I thought this may be an interesting little detail of interest to the community:

I noticed recently that the volume lights of my BM1900 were acting unevenly, i.e. the volume "down" light was always brighter than the "up" light, i.e. not even at full volume was equal brightness achieved. I assumed the correct way would be that at the lowest setting only the "down" light is on, while at the highest setting only the "up" light would be on.

While trouble shooting the stand-by problem (see last post) I took the opportunity to take a look at the volume adjustment circuit. I found that the previous owner had replaced one of the two transistors that "balance" the volume lights (3TR13) with a BC557B. The other transistor (3TR12) was still the original BC549C. The service manual prescribes a BC558B for TR13. Since I did not have this type at hand, but a BC547B, which is the listed complement for the BC557B, I exchanged the BC549C with it. This successfully changed the balance between the lights to the proper behavior. Now the "up" light goes full brightness when turning the volume up to the max, and vice versa the "down" light is at max when the volume is fully turned down.

I thought this an interesting phenomenon, since initially I assumed that these transistors are very similar, and should be interchangeable w/o problem...but apparently not so for this particular circuit where npn/pnp symmetry is important.

If anyone had some insight as to which particular transistor spec is responsible for the above phenomenon, I would be very interested to know. I am hypothesizing this may have to do with variations of the C-E resistance between the different transistor series...

On to finally fixing the veneer of the BM1900!

Rudy

I don't think the exact specs are that important. I lean more against a weak transistor.
If one of the balance transistors were already replaced because it had failed, it makes
sense to think that the other one suffered some damage too.
Maybe wrong wattage lamps were used at some point or something shorted.
The Beomaster 1900/2400 series is a bit sensitive to wrong lamps.

Martin

It did have "modern" lamps installed (although with a fairly high 26R resistance). I wish I had had a BC558 on hand yesterday to try what happens when exchanging the 557 with it...It did occur to me too that the exchanged BC549 transistor may have had a problem, but it shows nice rectifying I/V curves with a ~700 mV turn on...

Rudy

BC557 and BC558 are quite similar (PNP).
The same goes for BC547, 548 and 549 (NPN).
If I remember correctly, it's the self-induced noise in operation that decides
what markings the transistor receives. BC549 has better noise specs than BC547 etc.
The hfe specs (amplification) is usually given the same way but as a letter after
the number, ie. BC547C has a slightly higher hfe than BC547B.
The circuit is much more sensitive to wrong wattage lamps than unmatched transistors.
It's not an exact representation of the volume anyways, merely a quick guide so it's
normally not too critical when it comes to selecting transistors but may stall completely
or become unreliable with wrong lamps.

Martin

Thanks for the insight!! I always wondered what the letters in those transistor types meant etc... I realized I need to hit wikipedia a bit and ramp my transistor knowledge!...;-).

Rudy

Hi Martin,

Intrigued by your comments I decided to learn to do a circuit simulation in Pspice. I was able to get the volume lamp circuit running on the computer (see screen shot below; scroll to the right to see the voltage/current traces measured at various points in the circuit). You are absolutely right that the transistor type has not much influence. When short circuiting Q2 over CE in the simulation with a ~100R resistor I was able to get a similar behavior like described above, that the "up" lamp does go to full brightness. This indeed suggests that the BC549 that I exchanged had a problem.

Aside from the above, I also learnt this from the simulation: The voltages that one measures in-between the lamps depending on the volume setting (~5-10V) can only be properly obtained in the simulation when the lamps are approximated with ~146R resistors (my modeling software did not have incandescent lamp simulators installed-not sure if they even exist). The lamps that you sent me show only 16R in cold condition...that made me hook the spare one you provided to my power supply, and indeed, when they are glowing their resistance is almost 10x, i.e. between 100R (~5V) and 150R (~11V). 

This (I think) explains that the current traces 'measured' in the two lamp-resistors in the simulation are not symmetrical for the volume "up" and "down" branches of the input signal (green trace, which oscillates between 0 and -12 V, simulating the output from the A/D converter in the real circuit). As the lamps change their resistance due to their different brightnesses (temperatures) the circuit changes dynamically, which is not reflected in the simulation.

Rudy

Wonderful simulation ! 

The discriminator (tuning-) lights are balanced too and also quite sensitive
to correct wattage lamps. The reason is exactly as you've found out, that
the filament resistance is very low when cold and varies significantly with the voltage.

Martin

Nicely done :) I once spent a fortune on Tina (similar I assume to pspice) but couldn't remember how to operate it from my collage days so it has remained unused :(

I really respect you for doing this it is a great way of learning all the little things one might miss through simply stairing at a diagram!

Tonights homework: try introducing other faults and try to guess what the outcome might be before simulating 

Thanks for your encouragement! BTW: I used LTspice, which is free and comes with a nice library of standard parts (and of course the entire catalog of Linear Technology parts...;-). Here is the link: http://www.linear.com/designtools/software/. I think it is possible to add other manufacturers parts to the library as well.

Rudy