PRIMOVA - GKMX-33 Project - The ultimate-No US-20 bug 3in/3out GK switcher

Started by CodeSmart, March 25, 2014, 12:50:21 PM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

gumtown

For this project the PIC18F4550 might be a bit light on data memory (especially if programming in C),
I would recommend this family of PIC
http://ww1.microchip.com/downloads/en/DeviceDoc/39977c.pdf

I brought a PIC18F26K80 when I was doing my project when I realised the PIC18F4550 was running out of space quickly, but haven't used it yet.
Free "GR-55 FloorBoard" editor software from https://sourceforge.net/projects/grfloorboard/

Elantric

My employer uses PIC32MX340F512H MCU 32BIT 512KB FLASH 64TQFP for many of its current designs

smallest one is MICROCHIP# PIC18F46K20T-I/ML FLASH 32KX16 44QFN

CodeSmart

Quote from: gumtown on January 13, 2015, 12:17:57 AM
If you need some help with this, I am happy to lend a hand.
This project is similar to what I started 2 years ago, but I had problems trying to prototype with SMD chips (and my failing eyesight).

Gumtown, I thank you for your invite. And of course I have read you thread. Selection of analog components is crucial here. We can leave the digital side for now. It's somewhat "logical" even though I don't underestimate it at all. Currently with a little help of one of our senior moderators I've been struggling with the fact that the supply voltage of one of my elected components only allow +/- 4.5 or (+/-5V max rating)  power supply voltage (instead of +/- 7.5V normally found in 13-pin gear).

Today I borrowed a scope and did some measurements. I've tried all strings on 3 different guitars, one GK3 and two GK-Kit and the maximum Peal-to-Peak voltage I could ever physically accomplish (by abusing the string heavily, they almost broke...) was +/-1.5V (i.e. 3V peak-to-peak) on the low E.

If that's the worst case expected I'm ok with the chip elected. However I'm not a bass player and since I got most "bang" on the low notes I'm very  interested what worst case ramming a bass with hammer on a GK would be expected in Vpeak-to-peak, please!!!
But I got more gear than I need...and I like it!

Elantric

QuoteIf that's the worst case expected I'm ok with the chip elected. However I'm not a bass player and since I got most "bang" on the low notes I'm very  interested what worst case ramming a bass with hammer on a GK would be expected in Vpeak-to-peak, please!!!

Correct - That would be the worst case Max  Attack Transient Voltage "peak to peak" spike"  scenario for 13 pin systems.

of course I'm not a Bass player either ;)

But have worked on GK-3B Bass guitars and know they are capable of generating a higher peak to peak Voltage output!

If these Transients are clipped ( due to a +/- 4.5V power rail for the digipot) , it may impact the performance of COSM Modeling bass systems (V-Bass, VB-99, GR-55 in Bass mode) 

Also Early RMC Piezo guitars from 1999 had a very high output signal - much higher than the Roland GK pickups
Thats why you can read countless posts here where we suggest Piezo 13 pin users set their VG-99/ GR-55 / GP-10 GK String sensitivity to "0"  - to attenuate the higher voltage output from piezo guitars.

CodeSmart

Quote from: Elantric on January 13, 2015, 11:26:29 AM
But have worked on GK-3B Bass guitars and know they are capable of generating a higher peak to peak Voltage output!
Agree, but I still don't mind Gumtown throw me his numbers at me as Vpp. And as discussed there are ways to attenuate the signal before hitting the chip. I'm all ears but some boxes are very costly and some are more easily sold.
But I got more gear than I need...and I like it!

CodeSmart

Regarding the Microcontroller...I think I need USB feature to download new firmware. Can't expect something complicated as this goes without customer firmware updates from a PC. Also I need EUSART for MIDI IN/OUT and a lot's of I/O. Considering two or three MCP23S17 for IO and MCP4728 for 12-bit DAC (Analog LED backlight (PWM digital backlight ctrl probably cause noise) + VOLA,VOLB,VOLC out. The controller should have at least 6 ADC in, PIN9, GTRVOL, EXPRESSION PEDAL  A, EXP PEDAL VOL B, EXP PEDAL VOL C.
But I got more gear than I need...and I like it!

Elantric


CodeSmart

Quote from: gumtown on January 13, 2015, 11:06:47 AM
I brought a PIC18F26K80 when I was doing my project when I realised the PIC18F4550 was running out of space quickly, but haven't used it yet.

Yup want something like PIC18F26K80 memory wise, but also with USB onboard.
(Gumtown, please don't miss my request on Vpp on a bass).
But I got more gear than I need...and I like it!

gumtown

Yes I will check out the GK-3B levels + audio from my Fender jazz bass (with 18volt on-board active preamp), looking around for my Fluke 92 scope meter.
I guess you want the voltages unloaded/open circuit for maximum peak values.

the MCP23S17 I/O extenders looks like a chore to add in, an all-in-one chip would be easier (programming side).
I happen to glance on these chips "HV2661" in a 48 pin LQFP package, 8x3way (3:1 mux) analog switches with switch voltage ratings up to 200v.
One chip for 3 GK inputs and one for 3 GK outputs maybe?
http://ww1.microchip.com/downloads/en/DeviceDoc/hv2661.pdf
Free "GR-55 FloorBoard" editor software from https://sourceforge.net/projects/grfloorboard/

gumtown






Some initial results recorded as I work though the GK-3B test
on the 5 string bass in pictured in my avatar
<<<<<<

Multimeter on peak hold AC mV auto range
All strings strummed unreasonably hard
Normal PU @ full volume + onboard 3 band EQ all max boost open circuit load = 2.410 AC volts max RMS
Normal PU @ full volume + onboard 3 band EQ all max boost GR-55 input loaded = 2.631 AC volts max RMS

GK pin 2 (G) = 571.9 mV AC RMS
GK pin 3 (D) = 563.4 mV AC RMS
GK pin 4 (A) = 858.0 mV AC RMS
GK pin 5 (E) = 1317  mV AC RMS
GK pin 6 (B) = 339.1 mV AC RMS

Fluke 92 scope meter - GK with GR-55 terminated load
GK pin 2 (G) = 2.25 V AC P to P
GK pin 3 (D) = 1.95 V AC P to P
GK pin 4 (A) = 2.40 V AC P to P
GK pin 5 (E) = 2.40 V AC P to P
GK pin 6 (B) = 1.20 V AC P to P

the action on my B string is raised a little higher than the others, and I never noticed before the G string is a little high too.

Free "GR-55 FloorBoard" editor software from https://sourceforge.net/projects/grfloorboard/

CodeSmart

Quote from: gumtown on January 14, 2015, 01:54:03 AM
Some initial results recorded as I work though the GK-3B test
on the 5 string bass in pictured in my avatar

Thanks a lot. Looks like I'll be ok with +/-5V supply :)
But I got more gear than I need...and I like it!

CodeSmart

It's been a costly day. I went ahead and bought a brand new 100MHz oscilloscope. Never have owned one. This is what I ordered:
http://www.peaktech.de/productdetail/kategorie/digital-oszilloskope/produkt/peaktech-1245.857.html

There's no upper limit to what a good scope can cost. This one is very moderately priced at 539 Euro/610 USD (incl. 19% German VAT) and I hope it will be good enough for my projects.

I also ordered samples of all components I plan to use. Need to be 100% sure about all the footprints before start working with the PCB's. When it comes to SMD components there's so many different casings and I don't trust the dimensions until I see them in real life. I've had my fair share of footprint mismatch already.

So when this stuff arrives it's just a matter of finding the most difficult component..... TIME!  8)
But I got more gear than I need...and I like it!

gumbo

  "So when this stuff arrives it's just a matter of finding the most difficult component..... TIME! "


...Oh YES!!!!!!     ::)

Stay warm!
Peter
Read slower!!!   ....I'm typing as fast as I can...

Majiken

This reminds me of my days in the automobile industry, when I loved hanging out with engineers even though I could only understand maybe half of what they were talking about (I was in Purchasing & Supplier Quality). Watching this gibberish turn into a fine piece of craftmanship is a lot of fun! BTW, I hadn't been in the forum longer than a few days before I ordered something from Robert- and you can bet this piece is already on my wish list  ;D
Take what you need, put back a bit more, leave the place behind you better than it was before :-)

www.majiken.rocks

whippinpost91850

#39
Codesmart, I am so excited about your new project.  8)

CodeSmart

A small tiny power converter board completed - Check!
If powered by 12V DC adapter it provides GK +/-7V, analog +/-5V, digital 5V and LCD backlight 5V
If powered by GK +/-7V it provides additional analog +/-5V and digital 5V
Got EMI filters, reverse polarity protection, over voltage protection, resettable fuse, ferrites etc.
Planning to use this small board in multiple products.

Now time to complete the real stuff :-)
But I got more gear than I need...and I like it!

Elantric

Thanks for reporting the updated progress on your new GKMX-33 Project!

CodeSmart

Been designing the main PCB board for weeks now (on low speed). It's like a strategy game trying to fit everything, hours just disappearing. Going for only two layers at least for the initial PCB version. Four layers is about 2.7 times more expensive and at this stage without even knowing if the stuff will work I need to save bucks. I also have a few components so freakin tiny I'm not sure I'll be able to solder them by hand. May have to find larger SMD versions  ::)

After this I also need to make the three additional boards. Filter, front connectors and side connector board. Saving shipping cost by producing all boards at the same time.

Interestingly I foresee a few additional new products derived from GKMX-33 experience.
- GKMX-31 Guitar Selector (one of three 13-pin guitars)
- GKFX-A11 Per-string adjustable subsonic filter for any 13-pin instrument. Bass, guitar, violin etc. 22 steps between 21Hz (B0) to 405Hz (D5).

Recently been reading a very thick book to learn about USB. USB is heavy stuff. Would be way cool to develop a GP-10 USB/5-pin MIDI converter based on a Microchip USB Host. Just for the thrill. I guess one would need to buy a hardware protocol analyzer like Ellisys to be able verify the traffic. Hm...

Easy having ideas, but realizing them that's the quest.  8)
But I got more gear than I need...and I like it!

CodeSmart

I have the flu. Nobody wants me coughing my lungs out and snoring at the office  :(  So what can I do? Yup, been home eating pills and doing PCB as my weakened strength allow me. I hope to send them to production this week.  ;D

There's five PCB's in the GKMX-33. Main board, Power converter board, Input jacks with LCD board, Expression jacks board and Programmable Subsonic Filter board. This little box will be really stuffed. Not wiring any prototype, just hoping for the best with real boards. God knows how many hours spent.

What worries me is that I hear a constant echo of Gumtown telling me I need more than 64k, that I need to go 3.3V/16-bit/128k memory. Heck! I just need to prove I can write very efficient code. He mustn't be right. Going 3.3V would be a LOT of hardware/PCB redesign. Of course Gumtown is wrong... ::) (I'm repeating it like a mantra in my foolish head). We'll see. At least I need to try it before I bite the bullet.
But I got more gear than I need...and I like it!

Elantric

Its a 3V3 logic world today

Have not designed anything with 5V logic in past 10 years

Also the big issue is with 5V logic  - the voltage swings on high speed lines on PCB top layer can manifest as radiated EMI/RF - lower 3V3 logic allows faster speed and lower radiated EMI/RF - for FCC CFR47 part 15 compliance. 

Typically use 74VHCxxxx parts for Shift Registers, 138 Octal decoders, 244 clock buffers,  etc.

whippinpost91850


CodeSmart

Quote from: Elantric on April 08, 2015, 02:13:54 PM
Its a 3V3 logic world today
Have not designed anything with 5V logic in past 10 years

I'm from the dark ages (I studied a tiny bit of electronics 35 years ago) reawakening my hobby when kids left and try to use cheap simple mainstream components and popular processors with sample code available on the Internet. But agree everything eventually points in a 3.3V direction.
But I got more gear than I need...and I like it!

CodeSmart

Been looking at the 3.3 volt impact of all the chips and it seems going from 5V to 3.3V would not complicate much.
The boards are now ready for the first prototype production run. The specification has changed a bit during the design process but I'll guess it's a subject for the software. There's virtually no limitations in the hardware. But I really wish I had a GP-10 USB host in there using a PIC24J64 as well, but I have a heck of other stuff to sort out first. One step at a time. Fever is almost gone so I guess I have to drag myself to work on Monday  :(

Here's the boards I made for the intended Hammon enclosure (same enclosure I used for GKMX-42/41).
There's an awful amount of small things in the box...  8)






But I got more gear than I need...and I like it!

Elantric


CodeSmart

Thanks to Elantric I just added a second MIDI out.
MIDI THRU is gone. So now it's MIDI IN + 2xMIDI OUT.
Ehhh...The third MIDI OUT....  would be nice with a
USB GP-10 host output.
But that may take another six months after getting everything else working.

Having problem sleeping...reading the USB spec. and books about the subject.
I promise you it works... ZZZZZZzzzzzzz   ;)

But I got more gear than I need...and I like it!