DX-style Operator (Thru-Zero FM/PM oscillator - FM Ogre)

From circuitbending to homebrew stompboxes & synths, keep the DIY spirit alive!

Moderators: lisa, luketeaford, Kent, Joe.

Post Reply
User avatar
mxmxmx
Super Deluxe Wiggler
Posts: 2243
Joined: Thu May 31, 2012 1:54 am
Location: berlin

Post by mxmxmx » Tue Jul 28, 2015 6:33 am

diablojoy wrote:hmm not sure if I understand your request here though, my board is a single 12HP purely euro board, everything is on it- 5 pots, 8 jacks, 3 switches, 3 LED's and all of the main circuitry there is no wiring and
no allowance for a breakout to a control board as that's all covered.
oh, sure. what i meant was: let's assume i wanted three of those operators but don't want to or can't afford to waste 36HP; so one option might be to get, say, one fully featured module at 12HP (ie your layout) and two aux voices at 6HP (my layout); that's still 24 HP, but somewhat more compact. anyways, was just an idea, ie that i might try to adjust the panel pcb somewhat so that things look more coherent overall. this is how things look atm, about 35mm deep:

Image

the main board is fully self-contained (except the sync in), so there's some room for manoeuvre. anyways, i guess i'll just take a look at your .fpd when you post it and take it from there.
hey while I am at it I might have to buy a few of your 6HP boards when your ready
the smaller version could be useful to me. what controls/ IO did you drop
with the smaller form ?
sure .. i have 10 sets or so, which (it looks like) are ok +/- some silkscreen stuff but they should be ok for beta testing. i still need/want to try a/the firmware with the switches first ...

as for the IO, i didn't drop anything except for one switch and one LED, and there's no attenuators. i figured if all three switches are absolutely necessary one might still cook up some long-press workaround. also i simplified the circuitry somewhat to make it fit. (basically, the sync in and out don't use op amps but transistors, and i used MCP6002s not TL072s for the inputs, that got rid of the diodes etc). there's just one trimpot, for V/oct; and i used a lm4040 as a reference.

User avatar
wsy
Super Deluxe Wiggler
Posts: 2719
Joined: Wed Aug 24, 2011 10:20 pm
Location: near Boston Massachusetts USA

Post by wsy » Tue Jul 28, 2015 10:08 am

Well, the bug is that in phase modulation, the update rate is too slow for smooth PM and so there's a nonharmonic component up
at around 12 to 14 KHz (it varies according to the other frequencies involved, being a sort of sum-and-difference kind of thing).

From the point of view of the oscilloscope, it's a bug.

From the point of view of the listener's ears, it just sounds "sparkly".... it's not a glitchy bug by any stretch of the imagination.

There was also a second bug (now squashed) that had to do with the use of hard sync in as "freeze frame" - it was also resetting the
oscillator phase, just as you'd expect hard sync to do, but which doesn't make a lot of sense when you're doing granular.

That second bug is already squished....

- Bill
"Life is short. But we can always buy longer patch cords" - Savage

User avatar
diablojoy
Super Deluxe Wiggler
Posts: 1617
Joined: Fri Jan 22, 2010 9:03 pm
Location: melbourne

Post by diablojoy » Wed Jul 29, 2015 5:03 am

let's assume i wanted three of those operators but don't want to or can't afford to waste 36HP; so one option might be to get, say, one fully featured module at 12HP (ie your layout) and two aux voices at 6HP (my layout); that's still 24 HP, but somewhat more compact.
Ahh
ok now I get it - yes good idea :tu:
I was actually thinking of drawing up another board at 20HP in through hole except for the dispic itself but with some extra's - basically adding attenuators on the FM, PM and F/back inputs - same idea in mind , just larger , that and the through hole guys might like it more, the 44 pin dispic is pretty easy to solder itself and might not be a bridge too far for some or it could be done on an adaptor board that simply plugs in.
(there's an opening for some enterprising young person who likes soldering
44 pin TQFP packages and programing) :lol:
It probably wont happen today but if it does it definitely wont go smoothly.

User avatar
wsy
Super Deluxe Wiggler
Posts: 2719
Joined: Wed Aug 24, 2011 10:20 pm
Location: near Boston Massachusetts USA

Post by wsy » Wed Jul 29, 2015 3:07 pm

diablojoy wrote: Ahh
ok now I get it - yes good idea :tu:
I was actually thinking of drawing up another board at 20HP in through hole except for the dispic itself but with some extra's - basically adding attenuators on the FM, PM and F/back inputs - same idea in mind , just larger , that and the through hole guys might like it more, the 44 pin dispic is pretty easy to solder itself and might not be a bridge too far for some or it could be done on an adaptor board that simply plugs in.
(there's an opening for some enterprising young person who likes soldering
44 pin TQFP packages and programing) :lol:
If you're going to go to the trouble of adding attenuators in the analog FM, PM, and FB chains, you may as well go for attenuvertors.
Right now, the FM, PM, and FB are pots that are directly sampled by the DSP's ADC and you could skip them (or repurpose them, as in
sample inputs, all the time).

- Bill
"Life is short. But we can always buy longer patch cords" - Savage

User avatar
mxmxmx
Super Deluxe Wiggler
Posts: 2243
Joined: Thu May 31, 2012 1:54 am
Location: berlin

Post by mxmxmx » Wed Jul 29, 2015 5:44 pm

so... i've just tried with the leds and switches, and those check out ok, too *

Image

nice little thing. i like it.

anyways, now that i have it, i'm not sure i'm inclined to put much more work into this; ... nor do i know if anyone else feels inclined to build this in 6HP ... ? but if so, i'd prefer someone to beta-test before i make the gerbers available (?); pm me if you are so inclined (ie i have a couple of spare pcb sets).

* i didn't try with fmogre_switches_main.c yet, just the basic FM version with some additional lines of code. anyways, using non-latching switches seems to be fine.

User avatar
diablojoy
Super Deluxe Wiggler
Posts: 1617
Joined: Fri Jan 22, 2010 9:03 pm
Location: melbourne

Post by diablojoy » Sat Aug 15, 2015 12:15 am

realised I should have posted some pics well before now :doh:
just need to sort some documentation and release

Image
Image
Image
Image

ignore the use of a few 1206 caps on the front I ran out of 0805's and kludged those on instead.


mouser BOM - access ID fefb4fdec4
or project name - FM OGRE REV 2A - 12HP EURO VERSION

this will be missing the jacks ,switches , pots and led's
jacks can be kobicons or erthenvar's x 8
switches are NKK subminature SPDT x 3 (solder tag are fine the board has big enough holes to cope)
pots - your favourite B100K vertical 9mm snap in x 5 (recommend the thonk ones) edit : oh and knobs of choice of course x 5
and just any standard LED"s x 3 will do


OK here is the board and front panel files I will add more info here as I work it out.
and parts list by value
hmm not certain if F1 listed is the best choice if anyone has a better component suggestion in 0805 size available from mouser don't hold back
also adjusted mouser project missed a couple of 100nf caps and had too many resistors of one particular value
note R75 there is a choice depending on how fine the fine frequency control will be. 500K to 1M - higher value offers smaller control range on the fine frequency pot

EDIT: 06/08/2016 There was an error found in the gerber files C29 was incorrect this has been fixed in REV2B if you have built any REV2A boards please remove C29 from your board or do not install . can cause instability of the LP2985 2.8v LDO
You do not have the required permissions to view the files attached to this post.
Last edited by diablojoy on Sat Aug 06, 2016 4:02 am, edited 12 times in total.
It probably wont happen today but if it does it definitely wont go smoothly.

User avatar
wsy
Super Deluxe Wiggler
Posts: 2719
Joined: Wed Aug 24, 2011 10:20 pm
Location: near Boston Massachusetts USA

Post by wsy » Sat Aug 15, 2015 6:14 am

Wow yeah!

OK, I'm officially amazed. There are now three working builds of the FM Ogre,
all with different board layouts. But the amazing thing is that they exist.

Way cool... now, to see when the first "build for hire" hits the street.

- Bill
"Life is short. But we can always buy longer patch cords" - Savage

User avatar
apoisontree
Common Wiggler
Posts: 106
Joined: Sun Apr 22, 2012 6:25 pm
Location: Paris

Post by apoisontree » Sun Aug 16, 2015 6:22 am

Diablojoy, thanks for the share! (And thanks to wsy for this great project, of course)

In your fpd file, the left and right outputs are inverted compared to your mock-up panel (left is down, right is up), is it intentional?

User avatar
diablojoy
Super Deluxe Wiggler
Posts: 1617
Joined: Fri Jan 22, 2010 9:03 pm
Location: melbourne

Post by diablojoy » Mon Aug 17, 2015 4:38 am

In your fpd file, the left and right outputs are inverted compared to your mock-up panel (left is down, right is up), is it intentional?
oops haha nice catch, bottom most jack should indeed be Left
as per the fpd file although really it is fairly arbitrary what designation is used for the outputs.
bottom most jack is the phase modded output
L and R just come from the dispic outputs designations
A and B would be just as correct or whatever you want easy to change

has anyone tried the mouser project access code thing yet ?
not sure if I did it right at all.
still have to do-
1. redraw schematic
2. parts list
3 . build log and tips with pics - I need to do another build for this waiting on a parts order though
I now have five working though :tu: 4 in euro and 1 of bills original boards in my large custom format
It probably wont happen today but if it does it definitely wont go smoothly.

Dogma
Super Deluxe Wiggler
Posts: 4135
Joined: Wed Aug 28, 2013 10:46 pm

Post by Dogma » Mon Aug 17, 2015 7:04 am

diablojoy wrote:
In your fpd file, the left and right outputs are inverted compared to your mock-up panel (left is down, right is up), is it intentional?
oops haha nice catch, bottom most jack should indeed be Left
as per the fpd file although really it is fairly arbitrary what designation is used for the outputs.
bottom most jack is the phase modded output
L and R just come from the dispic outputs designations
A and B would be just as correct or whatever you want easy to change

has anyone tried the mouser project access code thing yet ?
not sure if I did it right at all.
still have to do-
1. redraw schematic
2. parts list
3 . build log and tips with pics - I need to do another build for this waiting on a parts order though
I now have five working though :tu: 4 in euro and 1 of bills original boards in my large custom format

Thank you so much for this - really.....

Im off to OSH!
look up!

User avatar
apoisontree
Common Wiggler
Posts: 106
Joined: Sun Apr 22, 2012 6:25 pm
Location: Paris

Post by apoisontree » Mon Aug 17, 2015 12:08 pm

diablojoy wrote:
In your fpd file, the left and right outputs are inverted compared to your mock-up panel (left is down, right is up), is it intentional?
oops haha nice catch, bottom most jack should indeed be Left
as per the fpd file although really it is fairly arbitrary what designation is used for the outputs.
bottom most jack is the phase modded output
L and R just come from the dispic outputs designations
A and B would be just as correct or whatever you want easy to change

has anyone tried the mouser project access code thing yet ?
not sure if I did it right at all.
still have to do-
1. redraw schematic
2. parts list
3 . build log and tips with pics - I need to do another build for this waiting on a parts order though
I now have five working though :tu: 4 in euro and 1 of bills original boards in my large custom format
Thanks for your answer. What is the central led for? You wrote HB (?) on your white panel but nothing in the fpd.

I did access your mouser project without problem.

User avatar
lamouette/rck
Wiggling with Experience
Posts: 471
Joined: Mon Oct 14, 2013 12:37 pm

Post by lamouette/rck » Mon Aug 17, 2015 1:47 pm

Looks absolutly perfect !

Waiting for your updates, if someone is ok to get some at OSH, i would be happy to get one !
Setup is never done.

User avatar
wsy
Super Deluxe Wiggler
Posts: 2719
Joined: Wed Aug 24, 2011 10:20 pm
Location: near Boston Massachusetts USA

Post by wsy » Mon Aug 17, 2015 4:17 pm

apoisontree wrote: Thanks for your answer. What is the central led for? You wrote HB (?) on your white panel but nothing in the fpd.

I did access your mouser project without problem.
Probably HB is HeartBeat.

In the default configuration, LED1 is normally clocked ON whenever the primary sine output is above 0 volts, so it acts like the Q106 LED
and flashes in time with the output.

However, that association is purely in software (and it's not PWMed anyway, maybe it should be!) and
one of the alternatives is to have it flip from off to on (or on to off) with every 1,000,000 "free" (not in interrupt) code of the DSP. That's a
help when doing timing and feature impact checks.

- Bill
"Life is short. But we can always buy longer patch cords" - Savage

User avatar
FingerTappin
Common Wiggler
Posts: 171
Joined: Fri Mar 11, 2011 6:03 pm
Location: New York

Post by FingerTappin » Mon Aug 17, 2015 9:01 pm

I have been keeping my eye on this thread. It is exciting to see how far this has come along! This module looks fantastic. I plan on building two.



Maybe I can find someone in the NYC area with a PICkit3. Wasn't there a website with a chip flashing service? Probably has a min requirement way above what I am looking for

Dogma
Super Deluxe Wiggler
Posts: 4135
Joined: Wed Aug 28, 2013 10:46 pm

Post by Dogma » Mon Aug 17, 2015 9:49 pm

I need to get a Pickit 2 for my Harvestman stuff - will it suffice with the Ogre?
look up!

User avatar
wsy
Super Deluxe Wiggler
Posts: 2719
Joined: Wed Aug 24, 2011 10:20 pm
Location: near Boston Massachusetts USA

Post by wsy » Tue Aug 18, 2015 7:42 am

Dogma wrote:I need to get a Pickit 2 for my Harvestman stuff - will it suffice with the Ogre?
Doesn't look like it... the PICkit 2 doesn't support most of the PIC32 line and the PIC33 DSP used is newer than the PIC32.

Sorry.....

- Bill
"Life is short. But we can always buy longer patch cords" - Savage

User avatar
burdij
Veteran Wiggler
Posts: 672
Joined: Mon Dec 28, 2009 4:38 pm
Location: Wisconsin US

Post by burdij » Sat Sep 05, 2015 3:07 pm

Here is the unit that I am fairly certain will be on display at our table at Knobcon next weekend.

Image

We are introducing this dual channel VCA, also, that will be a good companion for this module. It will allow modulation signal amplitudes to be controlled for dynamic FM synth effects.

Image

I will have some kind of a synth demo setup with the two modules.
Knobcon 7 - 2018 Photo Album
While website and store updates are in progress, access the Grove Audio Synth Store
Tesla Science Center http://www.teslasciencecenter.org

User avatar
wsy
Super Deluxe Wiggler
Posts: 2719
Joined: Wed Aug 24, 2011 10:20 pm
Location: near Boston Massachusetts USA

Post by wsy » Sun Sep 06, 2015 12:47 pm

(oops... should have PMed instead)

- Bill
"Life is short. But we can always buy longer patch cords" - Savage

User avatar
diablojoy
Super Deluxe Wiggler
Posts: 1617
Joined: Fri Jan 22, 2010 9:03 pm
Location: melbourne

Post by diablojoy » Mon Jan 04, 2016 11:16 pm

built another 2 last week
both same batch of boards and components
one is fine fired and programed first go
second is cursed it seems or I just had a bad build day
fixed a couple of faults - one short to ground taking down the 3.3v reg
second short to -v rail under a tl072 messing with the bias. both were solder issues
the third fault is proving harder
I am now getting all the correct voltages at test points however it will not program kept getting unexpected address 0x0 so couldn't program.
this eventually led me to checking the clock oscillator and yep no clock. in fact I see 2.8v on clk1 and 1.4v on clk0 pins.
Ok so I changed the 4mhz crystal- no change , changed the 2 x 20pf caps - didn't fix things , I changed the dispic chip itself - no difference I then even changed C3 . as there was a slight off hand reference to it on one forum as being a possibility
Now I am stuck for the moment.
anyone know of any other things I could check that would affect the clock ?
It probably wont happen today but if it does it definitely wont go smoothly.

User avatar
wsy
Super Deluxe Wiggler
Posts: 2719
Joined: Wed Aug 24, 2011 10:20 pm
Location: near Boston Massachusetts USA

Post by wsy » Tue Jan 05, 2016 8:34 am

diablojoy wrote:built another 2 last week
both same batch of boards and components
one is fine fired and programed first go
second is cursed it seems or I just had a bad build day
fixed a couple of faults - one short to ground taking down the 3.3v reg
second short to -v rail under a tl072 messing with the bias. both were solder issues
the third fault is proving harder
I am now getting all the correct voltages at test points however it will not program kept getting unexpected address 0x0 so couldn't program.
this eventually led me to checking the clock oscillator and yep no clock. in fact I see 2.8v on clk1 and 1.4v on clk0 pins.
Ok so I changed the 4mhz crystal- no change , changed the 2 x 20pf caps - didn't fix things , I changed the dispic chip itself - no difference I then even changed C3 . as there was a slight off hand reference to it on one forum as being a possibility
Now I am stuck for the moment.
anyone know of any other things I could check that would affect the clock ?
I've seen that before.

It's the RESET line. If you forgot to put in the pull resistor (or put in the wrong value) the DSP will never either enter program/reset, or
it will never exit program/reset.

At least, that's what I'd check first, because I've buggered that at least twice in my builds.

If that fails... hmmmm.... Check continuity and impedance on all of the J1 (programmer) pins right to the chip. (by impedance, both
diode drops with a LOW-current diode teste; ii.e. limited to 1 mA or less so it's within the safety bounds of the protection diodes on the
DSPs inputs ( check your DVM with another meter to be sure; Fluke meters tend to be 1 mA or less but that's not a very big sample space).

Failing that, check what the voltages are during initial program access on one of your good units vs. the failing unit. It might
be a voltage regulator or power supply diode that gave up the ghost after one of the other insults.

Failing THAT - get yourself a shotgun and some goose or varmint rounds..... :-)

- Bill
"Life is short. But we can always buy longer patch cords" - Savage

User avatar
wsy
Super Deluxe Wiggler
Posts: 2719
Joined: Wed Aug 24, 2011 10:20 pm
Location: near Boston Massachusetts USA

Post by wsy » Tue Jan 05, 2016 5:26 pm

Addendum to the above: as to the clock: Maybe it is or maybe it isn't; the clock driver circuit on the PIC is very capacitance-sensitive.
You should be OK with a 10X probe on the CLKO pin (clock out, pin 31), but IIRC even my good Tek scope has trouble looking at the CLKI pin (clock in, pin 30) without stopping
the clock with capacitance.

Actually, are you making these measurements with an oscilloscope with a low capacitance probe, or with a DVM (much higher capacitance, definitely will not work)?

- Bill
"Life is short. But we can always buy longer patch cords" - Savage

User avatar
diablojoy
Super Deluxe Wiggler
Posts: 1617
Joined: Fri Jan 22, 2010 9:03 pm
Location: melbourne

Post by diablojoy » Wed Jan 06, 2016 4:42 am

Actually, are you making these measurements with an oscilloscope with a low capacitance probe, or with a DVM (much higher capacitance, definitely will not work)?
my ancient BWD scope but as to whether the capacitance is low enough hmmm
I do see some signs of oscillation on one that works albeit it ísn't pretty.
definitely a discernable difference to the one that doesnt work 'anyway
I've seen that before.

It's the RESET line. If you forgot to put in the pull resistor (or put in the wrong value) the DSP will never either enter program/reset, or
it will never exit program/reset.
hmm would this prevent the clock from oscillating ? ok something else I can check anyway
thanks :tu:
It probably wont happen today but if it does it definitely wont go smoothly.

User avatar
wsy
Super Deluxe Wiggler
Posts: 2719
Joined: Wed Aug 24, 2011 10:20 pm
Location: near Boston Massachusetts USA

Post by wsy » Wed Jan 06, 2016 8:38 am

diablojoy wrote:
Actually, are you making these measurements with an oscilloscope with a low capacitance probe, or with a DVM (much higher capacitance, definitely will not work)?
my ancient BWD scope but as to whether the capacitance is low enough hmmm
I do see some signs of oscillation on one that works albeit it ísn't pretty.
definitely a discernable difference to the one that doesnt work 'anyway
I've seen that before.

It's the RESET line. If you forgot to put in the pull resistor (or put in the wrong value) the DSP will never either enter program/reset, or
it will never exit program/reset.
hmm would this prevent the clock from oscillating ? ok something else I can check anyway
thanks :tu:
Good point- and besides, it's not working when the scope isn't there, so that's a problem. And you swapped crystals... Did you try
swapping C13 and C14 (the two little capacitors next to the crystal)? Or just take them off and make sure there's no solder bridges
around them. Those two capacitors are to "help" the oscillation and are there because MicroChip's datasheet says to do it, but
my first units didn't have them and they worked fine. So, in my opinion, they're purely talismanic.

Yeah, try that, and look for any crud (burnt flux, anything) on either side of the board that might be parallelling C13 or C14 or the
crystal. Maybe even pull the crystal as well and give that whole area of the board a scrub to be sure.

Failing that (or maybe if nothing seems wrong) maybe there's an errant piece of copper between the crystal traces, so if you want a
last ditch effort, ditch between the pin 30 and pin 31 traces with an X-Acto knife to make SURE it's not shorting; cut it all the way
down to the fiberglass. Use a magnifier and take your time; there's not a lot of space there and a marginal etching job might leave a bit of copper there.

(if you remove the capacitors AND the crystal AND the DSP chip, you could test for such bridges with an ohmmeter between the
two crystal pins..

You might even be able to "zap" the bridge open with a car battery and a couple of jumper wires (that was an old technique
my buddies from DEC told me about - if an unstuffed PCB showed a short, they had a rack frame with two of their biggest linear
power supplies in the bottom (something like 500 amps at -5.2 volts) and really thick power lines, thick even for the ECL logic
in the KL CPUs. They'd put the badly etched board into the rack, by itself, and turn on the power, and all of the shorts would
light up like tiny light bulbs. When all of the "light bulbs" burned out, the board was "good" (or at least better!) and they'd retest it.
Sounded brutal, but hey, the board was going to go into scrap *anyway* if the burnout didn't work, so there was little to lose.
Apparently the technique worked out quite well; I've used it for the occasional patch cable that developed a short between
signal and ground - used a motorcycle battery for power. Worked great. And- the cable was bad anyway, so what's to lose?

If you try this, wear safety glasses! If you short the wrong things, you might get an arc and some weld spatter.

EDIT; A car battery is probably overkill. A motorcycle battery would probably be more to scale. Or one of those "car starter booster
boxes" that you can use to jumpstart a car with a weak battery.

- Bill
Last edited by wsy on Wed Jan 06, 2016 9:19 am, edited 1 time in total.
"Life is short. But we can always buy longer patch cords" - Savage

Jarno
Super Deluxe Wiggler
Posts: 2772
Joined: Mon Nov 15, 2010 6:34 am
Location: Rosmalen, NL

Post by Jarno » Wed Jan 06, 2016 9:14 am

Off topic, but a friend of my dad's worked at a cable factory (probably a pensioner by now).
When they had a cable which had a short, they hooked up an old welding transformer. The heat generated at the short would burn the cable jacket right open, so the cable could be cut it at that exact point, and production restarted.

Joonatin
Common Wiggler
Posts: 55
Joined: Tue Aug 11, 2015 9:30 am
Location: Finland

Post by Joonatin » Fri Feb 05, 2016 10:00 am

Has someone made too many pcbs and would be willing to sell me one or two? PM plz.

Post Reply

Return to “Music Tech DIY”