SSL Pre with JLM DI -> Big Hum

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

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

AW_music

Well-known member
Joined
Nov 4, 2007
Messages
148
Location
Italy
Hi everybody,
i need some help on this project.
I've finished my SSL pres and they are fantastic, so i've decided to add the JLM FET DI+ go between ( Phase, Phantom, Pad)
http://www.jlmaudio.com/JLMDI.htm

The Phase, Phantom, Pad are OK, but when i plug a jack on the input DI a BIG HUM (i think 50Hz) goes through the signal and inside the PRE.
SO it seems impossible to use a guitar direct with this pres.

I have used the star ground scheme, so i can't understand what's the problem.
Maybe the JLM DI cannot be used without an audio input transformer ?????
??? ???

Someone can give me a clue ???

Best regards


 
The Phase, Phantom, Pad are OK, but when i plug a jack on the input DI a BIG HUM

So the pre is perfectly clean until you use the DI Jack???

I'd EMAIL Joe @ JLM & ask him... then post his response back here to help others.

Joe's a cool dude & will definitely respond.

Make your title clear, like:
NEED HELP... HUM when using DI Jack!!!

I'd also send him a link to this topic.  ;)
 
Sounds like a ground loop to me. -I don't believe that there's necessarily anything wrong with it. -Does it hum if you ONLY connect the ground/shield connection?

-If so, it's a ground loop. -Your problem in that case lies in having two paths to electrical power ground, meeting in the audio ground connection.

Keith
 
SSLtech said:
Sounds like a ground loop to me. -I don't believe that there's necessarily anything wrong with it. -Does it hum if you ONLY connect the ground/shield connection?

-If so, it's a ground loop. -Your problem in that case lies in having two paths to electrical power ground, meeting in the audio ground connection.

Keith

sounds familiar.
have you checked continuity between the PSU board and chassis?is is isolated?
also your regulators (7818 and 7915) are bolded to chassis.are the heatsinks isolated?
I would check with the continuity meter all along to make sure there is ONLY one path to GROUND CHASSIS, like Keith says..

Mattia.
 
I'm hooking up my 9k right now & may see your "GND Loop" problem???

OK,
The JLM Go between gets connected to (0 Volt)
That same zero volt is also the "shield" for the INPUT XLR - Going from the input jack to the Go-between, then to the input on the 9k PCB's.

OK, from the looks of your picture, I THINK you have your "PIN 1" going to the "Chassis" too???

If so, then this is your double connection.

Lift your pin one from the chassis & keep it hooked up to the "Shield" of the input XLR.

If your Go-between is hooked up like JLM suggest, then that should work.
 
Ok I would think this will be a 48v power rail issue as the power supply kit you have is made to provide some phantom power for few mics from memory. The JLM FET DI runs on the 48v rail using 25 to 35mA each at idle. Is the 48v reg on a good heatsink? Does the 48v reg input pin have 55v unreg or more coming into the reg with everything plugged in?

Try unplugging 3 of the DI's so there is not as much load on the 48v rail to see if the hum goes away or is quieter in the one DI left plugged in.

Let me know what you find and I should be able to get you up and going quickly.
 
Ok, i have tried to workaround with the shield, but everything i do, i've got always the HUM.
The HUM is still present even if i disconnect the ground/shield connection.

I've done what JOE suggested, and when disconnecting the others DIs, leaving only one , the HUM is quieter in the one DI left plugged in.

I think this is the right direction, a "48v power rail issue" !  ;)

So what can i do now to eliminate it?  ???

 
Let me know what AC volts the windings on the transformer are? And direct me to the circuit for the power supply you are using? And let me know if you made any changes to cap values etc? And I will do the math :)
 
Silly question but you dont have the input jack wired backwards do you? The ground wired to the hot on the DI?
 
Hi Jim,

finally i've found the time to give you all the infos you nedd.
I've used a 0-220  to  0 -18V 1.39A  / 0-18V  , 1,39A  toroid .
and this is the schemo :


I hope you can give me "the solution" ....

Many Thanks

:)
 
Anything? What was the result of this? Fixed, moved on? I am having the same issue using only two JLM FET Di for a 4 ch SSL9K. Hum ony when the DI is switched on (by inserting plug into jack).
 
hello guys
i had a dual 9k with 2fet di and go between the signle problem was the 48v with the relay on the fe di , they caused a wrong noise when i pewered the whole on.
So they worked fine plugged and unplugged  exepted one with big hum and the problem was a bad stuff on this small board, try to put out the fet one by one,

if not try an another capacitors 220uf 100v on the psu could give a bad soundif there is a bad stuffed or if it is of bad quality.
 
yes, just one power supply.

I did try something though. Reading Joe's comments while trying to help AW_Music, I measured the voltage feeding the regulator for the 48v supply. It was measuring about 46v, and I was getting about 44v after the regulator. I disconnected power to the 2 DI boards, and then I was getting about 51v feeding the 48v regulator. I built a separate power supply for the DI to see if that was it. I had a 24v/400mA transformer laying around. Bridge rectifier with a 100uF cap gave me about 38v. Still have hum though. Now it is an octave higher though. More like a ground loop? I am not quite clear on why the 48v supply would produce hum. I am assuming it is too much for the supply, or more accurately, not enough for the DI, amperage, not voltage that is.

The higher pitched hum would seem to indicate a ground loop, where as before possibly AC inductance? I do not have the exact frequencies. The lower one was between 6th and 7th fret of a bass on the low E.
 
AW_music said:
Hi Jim,

finally i've found the time to give you all the infos you nedd.
I've used a 0-220  to   0 -18V 1.39A  / 0-18V  , 1,39A   toroid .
and this is the schemo :


I hope you can give me "the solution" ....

Many Thanks

:)

Ok as the circuit says 2 x 20vac is the minimum transformer needed and even that is a fine line and only ok for a couple of phantom mic loads on the 48v rail.

For the 48v reg to regulate it will need about 54v at it input pin under very light load. And about 59v or more under heavy load as the pump charge values for C3 and C11 of 100uF and storage charge cap C5 of 220uF is very small so ripple will be come large and the reg will go in and out of reg every half cycle.

OK here is the simplified math (Voltages below will be up to 10% higher under load due to less losses in the transformer. But also under load the doubler value of 2 falls more to 1.5 with large ripple.)

2 x 18vac transformer voltage to the 48v reg input (18 x 1.414) -1.4v(diode loss) x 2(doubler) = 48v into the reg input pin with no load means reg is not regulating ever.
NO GOOD for anything but phantom power will still seem fine and quiet if the match of the 6k8 resistors is good but the 48v reg is really not working. Add the Fet DI load to this and you will have a DI that HUMs

2 x 20vac transformer voltage to the 48v reg input (20 x 1.414) -1.4v(diode loss) x 2(doubler) = 54v into the reg with no load.
OK. So this will be just enough for running a few phantom mics fine with a properly regulated rail.

2 x 22vac transformer voltage to the 48v reg input (22 x 1.414) -1.4v(diode loss) x 2(doubler) = 59v into the reg with no load.
2 x 24vac transformer voltage to the 48v reg input (24 x 1.414) -1.4v(diode loss) x 2(doubler) = 65v into the reg with no load.
2 x 25vac transformer voltage to the 48v reg input (25 x 1.414) -1.4v(diode loss) x 2(doubler) = 68v into the reg with no load.
All Perfect. Should be enough voltage for the 48v reg to keep regulating with an extra load of the DI's. But the small 100uF values of C3 & C11 and C5 220uF could mean the rail will sag to much under heavy load like 4 DI's maybe.

OK before you all go out and buy the correct higher secondary voltage transformer see below solutions

Solution 1 if you have a 2 x18vac transformer and don't care that the 48v isn't perfect and just want the JLM Fet DI to have no hum. Short out the 2k7 resistor on the Fet DI and wire the +V of the DI to the +18v rail. Add heatsink to +18v reg. Done

Solution 2 if you have a 2 x 18vac transformer and want a well regulated 48v for perfect 48v phantom and to run the JLM Fet DI properly. Rewire pin 3 of BR2 by cutting its track and wiring it to the + of C1 to make the doubler into a tripler like our AC/DC kit can do. Giving the 48v reg about 72v to its input to work with. Add heatsink to +48v. Done. If you are trying to run 4 x JLM FET DI's and the last one or two start the hum to come back change C3 and C11 to 1000uF 50v types to get better pump charge to the 48v reg. C5 changed to 470uF or larger would also help.

Solution 3 if you have a 2 x 18vac transformer and want a well regulated 48v for perfect 48v phantom and to run the JLM Fet DI properly. Buy a AC/DC set it to tripler and set +48v and +/-18v and you done  :) (Our AC/DC uses 1000uF for these 3 pump change caps so will not sag under the 4 x Fet DI load.)

stanz said:
I did try something though. Reading Joe's comments while trying to help AW_Music, I measured the voltage feeding the regulator for the 48v supply. It was measuring about 46v, and I was getting about 44v after the regulator. I disconnected power to the 2 DI boards, and then I was getting about 51v feeding the 48v regulator. I built a separate power supply for the DI to see if that was it. I had a 24v/400mA transformer laying around. Bridge rectifier with a 100uF cap gave me about 38v. Still have hum though. Now it is an octave higher though. More like a ground loop? I am not quite clear on why the 48v supply would produce hum. I am assuming it is too much for the supply, or more accurately, not enough for the DI, amperage, not voltage that is.

The higher pitched hum would seem to indicate a ground loop, where as before possibly AC inductance? I do not have the exact frequencies. The lower one was between 6th and 7th fret of a bass on the low E.

Same as above problem and your measured voltages proves it :) Making a unregulated 38v rail is going to make the Fet DI hum and with only 100uF cap it is going to hum big time. It is the all important regulation bit that gives the perfectly smooth power that stops the hum being injected via the power rail.




 
Joe thanks for the tips.
I have to mention that i have tried and tried few different things...
Didnt help at all...i didnt used this psu but the JLM AC/DC 3 rails...nah didnt help...even a bigger toroidal.
Tried the Switching solution with a trex on 1 channel and still didnt found a solution.
What i did its get rid of the Fet DI...build my own DI and used a simple toggle DPDT.

Thanks.
 
Well, I tried solution #1, using +18v supply and jumper 2k7 resistor. Still have hum. 18v supply now measures 21v. Seems to be a common theme here, PSU problems. I had issues with regulators when first building it. Seems to me I need to deal with that first.

Joe, I have a 50VA 20VAC transformer with center tap (20 VAC from center on each winding). What is the recommended set up for one of your AC/DC kits in order to power 4 ch SSL9K and 2 of your FET DI boards?
 
1+
For the 48v reg to regulate ... snip
The regulator in this circuit is NOT set for 48V.
Assuming Vref of typ.1.25V (can be in range between 1.2 up to 1.3V), this LM317 is set for 1.25V *(1+8200/240) = 43.96V or about 44V.
For condenser mics this supply voltage most often is enough. If only one condenser mic is connected, this should draw at least 5mA, as the R1/R2 network only draws 5mA, but most regs (National,TI,STM,Vishay) want a min.load of 10mA (up to 12mA for Fairchild). With dropout voltage between 2.5 to 2.8V, this config will need at least 47V at its input pin.
 
3nity said:
Joe thanks for the tips.
I have to mention that i have tried and tried few different things...
Didnt help at all...i didnt used this psu but the JLM AC/DC 3 rails...nah didnt help...even a bigger toroidal.
Tried the Switching solution with a trex on 1 channel and still didnt found a solution.
What i did its get rid of the Fet DI...build my own DI and used a simple toggle DPDT.

Thanks.
We have seen about 10+ of these multiple SSL pre at the workshop with similar comments to 3nity about trying several things but no real technical info both with the normal power supply and our AC/DC and all of them with multiple Fet DI have worked fine with the above mods or had other simple wiring mistakes which we fixed. Main wiring problems have been to do with people lifting 0v points in the search for a ground loop problem that doesn't exist. And the old classic in and out XLR wired identical instead of mirror images of each other so one ground pin and signal pin are swapped. So not sure why you couldn't get yours to go.

stanz said:
Well, I tried solution #1, using +18v supply and jumper 2k7 resistor. Still have hum. 18v supply now measures 21v. Seems to be a common theme here, PSU problems. I had issues with regulators when first building it. Seems to me I need to deal with that first.

Joe, I have a 50VA 20VAC transformer with center tap (20 VAC from center on each winding). What is the recommended set up for one of your AC/DC kits in order to power 4 ch SSL9K and 2 of your FET DI boards?
Umm strange the voltage would go up but this can be a sign of the reg oscillating from overheating just before is shuts down as a digital multimeter will read this wrongly. Does the rail go back to 18v with no load or stay at 21v? Did you have the reg heatsinked well?

As always if having problems with one of our products send me by email some large well lit in focus macro on photo's of your PCB's and wiring (if you cannot see the components clearly in the photo's and trace the wiring which is hopefully not just one color then I will not be able to either). Give me a short description of the problem and what DC and importantly AC voltages you measure at the in and out pins of the regulators to the power supply 0v etc. (Be careful most cheap digital multimeter tell lies when measuring DC with high freq oscillation or measuring AC with a large DC component. You can sometimes prove if it is the meter by measuring a 9v battery in AC volts and see if you get the correct 0v or some crazy reading) I will usually ask more questions which i need answered so don't reply with it still isn't working as this isn't going to help me help you work out the problem :) It can take me some days to reply just because off the huge amount of email I deal with.

The Fet DI needs a fully regulated smooth 18v to 48v rail and draws up to 35mA. It cannot hum or oscillate on its own as it is a simple unity gain buffer.

Harpo said:
1+
For the 48v reg to regulate ... snip
The regulator in this circuit is NOT set for 48V.
Assuming Vref of typ.1.25V (can be in range between 1.2 up to 1.3V), this LM317 is set for 1.25V *(1+8200/240) = 43.96V or about 44V.
For condenser mics this supply voltage most often is enough. If only one condenser mic is connected, this should draw at least 5mA, as the R1/R2 network only draws 5mA, but most regs (National,TI,STM,Vishay) want a min.load of 10mA (up to 12mA for Fairchild). With dropout voltage between 2.5 to 2.8V, this config will need at least 47V at its input pin.
Yes harpo is correct you do have to watch not to draw less than about 12mA as well. When fixing one of these kits for a local customer with the above supply and a 100mA load we measured 10v ripple on the 47v going into the reg input which means the lowest voltage the reg could reg is the 47v - 5v half of the ripple - 2v reg drop = 40v. With 10mA the ripple was less than 2v. Changing the 2 pump charge and 1 storage cap all to 1000uF made the ripple stay at the lower couple of volts but the 48v reg had to be heatsinked to stop it thermalling in and out with the 100mA load.

Hope this helps and not just confuses everyone.

 
Joe,

Quote from: stanz on October 18, 2009, 02:42:38 am
Well, I tried solution #1, using +18v supply and jumper 2k7 resistor. Still have hum. 18v supply now measures 21v. Seems to be a common theme here, PSU problems. I had issues with regulators when first building it. Seems to me I need to deal with that first.

Joe, I have a 50VA 20VAC transformer with center tap (20 VAC from center on each winding). What is the recommended set up for one of your AC/DC kits in order to power 4 ch SSL9K and 2 of your FET DI boards?


Umm strange the voltage would go up but this can be a sign of the reg oscillating from overheating just before is shuts down as a digital multimeter will read this wrongly. Does the rail go back to 18v with no load or stay at 21v? Did you have the reg heatsinked well?


The voltage coming from the 18+ was 21v no matter what, both with DI and without DI. The regulator has a heat sink, but it is pretty small, as that was the only size I could use with them still on the PCB. Mounting them on the chassis was part of my initial issue with the PSU. I thought I had it well insulted, but alas, no. I purchased one of your AC/DC kits. I figured I need to get the power supply in order before I proceed with other issues, and by your account, the main issue with hum is one of a well regulated supply for the FET DI (or lack of one). My sense is that the power supply is on the edge of being able to handle all I am throwing at it. Clearly the DI works, it simply has the annoying hum and I trust that you know your design fairly well.  ;)
 

Latest posts

Back
Top