Echodeck - Quad PT2399 delay/reverb/chorus

[mth5044]

This is the block diagram ~arph had posted in one of my threads about the echorec with pt2399's. I had mine a little more than halfway done, but now that ~arph has finished his, I feel the need to change mine totally around to look more like his 

~arph's block diagram


Notes are that the compander/expander are gone and replaced with a downward expander.

Care to explain what you are using for F1/2, what is C1 and what you are doing for mixing? You have been so much more than kind with sharing all this info. I can't imagine the amount of time it will take to make the layouts and schematics.. this thing is HUGE  Luckly my head schematics weren't to different from yours. The switching still has me totally baffled though.

[~arph]

yes that diagram is what I started off with, and I have to redo it because it is not what I ended up with. F1 for instance is nothing but a simple low pass filter, and it is now afer C1, which is the feedback level control. The Comp stage is gone.
F2 is the SWTC2 from jack orman with values  10k 39k and 10nF and a 250k pot. Mixing is just a simple summing amplifier stage. Ala the mixer from GGG.

[alex frias]

Really great project!

[chris_d]

I love the sounds that you are getting out of this!

Very very cool project!

[snarblinge]

i'm up for a pcb if it gets to that stage. plus more info around the button matrix, I'm just not getting how that works.

such a nice job too. wow

[~arph]

Here is how the button interface works.

Each button switches between 0 and 5 volts. The output of that (0 volts when not pressed, 5 volts when pressed) goes into a voltage divider where all buttons are connected. Like this:

Code Select Expand



but 1 --[  1k   ]-----|
but 2 --[  3k9  ]-----|
but 3 --[  5k6  ]-----|
but 4 --[  8k2  ]-----|
but 5 --[  10k  ]-----|---- voltage out
but 6 --[  27k  ]-----|
but 7 --[  39k  ]-----|
but 8 --[  56k  ]-----|
GND   --[  10k  ]-----|                           


This way each button press will give a different voltage at the voltage out. So Depending on the voltage I can identify which button was pressed.
The voltage out goes to one of the arduinio's analog inputs. In the arduino I have eight voltage windows set up, by trial and error. Each read analog value is compared to this table of values to determine the pressed button.
This setup also means that you can only press one button at a time, but that's not a big deal to me.  You could use the digital inputs on the arduino to detect individual buttons too. this would be a bit more reliable.I chose this because of the fact that I want to control the leds separately too. Plus I have to do the switching. And finally I wanted to keep i/o pins for future MIDI control (easy too with the arduino)

The LED's are switched on and of using a CD4015 serial in parallel out shift register. I shift in the desired byte and all 1's light up an 0's dont. The same byte is shifted into the MAX395 to switch on the corresponding heads for playback and feedback. The only difference is that the MAX395 has a Chip select option so you can shift in the byte and then load the swithces in one pulse.
With the led's you can see the byte being shifted through the register (not obvious at all, because of the clock rate)
note that with the 4015 you have to connect one input and one output and share a clock line as this is actually two four bit sipo's in one IC.

The first four bits set the playback heads and the last four the feedback heads. So if I want to playback using heads 2 and 4 and feedback on head 4 I shift out this byte:  01010001
That is actually all there is to it. The clock pulses, data lines and CS signals for the max and 4015 come from digital outputs on the ardiuno.

One nice feature of the arduino is that it has an EEPROM where you can store bytes. So if the device is switched of, and later on again it will restore to the last selected head setting.

[snarblinge]

thanks, i didn't realise you had an arduino in there. a mate of mine is building a robot arm for animation with one of those controlling it.

I'm keen to build one of these, been looking for a larger project to sink my teeth into.
i guess the updated block diagram with all the bits labeled will make this a bit easier to organise, tho it will take me so long to put the pieces together I should just sit back and wait for developments. love your work.

[funkbass187]

i just bought an arduino and was about to make an echo base... i think my plans have just changed.

[Brymus]

I for one would love to see Arph put together a kit for sale.
I really doubt that was his intentions when he made this,but the result is just so cool.
A PCB,and the less common components altogether would be very nice to get all at once.

I know I would get one when the funds came available for me to do so.

[JKowalski]

Thanks a lot for giving us an overview! I see that you are going to optimize it for later versions, it looks like you can bring it down quite a bit.

I'm pretty sure you can bring the 5V regulators all together as long as you keep a small bypass cap near each PT2399 and digital chip, along with careful PCB layout like separating digital and analog grounds. Also, the buffering TL074's can be shrunk as well, do you really need that much buffering? The op amp in the PT2399 should keep the signal strong enough to go through the switching chips and spread into the mixers but I didn't build it so I can't say I have the experience there   Did you put the buffering in as a precaution or did you experience problems without it?

Also, for structural advice perhaps once the board is optimized you can use spacers to mount the entire PCB to the button board's 6 screws, so it is like a circuit sandwich. I'm not really sure how you did it since I can't see behind the perf. Looks like you can shrink the board space down to a final 30% or so of what you currently use, if you sacrifice the easy to work with nicely spaced layout you have. (maybe resort to upright resistors... or SMD!  )

For connecting the button boards to the main board, I see that there are four holes on three sides... There is a method I saw on a circuit sandwich before where they used long pin breakaway headers to connect the top board to the bottom board (think having a hole in the same place on each board and running a wire between them vertically, but with nice looking headers. It's not easy to take apart but it looks really nice. Too bad the button pins aren't all on one side, that would make it much easier to just run a 12-pin header....


Anyways I hope I don't come off as nitpicking or anything just giving suggestions. Like I said, you did an awesome job.

[~arph]

Thanks a lot for giving us an overview! I see that you are going to optimize it for later versions, it looks like you can bring it down quite a bit.

I'm pretty sure you can bring the 5V regulators all together as long as you keep a small bypass cap near each PT2399 and digital chip, along with careful PCB layout like separating digital and analog grounds. Also, the buffering TL074's can be shrunk as well, do you really need that much buffering? The op amp in the PT2399 should keep the signal strong enough to go through the switching chips and spread into the mixers but I didn't build it so I can't say I have the experience there   Did you put the buffering in as a precaution or did you experience problems without it?

Thank you for thinking with me. I checked the max current rating coming out of the 5v regulator and calculated that I could in theory power all pt2399's on a single regulator, but I want to be safe so I doubled all values so I would need one regulator for each two PT2399's. Also bear in mind that the arduino runs on 5V as well. Right now it shares a regulator with one of the PT2399's. So in the end I might cut it down to two/three regulators. OR: 1 regulator into a voltage follower, if that can supply more current.

As for the buffering, yes, I did try it without and let me explain why it is needed. When no buffering is used, the feedback and playback lines are tied together. This would mean that if you select for example heads 1 and 2 for playback, those two feedback points are connected too (through some resistors) This is fine if you then select head 3 or 4 for feedback, but if you select 1 or 2 you will allways select both 1 and 2. The feedback will self oscillate. Now imagine doing that with four playback heads (wich you will use a lot). So the buffers are there to ensure proper playback/feedback functionality.

Also, for structural advice perhaps once the board is optimized you can use spacers to mount the entire PCB to the button board's 6 screws, so it is like a circuit sandwich. I'm not really sure how you did it since I can't see behind the perf. Looks like you can shrink the board space down to a final 30% or so of what you currently use, if you sacrifice the easy to work with nicely spaced layout you have. (maybe resort to upright resistors... or SMD!  )

Oh yes there is lots of space to be won there, but the other two boards are really tight. I have to add four filters to the delay board too.
The perf is that big because of a few reasons. One being where the pins from the sparkfun boards come off. Second it houses the arduino. (you can see its little pin header sticking over the edge of the perf)

For connecting the button boards to the main board, I see that there are four holes on three sides... There is a method I saw on a circuit sandwich before where they used long pin breakaway headers to connect the top board to the bottom board (think having a hole in the same place on each board and running a wire between them vertically, but with nice looking headers. It's not easy to take apart but it looks really nice. Too bad the button pins aren't all on one side, that would make it much easier to just run a 12-pin header....

Yes it has four times four pins for the leds and switches, plus a few power supply/ground pins. I don't think I'll fix it to the main board using headers. I want to keep them mechanically separated. (same for the pots)  You can put the voltage diver resistors on the sparkfun board. So you'd only need two wires for the button detection and five for the leds.

I'm not sure if I will do kits, and if I do them I'm not sure if I will do kits including the uncommon parts, as they are as hard to find for me as anyone else (I got them of ebay).

I'm thinking about learning Designspark, as with eagle I'm limited now to the free versions board size.

[Taylor]

Just posting here to subscribe. I will be buying a PCB when they're available, for sure. Really cool project, congrats.

[dmc777]

Just posting here to subscribe. I will be buying a PCB when they're available, for sure. Really cool project, congrats.

Same Here! Though you're speaking a different language than me ha!!

[RonaldB]

great project Arph,

Coun't me in for a PCB.

Ronald

[boogietube]

count me in too...

[~arph]

Good to see there is interst in this. I'm still not sure if there will be pcb's available, bear that in mind.

[~arph]

Small update,

Fellow forum member Jarno was so kind to send me a few black imbus bolts. Now it looks even better.

[mth5044]

Sweet 

Hey I've got all the schematics together and most of the board layouts done.. maybe you could check if they 'look enough like yours' when they are done? 

[~arph]

Sure no problem!
You should breadboard it first, to check if it works properly. I've been pondering all week on how I can improve my current design I think I'll order some pt2399's again and start experimenting with filtering in the delay stages again. I want to get rid of the downward expander as that is a component pretty hard to come by. Also I want to improve the button response. Sometimes now when you press one button another will switch off and another two will switch on. Especially when the pedal is chained with other fx on the same supply (the button switching is voltage dependent)

Meanwhile check this 6 pt2399's in parallel reverb design. (it also holds hints on how the belton module works..)

http://sites.google.com/site/analogcustomkits/Home/reverb-digital

[AudioMime]

Sure no problem!
You should breadboard it first, to check if it works properly. I've been pondering all week on how I can improve my current design I think I'll order some pt2399's again and start experimenting with filtering in the delay stages again. I want to get rid of the downward expander as that is a component pretty hard to come by. Also I want to improve the button response. Sometimes now when you press one button another will switch off and another two will switch on. Especially when the pedal is chained with other fx on the same supply (the button switching is voltage dependent)

Meanwhile check this 6 pt2399's in parallel reverb design. (it also holds hints on how the belton module works..)

http://sites.google.com/site/analogcustomkits/Home/reverb-digital

Dang it Arnoud! I have to buy about 15 pt2399s now!