Is that a trick question? 😠 😜

If you pick just any transistor with arbitrary gain values and voltages, one would surely expect that the output is different than with original transistors, don't you think? Do you think that the B&O guys picked the BC547B and the BC182C for this opamp TR104/105 just to annoy you? (see footnote)

I would guess they did it on purpose and for some technical reason. Either they actually calculated the amplifier design or - more likely - they used a standard design and adjusted it to their needs.For today's standards, this is a run of the mill class AB amplifier with feedback to the driver stage. 
The base of TR105 is the non-inverting input of this opamp and carries the feedback from the final output signal. The base of TR104 is the inverting input and it carries the audio input signal. Changing hfe of those transistors will surely have an effect on the characteristics of the amplifier. If you must replace them, then use equivalents with similar characteristics.

Using 30V transistors instead of the original 50V ones is not a good idea. It may work if all other components are ok, but if TR108 shorts for whatever reason this will also send TR104 and 105 to their graves. B&O chose 50V versions; so, use the same or bigger.

------------------------

PS: Sorry for my initial sarcasm. I just find it a bit strange that you take the effort to hand-match modern transistors, but then just replace the original models with some other type without considering their function in the circuit.
The general rule in electronics is: If you need to replace obsolete transistors, you pick models that are as close as possible to the original. This way you have a high chance of success without having to analyze the whole circuit in detail. 
If you're good in electronics design, you can change the original transistor specs quite a bit based on what's going on in the circuit. It does leave you though with the risk that you're overlooking something that the original designers had already identified as a problem and fixed in their commercial product.

Oops typo, sorry for the confusion. The replacement transistors I used are BC547B, not BC548B. 

The originals were all BC547C even though the service manual lists BC547B for 6TR104/204 and BC182C for 6TR105/205.  

In the Fairchild datasheet I see following:

BC547B: hfe = 200-450
BC547C: hfe = 420-800

So yes, I'd consider that a substantial difference.
I didn't find any datasheet on BC182C, but BC182 has 125-500 gain. That is also quite different from 420-800.
You said that the transistors installed initially were all 547C. Did they seem factory-assembled or were they changed by some repair guy?
If they were factory-assembled, I'd stick with the same models. 15mV instead of 10 doesn't seem critical. That makes some 38mA bias current instead of 25mA. Your amp will run a bit hotter and a bit less "energy efficient" than the speced version, but it shouldn't be a problem for a 70W transistor (IC100). 

hfe for the BC547B's measured around 400, so I figured that might be ok (given the service manual listed a BC547B as well). But yeah, I can definitely swap them for BC547C's or just put back the originals.

It does not seem the BM1500 was worked on before, so the original transistors seem to be factory installed.

I guess the broader question I have is: what are the main contributing factors to DC offset (forgive my limited knowledge of AB amplifier circuit fundamentals)? My initial assumption was that dc offset is mainly dependant on the matching of transistor pairs 6TR104/105, 6TR204/205 and thus a relatively large DC offset while using matched transistors could point to an issue. But reading through some more threads it seems that DC offset can also be introduced by the power transistors? If the latter, would it make sense to experiment with some transistor pairs with different hfe matching values in order to reduce DC offset to below ~10mV?

25-30mV DC on the outputs of this amplifier is not a problem. I wouldn't worry about it.
Any (serious) imbalance in the differential amplifier would have a much larger effect (from 10ths of volts to several volts) on the outputs.

Martin

Oh goodness! I just realized that you talked about the DC offset on the speaker terminal, but I've been thinking of the 10mV setting of the idle current all along! 😠 So, sorry for that! I deleted my reply to avoid further confusion.

In any case, I still agree with Martin.

Granted, 0V at no input signal is the ideal value for a distortion-free output signal, but a minor DC offset of 25mV is quite certainly not audible.
In the service manual it shows a value of 0V +/-200mV at the speaker output. So that's what B&O considers "within spec".
They did not add any DC offset adjustment potentiometer like they did in the BM8000, which means they don't see a minor offset as problematic.

No worries Manfy and thanks to you and Martin for the quick help.

I did run into another problem with the BM1500 though...

When testing the output amps I noticed that the amplitude of the left channel is around 25% higher than the right channel. First I suspected a problem in one of the power amp circuits but after a number of measurements it seems that the problem is introduced in the preamp stage, in the treble/bass circuit to be precise.

In the pic below measurements I did with a multimeter in AC mode at various points in the preamp stages (after injecting a 1KHz signal from an audio signal generator through Tape 1, volume halfway).   

I've checked the treble and bass potentiometers for corrosion but they seem fine.

The electrolytic caps in the preamp circuit already have been replaced and I could not find any obvious faults when measuring the other components in that area (in circuit). No obvious cold solder joints as far as I could see.

Any ideas on next steps? Thanks!

Is that channel imbalance audible in the speakers?
How far do you have to push the balance slider to get a balanced signal on both channels (according to the multimeter)?

krais wrote:

I've checked the treble and bass potentiometers for corrosion but they seem fine.

How did you check that? Oxidation is not really easy to see with the naked eye. Did you try any sort of cleaning?