Just listen to the relays clicking. The clicking sound is easy to hear
I only measured the current (voltage) over one of the power resistors; the other one will (should) be more or less the same. So each rail will see the 30 mA iddle current.
the speaker delay/isolation relay does not need to be active.
Because you will have no load connected.
The only "load" connected is the NFB loop and that is in the kohms range.
To set your bias you need to remove the Mains Bulb Tester and power on direct from the mains.
You need to remove any "protection" resistors from the supply rails.
Short the input with a dummy zero ohms plug.
This is the time to check and adjust output offset.
Then you step up the bias voltage till you reach the Vre that is correct for your topology.
A EF output stage with only driver EF and no pre-driver stage will have a bias voltage of around 2400mVbias
The output emitter voltage Vre will be from 18mVre to 27mVre for each output emitter resistor. (D.Self measures across a series pair and states the doubled voltage for 2*Vre).
Monitor this as the output stage and the heatsink rise in temperature. Adjust the bias voltage to maintain your chosen Vre.
Recheck and adjust output offset.
The better way to arrive at your required bias voltage is to monitor the crossover distortion and adjust to minimise this distortion after reaching nominal operating temperature. This is referred to as "Optimal ClassAB bias".
Because you will have no load connected.
The only "load" connected is the NFB loop and that is in the kohms range.
To set your bias you need to remove the Mains Bulb Tester and power on direct from the mains.
You need to remove any "protection" resistors from the supply rails.
Short the input with a dummy zero ohms plug.
This is the time to check and adjust output offset.
Then you step up the bias voltage till you reach the Vre that is correct for your topology.
A EF output stage with only driver EF and no pre-driver stage will have a bias voltage of around 2400mVbias
The output emitter voltage Vre will be from 18mVre to 27mVre for each output emitter resistor. (D.Self measures across a series pair and states the doubled voltage for 2*Vre).
Monitor this as the output stage and the heatsink rise in temperature. Adjust the bias voltage to maintain your chosen Vre.
Recheck and adjust output offset.
The better way to arrive at your required bias voltage is to monitor the crossover distortion and adjust to minimise this distortion after reaching nominal operating temperature. This is referred to as "Optimal ClassAB bias".
Last edited:
Yes it needs to be !!
If do you first meassurement without the relays activated and the second one with (or visa versa) then your adjustments are wrong !!!!
Remember, the relays is on the main PCB and will automatically activated after serverall seconds. Just wait until it is activated to start your meassurements.
The output offset does not need adjustment. There is an opamp in the circuitry who does that actively.
Last edited:
The output offset should be set with the DC servo deactivated (IC socket empty).
The output relay does not need to be active since there should NOT be any load, other than the NFB.
Once the amplifier is correctly set up then you can remove the input shorting plug and insert the DC servo opamp. And recheck that all the operating votages are still correct.
And recheck from a cold start, more than once.
Would you like to explain why you think this is the case?
The output relay does not need to be active since there should NOT be any load, other than the NFB.
Once the amplifier is correctly set up then you can remove the input shorting plug and insert the DC servo opamp. And recheck that all the operating votages are still correct.
And recheck from a cold start, more than once.
Would you like to explain why you think this is the case?
We should be very carefull with our replies here Andrew. You commented on my meassurement method while the response is actually in regards to your meassurment method.
So why is this important in my method (using the two power resistors):
because in this method you're meassuring all currents; meaning the idle current + the current drawn by the rest of the amplifier, inlcuding the speaker relays.
Also, I would not advice to remove the servo IC because it is a vital part of the amplifiers design. Instead I advice everyone building the Q-Watt amp to read and follow the suggestions made by Elektor.
Taken (Verbatim) from the Q-Watt article: -
Q-Watt on test run Before you connect your Q-Watt amplifier directly to the power supply, you have to set the quiescent current of the output stage.
To do this, first connect two 47 Ω, 5 W power resistors in series with the positive and negative supply voltage terminals.
This prevents damage to the amplifier circuit if something is wrong, such as a short somewhere.
The worst that can happen is that the two power resistors go up in smoke.
Another option is to use a regulated power supply with current limiting, but most of us don’t have this sort of supply available for ±56 V.
Connect an ammeter in series with the positive supply line.
Before switching on the supply voltage, turn P1 fully counterclockwise, and remember to connect the secondary windings of the transformer to terminal block K7 on the circuit board.
After the power is switched on, the current through the positive supply line should be approximately 30 mA when the output relay is engaged.
Slowly turn P1 to the right (clockwise) until the current increases by 30 mA (60 mA total).
This relatively low quiescent current is more than adequate.
The quiescent current will rise slightly as the heatsink temperature increases.
However, it will normally remain below 90 mA. At very high output power levels, the junction temperatures of the two output transistors will rise much faster than the temperature of the heatsink, so the quiescent current transistor will not be able to fully track the change.
This causes the quiescent current to rise briefly to several hundred milli-amps, but it declines quickly when the temperature drops again.
That’s actually a nice extra feature with this amplifier, since the class A range of the power amplifier effectively increases with the output power level.
Q-Watt on test run Before you connect your Q-Watt amplifier directly to the power supply, you have to set the quiescent current of the output stage.
To do this, first connect two 47 Ω, 5 W power resistors in series with the positive and negative supply voltage terminals.
This prevents damage to the amplifier circuit if something is wrong, such as a short somewhere.
The worst that can happen is that the two power resistors go up in smoke.
Another option is to use a regulated power supply with current limiting, but most of us don’t have this sort of supply available for ±56 V.
Connect an ammeter in series with the positive supply line.
Before switching on the supply voltage, turn P1 fully counterclockwise, and remember to connect the secondary windings of the transformer to terminal block K7 on the circuit board.
After the power is switched on, the current through the positive supply line should be approximately 30 mA when the output relay is engaged.
Slowly turn P1 to the right (clockwise) until the current increases by 30 mA (60 mA total).
This relatively low quiescent current is more than adequate.
The quiescent current will rise slightly as the heatsink temperature increases.
However, it will normally remain below 90 mA. At very high output power levels, the junction temperatures of the two output transistors will rise much faster than the temperature of the heatsink, so the quiescent current transistor will not be able to fully track the change.
This causes the quiescent current to rise briefly to several hundred milli-amps, but it declines quickly when the temperature drops again.
That’s actually a nice extra feature with this amplifier, since the class A range of the power amplifier effectively increases with the output power level.
Last edited:
I class this as a bad settup procedure.
Introducing dropping resistors into the supply lines lowers the supply voltage to the amplifier.
This artificially reduces the bias voltage and the Vre as measured at the output emitter resistors.
Some amplifiers are reasonably tolerant of this arrangement. But many are not.
When the dropping resistors are replaced with shorting links or fuses the supply rail voltage to the amplifier rises and an intolerant amplifier increases the bias voltage and the Vre and this can result in further heating of the output stage which then increases the current and dissipation in the output stage and can result in thermal runaway.
You must check the bias using the Vre with the normal maximum voltage on the supply rails after the amplifier is fully warmed up.
Hi Andrew,
What is the Mains Bulb Tester?
Does this apply to Q-Watt? I'm really looking for some guidance specific to the Q-Watt topology. Where in the circuit do I measure the 'Vbias' on the 'EF output stage'. This is my first amp. Sorry for being dim.
What is the Mains Bulb Tester?
Does this apply to Q-Watt? I'm really looking for some guidance specific to the Q-Watt topology. Where in the circuit do I measure the 'Vbias' on the 'EF output stage'. This is my first amp. Sorry for being dim.
... don't quite understand what you mean here Andrew...
I answered that question in post #227.
I wonder what the designer (Tony Giesberts, Elektor Labs) might have to say about this comment?
There is a huge part of this all that is quite a bit subjective.
Andrew is correct when he says that the supply current drops while using drop in resistors. However, the change in current intrduced by those resistors is minimal. I wouldn't loose my sleep over it.
Elektor's recommendations work just fine.
Did you, I read your post and can't see where you explain that measuring the Vre gives the wrong output bias, compared to measuring the amplifier's quiescent current across the dropping resistor is the only way to get the correct output bias.
I am pretty sure that we would never come to an agreement.
he has stated his method and it is different from the norm. By coincidence his method is the same as Destroyer and I have told Members how wrong that was when he used to recommend it.
Delange sees the potential problem but has dismissed as
Do it properly. Measure Vre and recheck a few times as the amplifier warms up and for different mains voltages and different room temperatures.
some refer to it as a dim bulb and others have different names.
The most common name adopted on this Forum is Mains Bulb Tester.
After I championed it's use many years ago, many Members have come to recognise it's worth and they too are recommending it as the better/safer way to power on new or modified Mains powered projects.
T1 Vce and C4 have the bias voltage across them. You could measure the bias voltage at any of those nodes that have that voltage between them. Bias voltage is NOT the ruling factor for set up. Vre is your target voltage. But Vre does nothing for much of the adjustment range of the variable resistor. Bias voltage responds immediately for each small change in the adjuster. Two voltmeters clipped on to monitor both bias voltage and Vre, simultaneously makes for easier understanding of what your adjuster is doing.
The Qwatt is a dual EF output stage (i.e. no pre-driver stage and uses complementary drivers and outputs for the Push-Pull topology.
The bias voltage generator supplies the Bias Voltage to the driver bases. This is ~2400mVbias when the output stage is Optimally biased into ClassAB.
You set the adjuster to a much lower voltage for first start up. Usually around 1200mVbias to turn on the driver and not turn on the outputs. The Qwatt resistor values will give a minimum Vbias of ~1600mVbias. This allows you to check the Vbe of the drivers and of the outputs before any bigger currents are passed. It also allows you to check operating voltages of the other components to ensure no construction errors.
Once all this is complete, you remove the Mains Bulb Tester and power ON direct from the Mains.
Now you again check Vre of the drivers and of the outputs. Start adjusting the Bias voltage to bring the Vre of the outputs up to ~10mVre. Then proceed in smaller steps to bring it up as it warms up. Aim for ~18mVre and again let it warm up. Step up slowly towards your final chosen Vre.
For Re=0r22 I would aim for ~22mVre this will be pretty close to optimal ClassAB bias. But it will operate a lot hotter than the 30mA described in the extract posted by Calpe, where the output stage is severely underbiased @ ~30mA (Vre = 0.03*0.22= 6.6mVre)
The low setting of 30mA will give an excess of crossover distortion. But it does allow a much smaller heatsink to be used in return for that distortion (crossover does not sound nice because it contains many odd order high frequency components rather than low order even frequency components).
Last edited: