Reactive load - adding an XLR out and fan

[GainFreak]

Hi,

I decided to add a transformer coupled XLR out and a fan to my existing reactive load. After doing some search I found this schematic (the fan part added by me Powerbrake style).
Now here's the odd thing. When ground lift is on (speaker ground connected to chassis) the fan is working properly. However when speaker ground is lifted from chassis the fan stops working and the 100 Ohm resistor quickly heats up and burns. So what's going on here and is this the proper way to organize grounds?
Most of the time I've seen the ground is lifted from pin 1 of the XLR connector but in my case I followed the Jensen schematic which allows for balanced XLR/TRS and unbalanced TRS with mono jack using only one combo socket.

[Helmholtz]

Hi,

I decided to add a transformer coupled XLR out and a fan to my existing reactive load. After doing some search I found this schematic (the fan part added by me Powerbrake style).
Now here's the odd thing. When ground lift is on (speaker ground connected to chassis) the fan is working properly. However when speaker ground is lifted from chassis the fan stops working and the 100 Ohm resistor quickly heats up and burns. So what's going on here and is this the proper way to organize grounds?
Most of the time I've seen the ground is lifted from pin 1 of the XLR connector but in my case I followed the Jensen schematic which allows for balanced XLR/TRS and unbalanced TRS with mono jack using only one combo socket.

Wihout scoping I can't be completely shure. But it seems you're getting some scary oscillation (risky for the amp's OT) when the reactive load is not connected to the chassis.
The reactive load is passive and doesn't need a ground lift, so leave it connected to the chassis. You should try to wire the ground lift to the line-out circuit only.

[GainFreak]

Thanks. I'll try that.

[Enzo]

And what kind of amp are you driving this with?

[Chuck H]

My consideration was that by lifting the ground on the load rather than the XLR (after transformer isolation) is that there does remain a path for AC modulation through that 100R resistor. No 0V reference for the bridge arrangement that's rectifying the fan voltage is probably the problem. Though why the 100R resistor should be suffering when there's a load under ten ohms taking the brunt remains a mystery to me. Without a full and accurate schematic of all the parameters in play it's hard to diagnose.

[GainFreak]

And what kind of amp are you driving this with?

I'm driving it with a 4x6L6 power amp.

Without a full and accurate schematic of all the parameters in play it's hard to diagnose.

Sound path and connections are as per the schematic. Differences are XLR output is as on Jensen schematic and reactive load values are for 8 Ohm load.

[nickb]

I think we need to know how this has been constructed, wiring from OPT, conditions of test when this happens (i.e. what's on the output XLR/ jack) and amp input to understand what is going on. Since the ground lift just connects the speaker ground to the XLR ground then surely something has to be connected to the XLR to have any effect on the oscillation

But what I really wanted to say was this. I was looking at this and thinking about the relationship between speaker SPL and drive. It seems to me that it's the current through the voice coil (to a first approximation) that determines the SPL. Therefore, if you want an output on the XLR that sounds closer to the speaker ( or rather reactive fake) then you need to connect the Jensen circuit across the two 33 ohm resistors rather than across the amp output.

[Helmholtz]

But what I really wanted to say was this. I was looking at this and thinking about the relationship between speaker SPL and drive. It seems to me that it's the current through the voice coil (to a first approximation) that determines the SPL. Therefore, if you want an output on the XLR that sounds closer to the speaker ( or rather reactive fake) then you need to connect the Jensen circuit across the two 33 ohm resistors rather than across the amp output.

Depends on how the line-out is going to be used. I assumed it would feed a (small) linear power amplifier connected to a guitar speaker like with the Ultimate Attenuator. In this case taking the signal from the input ( = main amp output) makes sense.

I agree that some of a speaker's SPL will show in its current - at least at mid frequencies. But e.g. the bass resonance will show as a dip or notch as opposed to a peak in real SPL.
If speaker current would directly correlate with its SPL, the SPL would be linear when the speaker is driven from a CCS.

[GainFreak]

I think we need to know how this has been constructed, wiring from OPT, conditions of test when this happens (i.e. what's on the output XLR/ jack) and amp input to understand what is going on.

There isn't anything special about it. Power amp 8 Ohm output goes to the reactive load as per schematic. Power amp chassis is grounded via 3 prong cable.It's the same as a Slave Out found in many amps only instead of connecting a speaker you connect the reactive load so the Jensen circuit is connected correctly to the load.
The XLR out is supposed to connect to PC to be used with IRs. The problem occurs either way when the Jensen circuit is disconnected from PC or from the reactive load. I'll check with my PC scope if any oscillation is present when ground is lifted.

[nickb]

Depends on how the line-out is going to be used. I assumed it would feed a (small) linear power amplifier connected to a guitar speaker like with the Ultimate Attenuator. In this case taking the signal from the input ( = main amp output) makes sense.

I agree that some of a speaker's SPL will show in its current - at least at mid frequencies. But e.g. the bass resonance will show as a dip or notch as opposed to a peak in real SPL.
If speaker current would directly correlate with its SPL, the SPL would be linear when the speaker is driven from a CCS.

I was looking as it being used as a DI, as it seems it is. So, SPL isn't the current at LF but nor is it the voltage at HF. Therefore, the speaker simulator design seems fundamentally flawed.

[nickb]

There isn't anything special about it. Power amp 8 Ohm output goes to the reactive load as per schematic. Power amp chassis is grounded via 3 prong cable.It's the same as a Slave Out found in many amps only instead of connecting a speaker you connect the reactive load so the Jensen circuit is connected correctly to the load.
The XLR out is supposed to connect to PC to be used with IRs. The problem occurs either way when the Jensen circuit is disconnected from PC or from the reactive load. I'll check with my PC scope if any oscillation is present when ground is lifted.

I know it's obvious to you how it's constructed but we simply don't know. Is this built into it's own little metal box or internal to the amplifier? What about the secondary from the OPT; does it connect to any kind of ground or the chassis?

So, not having the XLR connected to anything has no effect on the problem. If this is in it's own box, as I had envisaged it, then that makes little sense. I'd conclude that this is built in to the amplifier and that the speaker output is not grounded. That would mean that the speaker output wires could have some common mode signal, how much depends on things like layout, transformer design and frequency. This means that when there is no OPT ground the OPT secondary wires could couple back to more sensitive section(s) of the amplifier.

The oscillation seems to be ultrasonic so your PC scope will need to have sufficient bandwidth, say >50Khz. In other words a scope using the audio card might not help.

[Helmholtz]

Therefore, the speaker simulator design seems fundamentally flawed.

Well it simulates a speaker load to the guitar amp. But requires the signal to be finally reproduced (at lower volumes) by a real guitar speaker. Or some filter circuit that emulates a speaker's response.

[GainFreak]

I'd conclude that this is built in to the amplifier and that the speaker output is not grounded.

The whole thing is built into its own separate from the amplifier enclosure made from 1.2mm steel.
OT secondary ground is grounded.

Therefore, the speaker simulator design seems fundamentally flawed.

The reactive load circuit (speaker equivalent circuit) is a classic from Aiken from a very long time and was built by thousands of DIY-ers. There's ton of info about it all over the place.

No oscillation is present at speaker's out (unit input) when ground is lifted and fan not working.

1/ Voltage across 100 Ohm resistor at onset of fan rotation ground NOT lifted 0.95VAC. Voltage at load 4.8VAC
Voltage across 100 Ohm resistor (at at same load voltage of 4.8VAC) ground lifted 4.6VAC.

2/Voltage across 100 Ohm resistor ground NOT lifted 2.4VAC. Voltage at load 20VAC
Voltage across 100 Ohm resistor (at same load voltage of 20VAC) ground lifted 18.7VAC.

Fan specs are 24VDC/0.68W (low noise).

[Helmholtz]

Power amp 8 Ohm output goes to the reactive load as per schematic.

With the component values shown, amp output impedance should be set to 16 Ohm. Otherwise power tube screens are at risk.

[Helmholtz]

The whole thing is built into its own separate from the amplifier enclosure made from 1.2mm steel.
OT secondary ground is grounded.



The reactive load circuit (speaker equivalent circuit) is a classic from Aiken from a very long time and was built by thousands of DIY-ers. There's ton of info about it all over the place.

No oscillation is present at speaker's out (unit input) when ground is lifted and fan not working.

1/ Voltage across 100 Ohm resistor at onset of fan rotation ground NOT lifted 0.95VAC. Voltage at load 4.8VAC
Voltage across 100 Ohm resistor (at at same load voltage of 4.8VAC) ground lifted 4.6VAC.

2/Voltage across 100 Ohm resistor ground NOT lifted 2.4VAC. Voltage at load 20VAC
Voltage across 100 Ohm resistor (at same load voltage of 20VAC) ground lifted 18.7VAC.

Fan specs are 24VDC/0.68W (low noise).

How did you measure AC voltages? Oscillation frequencies are often above 100kHz. What is the power rating of the 100R resistor?