Pultec clone - grounding problems

[morls]

Hi,

I've started learning about signal analysis, and in the process have found an issue in both of my Drip Electronics Pultecs. There is significant noise from 50Hz, the mains frequency here in Australia. In a previous thread I got some excellent help and advice which I've followed, but it hasn't rectified the issue. I thought it best to start a new thread.

Although both units have the same issue, I'm working on one only to try and track it down. The advice I have followed is to simplify the grounding, both mains and audio, and to twist the power leads from the transformer to PCB, in order to eliminate potential sources of this issue. I've done that, and while I was at it I replaced the old audio cables with new cable, including a drain wire.

The kit has 3 PCBs. The main PCB is grounded in 3 places: from the PSU ground point, from the audio ground point and from another general ground point on the main PCB. I have also grounded to chassis each of the two control PCBs.

The AC mains socket, power transformer and PCB power ground are all connected to the same point on the chassis. The audio ground is connected to the chassis by itself. The input and output XLRs have pin 1 grounded to the XLR lug only.

My test is to run a 1KHz sine at +4dBu through the Pultecs, with EQ bypassed. The output is analysed in Flux Pure Analyser. I've also built a Drip LA-2A, which I've tested using the same inputs and outputs, just to confirm it's not an issues with the DAW or interface.

I'll attach a few pics - one has the spectra of pultec and LA-2A for comparison. Another shows the pultec spectra before cleaning up the grounding and after - it's actually worse after! The third is a pic of the layout.

Any help would be much appreciated!

Cheers
Stephen

 

[Bo Deadly]

You cannot mount your capacitors off-board like that. The 0V lead from the cap must be as close to the ground plane as possible. So unsolder those filter cap wires, solder the neg to neg directly through a very short bit of thick wire, run a wire from pos to pos and then zip tie them to the board through those mounting holes. But you'll have to put some kind of rubber pad or foam under to keep the metal cases of those caps from shorting out on something. And make sure the caps are thoroughly drained before you start handling them.

Another problem is that the mains power switch wires are running diagonal all the way across the enclosure right under or next to various magnetics. Get some longer wire and run it down the back and up the side to the switch. Tape it down tucked into the lower corner. And again, make sure they're twisted.

Fix those two things and see what happens. Then we can talk about implementing a proper star ground and get rid of all of those chassis to 0V connections. But I am a little more optimistic than others about these "drip" boards. I don't see why they couldn't be made to work well.

 

[warpie]

Do you get similar noise figures on both units?

 

[abbey road d enfer]

The main PCB is grounded in 3 places: from the PSU ground point, from the audio ground point and from another general ground point on the main PCB.

How can you hope "ground follows signal" to work? Weren't there any directions given with the PCB's regarding how they should be grounded?

I have also grounded to chassis each of the two control PCBs.

Are these connected to ground on the PCB's? If yes, they should be connected in order to make the shortes continuity, not going to a distant "star" point.

 

[Winston OBoogie]

Weren't there any directions given with the PCB's regarding how they should be grounded?

PCB's are Drip Electronics.

 

[abbey road d enfer]

PCB's are Drip Electronics.

I went to the site and saw that it's "Made from vegetable-tanned leather for its durability and firm structure,"
I'm not sure the grounding is so good...
More seriously, what does being made by Drip explain about the grounding issue?

 

[Winston OBoogie]

I went to the site and saw that it's "Made from vegetable-tanned leather for its durability and firm structure,"
I'm not sure the grounding is so good...
More seriously, what does being made by Drip explain about the grounding issue?

Ah, OK. Because, you asked about "directions" and there have been numerous issues brought up on here with folks not getting instructions/directions with boards they bought from Drip..
Not even grounding, I'm talking stuff like a BOM so you know what C13 and R8 are supposed to be etc.

I haven't looked at pics of many of the boards recently, but the ones I looked at years ago seemed to have placed more emphasis on "looking nice' at the expense of proper signal routing etc.

None of this helps the OP though so, best I shut up.

 

[abbey road d enfer]

Ah, OK. Because, you asked about "directions" and there have been numerous issues brought up on here with folks not getting instructions/directions with boards they bought from Drip..
Not even grounding, I'm talking stuff like a BOM so you know what C13 and R8 are supposed to be etc.

OK, so I did get you right. I didn't want to offend anyone, but you've done it for good!
Actually, the buyers should be thankful they know what they are supposed to build.

 

[morls]

Fix those two things and see what happens. Then we can talk about implementing a proper star ground and get rid of all of those chassis to 0V connections. But I am a little more optimistic than others about these "drip" boards. I don't see why they couldn't be made to work well.

Thanks, I'll sort this out.

Do you get similar noise figures on both units?

I do, yes, which gives me hope that once I get my layout right there will be improvement.

Are these connected to ground on the PCB's? If yes, they should be connected in order to make the shortes continuity, not going to a distant "star" point.

Yes, they're marked as ground on each PCB.

Ah, OK. Because, you asked about "directions" and there have been numerous issues brought up on here with folks not getting instructions/directions with boards they bought from Drip.
I haven't looked at pics of many of the boards recently, but the ones I looked at years ago seemed to have placed more emphasis on "looking nice' at the expense of proper signal routing etc.

I bought these, and other Drip boards, knowing that it's just the PCB that Drip offers. Gregory makes the point of stating in several places on his website that he doesn't offer build guides or BOMs. I'm currently building the dual STA-Level, and it's a great board to work on. It's faithful to the original circuit, so the original schematic is a huge help. I think I learn more from doing a project like this, where I have to look into each part of the build rather than just ordering a BOM and following instructions. The LA2A sounds fantastic, as do these pultecs even though there is this noise issue, which is reflects my inexperience rather than the quality of PCB or documentation.

 

[Audio1Man]

Hi morls

The GREEN WIRE grounding of the PCB is not the problem. Turn OFF the signal and run the FFT looking @ the Power supply HUM & NOISE. I would think that the PCB LAYOUT is poorly done as EVEN & ODD LINE HARMONICS are showing along with lack of critical damping / snubbing of the rectifiers need to be resolved first.

Bo Deadly comments are good, however I believe they will only give mild improvements.

Duke

 

[Bo Deadly]

The GREEN WIRE grounding of the PCB is not the problem.

You mean the multiple chassis to 0V connections? Sure it could be a problem. If those are returning supply, that violates the ground follows supply rule.

Bo Deadly comments are good, however I believe they will only give mild improvements.

Clearly there are multiple problems going on so it's really hard to say which fixes will yield the most improvement.

At this particular moment I'm betting that the power switch is actually the primary source of mains noise. The clue is that the noise got worse when those wires were twisted together. That's because the twist shortened them so much they're pulled up close to the bottom of the board such that mains goes directly under the high Z primary output transformer and right next to the EQ inductors.

But mounting the caps away from the board over wires that make a big loop is a close second.

I find this build to be quite interesting because just about all of the grounding rules were violated. It's just an effen disaster. Sorry morls!

Incidentally this looks like the documentation for this:

 

[Audio1Man]

Do we have a picture of the back side f the PCB?

 

[morls]

Clearly there are multiple problems going on so it's really hard to say which fixes will yield the most improvement.

I find this build to be quite interesting because just about all of the grounding rules were violated. It's just an effen disaster. Sorry morls!

Incidentally this looks like the documentation for this:

No problem Bo, I'm here to learn...

I'd just dug up that same documentation, and was about to post. In my own defence, I followed the grounding in this (except for the motor run caps, of course). The documentation clearly states to ground audio and XLRs pin 1 to the chassis together, and to ground 2 points from the PSU section to the transformer star ground. This is what I originally had.

There's no mention of grounding the smallerPCBs in the documentation. I've grounded each from points marked on the PCB to a separate point on the chassis and you can see in the photo I uploaded previously at the top rght there is a green wire from each of these smaller PCBs. Does this look OK?

I've just ordered some axial electrolytics for the PS, and will replace the motor run caps. I'll sort out the wiring from the power switch when I install these new caps, in a day or two. I'll also take a pic of the underside of the main PCB and post.

Thanks to all for the help so far
Stephen

 

[morls]

I've just noticed something on p7 of the manual. There is a section that mentions grounding the main PCB to the top control board. During the build I assumed this was referring to the ground of the shielded wire used to connect the two. Looking at it now, maybe it's actually connecting the PCB grounds? I've attached a grab, and have drawn in what I mean.

The same directions are given on p8, for the lower control board.

I'm confused by the marking on the lower board "main ground". Is this the chassis ground for both control boards? Do I need to connect upper and lower to main PCB, upper board to lower, and then from the main ground on lower to chassis?

 

[abbey road d enfer]

I don't know how to interpret this lack of/contradictory instructions, but I certainly would start with disconnecting the "grounds" on the smaller PCB's from the chassis. The connections via the shields should be enough in a well designed system.

 

[Bo Deadly]

This is how I would do it:



At least this is how I would try it to start. Thick black lines are 0V / chassis. Notice how you can trace any 0V line and not make a loop. This is sort of like a star or more accurately branches of a tree.

However, the one catch would be if the ground points on those sub-PCBs is actually not connected to the ground point next to the signal wires of the same PCB. So when you disconnect those wires, use your meter to check. I cannot imagine why they would not both be connected to ground plane but I have to wonder why those instructs say MUST in ALL CAPS that you have to connect those points to the "star ground". Some of the instructions are questionable so I'm going to assume those ground points are connected and therefore, like Abbey said, the shields will connect the sub-PCBs to the main PCB. There aren't even any currents in those lines since the EQP1 filter circuit is not powered.

Note that motor run caps are actually super good. I think they're effectively like non-polar film. Yeah they're a little clunky in this case but I would use them in a heartbeat. When all is said and done nobody is going to be looking at the inside of this thing. The only problem is insulating the metal cases from touching the board. You need a little piece of rubber or something to insulate.

 

[morls]

Wow, thanks Bo. I’ll give this a go and post the results.

I remember selecting the motor run caps based on the points you made about being like film caps, and thought I’d try them out. The specs are great. I just wasn’t aware of the issues created by mounting this way.

I've also found that my wiring of the power switch wasn't quite so disastrous - I tried to avoid interference issues by using the PCB mounting legs to run the wires around and away from the audio section of the board. They were touching the aluminium mounting legs though, so could this have contributed to the noise?

This pic shows the path I used for the power switch wiring.

 

[abbey road d enfer]

This pic shows the path I used for the power switch wiring.

Potential issue there. The power switch should be on the same side as the mains connector, and the wires maintained close to the chassis and far from any sensitive node.
Can you rotate the rear panel?

 

[morls]

I don’t think so. The power transformer is positioned right at the PCB connections, and the leads have been cut to fit.
I could drill a new hole in the front panel and put the switch on the other side. This would simplify the power switch wiring very nicely.

Bo, I notice in the grounding layout you drew that the only connection from pin 1 of both XLR sockets is to the audio ground, while at the PCB connection you have pin 1 connected to shield/drain of the cable. I have connected pin 1 at the PCB end AND at the XLR end. Pin 1 is also connected to the lug on the XLR. Should I remove this connection from XLR pin 1 to audio cable shield, and just connect the shield at the PCB end? Also, should XLR pin 1 be connected to the lug, or just audio ground?

I’ve tested the control PCB grounding, and the ground points are all connected, including the cable shield. There is also a ground connection between the two control boards, via points 8 and 10 of the wiring between the two.

 

[warpie]

I don’t think so. The power transformer is positioned right at the PCB connections, and the leads have been cut to fit.
I could drill a new hole in the front panel and put the switch on the other side. This would simplify the power switch wiring very nicely.

Bo, I notice in the grounding layout you drew that the only connection from pin 1 of both XLR sockets is to the audio ground, while at the PCB connection you have pin 1 connected to shield/drain of the cable. I have connected pin 1 at the PCB end AND at the XLR end. Pin 1 is also connected to the lug on the XLR. Should I remove this connection from XLR pin 1 to audio cable shield, and just connect the shield at the PCB end? Also, should XLR pin 1 be connected to the lug, or just audio ground?

I’ve tested the control PCB grounding, and the ground points are all connected, including the cable shield. There is also a ground connection between the two control boards, via points 8 and 10 of the wiring between the two.

I'm not Bo but as I mentioned in a previous post Pin 1 on both the input and output XLRs shloud be connected only to the chassis with the shortest possible path (i.e to the XLR lug).

The shlied/ drain of the cable should be connected only on one side to avoid ground loops. Not sure whether it should be connected to the chassis or to the audio ground although I believe it's better to connect it to the audio ground. So, I believe Bo's layout is correct.