mode, aftertouch responds smoothly but requires more finger pressure to reach full scale deflection (MIDI value 127). In Hi mode however, aftertouch gets to full scale deflection a little quicker and of course, requires less finger pressure to achieve this.

FAI AT-AJ-2 MIDI Responses
With hits of around MIDI velocity 95 and subsequent average finger pressure, you can see the difference between FAI's Norm and Hi modes.

In the AJ-2, aftertouch can be assigned to DCO, VCF or VCA and you have scope for further adjustment of the affect of aftertouch on the respective parameter.

Like all of my replacement aftertouch sensors, the AT-AJ-2 employs two FSRs and it's not practical to simply connect them to the original aftertouch wires and so FAI also makes life easy by passively combining the signals off the FSRs.

Although quite small, again with limited space inside the chassis, the positioning of FAI had to be carefully considered. FAI couldn't go behind the pitch-bend assembly like my AT-JX-8P and AT-JX-10 installations.

FAI installed in Alpha Juno 2
FAI mounted close to the main-board and not behind the pitch-bend assembly as with other FSR-based aftertouch systems I've done.

Different to previous FSR-based aftertouch projects, designing and installing my AT-AJ-2 presented some unique and interesting challenges. It wasn't quite the 'walk-in-the-park' I thought it would be, that's for sure. The Alpha Juno 2 was for example, particularly fussy about the height of its aftertouch strip. Having already designed a couple of FSR-based replacement aftertouch sensors, the height of the whole sensor assembly is of course, one aspect I pay a lot of attention to. If the sensor is too high, when a note is pressed, contact with the second carbon nipple in the contact bubble might not happen and you'll get no sound. If the sensor assembly is too low, you'll potentially wear out the key contacts quicker. Seriously, a fraction of a millimetre can make a significant difference.

Wanting this system to be available to all AJ-2 owners however, meant that not only did I have to overcome these challenges but I had to do so in a way that any technically competent person could easily follow. Operating on vintage synthesisers requires patience and an appreciation of how the manufacturer did things back in the day. You can't just take things for granted!

"So what does my Roland Alpha Juno 2 replacement aftertouch sensor sound like?"

Well, that's actually the wrong question! With a dynamic range that's quite literally in a different league to the second-generation transducer that Roland used in the Alpha Juno 2, my AT-AJ-2 feels just magic! The new system delivers a conversion of human touch to sonic expression that is perhaps what Roland had in mind in the mid-eighties but was unable to realise at the time.

If you want to know more about the fabulous Roland Alpha Junos, then you might find this interesting reading.

UPDATE - 10th July 2023

Yesterday my friend Guy Wilkinson came to see me to collect Johnathan's Alpha Juno 2 and to catch up about a whole bunch of other stuff. Not knowing too many people who speak my language, it's great having Guy over. When we meet up, it kind of gets a bit geeky 😀 but it's so much fun!

Anyway, Guy messaged me later in the evening with a link to his home page. Here's what he had to say:

"Repeat 9th July 2023: Alex did it again for the Alpha Juno 2, such a sublime upgrade because the synth is such a lovely sounding and traditional polysynth (less is more)……a great Aftertouch performance really makes it shine expressively. I have spent a lot of time with it, enjoying the synth in a new way. Shame I have to hand it back 😁

If you are interested in one of these fabulous kits with an extremely long life modern sensor, then check out Alex’s site pages:  JX-10 ,  JX-8P & Alpha Juno 2.

Bring the expression back into your Vintage synthesizer."

Thank you for your kind words, Guy.

A special  T H A N K  Y O U  to Jonathan W for lending me his Roland Alpha Juno 2.

UPDATE - 28th July 2023

My AT-D-50 FSR-based replacement aftertouch sensor for the D-50 is a go! You can read all about here.

Roland D-50

I'm sorry to say this but I'm not exactly a big fan of modern guitar amps! 🙁 So sorry but I think there's far too much stuff going on for something that should in essence, be quite simple with the important stuff being well... seemingly, not so important! The fact that I had to do this EVH 5153 Mk3 screen grid circuit rebuild kinda says it all.

This amp came to me because it was blowing fuses. The power valves were fried but it didn't look like 'normal' wear was responsible so I decided to open up the amp.

Unfortunately, I didn't take any pictures of this prior to changing components. The image below however, will give you an idea of where the power amp circuit board is (top right). On the other side of this small PCB is a pair of 6V6 valves.

Inside of EVH 5153 Mk3

The schematic marks resistors R139, R140 and R157 as 'fp' meaning 'flame proof' which is rather ironic as they were burnt out!

EVH 5153 Mk III Schematic showing crucial screen grid, cathode and plate components
EVH 5153 Mk III Schematic with crucial screen grid components highlighted in red and cathode to ground and V8 plate components shown in blue.

The reason that R139 and R140 were failed (in my humble opinion) was that they were underrated. I therefore chose to replace the original 1W resistors with 5W wire-wound versions. Luckily, there was enough room on that little PCB to take them. I also mounted the new components a few millimetres off the PCB.

R157 is interesting. In many amps, this resistor doesn't exist and the connection is a short. Similar to a fuse however, a 1Ω resistor in this position (between the valve cathodes and ground), helps protect the output stage of the amp, in the event of a valve failure. The current flowing through the resistor is less than 50mA and substituting the original component with one of a higher power rating would negate this cool little safety feature. Hence, I decided to drop in a similar 1Ω / 1W device.

The schematic doesn't specify the tolerance of R157 but here's why I chose a high tolerance substitute:

A high tolerance 1Ω resistor (say 1%) in place of R157, also makes bias current measurements easier. Whilst not mentioned on the technical literature, you can clearly see that 'BIAS' is actually written on the PCB next to R157. By measuring the voltage drop across the resistor and then using Ohm's law (V = I x R), it's easy to suss out the bias current. Since R = 1, the current will be the voltage that's read across R157 because I = V / R or I = V / 1).

D29 was also fried. This diode is a plate diode for one of the valves. It may have burnt out because R157 which sits right next to it, overheated or because V8 started pulling more current due to the failing screen grid resistors. Now that the screen grid resistors have been changed however, I think this'll be okay with a standard 1N4006 direct replacement.

If you have one of these amps and are considering a similar EVH 5153 Mk3 screen grid circuit rebuild, please take all precautions. Valves amps bite!

EVH 5153 Mk III screen grid, cathode, plate circuit with upgraded components
Highlighted in red are uprated (5W) screen grid resistors. Highlighted in blue are replaced 1Ω / 1W resistor R157 in cathode circuit and 1N4006 (D29) plate diode of V8.

It's difficult to buy a newer amp that doesn't have a printed circuit board (PCB) but I feel that designers need to take care.

While currents are low, valve amps have hundreds of volts all over the place and to route these voltages on PCB tracking worries me a little, especially when the layout is pushed for space. High-voltage carrying tracks often end up being very close to other tracks which in my humble opinion, isn't good.

If we take a closer look at this 5153 Mk III for example, carrying several hundred volts, the track on the positive side of D29 (that's going to the plate of V8) really doesn't seem very substantial and on top of that, it seems a little too close for comfort, to the ground side of R157. 🙁

Sorry for the rant!

Welcome to my Galaxy! Initially a proof-of-concept, Galaxy ended up being the ultimate Behringer DEQ2496 replacement power supply!

Galaxy replacement power supply for the Behringer DEQ2496
Galaxy replacement power supply for the Behringer DEQ2496.

I’ve been a big fan of this processor since Behringer launched the 2U DSP8000 Ultracurve which soon evolved into the DSP8024 Ultracurve. It's no surprise therefore, that I'm currently using six Behringer DEQ2496 Ultracurve Pros in my studio. An excellent processor which in its current version, offers features and benefits such as

  • Compact 1U format
  • Ergonomic front-panel layout
  • Intuitive GUI on a large display
  • Simultaneous multi-function processing
  • Huge audio connection options, both analogue and digital
  • Professional interfaces including Word Clock
  • Excellent audio quality
  • Analogue true bypass
  • Internal memory
  • MIDI to allow remote control and saving of internal memory to a computer
  • Amazingly cost-effective solution
One of six Ultracurves in my studio
One of six Ultracurves in my studio together with another couple of Behringer 'Swiss army knife' type processors.

The Behringer DEQ2496 joins a selected few processors that can be found in recording studios, broadcast studios, night clubs, theatres, in the filed, audiophile set-ups and home theatre systems. That in itself is a big deal.

So why are we here? What’s wrong with the DEQ2496?

Well, nothing really but despite its huge feature list, the fact that at the time of writing, it’s still in production and retails for a very acceptable price, it does have one or two snaggy annoyances.

Made in the Far East, quality control used to be and sometimes still is an issue. Problems with internal cables for example, have been well documented. In fact, I’ve repaired a lot of Ultacurves over the years and indeed many have had simple inter-board connection issues.

Great but that’s still not why we’re here!

With thousands of Ultracurves across the world now getting quite old, many DEQ2496 power supplies are starting to fail. Symptoms are various ranging from erratic or unpredictable behaviour, to the DEQ2496 simply not switching on.

Yeah but hang on a second... is the power supply in the DEQ2496 actually a Behringer power supply? 

Good question! Labelled as a Behringer Model PSU2496, it would appear that it's actually made by Eton, a well respected manufacturer of switched-mode power supplies and is from the ET166 range.

Behringer PSU2496 or Eton ET166
The Behringer PSU2496 power supply or is it an Eton ET166?

The PSU2496 is a very cleverly designed little power supply. Labelled as a switched-mode power supply (SMPS), it's actually a hybrid. The voltage for the +/-15V analogue supplies for example, comes off a pair of 78 / 79 series regulators. The input to those regulators however, does come from a very fast-switching back-end. It's a trick I've used myself, to make guitar pedal power supplies and it works really quite well. Using LDO regulators on the back-end of a SMPS is an excellent way to reduce EMC and ripple but is expensive for manufacturers to implement.

Behringer PSU2496 uses linear voltage regulators for main audio supplies
The Behringer PSU2496 uses linear voltage regulators for+/-15V audio supplies.

The PSU2496 is packed with a bunch of similarly neat little features including proper capacitive ground / earth decoupling which I have incorporated into the Galaxy design.

From my experience, these things don't fail because of poor design, sub-standard components or build quality. They fail because they're old. A lot of heat is generated by the power supply and devices will eventually succumb to thermal stress. It's that simple. When it comes to current provision, it's difficult to tell what margins exist between the PSU2496 rated currents and what the DEQ2496 pulls off each supply. One could argue that the reason the PSU2496 runs so hot is because there isn't much margin at all and that the power supply is actually running near to the edge of its limits.

In the event of PSU failure, swapping out the through-hole electrolytic capacitors for high-temperature rated, low ESD equivalents, can help but sometimes other components fail and fault-finding a switched-mode power supply can be quite challenging, even with the appropriate technical literature. The wire-wound components are of particular concern as it's virtually impossible to acquire these as spare parts.

Initially a proof-of-concept, my Behringer DEQ2496 replacement power supply wasn’t intended to become a commercial product. Once it was up and running, installed in two of my own Ultracurves and having worked flawlessly for a couple of months, I mentioned the project to some friends and customers who I knew were DEQ2496 owners. Within a couple of weeks, three customers had brought me their Ultracurves asking if I could install my Behringer DEQ2496 replacement power supply. That’s when things started to get serious and Galaxy was born.

Galaxy replacement power supply for the Behringer DEQ2496 fits like a glove
Using the same mounting points as the original PSU (and a couple of PCB stand-offs), Galaxy fits like a glove.

Had I designed a power supply similar to the Behringer PSU2496 (or Eton ET166), I don't think I'd have come up with anything much better, to be honest. As I've already commented, it's actually a good power supply! Anyway, fortunately there's enough room in the DEQ2496 case to develop something a little special and what this awesome processor really deserves.

Galaxy is also a SMPS but with several differences to the Behringer PSU2496.

Unlike the PSU2496 with all its interdependencies, Galaxy offers fully independent voltage supplies, with only pre-filtered mains, common to all.

Using low-leakage AC / DC converters which are both over-voltage and over-current protected, the heat generated is much, much lower than the original PSU2496. During tests, I measured the temperatures of the converters an hour after power-up and with the top-case on. The hottest measured only 38°C. To put things into perspective, one of the processors was about 42°C.  It's quite reassuring that Galaxy is running cool and not significantly contributing to the heat build-up inside the Ultracurve as was its predecessor.

Galaxy replacement PSU for the Behringer DEQ2496 is an easy installation
I wanted Galaxy to be an easy installation, something that anyone with a little technical competence could achieve.

Switched-mode power supplies shouldn't generate hum like linear power supplies. Due to the nature of operation however, they can / do generate very high-frequency noise. When designing switched-mode power supplies intended for use in audio electronics, comprehensive noise filtering is paramount. To that end, Galaxy has precision-designed filters on the back-end of each AC / DC converter, including those that supply 5V and 3.3V for the digital side of things (something you don't see too often).

Galaxy has precision designed filters on all supplies
Galaxy has independent voltage supplies and each is individually filtered.

Each supply also has its own status LED giving a simple visual indication that respective supply lines are working.

Galaxy status LEDs let you know what's going on
Galaxy glows in the dark! Independent status LEDs show you that things are working. It's going to be bright in that box...

Voltages are directed to a rather over-sized 2.54mm (0.1 inch) pitch Molex KK 259 header and Galaxy is supplied with a connection cable to go to the Behringer DEQ2496 main-board.

Molex to JST connector from Galaxy to DEQ2496 main-board
7-way Molex to JST connector from Galaxy to DEQ2496 main-board is included with the kit.

Like all my modular switched-mode power supplies, there’s very little exposed mains voltage on the top-side of the PCB.

Galaxy mains filtering
Galaxy has minimal exposed mains on the top-side of the PCB. There's no need to upgrade the Ultracurve's fuse and filtering ensures a clean and healthy mains AC is fed to the AC / DC converters.

In fact, safety always comes first so again in common with all my modular switched-mode power supply designs, Galaxy features a bleed resistor across the already fuse protected mains supply.

Galaxy bleed resistor is underneath the board
To reduce exposed mains on the top-side of the PCB, the bleed resistor is mounted underneath the board.

With safety still in mind, Galaxy is secured to the inside of the DEQ2496 chassis using the same mounting points as the original power supply. Stand-offs around the two connectors, reinforce the PCB and prevent the PCB from bending when connecting the mains supply from the switch and the connector from Galaxy to the main processor-board.

Galaxy mounting to DEQ2496 chassis
Galaxy uses the same mounting points as the original power supply and is very secure.

After measuring the current consumption of each voltage line, I wanted my Behringer replacement power supply for the DEQ2496 to have as much headroom as possible. One of my initial design objectives therefore, was that each voltage supply of Galaxy should be able to offer at least twice the current that is required.  Taking advantage of the space available in the DEQ2496 case, I was able to fulfil that objective. Not stressed and with ample headroom, this is one reason why Galaxy runs so cool; the power supply can actually deliver considerably more power than is needed.

Galaxy uses British made Vigortronix AC DC converters
Galaxy uses high-specification, low-leakage, British made Vigortronix AC / DC converters with built-in over-current and over-voltage protection.

I'm so sorry to have to say this but one potentially big issue with Far Eastern manufactured equipment that's powered from mains, is the lack of standardisation, compliancy and safety of the power inlet. Quite frankly, it's all over the place. Live and neutral are often swapped around and it's very common for the power switch to be connected to the neutral line. This is really frustrating when it comes to confirming which line the fuse is in, for example.

"So what's the big deal?" I hear you ask.

Hey, don't take my word for it. The following was taken from some PAT testing guidance I found on line here.

Under normal circumstances, the current will go to the appliance down the live wire first passing through the plug fuse (and any internal fuse).  If the appliance has a fault and draws too much current, the fuse would detect this and blow.  If the live and neutral wires are crossed over, the current passes down the neutral first.  The result is the equipment user could be at risk if the appliance has a fault.

I therefore include a couple of bits with Galaxy like an insulating boot that fits over the back of the IEC socket. I also recommend that while the box is open, you check the input mains polarity and correct it if it's wrong.

Mains power input to the Behringer DEQ2496 can be non-standard
Galaxy includes a kit to help make the DEQ2496 a little safer. All required information is included in the installation manual.

Behringer has a well established reputation for making affordable audio gear. It doesn't always mean however, that quality needs to be compromised. The DEQ2496 Ultracurve Pro epitomises that fact. Competitively priced, it's difficult to find something similar. My Galaxy replacement power supply for the DEQ2496 will ensure that at least the power side of things will run smoothly and last for a very long time. 🙂

Replacement power supply for the Behringer DEQ2496 Ultracurve Pro
Replacement power supply for the Behringer DEQ2496 Ultracurve Pro.

Installing Galaxy is amazingly straight-forward for anyone who has experience with opening up 19-inch rack gear. Personally, I would check the polarity of the mains live / neutral, correct them if they're wrong and check that the ON / OFF switch is switching live and not neutral. Other than that, there's no desoldering / soldering, just four screws to remove, four replacement (longer) screws to fit, a couple of connectors to pull off the old PSU and main-board and then connect Galaxy.

Plasma Music Limited - I'm deeply concerned about the environment and the exploitation of labour and so  I always use local manufacturers in preference to the Far East, with the following in mind:

  1. I can be confident that workers are treated fairly and earn a proper wage.
  2. I can be confident of the standard of quality of each item that is delivered to me.
  3. Communication is important and using local manufacturers, all correspondence is quick and understandable.
  4. I believe in supporting the local economy.
  5. I can be confident that the disposal of manufacturing waste is managed properly and in accordance with national and EU law.
Galaxy replacement power supply for the Behringer DEQ2496 is made in the UK
Galaxy replacement power supply for the Behringer DEQ2496 is made in the UK.

Using local manufacturers isn’t the cheapest option but the above points are important to me. I hope that they’re important to you too.