Beomaster 4400 (2419): Amplifier 50W Output Check

As a burn-in test I left the Beomaster playing the FM tuner for 24 hours. I varied the volume level every now and then but always kept it at normal listening levels (soft to loud). The output transistor heat sinks remained cool to the touch all during the burn-in testing and of course the sound was good.

Now it is time to see how the Beomaster does when full output power is requested.

First the test setup. My source input voltage is a 1KHz sine wave at 2Vp-p into the Beomaster 4400 Tape 2 DIN plug.

The output of the Beomaster will be measured using the speaker 2 output connected to my 8Ω dummy speaker loads. These are power resistors with heat sinks mounting on an even larger heat sink as the resistor loads will get pretty hot when full power is applied.

Here is the Beomaster with speaker 2 selected, Tape 2 selected and the volume level set to output 50W (the rated power output into 8Ω as listed in the Beomaster 4400 manual).  I used my DMM to monitor the speaker load to check when 50W was reached. The QA400 analyzer connects to the dummy load as well via a differential probe.

The power test worked nicely. I ran the volume up until I got to 20 V on the DMM. A little bit higher and the Beomaster overload lamp came on as I would expect. I backed the volume down a bit to measure the distortion of the 1KHz signal across the dummy loads.

Both channels measured quite good. Just below 0.02% THD for a 1KHz signal at maximum power.

As with previous amplifier measurements using this setup... I can't really do a direct comparison with the B&O published specifications as their test equipment and methods are different from mine. The best I can do is repeat my own test setup with the amplifiers I restore and collect that data for my reference. In comparison to my recent Beomaster 8000 output tests the THD on this Beomaster 4400 also measures very good. Well below what human ears can hear.

Beomaster 8000: Wrapping Up The Performance Tests

The current Beomaster 8000 restoration project has been going through quite a bit of testing in the last few weeks. Its owner mentioned that this Beomaster would occasionally come on by itself. Sometimes in the middle of the night it would come out of Standby mode and start playing music. It's always nice to listen to a Beomaster 8000 but you do want it to obey your commands.

I was confident that the restoration work I performed would address that issue but to test it I left the Beomaster plugged in on Standby mode for three full days. No hiccups and it never went into play mode by itself.

It was back to more listening tests while I set up the bench for some performance tests.

As Beolover recently posted, we don't really have the specific test information and test equipment that Bang & Olufsen used when they tested these Beomaster 8000 amplifiers with respect to what is in their technical specifications. On amplifiers there are usually only two service manual adjustments. No-load current (idle current) and DC Offset. That is the case with the Beomaster 8000. Other than those two adjustments a restored amplifier is expected to perform at the manufacturer listed specs for their design. We can use their published specifications as a guide as we collect our own measurements using test equipment available to us today. With these new measurements we will have measurable and repeatable values to compare our amplifier restorations.

Both Beolover and I use the QuantAsylum QA400 Audio Analyzer to make measurements. We also both use 8Ω fixed loads (power resistors mounted to large heatsinks) as dummy speaker loads during the tests. A small 0.08Ω to 0.1Ω resistor is added in series on each speaker load to allow direct measurements to the QA400 analyzer.  That is because the full voltage across the 8Ω load would damage the analyzer inputs. I have a pair of  QuantAsylum QA190 low noise, differential probes that allow measuring directly across the 8Ω (actually 8.08Ω in my case) dummy load. It should be noted that the updated QuantAsylum QA401 analyzer has built in differential inputs so the external probes are no longer necessary.

Here are some pictures of my test setup.

To run a direct measurement to the QA400 analyzer inputs from the 0.08Ω series resistor I use a coax cable with BNC connections for the analyzer and mini-grabbers for the 0.08Ω resistor.

Here is the Beomaster under test with the QA400 analyzer in the background and its measurement screen on my lab computer.

The QA400 analyzer, as described in Beolover's post, outputs a tone burst test signal for single frequency stimulus tests and a square wave impulse for frequency response test stimulus. I can run both left and right channel measurements at the same time as the analyzer provides the input to the Beomaster Tape 1 or Tape 2 left & right channel source inputs.

This test setup is easy to use and I can connect a bench DMM to the speaker loads to measure the actual voltage put out by the Beomaster output amplifier. One problem I run into is occasional noise that mucks with the measurement device sometimes. Usually down at the line 60Hz frequency. In my test setup the direct measurements across my 0.08Ω series resistor just don't provide as clean a measurement as the QA190 differential probes. I also have to monitor the QA190 probes though as they are battery powered (so the batteries need to be fresh).

For these Beomaster 8000 performance tests I decided to first grab a THD measurement for a 1KHz input signal right before it goes to the Beomaster Output Amplifier board. The audio signal goes from the input source (Tape 1 or 2) through the Preamplifier then to the Filter & Tone Control board. The path from the Filter & Tone Control board to the Output Amplifier is via connector 4P24. For this low voltage test I did use some direct probes with alligator clips to the QA400 inputs and not the QA190 probes. That is because of the space available to connect probes to.

Here is the THD measurement for a 1KHz input signal at 4P24.

The left and right channel THD is very low after going through all of the preamplifier and tone control circuitry.  Though it is very low (nearly -100dB) you can see the effect of some 60Hz noise picked up by the measurement cable.

Next was the 1KHz signal THD measurement after the Output Amplifier. This is measured across the 8.08Ω fixed speaker load at different output levels. It should be noted that during these tests I took the opportunity to calibrate the Tape 1 and Tape 2 input level trimmers again so I could make sure they are equal. I also tried to set them where I could set the Beomaster volume control on a level that produces 100W across the 8.08Ω load.  The result of that is a Beomaster volume level of 5.4 producing my 100W test output. The next volume level up (5.5) with this setup causes the clipping light to illuminate.

Here are left and right channel THD and SNR measurements at volume levels of 5.4 (~100W) and 5.3 (~67W).  I also turned the analyzer A Weighting mode on as the B&O specs use that.

The next volume level up (5.5) on the Beomaster puts the output amplifier into clipping so THD of 0.01% at the rated maximum power into 8.08Ω is very good. The next volume setting down (5.3) drops the THD to 0.008%.

Next is a check of the Beomaster frequency response. Beolover also describes this measurement with the QA400. I am showing the measurement picture produced by the QA400 measurement software (with a few of my own text comments added). It has frequency markers for three points I wanted to focus on (1KHz, 10KHz and 20KHz).

My results are pretty similar to what Beolover got although it appears the QA190 differential probes helped my measurements out on the low frequency noise. The drop at 20KHz was right around 1dB which is more than what the B&O published specs state but then our measurement methods are not exactly the same. Between these Beomaster performance results and the last few we have done it compares with very similar results. The results are definitely a pass for this Beomaster receiver.

Beomaster 8000 - Output Amplifier Bench Testing Review

I felt like the left and right channel output amplifier boards were in good shape after the restoration and I was successful in presetting the no-load current adjustment during the bench testing.

However, I was left disappointed that I wasn't able to preset the DC offset adjustment of the boards during my bench test. I tried doing the DC offset adjustment but when I connected a dummy 8Ω speaker load for the adjustment procedure the current draw on the ±55V supplies was not balanced. There was quite a bit more current on the -55V supply than on the +55V supply. To go along with that I could only adjust the DC offset to around -380mV across the speaker load.

Here is a picture of the bench power supply load when attempting the DC offset adjustment.

The current draw imbalance occurs as soon as the 8Ω speaker load is applied. Even when the output amplifier is in the off state (+15V supply at zero volts).  Turning the output amplifier on (+15V supply above +14V) raises the current draw on the ±55V supplies but there is still a 50mA difference.

Why was that?  The DC offset adjustment procedure calls for the speaker connection to be on and the volume level to be at zero.

I have the 8Ω speaker load connection and I have nothing applied to the audio input connector on the output amplifier board.

Here is the schematic marked up with my test connections.

You can see that I have the audio input pins shown as having the option of being open or shorted together. I first thought that shorting them together would solve the problem because that should be the same as zero volts per the service manual instructions.

That was not correct. My next thoughts were that this board must have some fault preventing the DC offset adjustment. When I tried three more output amplifier assemblies I got the exact same result on all three. No way that I have four faulty boards. Especially with the exact same fault.

Here is a picture of my fourth output amplifier assembly DC offset adjustment test.

This meant that my stand alone bench test is not correctly representing what the board sees when it is properly installed in a Beomaster.

At this point I turned to Martin Olsen for advice. I get a lot of parts from Martin and he is an experienced Bang & Olufsen tech who has provided me with some great information on repairs over the years.

After thinking about the problem of attempting this stand alone bench test procedure Martin suggested that to be accurate in my audio input setup the two input pins should be grounded.
That was exactly the problem!

From the schematic picture above you can see how I thought shorting the two audio input pins would ground the signal so I would have zero volts. That turns out to not be sufficient for this stand alone board bench test. Looking at the output amplifier trace side and the service manual schematic again I can see that the audio input low signal is shown to be going to signal ground. Not board ground. There is no real signal with the output amplifier sitting on the bench by itself. Signal ground is really just undefined in this test environment.

Following the board traces that P39-1 connects directly to I can verify all of the signal ground connections indicated on the schematic. I have marked them up on my schematic below as the green, dotted lines to clarify those connections.

Here is the clarified schematic and my new test setup.

So connecting the audio inputs (P39-1 and P39-2) together is pretty much the same as leaving them open on this stand alone bench test. That is what I observed in attempting to perform the DC offset with the 8Ω speaker load. With the board being stand alone on the bench it requires P39-1 and 2 connected to each other and to board ground so there is a true reference to the return in this environment. That sets the input at a defined, 0 VDC level and the amplifier stable for the DC offset adjustment.

Marking up the Beomaster relevant schematic parts that show the output amplifier board connected to the rest of the Beomaster I can see the required ground connection I missed. The audio signal ground between the Filter & Tone Controls board gets routed to chassis ground. Chassis ground connects to the output amplifier board ground post, P37, which connects to the main ground.

Connecting up output board P37 ground post to the P39 input pins, I immediately noticed the result on the bench power supplies.

The current draw on the ±55V supplies is now close to the same. This is what I see with good output amplifier boards in a Beomaster 8000 receiver. The DC offset adjustment is now able to dial in to the expected 0 ± 0.5mV range.

Here are some pictures of the output amplifier board and the test setup.

Here are a few summary notes regarding the stand alone bench testing of the Beomaster 8000 output amplifier board.

1. Start with the audio input connector pins shorted to the ground pin on the output amplifier board.

2. Have the speaker load disconnected

3. You need two DC power supplies capable of +55 VDC and have current limiting protection. I set mine to stop at 150mA. This will prevent a bad situation should the output amplifier have a short somewhere.

4. You need a third DC power supply capable of +15 VDC to enable the output amplifier to turn on.

5. Dial up the ±55 VDC supplies first. Increase their voltage slowly from 0V to 55V. Observe their current draw as the voltage is increased. When new parts (including trimmer resistors) have been installed on the output amplifier board it is very common for the supply current to go above 130mA way before 55 volts is reached. When this happens you must dial down the supply current by adjusting the no-load current trimmer on the output amplifier board. Use that trimmer to keep the supply current around 30mA until ±55 volts are reached by the supplies.

6. Turn the output amplifier on by slowly increasing the +15 VDC supply from 0V to +15V. Observe the current draw on the ±55V supplies while doing this. The ±55V current should not go above 120mA. The typical current values I have seen are 100mA to 110mA.

7. With the output amplifier powered up and on, adjust the no-load current to the desired 18mV value across the emitter resistors.

8. Power down the +15V supply to turn the output amplifier off.

9. Connect the speaker load

10. Reapply the +15V to turn the output amplifier back on and observe that the ±55V supply current is in the 100mA to 110mA range.

11. Measure and adjust the DC offset trimmer to get 0±0.5mV across the speaker load.

That should do it. The output amplifier assembly is nicely checked out prior to installation. If there is a failed component on the board it should reveal itself during these tests and should be easier to trouble-shoot with this test setup.