Grounding micpre

[pazduha]

With paint completly removed from the parts touching each other. I get lower volume, but more nasty mid freq.

 

[Voyager10]

Can you try with the input to HJFP1 shorted directly to ground, e.g. by disconnecting the wire from the DI socket and soldering the ends together?

If it still hums, we can eliminate the input XLR, pad switch, transformer, and DI input from the list of possible causes.

 

[ruffrecords]

With paint completly removed from the parts touching each other. I get lower volume, but more nasty mid freq.

Are these tests with the input XLR pins 2 and 3 shorted together?

Cheers

Ian

 

[ruffrecords]

The way you show Chout wired in your diagram you have the XLR pin 2 connected to 0V. If you are wiring an unbalanced signal to an XLR then you should make pin 3 ground and pin 2 signal. It is possible that what you are connecting it to does not like pin2 grounded and singal on pin 3.

Cheers

Ian

Cheers

Ian

 

[pazduha]

In order of tests:

Are these tests with the input XLR pins 2 and 3 shorted together?

Sorry for the inconsistency . No. Bellow is the shorted version i redone (boxgoodconnections_2&3short). More lower fq problems.

Can you try with the input to HJFP1 shorted directly to ground

Yes. It remove some of the upper fq problems. (DI-shorted.png). If u look on my first layout pic there is still ground runing to the input transformer inside the veroboard, is that a problem?

you should make pin 3 ground and pin 2 signal

with pin 3 ground and pin 2 signal there is not much difference.

 

[Khron]

Since the main peak is at 100Hz, that pretty much rules out magnetic coupling; the issue is downstream of the rectifier, so some sort of grounding issue / ground loop.

 

[pazduha]

I suspect the problem might be with ground running from psu to hjfp2 to hjfp1 to input transformer and also to the caps and pots at the same time? That would make ground not be a single line but branch out and return together 3 times. Should the wiring in layout I proposed in layout_big_worm.png fix this?

​

 

[Khron]

https://sound-au.com/earthing.htm

 

[ruffrecords]

Dominant 100Hz implies a smoothing problem. Can you check the ripple on the 24V supply?

Cheers

Ian

 

[pazduha]

I noticed something:
For my regulator:
https://si.farnell.com/stmicroelect...13?CMP=e-email-sys-orderconfirmed-GLB-Product
Datasheet says (which I should be reading more):


Measured voltages with DVM under load:
21.34VAC ... before the diods
26.25VDC ... after diods
23.92VDC ... after L7824

Should I put more V out of transformer?

 

[pazduha]

Can you check the ripple on the 24V supply?

I will, but it will take a while, since I dont have a oscilloscope.

 

[Khron]

I will, but it will take a while, since I dont have a oscilloscope.

100Hz is low enough that you can measure with the AC range on your multimeter.

Datasheet says (which I should be reading more):

If anything, the spec to look at is a few rows lower, the "Dropout voltage", typically 2V. So for 24V out, you'll want a solid (at least) 26V input, ideally 27-28V (to account for some ripple).

 

[pazduha]

100Hz is low enough that you can measure with the AC range on your multimeter.

My UNI-T UT890C jumps all over the place when set on AC 6V. 4V, 130mV... wildly differnet numbers changing all the time anywhere on psu behind the diods. With 60VAC range its ok before the diods.

 

[Khron]

My UNI-T UT890C jumps all over the place when set on AC 6V. 4V, 130mV... wildly differnet numbers changing all the time anywhere on psu behind the diods. With 60VAC range its ok before the diods.

Probably safe to assume there's plenty of ripple voltage there then Can you get a more stable reading if you disconnect one of the preamps?

 

[pazduha]

I tested the phantom psu with no load. Same thing, wildly jumping numbers. Usualy if held longer scraching against the wires it shows 0.000V. Then completly disconected showing 0.167V or any number. Then going to 50mV. If I take the meter far away its 0.000V.

 

[rh001]

Hi!

There must be at least 28vdc at input of 7824 regulator. You can measure that with DMM on vdc setting which will mostly or completely ignore the ripple. so ripple voltage will not confuse the DMM's vdc reading. The ripple voltage at input to 7824 should be no more than 2vac rms. I looked at your posts...seems like custom power transformer is not putting out enough voltage.

It is very common for power transformer to induce hum in high gain circuits when located close to circuit as you have done. That transformer should be mounted in a separate metal enclosure with at lease 4 ft of wire connecting mic pre to power transformer enclosure. The ac mains power cable should connect at the power transformer enclosure and so obviously the mains power switch and fuse should be on that enclosure.

rectifier and power supply filter caps can be in the power enclosure but that's probably not necessary for hum/buzz avoidance, although it might be...no way to be sure without trying. Voltage regulator and post regulator capacitor (470uf to 1000uf believe it or not) can also be in the power enclosure.

AC mains ground wire first connects to power enclosure. Then wire from that point is included in 5 conductor wire joining power enclosure to preamp enclosure. In preamp enclosure that mains ground wire is connected to chassis near where it enters chassis. Now that is the star ground point for the preamp.

5 wires (assuming rectifier and voltage regulator is in external enclosure):
1) ac mains gnd
2) voltage regulator gnd
3) +Vdc preamp
4) -Vdc preamp
5) +48 vdc phantom power

If power supply components ( like rectifier, filter caps, regulator and such) are located in preamp chassis that star ground point is their ground point.

mic transformer and preamp circuits also use that star ground point. Star does not have to be super perfect...there can be a little cheating, for instance low side of mic transformer secondary can go to preamp circuit ground near it's input instead of to star ground point. but there should be a ground wire from star ground point on preamp chassis to preamp circuit ground. so yes, preamp circuit gnd will get gnd wire #2 (see above) AND wire to preamp star gnd point.

Pin 1 of xlr connectors should go to star ground point, not chassis. I know that sounds unusual but it works best, especially with phantom powered mics.
Pin 3 of output xlr can just be joined to pin 1 and it's wire to star ground point.

Output xlr pin 2 is the hot pin.

The ground for pots can go to preamp circuit ground near where the wiper goes.

Preamp output to pot, then to output xlr is to be avoided unless 1kohm or less, like 600 ohm. If due to ouput pot rotation position the impedance (resistance) is more than 1 khom the air and even some impact will be lost and that loss will depend on length and capacitance of cable connection output to next piece of gear.

Note: within preamp chassis all circuit ground wires should be neat and as much as possible run together....can be twisited together or kept
together with cable ties.

Once you do as described here you will have no hum contributed by power supply circuit or lack of sheilding. You can then try cheating to see what you can get away with. For instance, maybe you can put toroidal power transformer in mic pre chassis located away from mic transformer and preamp circuits. Try toroidal transformer connected with 14 inch wires so you can try different positions and rotations...even on edge (that is, standing up instead of lying flat).

I am attaching pics showing mic pre card in 1U rack chassis with toroidal power transformer. Note toroidal transformer is as far as can be from mic pre circuit. Even so it had to be rotated to find lowest hum/buzz position and wrapped in a band of mu metal. Voltage regulator gnd on it's pcb had to be connected to chassis at mic preamp rather than more direct route on same pcb running past transformer to same point as mains gnd is connected which is connected to chassis.

The pdf sketch is not complete...just conceptual.

 

[ruffrecords]

100Hz is low enough that you can measure with the AC range on your multimeter.

Via a film capacitor, not directly.

If anything, the spec to look at is a few rows lower, the "Dropout voltage", typically 2V. So for 24V out, you'll want a solid (at least) 26V input, ideally 27-28V (to account for some ripple).

I agree. It looks like a supply ripple problem.

Cheers

Ian

 

[ruffrecords]

I noticed something:
For my regulator:
https://si.farnell.com/stmicroelect...13?CMP=e-email-sys-orderconfirmed-GLB-Product
Datasheet says (which I should be reading more):
View attachment 134303

Measured voltages with DVM under load:
21.34VAC ... before the diods
26.25VDC ... after diods
23.92VDC ... after L7824

Should I put more V out of transformer?

More voltage from the transformer would help. 24V should be sufficient. In the mean time you can reduce ripple by increasing the smoothing capacitor value. I suggest you try 2000uF.

Cheers

Ian

 

[Khron]

Via a film capacitor, not directly.

Doesn't the AC range on multimeters already have a DC-blocking capacitor on the input?

 

[My3gger]

The old "buzzing transformer" isn't buzzing under loads up to 180mA at 24vdc, or 60mA at 24V in this circuit, didn't try it with more than 180mA. It puts out over 30vdc after full wave rectifier loaded with 180mA of current. Noise in recordings sent by Miha is below that of my monitors, it can and should be improved, the real problem is how recordings sound compared to good HJFP op amp with 2N5457. We went though what you guys are saying, several times.