Noise issues with Gen3 and Syntonie

Helpdesk
1

Hey,

In reorganizing my system to reduce cable lengths, I encountered some power and/or RF noise issues. Power sources: Doepfer PSU3 & DC Distro (5 amp version).

@syntonie VU005 Dual Waveshaper did not like sitting next to Proc. Both VU005 and Proc powered from PSU3. Incredibly massive noise coming out of the VU005. Just moving them 4HP apart from one another completely resolved the issue. But my layout didn’t allow that as a permanent solution, so I ended up having to move the VU005s to a different case. Not the optimal workaround from an ergonomic standpoint.

Syntonie VU002 Frequency Doubler did not like sitting next to Stairs. VU002 powered from PSU3, Stairs from DC Distro. Visible noise coming out of VU002. Not nearly as bad as the VU005, but still unacceptable. Again, moving the VU002 resolved the issue, but again, it’s not optimal in terms of physical layout.

Stairs itself was super noisy when powered from PSU3. Independent of placement relative to other modules, all outputs of Stairs looked like hell. Switching to DC Distro helped. Outputs 2 through 8 now look fine. But outputs 16 through 64 are not up to my standards. Maybe it would be less of an issue if I was working in SD rather than HD, but that’s not an option artistically. I guess Stairs is another one of those high gain circuits, like Topogram, that’s just super susceptible to power noise. But it’s disappointing. The 16x output of VU002 is cleaner than that of the 16x output of Stairs. That was a surprise, I assumed that Gen3 would be cleaner. Maybe there’s an issue with this particular Stairs?

So, thoughts? Is this normal? Is Gen3 leaking a ton of RF? Are the Syntonie modules especially susceptible? Is there a possible solution other than physical separation? Maybe shielding? I did a search for RF shielding tape and foil, but everything I saw was conductive. I wouldn’t want to use that for fear of shorting out some delicate component. Is there some product that blocks RF but does not have a conductive outer surface? Like copper shielding in an insulator sandwich?

Thanks

2

Are you running HD signals through the VU005? I don’t think that circuit was optimized for HD. Surprising that the Syntonie doubler appears clearer than Stairs. Maybe it comes down to component tolerances, or maybe the phase?

As far as shielding, you could potentially sandwich the foil between layers of wax paper DIY style. Where that falls between a Farraday cage and a capacitor that bleeds high frequencies, I can’t say.

3

Noise issues are always a bit hard to pinpoint since it can be conducted from the PSU to the modules through the power cables, in this case, heavy filtering of the PSU outputs and module input will help. However there is also radiated noise, which is harder to mitigate, as it requires shielding of the PSU, though one small opening on the shield will make the noise leak again.
One first guess is that maybe there was already some noise coming from the PSU, though it was “hidden” by using longer cables which were filtering higher frequencies.

The system I’m using right now is housed in a Doepfer case with PSU3 and I got all the Gen3 modules into a 4MS pod powered through DC barrel and I’m pretty satisfied with the noise floor, then since those are two separated cases, I’m often using up to 60cm cables to patch them together, which may help with noise, but as discussed in the other thread about cables, will result in a softer output too. I don’t have a Proc, however I got a Stairs and will try to reproduce the issue you have.

Then when it comes to rectifier/frequency doubler circuits as Stairs and VU002 are based on, they have a bit of gain at each stage, which may increase the noise floor. Another thing to note is that, since it brings the input signal to higher frequencies, noise will be more visible with more details than with a non-rectified ramp for example. About the output of VU002 seemingly being cleaner, it may comes from the fact that it is a bit more limited in bandwidth than Stairs (?). Since VU002 is built from through hole components, with multiple boards stacked with connectors connecting them, the output looks a bit softer than Stairs from what I’ve tried. Then haven’t done proper bandwidth tests, mostly comparing the output of the synth using both modules, as I’m not sure what would be the best way to test it, since it is usually done by sending a square wave and measuring the rise and fall time at the output, though rectifier doesn’t really work with square wave, since it needs a slope to effectively rectify.

Unfortunately, I don’t think there is much other solution than try to move things around and see if it improves, and as I said in the cable thread, seems like it will always be a tradeoff between noise immunity vs bandwidth.

It’s true that at the time I did VU005, I was running my system in SD so it wasn’t thought for HD, though it looks pretty decent in 720p60 from what I’ve tried. Then since the goal of a sine shaper is to soften the edge of a triangle to produce the desired sine shape, it inherently reduce the bandwidth a bit.

5 Likes
4

FWIW, The VU005 is one of my all time favorite modules. :clap:

6 Likes
5

I would try powering the Gen3 modules directly from DC barrel distribution and a separate power brick (rather than use the 12V rails of the EuroRack power supply) and see if that makes a difference.

I would make sure all modules are mounted on the rails with mounting screws – if some are left disconnected, that can definitely introduce more EMI issues.

1 Like
6

The PCB planes and rear panel on the Gen3 assembly are intended to shield the circuit (each board has at least one full coverage ground plane, most of them have 2). You could certainly experiment with some copper shielding, but I don’t know if there is some quick fix that resolves it. Maybe a thin piece of copper sheet just to experiment with could help you determine if shielding helps or not. High gain circuits will definitely amplify whatever noise is already present in the circuit, so it is normal to see the relationship of higher gain = more noise. I can definitely file this as an issue and investigate more – it would be good to publish some datasheet style measurements.

2 Likes
7

VU005 works great with HD, I’m not seeing any noticeable high frequency loss. And I have to echo the statement, this is one of my favorite modules. Just got two more in the mail today! So I’ll be able to process two channels of RGB+Y once TBC lands!!

I thought about powering Proc from the DC Distro rather than the PSU3 to deal with the VU0005 noise. But to be honest, I panicked! Just took the easy way out and moved the Syntonie modules to a different case entirely.

Interestingly, I haven’t seen any issues at all with VU009, VU008, or VU006. But I haven’t done an exhaustive analysis of which modules don’t like each other, and I’m not going to. Right now, VU009 sawtooth oscillators are installed next to DSG3, mainly for convenience with sync cables, and I’m not seeing any noise.

All modules are screwed in properly. I realized that these Doepfer rails are kind of shallow, and depending on the thickness of the module panel, a 6mm M3 screw can be too long. Switching to shorter M3 screws has secured everything much better. But the noise issues I saw came after that.

3 Likes
8

Interesting, side note, my Mapper got interference patterns in the output when I installed a Chronenberg / Scanners next to it. Going to test removing the Scanners soon. But am sure I had clean Mapper output in HD before Scanners went in. Edit: turns out it was a power distribution remix that helped Mapper lose it’s noise problem

9

“Interesting”, indeed. My shiny new DWO3’s are way noisier than I expected. I thought they were going to be ultra super clean like DGS3 and Proc. But I know that video rate oscillators are a technical deign and manufacturing challenge, especially in HD.

They’re getting power from DC Distro. Installed next to BSO Crossfade. I will try taking the Crossfade out to see if that makes a difference.

1 Like
10

are they noisy all the way through the frequency range or mostly up toward the top end?

would you mind taking a screen capture I’d like to get a better idea of what is showing up

11

Can you characterize the noise you’re seeing or post a screenshot? There are many places in a video oscillator’s frequency sweep (like sync position #2 on the dial, and any asynchronous high frequency ranges) where the signal will be astable. But you should be seeing clean waveforms in any of the sync’ed settings (positions #1 and #3 on the dial.) It’s possible you could see noise embedded in an external signal, if that signal is used to FM the VCO.

1 Like
12

There’s a video demo available. Towards the end of the video (17 minutes in), you can see the more astable settings being used to intentionally create texture. Towards the beginning of the video, you can see the clean waveforms (6:15 is a nice pattern.)

13

Installed DWO3 today, still working on testing it, but seems stable and clean so far. I removed Chronenberg / Scanners and the noise in my Mapper was gone, reinstalled it and it’s back. Using rail power from the TipTop Mantis case, but it shouldn’t be at fault here. I do have the DC module and can try that with some case remixes, but feel like something is amiss somewhere on the module, as I’ve never seen interference this bad with anything else. (note that Visual cortex is in the same power group and don’t see any interference in it’s outputs, I believe there should be no issue with amount of power, but haven’t double checked yet)

Here is a clean mapper

and the diagonal interference once Chronenberg is installed next to it on same power system (not being used at all in patch)

it’s a bit hard to see when recompressed by Vimeo, here is screenshots

Screen Shot 2023-01-24 at 11.43.20 AM
Screen Shot 2023-01-24 at 11.43.20 AM1920×1079 81.5 KB

Screen Shot 2023-01-24 at 11.43.28 AM

14

Installed Shutter next to mapper and moved Chronenberg to the other case, but still seeing interference in Mapper… Tried removing shutter and now interference is still there when it was gone before… hmm maybe something about Mapper? super weird, doesn’t seem to be specific to other third party module combos

Note that I had tested this with Mapper going direct to ESG3, and when there was interference present, in previous tests I found using the input source ramp that was going to mapper direct to esg3 was clean, thus Mapper is where it’s happening.

I moved Mapper to a different power bus and it’s looking better for the moment :sweat_smile: :man_shrugging:

15

I am interested to hear which power brick you have and how many modules/current draw of said modules. Do you have the DC Distro 3A or 5A? I can’t quite make out which one you have from the pic on discord.

16

Thanks @Fox , all my tests so far are using TipTop Mantis power. Seems to be something about the way Mapper plays with others on shared power. I don’t think it’s your or VHS modules at fault. I have it on a different power bus with a bunch of random makers modules with lower power draws and it seems ok for now.

17

I’m just concerned with how the current draws add up across all of the modules and power bricks.
Is there one power brick for the mantis power another power brick for the GEN3 modules?

18

Here is the spec for their power. I don’t think any of my combos have gone over the Positive limit, just putting it all in my private mod grid now to check about negative too. LoFiFuture T-420 is missing from their site I’ve just found.

Bus board: Zeus Powered Mantis series with 36 Eurorack keyed connectors Input Voltage 13VDC
DC Regulator model: Zeus DC4000
DC4000 maximum output current:

+12V@3A with split zone topology (1A per zone, 3 zone total)
-12V@1.1A
+5V@300ma
Power Adapter: Tiptop Audio 3000ma Mantis Universal Power Adapter

19

Found t-420 on mod grid under other maker, not sure if the minus draw 0 spec is correct, but if so I’m under all power limits
2419 mA +12V | 902 mA -12V

Funny enough my bottom row ads up 1505 mA on one 1A zone and has been operational. Maybe I should move some modules power to upper buses too. :sweat_smile:

20

Rearranged power to all be proper limit per zone, looks good, I still see a bit of interference on high contrast gradient edge areas out of mapper sometimes, but it’s much more subtle and not always noticeable.