0

Power Supply Modding
Moderators: altitude, adafruit_support_bill, adafruit, phono, hamburgers

Please be positive and constructive with your questions and comments.

Re: Power Supply Modding

by antto on Sun Jan 21, 2018 5:13 am

so far the circuit above (if it works) should be compatible with the x0x mainboard
after a bit of "rough" simulations in falstad, it seems 2x 680uF is a better choice for the reservoir caps after the rectifier
i measured my 9VAC/0.6A adapter when it has no load, and i get 11.4VAC from it

so in the simulator i added a 50Hz sine wave voltage source with 11.4VAC, then a series resistor of 4 ohms and a "RMS" current meter
when i put a 15 ohm load - the voltage drops to 9VRMS, and i get 599.9mA(rms)
i hope this is a "not too poor" representation of an 9VAC adapter
then i remove the 15ohm resistor and add a rectifier, reservoir cap, and a load..
2x 330uF results in a ripple between 6.29V and 10.12V
2x680uF ripples between 7.43V and 9.4V

i thought about adding an option to skip the FTDI and expose a connector for external UART, but there are a few questions
- is there a common standard pinout for such connector?
- should it be RX/TX/GND or also 5V?
- should the external UART provide its 5V? then we could actually use the two opto-couplers in this circuit

and, on the other hand.. i could try to get the exact locations of the audio/din/usb/barrel jacks, so that we can hopefull reuse the rear panel, but i want to add some more things
n0nx0x2 has two 5V logic outputs which can be used to sync some devices (boss dr-220, korg volca, etc..)
so i'm thinking about using some of the spare channels from the 4050 to buffer those, and add a few more MJ-3536N jacks
n0nx0x2 (the dev version) also has something like a "hard disk busy" LED output, so i'm also thinking to add a LED or two which could be seen on the rear panel
these will be available on additional connectors, and using them will be optional and fully up to the user

antto
 
Posts: 1618
Joined: Thu Apr 15, 2010 3:21 pm
Location: 127.0.0.1

Re: Power Supply Modding

by 3phase on Thu Jan 25, 2018 4:49 pm

what is with theese minatur dc/dc boards youget theese days in all szises? I got a few really small ones for a 606 modding project where the sumida coil got removed to do duty in a xox box.

https://www.ebay.de/itm/MT3608-DC-DC-St ... Swjc5Z32CH

They should have some similar features or high freq bleed as the original 303 dc/dc converter,,so it would be possible to use one or two of them and than just the original 303 circuit where needed.. => the audio and refference DC

I also found a lot of bleading of the oscilators into the filter of the original 303. i would think that comes thru the refference voltage supply..

Dont have a xoxbox anymore..is that wave bleed present in the xox filter too? on my 303 that bleed onto the cntrol voltage is rather strong..in 5-10% of the voltage swing range.. that is fm modulation of the filter.. is that because of weakness of the original 303 power supply? if so it shouldnt be as visible on the xox
3phase
 
Posts: 203
Joined: Wed Apr 22, 2009 2:06 pm

Re: Power Supply Modding

by 3phase on Thu Jan 25, 2018 4:57 pm

303filt cv - 1.jpg
acc to filter CV at summing point
303filt cv - 1.jpg (25.89 KiB) Viewed 713 times


division 0,5v Testpoint collector Q11
3phase
 
Posts: 203
Joined: Wed Apr 22, 2009 2:06 pm

Re: Power Supply Modding

by Nordcore on Sat Jan 27, 2018 3:52 pm

AGAIN: there is no "weakness" in the supply. Everything that is there is in the schematic!
And that is:
R172: 100Ohm.

The DC/DC does not have any (HF) bleed. In fact, it is so clean that the 606 can use it w/o R172*C69[1] filtering, directly as 15V supply for the analogue part.

What *does* happen is, that the current through R172 causes a little to much voltage dropout, so the AN1358 looses regulation. If this happens at all and how large the effect is, depends on the actual value of the 15V supply, mainly determined by D42 and its tolerance.
Antto had made the 15V adjustable in his new schematic, so that works out well for cloning the behavior. (... and not the optical appearance ... )
The other influence is the actual value of the 12V, which is set by two 5% resistors R178 (2k2) and R179 (1k8) . I would suggest to add in trimmer here, so one can adjust the 12V a little. (may be: R178=2k, R179=1k5, 500Ohm Trimmer in between them, trim slider to pin2 U2A. )
This has a relative large influence on the maximum setting of the accent. And of course, of the "lost regulation" behavior.

And again: dropout voltage of the AN1358 is 2.5V. Current consumption is around 20mA.


The H11L1 is pretty expensive (and *a lot* faster than needed), the PC900 is obsolete. The 6N138 is currently the usual selection for MIDI.
The 74HC4050 is not that much needed (.. I'd select a 74HC14, of course two gates needed per buffer, as it is inverting ... and no, there is no real reason for that... ) except for the midi through buffer.
The ATMEL could drive 20mA, the 74HCxx does 6mA. both enough for the 5mA MIDI uses up.

Rest of the circuit seems to look OK at a first glance, but I'll dig in a little bit more to check for the details ( ... around the FTDI ...may be one or two LEDs? ... )

The DIN-SYNC has 2k2 resistors, Roland has used 1k often, which I would prefer. (Again, just a gut feeling...)
Usually only two of the signals are used (at least my software does so), so one might think of a splitter cable.
A "compatible" solution would be to have two of the signals as pads on the PCB, some jumpers to select where the signals go, and the builder might drill another hole in the case and wire the jack to that pads.


_____
[1] 47µF in the 303, so 100Ohm *47µF=4,7ms. In the x0x it is 100Ohm*100µF=10ms. Which does not do any audible difference until you reach the "lost regulation" mode of operation.

Nordcore
 
Posts: 93
Joined: Sat May 23, 2015 3:14 pm

Re: Power Supply Modding

by antto on Sun Jan 28, 2018 5:28 am

after i added the stuff to the PCB, it turns out there is (of course, what was i thinking) not too much space on the rear to include extra audio jacks
if i squeeze everything as tight as possible - i could barely fit two extra 3.5mm audio jacks
unless i change some of the jacks to other models which take up less space on the board
panel-mounted jacks are another option

all 4 of the dinsync signals may be used, the 303 uses 3 of them (iirc)
the resistor values should not be a problem, as long as they are there.. the values in the schematic and BOM can always be changed later

i added the 4050 buffers because i remember something about "20mA max per port" .. or that might be for PICs and not AVRs
also, the 4050 can play the role of a level shifter

i wouldn't think the exact value of the 12V would make a significant difference, and so far i think it is usually around 11.84 to 11.89V
but i can add another trimpot, in a way that you could populate yours, while i'll skip it (replace it with a jumper) and just go with the basic values for the two resistors

i'll change to 6N138 then..
what worries me around the FTDI is whether the UART from it is 3.3V, because that's not too clear from the datasheet
also, i've found that the two UARTs in the x0xb0x (for MIDI and USB) interfere with each other and data gets corrupted, i don't remember the details, but that's another reason why i wanted to have some kind of logic buffer

another thing i've been thinking about..
to use SMD for some of the passive elements.. the simpler ones like resistors/capacitors (two legged SMD 1206 should be easy to solder/desolder with just a normal soldering iron, right? )
i will try to only use that for where there are at least 2 or more components of the same value, e.g. 100nF caps would be candidates for SMD, 2.2k/100k resistors on the dinsync jack too, 220 ohm series resistors on the MIDI probably too
what do you think about that?

antto
 
Posts: 1618
Joined: Thu Apr 15, 2010 3:21 pm
Location: 127.0.0.1

Re: Power Supply Modding

by antto on Sun Jan 28, 2018 5:50 am

3phase: afaik, it's not the voltage that determines the cutoff frequency of the filter, it's the current thru R69
my "software model" of the filter doesn't have cutoff FM and it matches
my other "software model" does have it (similar to the VCO->VCF FM mod) and that's visible on the spectrograph even when it's not very audible

what about those switching supplies on ebay?

antto
 
Posts: 1618
Joined: Thu Apr 15, 2010 3:21 pm
Location: 127.0.0.1

Re: Power Supply Modding

by 3phase on Sun Jan 28, 2018 6:17 am

Nordcore wrote:AGAIN: there is no "weakness" in the supply. Everything that is there is in the schematic!
And that is:
R172: 100Ohm.


and how does that differ to the false old school statements and advices for a power supply mod of the xox box?

havent that been proved to not cure the stiffness again and again? Do you ever heard a real 303 or tried a xox with 303 supply?

When it would be that simple, why is it still an issue? And are you 100% sure that there cant be any detail that has escaped your analyses? Nothing again your expertise, but in the pro audio teams i ve worked we had some problems where some of the best audio circuit developers of my country was consulted, actually guys with an international known and respected name, that have their own companies with a handfull of employes and big rhode and schwarz analyzers.. real lab guys.. With similar projects where ancient 1950 cult audio technology was ment to be cloned.. and again and again this guys was prooven wrong by the ingenuity of the 1950´s developers, that used some tricks that never made it in any school book. they also claimed one resistor would be enough to do this and that,.. but.. nope..it wasnt....once we proved the situation to them they just needed 30 seconds to explain what was going on..they actually had the knowledge allready,,but they didnt expected anything hidden..and wasnt thinking in that direction.. This dot happens often, just had it 3 times in the last 20 years that electronic specialists can be wrong when it comes to sound influences of circuit design. But thats enough to question over simplified solutions a little. When the best guys in the industry can go in the trap of taking the own cemented knowledge as granted when it comes to sound phenomenas, it can happen to any circuit developer.

which resistor? r172? its years ago...but thats the one i tried, that dont works..just weakens the peaks..but dont adds elasticity.. When that failed i rebuilded the original 303 circuit..that gave a huge sound improvement.. more in the direction to get to the singing quality on accented high resonance lines..and more elastic, havent i posted a sound example here? anyway.. you might have noted that there are 2 clone versions out there that use a sumida coil.. in your world just for superstitious reasons. But aslong you dont allow yourself to look at the problem from another angel its not likely that you will find the reason why that coil has a usefull impact on the sound.. no hf bleed? ok.. than what? 20 mA..constant? or with too fast fluctuations to be seen on simple meters? You hear fluctuations on the real 303--ts allmost like a pumping compressor.. with batteries that get empty the machine actually sounds the best.. In the early 90´s we allways had a batch half empty batteries in the studio to have them at hand for the final mix run downs.. So reduced voltage before the coild does something good.. increases the breathing fx.. also modulates the tuning a little, buts not just the tuning.. anyway, i gues you guys are the wrong adress for the issue anyway. But maybe its good for future investigators into the topics to read that remarks out of the world of practical 303 application.

keep xoxing, good luck
3phase
 
Posts: 203
Joined: Wed Apr 22, 2009 2:06 pm

Re: Power Supply Modding

by antto on Sun Jan 28, 2018 6:31 am

3phase, are you even reading the conversation?
it's right there, with an adjustable supply for the analog section, you can configure your ~15VDC and "starve" the opamp in the x0xb0x
Nordcore already explained how the x0x supply works, and what the main difference is with the original one, read the first page!
he also explained why changing the 100ohm resistor in the x0xb0x is not too equivalent, since the voltage before it is about 15VDC in the 303, while much higher in the x0xb0x (due to the use of a voltage doubler)

maybe you don't like reading, then just look at the schematics we posted, and comment those instead
do you have anything to say about them?

antto
 
Posts: 1618
Joined: Thu Apr 15, 2010 3:21 pm
Location: 127.0.0.1

Re: Power Supply Modding

by 3phase on Sun Jan 28, 2018 6:42 am

antto wrote:3phase: afaik, it's not the voltage that determines the cutoff frequency of the filter, it's the current thru R69
my "software model" of the filter doesn't have cutoff FM and it matches
my other "software model" does have it (similar to the VCO->VCF FM mod) and that's visible on the spectrograph even when it's not very audible

what about those switching supplies on ebay?


i just use similar ones now on a modded 606 to run the sound stages on a higher voltage.. there are really small versions, but i havent found them on ebay yet..just posted an example picture.. maybe they also could be beneficial for alternative xox supplies.. no idea if they share sound modulating characteristics with the sumida coil design.. they are closer related than a standard voltage regulator. and you get step down and step up versions for whatever task at hand.. and best of all..as smaller as cheaper they get.. the boards i got just 2 euros each..

and regarding the control voltage on the filter.. i know that such filters are current controled, but the oscilator bleed i see in the voltage shoudl be present in the associated currentflow aswell.

Question ..where does that osc beeld comems from? maybe just from inside the filter bleeding thru R69? probably most likely, or?
Anyway..dont explains the power supply "myths".
3phase
 
Posts: 203
Joined: Wed Apr 22, 2009 2:06 pm

Re: Power Supply Modding

by 3phase on Sun Jan 28, 2018 6:46 am

antto wrote:3phase, are you even reading the conversation?


sure i do, but why is he referring to r172 than?
3phase
 
Posts: 203
Joined: Wed Apr 22, 2009 2:06 pm

Re: Power Supply Modding

by 3phase on Sun Jan 28, 2018 7:09 am

antto wrote:maybe you don't like reading, then just look at the schematics we posted, and comment those instead
do you have anything to say about them?


i already did with pointing you to the ready made boards that safe you from handling a magical coil since its all pre installed already.
3phase
 
Posts: 203
Joined: Wed Apr 22, 2009 2:06 pm

Re: Power Supply Modding

by antto on Sun Jan 28, 2018 7:24 am

that resistor is put there by design
together with the 47uF capacitor after it, it forms a sort of Lowpass filter, or a reservoir capacitor which can give out current quickly to the analog section, but for short amounts of time, since it cannot charge back too quickly due to the 100 ohm resistor
if i have to guess what its role is, i'd say it does a combination of a few things:
1) limits the average/max current that the analog section can draw
2) attenuates high-frequency content from the switching supply
3) serves as a reservoir that can give "decent" amount of current but only in periodic bursts

look at the 808 schematic, every drum section has its own version of the +15/-15 power rails that pass thru a resistor and then a reservoir cap, iirc they may even be the same values as here - 100ohms and 47uF

your "questions" were chaotic, and you got an answer from Nordcore to just one of them, then you couldn't make sense of it
the fact is, his explanation of the 303 power supply is the first thing i read that makes sense
it can be explained so simple:
the positive power rail to the opamp that is used as a regulator for the analog section, is passed thru a lowpass filter (100ohm + 47uF)
the first difference here is that the capacitor in the x0xb0x is bigger (100uF)
but the second difference, which is much more significant, is that the voltage before this resistor is different between the two
in the x0xb0x that voltage is high and somewhat not very stable (it comes from a voltage doubler)
in the 303 it comes from the magical DC-DC supply, and it's probably more stable, BUT it turns out to be ~15VDC, which is close to the bare minimum required for the opamp (which is not rail-to-rail) to manage to produce 11.87V under stress, especially since the 100ohm resistor

so, in other words, the difference is that the opamp in the x0xb0x is mostly running "stable" .. it's power rail may go up and down, but is usually higher than the bare minimum required for it to make 11.87V
while in the 303, the opamp power rail is already close to the bare minimum, and thus the opamp is more likely to be "starved" when it has to give bursts of current (when notes retrigger as mentioned earlier)

so it smells like the sumida coil maybe isn't so magical
what seems to be important is the voltage before the 100ohm+47uF structure
and as you can see, we are talking about making this voltage more similar to that in the 303, and also adjustable with a trimpot
from page one!

antto
 
Posts: 1618
Joined: Thu Apr 15, 2010 3:21 pm
Location: 127.0.0.1

Re: Power Supply Modding

by antto on Sun Jan 28, 2018 7:33 am

3phase wrote:
antto wrote:maybe you don't like reading, then just look at the schematics we posted, and comment those instead
do you have anything to say about them?


i already did with pointing you to the ready made boards that safe you from handling a magical coil since its all pre installed already.

you should've said that
well, i checked a bunch of them out, most of them were supposedly "2A" and very very poorly documented
they also appear similar but offer step-down and step-up variants
so tbh i'd rather just include the circuit on the board instead of hunting down "the right kind" of blue-ish no-brand DC-DC converter from ebay
we need something like 20mA from it, and the one i've put is supposedly calculated to give 100mA, so it should be fine

antto
 
Posts: 1618
Joined: Thu Apr 15, 2010 3:21 pm
Location: 127.0.0.1

Re: Power Supply Modding

by Nordcore on Sun Jan 28, 2018 11:08 am

The FTDI can be used with 3.3V I/O mode (as drawn), but than the 1k resistors are too large.
OR you could hook up pin 3 (VCCIO) to pin 12 (VCC from the USB) which would give (about) 5V I/O.
As the 6N138 is "proven reliable" in 5V applications and we could easily use 5V, I don't see any reason to gather new/own experience on how it runs with 3.3V...
( ... it should be no problem. )

Regarding space: may be resistor arrays help a little?

Nordcore
 
Posts: 93
Joined: Sat May 23, 2015 3:14 pm

Re: Power Supply Modding

by antto on Sun Jan 28, 2018 1:44 pm

okay, here's the update:
Image

seems even if the FTDI "VCCIO" is 3.3V - the thing should work since afaik the transistor in the 6N138 drives low, while otherwise it's the pullup resistor that gives the logic high, which is taken from there
so i hope that works

resistor arrays? SMD?

antto
 
Posts: 1618
Joined: Thu Apr 15, 2010 3:21 pm
Location: 127.0.0.1

Please be positive and constructive with your questions and comments.