Hacking the Perc Pro

[BJG145]

I've been interested in trying out the Polyend Perc Pro ever since the SOS review in August 2017.



The RRP of £1049 was a bit steep, but no-one's ever doubted that it's well engineered and fun, with the potential to play impossibly fast and intricate patterns on anything you stick in front of it.

(MIDI-driven percussion is pretty thin on the ground, the only other examples I'm aware of being the Cabot cajon which is still under development, and the Automat Toolkit by Dada Machines.)

Polyend launched V2 of their drumming tech not all that long ago, and instead of a box driving three beaters, you now have individual boxes with a single beater which sell for £400 each.



This stuff doesn't turn up secondhand very often, and isn't cheap when it does, but I recently found a V2 system on eBay for about £160 and decided to splash out.

Sadly, when it turned up, the box contained only a V1 base unit, with no beaters.

The seller didn't have them. However, he sent an apology and a full refund...and doesn't apparently want it back. (Sounds like he was just selling it on and didn't check what it was.)

I contacted Polyend who said that they didn't usually sell the beaters separately, but offered to supply them for about £270 each.

Too rich for me, though, so I'm wondering if I can hook up some solenoids to it instead.

Polyend don't supply any technical info or schematics so I'm looking for advice on where to start.

The beaters are connected via five-pin XLR-style connectors:







The electronics looks like obscure surface-mounted stuff. I only have a multimeter to use on it. I guess I could start probing pins...? I don't really know what I'm doing, but I guess I could connect one to Earth and the other to each pin in turn while playing a MIDI pattern into it? What's the easiest way to find Earth in this instance? I don't want to accidentally damage anything.

[BJG145]

...just some extra pics...


(MIDI USB/in/out/thru on the right, three five-pin beater XLRs on the left.)







The last one shows the five contact points for two of the XLR beater connections. I'm wondering if one or more is unused. The beater electronics is a mystery, but I wouldn't have thought it needed more than on/off/velocity and a power source. I'm thinking it shouldn't be rocket science to make one out of an Arduino and a solenoid but I don't know how much processing they do. I'm hoping they're relatively "dumb".

Not sure what I'm looking at, or looking for, but happy to answer questions.

[BJG145]

...just spent a little longer peering at this thing, identifying some of the components such as the Teensy 3.2 CPU and the three power transistors.

The PSU delivers 24V/10A, with connections for +V, -V, Earth, N, L.

Polyend confirmed that the beater contains a solenoid, though I don't know what else. I'm hoping that this controller can be persuaded to deliver bursts of power to the XLR outputs at 12V or 24V to drive a solenoid in sync with the MIDI input, rather than supplying the beaters with steady power and messages for them to interpret. But before testing that I just want to make sure I'm not going to blow anything up.

If I set the multimeter to measure voltage and connect one probe to the Earth connection on the PSU, and the other to the various XLR output pins, is that going to be safe in terms of avoiding damage...?

[Martin Walker]

Hi BJG145,

Tricky - a good bargain, but obviously some work involved.

I've got one of the Dada Machines Automat Toolkits, and the main PSU limitation is simply down to which solenoids you choose.

The Automat control unit can be supplied with anywhere from 9v to 24v (again, depending on what solenoids you intend to power, and I did some research into the possibilities that I posted here in the Dada forums:

https://forum.dadamachines.com/t/what-p ... er-need/23

Here's what I found, which my give you a guide to how to approach your Perc Pro choices:

"This needs to provide a regulated DC output via a 2.5mm by 5.5mm plug, with center positive.
Its output voltage can be anywhere between 9V and 24V, depending on what voltage solenoids you want to use.

However, the circuit board has a 6A fuse, so if you want to plug in the full complement of 12 solenoids, each one can only safely consume 0.5A (12 x 0.5A = 6A)

24V solenoids tend to consume less current that 12V ones (for instance), so if you want to use all 12 then buying 24V solenoids that each use 0.5A or less is the best solution. You will therefore need a PSU with a 24V output capable of 6A, which equates to 144 watts.

If on the other hand you only ever anticipate running a maximum of six solenoids then you can instead go for 12V 1A models, requiring a PSU with a 12V 6A capability (remember, you won’t be able to run any more of these solenoids, because you’ll burn out the Automat’s 6A fuse.

However, 12V 1A solenoids may be considerably cheaper, and a 12V 6A (72 watt) PSU will also be considerably cheaper than a 24V 144 watt one!"

So, look for a fuse and any voltage markings on your circuit board, and of course any spec details in the Perc Pro manual. Contact other Perc Pro users for more info if you don't have a manual, and I suspect spending some time on the Dada forums may glean you some further details.

As for choosing solenoids, you might just persuade a fellow Perc Pro user to see if they can see any markings on theirs, but ultimately your choice will probably be down to desired solenoid travel, and how 'hard' you want to hit your percussive devices.

Hope this helps!

Martin

[Martin Walker]

OK, just re-read your posts:

The PSU delivers 24V/10A, with connections for +V, -V, Earth, N, L.

If I set the multimeter to measure voltage and connect one probe to the Earth connection on the PSU, and the other to the various XLR output pins, is that going to be safe in terms of avoiding damage...?

You should be OK doing that. Have to think on about the +V and -V though - I would have thought that a solenoid would simply require one voltage, so I suspect that the LNE are for mains, and you get 24 volts across the +V and -V connections. It may be that the control circuitry is operating at a lower voltage, but the solenoids get their own higher voltage supply.

Martin

[BJG145]

OK! Thanks Martin.

The thing has a button and LED for each of the three channels. Pressing the button is like a "MIDI learn" function which assigns the incoming MIDI stream to that channel, and the LED then flashes in time. I just got that working on three separate drum tracks, so it's doing something at least...

I tried setting my UT30B to 20V and testing the various pins via a probe on its "Unfused 10A Max" connection, but got nothing except an electrical spark on one of the pins.

[Martin Walker]

Here's a link to the spec of the power transistors driving the beaters:

https://www.infineon.com/dgdl/Infineon- ... 9f6a9d4d02

Huge current rating!

Martin

[Sam Spoons]

Would velocity be voltage related? The acceleration of the solenoid iron (and so it's kineic energy) would be proportional to the voltage applied?

[Martin Walker]

Would velocity be voltage related? The acceleration of the solenoid iron (and so it's kineic energy) would be proportional to the voltage applied?

Yes, that makes perfect sense Sam

BJG145 - I've just tracked down the Perc Pro PDF manual, and it state that the solenoid movement is 5mm (fairly typical methinks), so this is starting to narrow down the sort of solenoid you need.

Martin

[BJG145]

Would velocity be voltage related? The acceleration of the solenoid iron (and so it's kineic energy) would be proportional to the voltage applied?

...ay, the only control you have over a solenoid is power and duration, so velocity must be some combination of those.

I've been interested in robotic percussion for ages and previously messed around with an Arduino and some solenoids I scrounged on this forum back in 2013. I think I got as far as tapping out a rhythm...but it was a hassle, and I didn't get onto velocity, so I'm interested in seeing how this is implemented.

In particular, I seem to remember that Arduino is all based around 5V and small currents, and it was a bit of a PITA for an electronics noob like me to get it working with the higher voltages/currents that solenoids yearn for. As well as supplying the necessary juice, this Polyend box takes care of the MIDI logic quite nicely, so it might be a useful shortcut for a DIY system.

I tried sounding out one of the Polyend techs on how complex the beaters are, but understandably they're keeping the details under wraps. The only clue was that it was an "optically steered solenoid". I'm not sure exactly what that means...the Polyend beater seems pretty tightly enclosed, so any optical components in there must be internal. I noticed that the electronics in the base unit included some "optocouplers" / "photo couplers"; I'm guessing these provide some kind of electrical isolation.

It's tricky to follow the traces on the double-sided circuit board; I might try sticking it under the office photocopier tomorrow. But it appears that one of the XLR pins is connected in common to the "+V" on the PSU. (Although "-V" on the PSU is also wired to the circuit board, I can't see any onward connections from there.)

So...I'm now thinking that the beaters are provided with a steady power level, and some kind of control signal, which I'll investigate further. Without knowing the beater components, I guess this could be as simple as an on/off switch, or as complicated as a digital message, or perhaps involve a control voltage. (The base unit has inputs for Gate/CV as well as MIDI.)

Thanks for your thoughts so far, and I'll post up any further findings; I like trying to puzzle it out.

Now I've identified the...



...pin, I might revisit the voltage readings on the others.

[BJG145]

I've got one of the Dada Machines Automat Toolkits

(PS I missed this post before - thanks for all the info. I'd like to get hold of an Automat - currently out of stock but it looks like they might be available soon.)

[Sam Spoons]

Would velocity be voltage related? The acceleration of the solenoid iron (and so it's kineic energy) would be proportional to the voltage applied?

...ay, the only control you have over a solenoid is power and duration, so velocity must be some combination of those.

Yeah my thoughts exactly, can't see how the duration could affect velocity (though it might be feasible to damp the head with the beater). As the coil resistance is going to be fixed voltage is the only thing that can give the solenoid more energy.

Good thread BTW

[BJG145]

Just been staring at this through a magnifying glass until I got eyestrain and gave up for now.

...so, each XLR has five pins and a "shield"/"case" (?) connection:

Pin 1 - Can't see a connection, don't think there is one
Pin 2 - Each is common, to V+
Pin 3 - Each is connected to one of the three power transistors
Pin 4 - Common, but can't see where they go
Pin 5 - Can't see a connection
Shield - Can't see what happens

(There are some traces running beneath the sockets, so I can't tell if they're connected or not. I find it pretty baffling, as I can only see the one connection to each of those transistors. I'll see if I can post up a pic tomorrow so someone can explain that.)

[Folderol]

That looks like a fairly modern PCB so it's quite possible it's multi-layer. This makes it an absolute pig trying to work out the track routing

[BJG145]

...it's quite possible it's multi-layer. This makes it an absolute pig...

...ah, I was just starting to wonder if it had another layer; I didn't know if that was a thing. Thanks Folderol.

[zenguitar]

Is the Teensy the only microprocessor you can spot?

The Teensy can be set up as a MIDI device so it is possible that it might be focused on the MIDI functionality rather than the broader functionality.

Andy

[BJG145]

...spent a bit more time testing between each pin and the shield this morning.

Pin 1 stays at 0V, while Pin 2 is connected to V+ at 24V.

Pin 3 shows the most promise. This delivers around 15V with a voltage drop on a MIDI pulse which varies with velocity. The highest MIDI velocity causes a drop of around 0.2V, and the lowest around 2V. The LEDs for each channel flash with a brightness which corresponds to velocity.

Pin 4 stays around 0.03V, and Pin 5 at 0.3V

I don't think I'm going to get much further with this though, so I reckon I'll wait until a secondhand one turns up one day, then attack it with a hacksaw and find out that way.

I just tried connecting up an old solenoid I had lying around to Pin 3 and it did nothing. Then I hit it with Pin 2, which certainly made it move, before it overdosed.

Dunno how complicated it would be to wire a circuit that took a 24V power supply and a signal that varied between 15V and 13V, and translated it to an output that could drive a solenoid between 0V and 12V...?

I can't seem to measure a current from the Pin 3 signal.

[BJG145]

I posted up a description on an electronics forum and someone replied that the voltage dip I mentioned sounded unlikely - which made me think that perhaps a multimeter isn't going to respond fast enough to evaluate the signal accurately, and I'd need to use an oscilloscope. I've blagged one off the Physics Dept at the University where I work, and I'll try hooking it up when I get a chance.

Meanwhile...I just had a notification that the Dada Machines Automat range is back in stock, and decided to pick up a controller before they run out again. I didn't think I'd need to get any of the solenoid / "element" packs as they seemed rather expensive, but...I dunno. Would it be wise to pick up any of their standard peripherals, or can I happily do without...?

[Martin Walker]

I posted up a description on an electronics forum and someone replied that the voltage dip I mentioned sounded unlikely - which made me think that perhaps a multimeter isn't going to respond fast enough to evaluate the signal accurately, and I'd need to use an oscilloscope. I've blagged one off the Physics Dept at the University where I work, and I'll try hooking it up when I get a chance.

Meanwhile...I just had a notification that the Dada Machines Automat range is back in stock, and decided to pick up a controller before they run out again. I didn't think I'd need to get any of the solenoid / "element" packs as they seemed rather expensive, but...I dunno. Would it be wise to pick up any of their standard peripherals, or can I happily do without...?

1. Good luck with the scope - sounds like a suitable approach, so short pulses makes perfect sense to me - perhaps pulse width determines the 'velocity' of the strike, rather than the voltage of the pulse.

2. Sensible to get the Automat - handy and cost effective little unit. I tend to agree about the peripherals - if you're handy as a DIYer you can make your own to perfectly do the job you require. I also seem to remember that the Automat designer offered 3d printer files for anyone who wanted to go that route and make their own.

Martin

[Martin Walker]

Also, there are very detailed assembly plans of the various Automat 'Adapters and Elements' that make it quite obvious how everything works.

Here's one: https://dadamachines.com/getstarted/aut ... -assembly/



Martin

[BJG145]

In the end I decided to take the easier option of flogging the Polyend controller on eBay and ordering a Dadamachines Automat instead. Still waiting for that, but I was interested to get a notification today about a new software update that now implements velocity for it, apparently by setting the gate duration.

https://forum.dadamachines.com/t/automa ... -2-0-0/326

Looking forward to finally getting my hands on some robot percussion. The new batch should start shipping next month with any luck...

[BJG145]

* another necro-bump *

I'm not very good with oscilloscopes. So I tried to sell it for a while; last time the buyer resolutely missed deliveries (and ignored emails) until it got returned to me, for a charge. (What a dick.)

But it's a case of, once SOS-bitten, forever smitten. So I've kept a lookout from the the crow's nest, biding my time...

...and I just pulled the trigger on another one on eBay; just the single-beater version. (Well, they hardly ever turn up secondhand .) It hasn't turned up yet, but I'm already agonising over whether to play it cautiously, keeping it wrapped in cotton wool when not in use...or take a bandsaw to this spheroid black box in the name of science.

(...watch this space...)

[BJG145]

...so, I bought this thing from a seller in Austria just as they went into lockdown. They were concerned that they might not be able to get it shipped out, but it turned up today. I just scissored it open, binned the outer packaging, and smothered myself in antibac.

I'm sorry, I had actually thought I might crack it open to see what was inside but...no fricking way. It's gorgeous. The beater is ridiculously heavy, beautifully engineered, and all in mint condition.



Can't wait to see what this thing can do, but the internal workings will remain a beautiful mystery.

[Martin Walker]

Always uplifting to hear some GOOD news in these troubled times.

Enjoy your new purchase BJG145!

Martin

[damu]

...so, I bought this thing from a seller in Austria just as they went into lockdown. They were concerned that they might not be able to get it shipped out, but it turned up today. I just scissored it open, binned the outer packaging, and smothered myself in antibac.

I'm sorry, I had actually thought I might crack it open to see what was inside but...no fricking way. It's gorgeous. The beater is ridiculously heavy, beautifully engineered, and all in mint condition.



Can't wait to see what this thing can do, but the internal workings will remain a beautiful mystery.

Thank you for this fantastic thread. I'm trying to make some of of my own solenoid beaters with velocity sensitivity.

Don't suppose you ever got any closer to discovering how the innards of the Polyend Beater work? I'm considering having some custom solenoids built. It would really help my project to understand what makes these so good.