My designs are always noisy :(

Hardware questions and issues with the FV-1

Moderator: frank

iampoor
Posts: 27
Joined: Sat Nov 15, 2014 3:44 pm

My designs are always noisy :(

Post by iampoor » Sat Feb 13, 2016 4:50 pm

Hi

I have been through quite a few different iterations of FV-1 based PCB designs. Unfortunately, they are always too noisy to be useful.

I think part of this has to do with my layouts (Might be my ground plane too), but I think I have finally figured out a solution. I was trying to cram everything onto 1 PCB an it was just getting messy. I have decided to use a few different "breakout" PCB's, and this will not only allow for more flexibility, but should help isolation the issues Im having.

Here is the current layout and filters I am planning on using. Before I was relaying on just the passive filters recommended in the datasheet, but I was getting far too much high frequency noise. I am planning on using a MFB filter with a -3db point at 12khz.

How do these boards look? Is there anyway I can improve the layout? I have laid out the input, 1 channel of filter, and the mixer on the analog board. How does this look as a starting point? What can I improve?Should I use a ground plane? (I think my ground plane has caused issues in the past, but I am not certain). I am well aware that some unnecessary caps are in the schematic, I use them for test points and for testing.

I know this is alot of questions at once, Im just really starting to get frustrated at this and want to figure out what Ive been doing wrong! :D

Here is a link to the schematics and layouts I have so far

http://imgur.com/a/xw6mr

Thanks!

frank
Posts: 991
Joined: Wed Oct 19, 2005 12:26 pm
Contact:

Post by frank » Sat Feb 13, 2016 10:09 pm

There are lots of possible causes of noise, how clean is your power source? If possible try a battery to power it to eliminate the power supply as the source of noise.

I looked at the schematic and boards and they don't seem to match up, the top schematic seems to have only one header but the small board as two.
Frank Thomson
Experimental Noize

iampoor
Posts: 27
Joined: Sat Nov 15, 2014 3:44 pm

Post by iampoor » Sun Feb 14, 2016 12:31 pm

frank wrote:There are lots of possible causes of noise, how clean is your power source? If possible try a battery to power it to eliminate the power supply as the source of noise.

I looked at the schematic and boards and they don't seem to match up, the top schematic seems to have only one header but the small board as two.
Hi
Power source is clean, I have noise issues with or without batteries. Its a mix of white noise (which seems to be coming from the outputs) and some 20-30hz oscillations usually. The wideband noise is objectionably loud, and seems to be increased when I sum multiple outputs. Hence I think I might need better filtering.

The top header on the larger PCB is for the control pots, the 2 bottom headers surrounded by the box are for the plugin board.

frank
Posts: 991
Joined: Wed Oct 19, 2005 12:26 pm
Contact:

Post by frank » Mon Feb 15, 2016 9:07 am

I don't see the filtering there to remove the high frequencies from the delta sigma converters, i.e. there is no 1K/2.2n cap on each of the outputs. Also, any filtering should be on the small board, as close to the pins as possible.
Frank Thomson
Experimental Noize

iampoor
Posts: 27
Joined: Sat Nov 15, 2014 3:44 pm

Post by iampoor » Mon Feb 15, 2016 6:23 pm

frank wrote:I don't see the filtering there to remove the high frequencies from the delta sigma converters, i.e. there is no 1K/2.2n cap on each of the outputs. Also, any filtering should be on the small board, as close to the pins as possible.
Hi, its there. R7/C4 and R12/C7 on the larger PCB.
I will move the filtering to the smaller board.

I am doing an order of these PCB's in a few days, so I will report back.

I am also including a 2nd order MFB filter, to see if I can get even better performance. I will report back. I think the issue in the past has been a messy layout with a not particularly solid ground plane.

iampoor
Posts: 27
Joined: Sat Nov 15, 2014 3:44 pm

Post by iampoor » Tue Mar 15, 2016 9:13 pm

Hi

Finally got the opportunity to build this

I used my soundcard (Mbox 2) to pickup the output noise on the main card the fv-1 is on. Here are are the noise plots Im getting. Not sure what it is so different for each channel. This is straight off the outputs after the passive filter recommended in the datasheet.

Got any ideas Frank? This progress is good, but it could be better. :D



Image

iampoor
Posts: 27
Joined: Sat Nov 15, 2014 3:44 pm

Post by iampoor » Tue Mar 15, 2016 9:18 pm

Here is the schematic and board layout I used too. It is slightly different from my original post. :wink:

http://imgur.com/a/uALhy

Im almost at the point of buying a development board to compare against. Ive been having so many noises issues with these designs. Not sure what Im doing wrong, or if my workbench is surrounded by emi...or...!?!?

I bought twenty chips and havent been able to use them yet. Sigh. Ill stop complaining.

Anyone in Central California have a demo board I can use for 1 day? :D

frank
Posts: 991
Joined: Wed Oct 19, 2005 12:26 pm
Contact:

Post by frank » Wed Mar 16, 2016 9:17 am

Please post complete final schematic of all boards, etc. Also higher res images of the boards as I cannot zoom in on the current one.
Frank Thomson
Experimental Noize

iampoor
Posts: 27
Joined: Sat Nov 15, 2014 3:44 pm

Post by iampoor » Wed Mar 16, 2016 11:19 pm

frank wrote:Please post complete final schematic of all boards, etc. Also higher res images of the boards as I cannot zoom in on the current one.
http://imgur.com/a/oNaeg

Here is an album, first 2 pictures are of the plugin board the FV-1 is on. It has a ground plane on both sides.



The second two pictures are the "analog" breakout board it is plugged into. I added additional filtering for experimentation on this board, but did not use it. I am tapping off the outputs at C1 and C6 on this board to make the measurements.

My Fv-1 program is taking a mono input. I am not using the unterminiated input.
Let me know what else I can provide. :)

frank
Posts: 991
Joined: Wed Oct 19, 2005 12:26 pm
Contact:

Post by frank » Thu Mar 17, 2016 11:05 am

OK, a few items to check:

You only have a 1nf on the AVDD/DVDD pins on the FV-1, that should be 100nf (0.1uf).

You don't have a bypass cap on the EEPROM, add a 0.1uf.

Do not tie anything to pin 3 of the FV-1 other than the 1uf cap!!!! It looks like you have pin 2 tied to it in the schematic. If only using 1 input then tie the unused one to the used one, drive both with the same signal.

Add more ties points between the upper and lower ground planes.

The ground on the large board is running all over the place and loops. Make a ground plane.
Frank Thomson
Experimental Noize

iampoor
Posts: 27
Joined: Sat Nov 15, 2014 3:44 pm

Post by iampoor » Fri Mar 18, 2016 4:19 pm

frank wrote:OK, a few items to check:

You only have a 1nf on the AVDD/DVDD pins on the FV-1, that should be 100nf (0.1uf).

You don't have a bypass cap on the EEPROM, add a 0.1uf.

Do not tie anything to pin 3 of the FV-1 other than the 1uf cap!!!! It looks like you have pin 2 tied to it in the schematic. If only using 1 input then tie the unused one to the used one, drive both with the same signal.

Add more ties points between the upper and lower ground planes.

The ground on the large board is running all over the place and loops. Make a ground plane.
Okay, I made alot of progress today.

Added a bypass on the eeprom, and change the wrong bypass value. Also tie the inputs together It seems to help somewhat.

Dont worry the input wasnt tied to pin 3, I just moved around some parts on the schematic and eagle added that.

I took ome noise plots and saw an improvement!

I will fix the grounding scheme on my next revision. I am currently not using the analog board for anything other than input buffering and power supply.

Now this is where things get weird....
I realized the noise is coming from multiple sources. My bench power supply is putting out some 2nd 3rd an 4th harmonic ripple. It also has a high pitched whine...That was one of the noise sources. I have been running it off battery power to eliminate those noise sources.

The next one I am confused about. The spectrum analyzer software I am running is on my PC through an Mbox. If I connect the ground lead to the ground (with both battery and with my bench power supply) the noise is dramatically reduced. If I then plug the mbox into my computer the noise is reduced even further. My laptop is running on charger with no earth connection (its a 2 prong). In fact the circuit is now very quiet!

Any ideas on what could be causing this behavior?

If I copuld get this sorted out, the circuit is now useable! I just have a few other minor issues to work out.

1. The circuit hums when I roll my guitar volume down.

2. There one other minor noise source I havent been able to pinpoint, but I am not concerned with it yet, its a low frequency oscillation. I think I need larger bypassing on my main supplies.

Otherwise, I can hear and detect is the white noise from the converter, and that is not objectionable at all. :D

frank
Posts: 991
Joined: Wed Oct 19, 2005 12:26 pm
Contact:

Post by frank » Sat Mar 19, 2016 1:04 pm

Tracking down noise sources can take lots of time and can be from external things like your power supply as you discovered.

"Ground" can be different things, the "ground" on a 3 prong plug is a chassis ground for safety and may not be the same as signal ground so you need to make sure all the signal grounds are connected together when making measurements. Do not connect the signal ground to chassis ground.

Hum can come from ground loops, bad cable, bad connector, etc.
Frank Thomson
Experimental Noize

iampoor
Posts: 27
Joined: Sat Nov 15, 2014 3:44 pm

Post by iampoor » Sat Mar 19, 2016 2:29 pm

frank wrote:Tracking down noise sources can take lots of time and can be from external things like your power supply as you discovered.

"Ground" can be different things, the "ground" on a 3 prong plug is a chassis ground for safety and may not be the same as signal ground so you need to make sure all the signal grounds are connected together when making measurements. Do not connect the signal ground to chassis ground.

Hum can come from ground loops, bad cable, bad connector, etc.
Yes, it has been quite the experience! :D
This is the first time I have ever had serious noise issues in my engineering career. Trying to not pull my hair out.

I usually refer to "chassis ground" as earth.

That is whats confusing me. My power supply, and laptop are both isolated from "earth". Yet when I plug in my audio interface and ground the circuit (weather on battery or my power supply) the noise basically vanishes. I need to do more testing. I wonder if it has something to do with the interface input resistance/impedance.

frank
Posts: 991
Joined: Wed Oct 19, 2005 12:26 pm
Contact:

Post by frank » Sat Mar 19, 2016 6:18 pm

What do you mean by "Yet when I plug in my audio interface and ground the circuit" ? Are the audio cables not grounded? Are you adding a separate ground?
Frank Thomson
Experimental Noize

iampoor
Posts: 27
Joined: Sat Nov 15, 2014 3:44 pm

Post by iampoor » Tue Mar 22, 2016 2:22 pm

frank wrote:What do you mean by "Yet when I plug in my audio interface and ground the circuit" ? Are the audio cables not grounded? Are you adding a separate ground?
Hi

What I meant wa, when I plug in my audio interface and then plug my laptop power supply in ("grounding" it). Then the noise goes away. I will continue to experiment with this.

Post Reply