So close and yet so far.... The "Heavy" preamp mess continues

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

CurtZHP

Well-known member
Joined
Mar 21, 2005
Messages
634
Location
Allentown, PA
Got it all hooked up again tonight.

Nasty ground-loop-type hum going on.  It gets far worse when I switch on the phantom power, but at least this time I can actually hear the mic.

A few questions for the time being....

1.  Where should the phantom power ground be connected in relation to the other grounds?

2.  All my connections from the audio circuit to the panel controls are shielded cable with the shields connected at one end only.  Does it matter which end?

3.  Should the case grounds on the transformers be connected to anything?  I currently have them connected to the same ground as the rest of the audio circuit.

4.  What considerations should I take into account regarding the location of the first stage grid resistors?  The first stage in general with regard to the power supply?
 
CurtZHP said:
1.  Where should the phantom power ground be connected in relation to the other grounds?
If the phantom supply is floating then its 0V should be connected to pin 1 of the mic XLR whiich in turn should be connected to the chassis at the connector. If the supply is not floating i.e. its 0V is already collected to analogue oV then there is nothing else to do. Personally I am concerned about your multiple bridge and voltage tripler? power supply. I think I have an small ols 50VAC toroid I bought ages ago for a ophantom supply. I'll happily send it to you
2.  All my connections from the audio circuit to the panel controls are shielded cable with the shields connected at one end only.  Does it matter which end?
Depends on what the controls are doing. If they need analogue 0V then you can usually use the screen fir that in which case it is connected at bth ends. If the screen is mean just to screen then it should be connected to chassis. You can connect it at either of both ends.
3.  Should the case grounds on the transformers be connected to anything?  I currently have them connected to the same ground as the rest of the audio circuit.
Some transformers have internal screens as well as connections to the case itself. Strictly speaking the screens should go to analogue 0V and the case to chassis but I always connect mine to analogue  0V because it is usually more convenient and I have not notice any performance degradation from doing it that way.
4.  What considerations should I take into account regarding the location of the first stage grid resistors?  The first stage in general with regard to the power supply?
Grid resistors should always be close to the grid - preferably wired right across the tube socket. If the transformer secondary is connected directly between grid and analogue 0V then this matters less as the low dc resistance of the transformer secondary will swamp the grid resistor. Obviously the first stage should be as far away from the power supply as possible. Definitely use a screening can for the first stage tube and orient the input transformer for minimum  pick up from the mains transformer(s).

Cheers

Ian
 
Thank, Ian!

Another observation...
Once I removed the pile of papers (schematic, notes....), I noticed that the mains AC cord was running pretty close to the tubes.  That might explain why it got a little quieter when I picked up the perf board it was on.

I've got two channels.  The left input XLR has Pin 1 attached to the chassis ground lug on that jack, as I've been advised to do.  Should I do the same on the right channel?  I have done.

The input stage grid resistor for each channel is certainly an issue, based on what you said.  They are currently the farthest components from the tube, relative to all others!  I'm surprised I'm not picking up the BBC!

I'll have to work up a drawing or diagram to show you my proposed control layout.  What I've done so far is take a blank 3RU rack panel and drill it out to mount the controls and jacks where I'd like them.  Also makes life easier when wiring things on the bench, and you don't have to worry about switches and pots wandering around.

 
CurtZHP said:
I've got two channels.  The left input XLR has Pin 1 attached to the chassis ground lug on that jack, as I've been advised to do.  Should I do the same on the right channel?  I have done.

Basic rule of thumb: All XLR pin 1 go direct to chassis. Despite what some people might tell you, pin 1 is NOT a signal pin; it is a SCREEN and as such should be connected straight to chassis at the earliest opportunity.

Cheers

Ian
 
ruffrecords said:
Basic rule of thumb: All XLR pin 1 go direct to chassis. Despite what some people might tell you, pin 1 is NOT a signal pin; it is a SCREEN and as such should be connected straight to chassis at the earliest opportunity.

Cheers

Ian


Wow, I did something right!

So, two things to do this weekend....
1.  Get the grid resistors  and inputs closer to the tube.  Like you said, I'll attach them right to the socket.
2.  Move that whole side of things away from the power supply.  Right now, they're practically right next to each other.

As for the phantom supply...
I did pretty much the same thing that Jakob did by the time I got this far.  Except I used a TL783 instead of a TIP121.  Are those voltage tripplers that much more susceptible to noise?
 
The other thing I should do is completely dismantle at least one channel and strip it back to its barest essentials.  No pad switch, no polarity switch, no HPF.  Ditch the instrument input.  Just get the basic preamp to work first, then add all the "bells and whistles," and see what gets me into trouble.

The other big construction challenge here is that I intended for this unit to be portable.  (Think Focusrite's ISA One or Scott Hamton's Hamptone...)  So the standard 2RU rack chassis is out, unless that's the ONLY way it will work.
 
CurtZHP said:
As for the phantom supply...
I did pretty much the same thing that Jakob did by the time I got this far.  Except I used a TL783 instead of a TIP121.  Are those voltage tripplers that much more susceptible to noise?

The TL781 is what I use for phantom power but it does like a good 20V across it to work properly. 12VAC tripled is only 50V.

Cheers

Ian
 
CurtZHP said:
The other thing I should do is completely dismantle at least one channel and strip it back to its barest essentials.  No pad switch, no polarity switch, no HPF.  Ditch the instrument input.  Just get the basic preamp to work first, then add all the "bells and whistles," and see what gets me into trouble.

I find instrument inputs a PITA because they nearly always involve a considerable length of high impedance screened cable. The "normal"way to do this is to wire the output of the mic transformer to the instrument socket and then wire from there to the grid of the first tube. That length of HI-Z cable is asking for trouble. In my Eurochannel design I placed a relay close to the first grid and placed the input transformer right next to it. That way, when the instrument input is not used, the signal path is very short. It is only when the instrument is used and the relay is operated that the HI-Z cabling has any length.
The other big construction challenge here is that I intended for this unit to be portable.  (Think Focusrite's ISA One or Scott Hamton's Hamptone...)  So the standard 2RU rack chassis is out, unless that's the ONLY way it will work.

You are squeezing a lot of functionality into a small box. This increases the heater supply which makes the mains transformer bigger and more likely to induce interference. HT requirements are much more modest. I recommend you consider replacing the heater supply with a switched mode power supply (SMPS). I am doing this as a matter of course in my design now. I use Meanwell SMPS which have an excellent reputation and which I have found to be extremely quiet. Always use a fully enclosed version.

Cheers

Ian
 
ruffrecords said:
I find instrument inputs a PITA because they nearly always involve a considerable length of high impedance screened cable. The "normal"way to do this is to wire the output of the mic transformer to the instrument socket and then wire from there to the grid of the first tube. That length of HI-Z cable is asking for trouble. In my Eurochannel design I placed a relay close to the first grid and placed the input transformer right next to it. That way, when the instrument input is not used, the signal path is very short. It is only when the instrument is used and the relay is operated that the HI-Z cabling has any length.


I've got the instrument input located right under the mic input, and if all goes according to plan, the input transformer is right under that.  Then it's a short hop (less than four inches) to the preamp circuit.  But I'm certainly not married to the idea of having an instrument input.  (After all, that's why they make direct boxes!)
 
ruffrecords said:
The TL781 is what I use for phantom power but it does like a good 20V across it to work properly. 12VAC tripled is only 50V.

Cheers

Ian

I actually have a 15VAC secondary feeding a voltage quadrupler, so the input to the regulator is about 63VDC.  (At least that's what I measured...)

(Did my schematic say 12VAC?  I might have to correct that.)
 
CurtZHP said:
I actually have a 15VAC secondary feeding a voltage quadrupler, so the input to the regulator is about 63VDC.  (At least that's what I measured...)

(Did my schematic say 12VAC?  I might have to correct that.)

No I just ASSumed.

Cheers

Ian
 
ruffrecords said:
No I just ASSumed.

Cheers

Ian

;D

Another question...
I'm not sure what you meant by...
You are squeezing a lot of functionality into a small box. This increases the heater supply...
What does the physical size of the box have to do with supply demands?  Or am I completely misunderstanding you?
 
CurtZHP said:
;D

Another question...
I'm not sure what you meant by...What does the physical size of the box have to do with supply demands?  Or am I completely misunderstanding you?

I did not explain myself properly. If I remember correctly, as well as a mic pre, this box contains EQ and a compressor. So you need more tubes than you would just for a mic pre. More tubes means more heater current which mean bigger transformer which means bigger magnetic field.

Cheers

ian
 
ruffrecords said:
I did not explain myself properly. If I remember correctly, as well as a mic pre, this box contains EQ and a compressor. So you need more tubes than you would just for a mic pre. More tubes means more heater current which mean bigger transformer which means bigger magnetic field.

Cheers

ian

Nope.  That's someone else's baby.  This is just a mic preamp.  The only "EQ" is a HPF.  I only have to feed one 12AX7 and one 12AU7.

Incidentally, for what it's worth, I'm using toroidal transformers to hopefully keep the EMI to a minimum.

I tried to attach the schematic, but I'm getting an error message about the upload folder being full.  Here's a link to a previous thread on it....

https://groupdiy.com/index.php?topic=68312.0

It may not reflect the latest changes to the power supply, but it'll at least give you an idea of the signal path.

Anyway....

I just finished completely tearing it down and putting one channel back together with just bare essentials.  No pad or polarity switch, and no HPF.  The only switch is for phantom power.  I also eliminated the instrument input.  I also moved the grid resistor to the actual grid pin on the tube socket and attached the input directly there as well.  I also reoriented the whole thing in relation to the power supply.  That seems to have quieted things down considerably, until I switch on phantom power.  Then it's Hum City!
 
My bad; there are so many people on groupDIY design tube mic pres that I get confused - not that I am complaining mind you.

I checked out the schematic It seems basically OK. I seem to rememeber you said the phantom is now taken from a separate winding. If that is right then it should work OK and I think you increased the caps in the tripler to 100uF. If you look at the TL783 datasheet you will see that nearly all is specs assume 25V across the device so for the phantom power you need at least 70V raw dc at the input:

http://www.ti.com/lit/ds/symlink/tl783.pdf

How are you managing the analogue 0V inside the chassis.

Cheers

Ian
 
ruffrecords said:
My bad; there are so many people on groupDIY design tube mic pres that I get confused - not that I am complaining mind you.

I checked out the schematic It seems basically OK. I seem to rememeber you said the phantom is now taken from a separate winding. If that is right then it should work OK and I think you increased the caps in the tripler to 100uF. If you look at the TL783 datasheet you will see that nearly all is specs assume 25V across the device so for the phantom power you need at least 70V raw dc at the input:

http://www.ti.com/lit/ds/symlink/tl783.pdf

How are you managing the analogue 0V inside the chassis.

Cheers

Ian

No worries.  I know what you mean.  After a while they all start blurring together.  :eek:

My phantom supply topology might need some work, but I don't that's the real problem at present.  I've almost certainly got a wicked ground loop going on that only rears its ugly head when the phantom supply is actually sent to the inputs.

After changing things this past weekend, I made an interesting, and encouraging, observation.  Without phantom power, the unit is fairly quiet.  Mind you, I haven't yet shielded the tubes, and the whole thing is still exposed on the bench.  But, I picked up the audio circuit and actually moved it closer to the supply.  (They are separate circuits, so I could get the supply built and tested prior to introducing the audio portion.  Fat lot of good that did me!)  I didn't get any appreciable hum until I was practically touching the glass of the tubes to the power transformers.

So, we're back to hunting down a ground loop.  I'll draw up what things look like at this point and see if I can post that.

Took a break from it today.
 
If you are building it point to point on a metal chassis you can do worse than take a leaf out of the RCA book for arranging the 0V (and keeping it separate from the chassis except at one point. The technique they used was to run a thick copper bus bar (15 amp household copper will do) from the PSU towards the input. The bus bar is connected to the power supply 0V at one end. The output stage 0Vs are connected next along the bus bar and then the input stage 0Vs are connected to the bus bar. The at one point, and one point only, the bus bar is connected to the chassis. This means signal currents run only in the bus bar and interference currents run only in the chassis. In unbalanced gear the bus bar was connected to the chassis at the unbalanced input. In balanced gear you can connect it to the chassis where the power enters the chassis - in your case the mains inlet socket.

Getting an organised separation of 0V and chassis is the first line of defence against hum loops.

Cheers

Ian
 
Definitely looking very closely at the grounding on this.

But one other thing...
Is it possible that I'm getting some sort of nasty interaction between my phantom supply multiplier stage and the step-up transformer for the plate supply?  They are both fed from the same 15VAC winding of the other power transformer.
 
CurtZHP said:
Definitely looking very closely at the grounding on this.

But one other thing...
Is it possible that I'm getting some sort of nasty interaction between my phantom supply multiplier stage and the step-up transformer for the plate supply?  They are both fed from the same 15VAC winding of the other power transformer.

I thought you had them connected to separate secondaries - or was that a different thread???

Anyway, there is no way you can have two separate bridge rectifiers operating off the same winding - it will cause horrible hum.

Cheers

Ian
 
Hi...

May I ask what kind of power (toroidal?) transformer you use? I know that there are Swedish made Noratel that are being used for Gyraf G9's circuit, but they are terrible for audio since they have quite poor EMC.

I have faced this problem before.

What you say implies that you might have a ground loop so use Ians comments as a guide line for fixing that.

I usually divide my grounding in three blocks. One for the power supply including the heater arrangement, One for the phantom power and one for the signal path. As Ian says... connect the phantom 0V at the input XLR (pin 1) Connect the power supply as close as possible to the main inlet ground. (in Sweden there is a rule that the main inlet ground must be on its own... end) but you could use a bolt and nut to connect the power supply 0V a centimeter from the main inlet ground. And finally connect the signal path 0V on the same bolt as the power supply 0V. Do not connect the signal 0V to the power supply 0V as a "shortcut" or the other way around. So three 0V cables from these three blocks.

The other thing seems to be you phantom power... triple or quadruple voltages might be a source for problem. Since the phantom power is injected directly at the input transformer it is a sensitive thing to handle.

There are in fact very good toroid transformers on the market that will have secondary voltages that are suitable for creating HT, phantom and heater arrangements. They are more expensive than a single transformer with one voltage secondary, but your design demands two of those transformers. Perhaps the solution can be to get a proper transformer from the beginning?

I actually built the Gyraf G9 some years ago and I could not get it to work due to a lot of hum. It proved to be the output transformer that were picking up a lot of stray from the (cheap swedish) power transformers and the only solution was to get rid of the output transformer and use the preamp with unbalanced output signal or get rid of the power transformers and get a descent one.

Nowdays I buy my power transformers (toroids) from don-audio in Germany. That might not be the best place for you to buy components but take a look at their website and see what you could use for your design.

Hope it will work out for you

Best regards

/John
 

Latest posts

Back
Top