Jump to content

AINSER + CLCD display = more analog jitter


tago
 Share

Recommended Posts

Hi,

i built an AINSER8 Module on perfboard. When using it as the only module connected to the core the jitter seems reasonable.

Now if i connect a CLCD (OLED/Winstar 40x2) to the core the great jittering begins. (7bit alternates +/- 1)

Next i tested a standard 128x64 OLED which seems not to induce additional jitter.

Has someone experienced something like this? Any ideas how to solve such issues?

 

Thank you very much!

Edited by tago
Link to comment
Share on other sites

Try to power the AINSER from a separate, dedicated PSU. Or try to power the display from it’s own PSU or regulator. Some older systems tried to keep these issues under control by providing VCC for digital stuff, well regulated V+- for analog electronics as well as AGND, plus a separate dirty DGND for digital, switching stuff.

The Preen FM2 needs some care with shielding the OLED coil, or replacing it on the OLED PCB with a quieter model, that’s also worth investigating.

Edited by jojjelito
Autocorrect...
Link to comment
Share on other sites

Thanks @jojjelito

I have some further questions if you don't mind, since i don't have much experience in electronics.

I'd like to power the system via a single USB port.

Does a separate regulator for the display or AINSER decouple each other in a sufficiently manner?

How do i get a separate AGND with CORE_STM32?

Do you mean by 'other OLED PCB', the pcb on the display itself?

Edited by tago
Link to comment
Share on other sites

Hello

Dedicated regulator for AINSER8 Vref help a lot...

Unfortunately it will be difficult to regulate 5V from a 5V USB power (usual voltage drop at regulator)

With the proper design it's possible to run the AINSER8 at 11bit without jitter and 12bit low jitter.

As already mentioned, 0V path is also a key to achieve this result.

Best

Zam

Link to comment
Share on other sites

Thanks @Zam

So in my case i need a separate voltage regulator for the display, because AINSER needs the full 5V and therefor should be left alone? The displays works from 3.3V to 5V and i suppose a small voltage drop is not much of an issue. What type of regulator should i use?

I read about bulk or decoupling capacitors. Do you think they could be of help?

Link to comment
Share on other sites

A single USB won’t give you that much wattage to play around with. Plus, the USB power itself can be assumed to be quite dirty if derived from a cheap switcher shared with computer components.

Your best bet would be to derive multiple voltages from 9VDC, or even 9VAC if you need bipolar voltage and no fancy solution for -12V from +9VDC (inverter or DCDC converter), then use multiple 78xx or 79xx with the addition of a 78/79xx compatible switcher for the digital VCC and OLED VCC.

The Preen FM2 OLED saga is described here: http://ixox.fr/forum/index.php?topic=69159.0

Basically, shielding works. Separate power will also help.

Link to comment
Share on other sites

The solution is easy:

Go and get a MCP1541. You connect USB 5V to the input and it's output you connect to the VREF input on the AINSER8 (of course you need to cut the trace between VREF and +5V on the AINSER) and to the all analog controls that are connected to the AINSER8. 

This little IC solved all my problems I had with jitter. Especially when running just at 7bit, this should do the trick.

Link to comment
Share on other sites

Thanks @FantomXR

I really like your solution.

The MCP1541 datasheets says 4V output. Is really enough for Vref? What effect has a 1V drop for the A/D converter? Does it mean -20% resolution? If i understand your solution correctly the quality of the Vref is more important than the voltage range.

Do have to branch the USB power before it goes into the core or can use the 5V of J19 as input voltage for the MCP1541?

 

Edited by tago
Link to comment
Share on other sites

I use this method a year or so on different cores with different configurations. My last build was a faderbox with 9 100mm faders which can send 10bit resolution with USB-power. If you only need 7bit, there shouldn't be any problem.

Yes, it has 4V Output. Of course 5V would be better. But as @Zam already said: You can not really clean up 5V USB and end up with 5V. You will loose something.

I'm okay with it, because it works perfectly. Even on breadboard. No need for a high-precision PCB-design to get this working. Just try it. It costs just a few cents. 

I have a few capacitors connected to the USB-5V-rail and an inductor in front. But even that is not necessary. My earlier designs didn't have that filter-caps and worked very well too. 

Link to comment
Share on other sites

Hello

21 hours ago, tago said:

The MCP1541 datasheets says 4V output. Is really enough for Vref? What effect has a 1V drop for the A/D converter? Does it mean -20% resolution?

No, you still have 100% res , but steps is lower, 4v is more or less 1mV per bit at 12bit (4096 steps) wile 5V it's about 1.25mV

in fact the MCP1541 is designed to produce 1mV/bit resolution at 12bit ! (data say 4096mv Vout)

This won't change that much for ripple, noise and jitter from external interferences.

22 hours ago, tago said:

Do have to branch the USB power before it goes into the core or can use the 5V of J19 as input voltage for the MCP1541?

Use the digital 5V close to the MCP3208 (coming from J19) and don't forget to use the same Vref for your pots or faders, be careful about load as this regulators don't offer that much mA at output (don't drive anything but ADC and pots)

 

Have a look at MCP3208 datasheet page 19 for design notes and coupling with the precision regulator

Best

Zam

Link to comment
Share on other sites

1 hour ago, FantomXR said:

Well... at least for me it solved all problems :-) 

I mean 1mV per bit or 1.25mV per bit (according to different Vref, 4 or 5V) won't change that much if you have a regulator that clean the supply below or around mV noise/ripple, you'll still have jitter at 12bit but clean at 11bit or lower, which is way better than jittery 7bit for the OP !

Best

Zam

Link to comment
Share on other sites

12 hours ago, tago said:

Yeah, i couldn't figure out the needed input voltage either.

The whole datasheet define it more or less like a zener, so I suppose there is no voltage drop, but a "voltage limiter" behaviour.

Just my guess, but you better try a 4.1 version (4.096V) because if your 5VDD go down below 5V for any reason, your "precision zener" at 5.000V will follow. With the 4.1V version you have 20% margin before issue at analoge Vref.

Best

Zam

 

Link to comment
Share on other sites

On 6.4.2019 at 7:51 PM, FantomXR said:

The solution is easy:

Go and get a MCP1541. You connect USB 5V to the input and it's output you connect to the VREF input on the AINSER8 (of course you need to cut the trace between VREF and +5V on the AINSER) and to the all analog controls that are connected to the AINSER8. 

This little IC solved all my problems I had with jitter. Especially when running just at 7bit, this should do the trick.

Just to be sure. Vref is also used on the analog inputs J6 (AINSER8) instead of Vd (5V), right? If i connect pots to J6 they will provide max 4.096V (Vref) to the analog ins.

Edit: i wonder why the external psu on AINSER64 is only connected to Vref of the ADC, but not to J6-13.

Edited by tago
Link to comment
Share on other sites

1 hour ago, tago said:

Just to be sure. Vref is also used on the analog inputs J6 (AINSER8) instead of Vd (5V), right? If i connect pots to J6 they will provide max 4.096V (Vref) to the analog ins.

Edit: i wonder why the external psu on AINSER64 is only connected to Vref of the ADC, but not to J6-13.

Hello

Vref is your ref analog voltage, you need to provide the pot with the same voltage.

Vref is the 100% value for analog inputs whatever voltage you feed.

 

Please check again the AINSER64, Vref connect to ADC as J6 to 13 for pots (and 4051 multiplexed analog switches)

 

Best

Zam

Edited by Zam
Link to comment
Share on other sites

Hello

We can't call a resistor a decoupling device but you use " so it's ok.

In this case I think it's a small voltage drop to ensure Vref stay below VDD, to avoid error at full scale, as I nether won't call 10ohm resistor a current limiting for Vref input with half an amp at 5V...

If you use separate regulation you don't need this resistor.

Best

Zam

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...
 Share

×
×
  • Create New...