Sign in   |  Join   |  Help

Taking on the BM8000

rated by 0 users
Not Answered This post has 0 verified answers | 168 Replies | 1 Follower

krais
Top 500 Contributor
The Netherlands
134 Posts
OFFLINE
Bronze Member
krais posted on Sat, Dec 28 2019 11:56 AM

Hi all, 

I've recently started restoring an old BM 8000 (type 1901) that I bought second hand years ago. The seller told me that the BM would not turn on (repeatedly clicking relays). I did not try to power the unit up to avoid blowing the output stages so not sure what the actual symptoms are.

After not being able to find a B&O technician in the Netherlands that is willing to take on a BM8000 due to the labor involved I decided to have a go at it myself. I know, I know, these are complicated amplifiers, not really stuff for an inexperienced hobbyist with little knowledge of electronics... I just cannot resist the challenge so decided to give it a try with the help of information available online (beolover website). 

Any help along the way would be appreciated. I love this design so would be awesome to get it working again (or at least take care of the laborious tasks such as recapping and rebuilding the LED displays).


All Replies

krais
Top 500 Contributor
The Netherlands
134 Posts
OFFLINE
Bronze Member
krais replied on Mon, Apr 13 2020 11:01 AM

Thinking this through some more... I could unplug P76 from the uProcessor board so that there's no power going to that board. This way the uProcessor will not activate and therefore cannot turn off the +/-15V lines. With the 7RL1, 7RL2 relays bridged I could then slowly raise the voltage on the variac while monitoring the current draw without even having to switch on the BM8000.

Anything I'm overlooking here?

 

manfy
Top 500 Contributor
142 Posts
OFFLINE
Bronze Member
manfy replied on Mon, Apr 13 2020 12:33 PM

But P76 is a 10-way plug. What do the other pins do?

Seems like a hassle - and it requires that you can read and understand circuit diagrams very well.
Even if you pull only pin P76-7 to disable 15V rail switching, you should be very sure that this pin ONLY performs this function!! Additionally, free-floating pins can cause the strangest problems. When the uP sets a low signal it is switching its output to GND. Some of the electronics connected to that signal might depend on that!

I think I'd just do a two way test:

1) keep P51 pulled. This way you disable the main transformer. Turn the voltage up to 120V or whenever you can power on the BM8000 via selector buttons. Increase voltage to 230V and keep monitoring the current.
If that works fine you know that the standby transformer and all rails it supplies are ok.

2) jumper the relays and slowly turn the voltage up to 230V while monitoring the current. Do NOT power on with selector buttons.
If that's ok you know that the main transformer and all its rails and connected electronics works.

By adding the mains current of both operations you know the idle current of the unit.
If both subsytems work fine individually, there's a high chance they will also work in normal operation. Of course it's not a guarantee, because the electronics on the standby transformer might activate some parts of the output amp that can cause a problem. But that seems farfetched.

If you're excessively worried, you could also replace F2 with 1A rapid-blow for the final startup test with both transformers. If something is really wrong this could limit the current on the secondary side and could avoid serious damage.

sonavor
Top 25 Contributor
Texas, United States
3,701 Posts
OFFLINE
Gold Member

Hi, I apologize as I probably wasn't clear in using the variac and dim bulb tester. Especially on the Beomaster 8000. They are useful tools for an initial check to make sure something isn't shorted when first powering on and you thought there was a problem (smoke) coming from the start up circuit.

Once you are past that the variac isn't very useful on the BM8000 as the Beomaster will just come up in standby when fully powered from the line voltage (as it should).  The variac could be useful in monitoring the current when the Beomaster switches to full on but it sounds like you are also past needing that. You also don't need the dim bulb tester in the circuit now. I agree with Jacques that at this point those devices can confuse what is really going on. 

I suggest returning to connecting power to the Beomaster 8000 by just plugging it into line power and then measure the power supplies. If activating a source to turn on the Beomaster causes the protection circuit to remove power then start disconnecting power from some of the boards (i.e FM, preamp). 
If you suspect a problem with the output amplifiers you can pull the rail voltages from them before plugging the Beomaster in (being careful they aren't left where they can touch something conductive). However I remember you checked the output amplifier boards pretty thoroughly earlier.

Leave the power supply board connections in place as those are need to check out all of the other voltages.  Let's see what the present state of the Beomaster 8000 standby and power on state is - what DC voltages are present or not present.

-sonavor

krais
Top 500 Contributor
The Netherlands
134 Posts
OFFLINE
Bronze Member
krais replied on Thu, Apr 16 2020 8:43 AM

sonavor:
Hi, I apologize as I probably wasn't clear in using the variac and dim bulb tester. Especially on the Beomaster 8000. They are useful tools for an initial check to make sure something isn't shorted when first powering on and you thought there was a problem (smoke) coming from the start up circuit.

No worries Sonavor, an isolation transformer and variac were both already on my wishlist. So the device is a welcome addition to the workshop even if not that useful for the BM8000.

Update: I did a power up test with the relays bridged and found no further issues. The 55V rails measured ok and there was no high current draw. 

After that I proceeded to power up the BM8000 without the variac and dim bulb tester. At first glance all was well: the double relays are now working and the uProcessor switches on without issues.

However, I did check the idle current on both output boards and while the iniitial values were as expected (~16mV) the voltage quickly started to drop towards 0V on both boards. This is not something I noticed on my earlier tests so not sure what's going on. I'll double check the wiring and do some more voltage checks this evening.

manfy
Top 500 Contributor
142 Posts
OFFLINE
Bronze Member
manfy replied on Thu, Apr 16 2020 3:20 PM

krais:
However, I did check the idle current on both output boards and while the iniitial values were as expected (~16mV) the voltage quickly started to drop towards 0V on both boards. This is not something I noticed on my earlier tests so not sure what's going on. I'll double check the wiring and do some more voltage checks this evening.

It's unlikely that both output amps suffer from the very same fault, so I'd guess that this rapid decrease in idle current on both amps at the same time is caused by some controlling circuitry that acts on both amps. As per block diagram there's only the fault switch and the clipping circuitry that's doing that.
If the supply voltages to the amps are ok, I'd focus my troubleshooting on the fault switch and the clipping circuits next.

 

krais
Top 500 Contributor
The Netherlands
134 Posts
OFFLINE
Bronze Member

Looks like you were right, Manfy.

Yesterday I started checking the base voltages and the 5V/15V/-15V lines all measured ok. The +/-55V lines are being cut off immediately when switching on. I measure the relay control voltage on P51 and indeed that voltage drops when powering on, which causes the relays to switch off.

Next I did some voltage checks on the fault switch circuit on PC6. Values in yellow are with P51 detached (relays not activated) but uProcessor switched on, values in green are with relays activated and disengaged. Definitely something that is triggering the fault switch...

So I figured there were 2 possibilities. Either #1 there is an issue on one of the output boards that is sending a trigger signal to the fault switch circuit or #2 something is wrong with the fault switch circuit itself. To eliminate #1, I began testing with one of the output amplifier boards disconnected. Starting with the left board disconnected, the Beomaster powered on and the relays did not disengage anymore. 

So it looks like there is an issue with PC5L... Will do some more troubleshooting this weekend.


manfy
Top 500 Contributor
142 Posts
OFFLINE
Bronze Member
manfy replied on Fri, Apr 17 2020 9:40 AM

Excellent! Good progress!

So, there's definitely a problem in the left output amp. The fault switch on PCB6 works as it should.
Now I see 2 potential sources of the problem:

#1) The output darlingtons really overheat, which causes the PTC R251 to rapidly increase its resistance and activate the fault switch correctly. You can check that by touching the output transistors near the thermistor R251. If they are hot then R251 is doing its job correctly and you have to continue troubleshooting in the output transistor path.

#2) If the transistors are cold, then there is something pulling the voltage at the loudspeaker output towards a negative DC value and that's what's sending the fault switch signal (between R253 and R254) to the 12.5V you're measuring. If so, you should look at the offset circuitry (TR201/202/203).

krais
Top 500 Contributor
The Netherlands
134 Posts
OFFLINE
Bronze Member

Update on PC5L troubleshooting: I removed the board from the chassis and hooked it up to external power supplies to do some voltage checks. Immediately, I noticed an issue when powering up the board. The idle current voltage was at -10mV without the 15V rails being applied and the -55V rails was drawing unusually high current (~70mA). 

After that I spend hours checking pretty much every component on the board as well as traces and solder connections around the offset circuit but I could not find anything wrong. Finally I noticed this...

A trace connecting ground pin P37 with the ground plane had been completely burnt up! That explains the high DC offset leading to the fault switch circuit kicking in. After restoring the board by adding a jumper wire I completed the idle current and DC offset procedures and the output amp looks in good condition again.

However, you have to wonder what could have caused such a wide trace to burn up. In fact, it seems likely to me that this was the cause for the smoke appearing when testing the BM8000 a few weeks ago (instead of the resistors on PC7 overheating). I'm really struggling to understand what could have caused such high current going through that trace.


sonavor
Top 25 Contributor
Texas, United States
3,701 Posts
OFFLINE
Gold Member

I am kind of surprised none of the transistors gave out. It's hard to say what happened considering you ended up also finding some bad connector wiring after the smoke incident. Nice to be able to check out the output amplifier assembly outside the Beomaster. 

When you re-install the left channel output amplifier carefully check the heatsink mounts and clearances on the power transistors with the BM8000 metal frame. Check the right channel again too while you are there.

-sonavor

krais
Top 500 Contributor
The Netherlands
134 Posts
OFFLINE
Bronze Member
krais replied on Sun, Apr 19 2020 9:23 PM

Yeah, I forgot to mention that heatsink mount on TR206 was almost touching the leads. The heatsink mount must have moved when mounting the heatsink to the chassis. Very well possible that at some point the leads were shorted.

Another thought: when the smoke appeared I was testing an audio input signal with headphones plugged into the BM8000. Could this have something to do with it perhaps? I noticed on diagram 2 that the headphone signal ground is connected to pin 37 on PC5 (just next to the burnt trace).

sonavor
Top 25 Contributor
Texas, United States
3,701 Posts
OFFLINE
Gold Member

I would recheck all of your ground connections to make sure you having missed any.  Are you headphones still working?  They aren't shorted to ground? 
The headphones get the audio signal the speakers would get except the headphone path goes through a voltage divider.  That cuts the signal about in half.  The ground for the headphones and speakers should be the same ground. It gets confusing because in the board and chassis connections involving the output amplifiers ground, signal ground and chassis ground all tie together.  Go through that part of the schematic and verify the grounding connections at the points shown. Make some measurements with the headphones plugged in (and the Beomaster not plugged in).

-sonavor 

krais
Top 500 Contributor
The Netherlands
134 Posts
OFFLINE
Bronze Member
krais replied on Sun, Jul 5 2020 9:02 AM

Hey guys, finally another update on the BM8000 project. I have been quite busy the last 2 months and needed to finish a few side projects so did not have a lot of time to work on it.

So let's continue where I left off, which is putting everything together including the repaired left output board and doing some basic tests. Immediately I found another issue: the idle current on the left channel was rising much more quickly than on the right channel (ramping quickly from 12mV to 20mV within a couple of mintues). Yet another issue with the left power amplifier board it seemed...

I removed the output boards from the Beomaster for the 3rd time ( Sad ) to investigate. When testing the idle current on the individual board with external power supplies, I noticed that there was an initial voltage of about 0.8V across the 0.18ohm emitter resistors (without the 15V rails being applied) on the left board vs 0.0V on the right board. I suspected one of the emitter resistors (when measuring I noticed that resistance seemed to drift) and decided to replace these. This solved the issue of the initial voltage on the left channel.

So... after assembling the Beomaster again I fine-tuned both the idle current and DC offset trimmers. Good news: the problem on the left channel has indeed disappeared and the idle current on both channels is now quite stable. I adjusted the idle current trimmers so that the initial voltages across the emitter resistors is 16mV and stabilize after warming up for an hour around 21mV.

 

 

krais
Top 500 Contributor
The Netherlands
134 Posts
OFFLINE
Bronze Member

Another improvement: replacing these old speaker switches (one of which was faulty) with modern encapsulated switches.


krais
Top 500 Contributor
The Netherlands
134 Posts
OFFLINE
Bronze Member

I took the base PCBs from the old switches and used a rotary tool to create new slots to fit the new switches. The switches have been secured to the base plates with some blobs of CA glue and baking powder.  


krais
Top 500 Contributor
The Netherlands
134 Posts
OFFLINE
Bronze Member
krais replied on Sun, Jul 5 2020 9:29 AM

Next: it was finally time to test whether there's an audio signal on the speaker outputs. I hooked up a 0.5V sine wave from an audio signal generator to TP1, connected 8ohm dummy loads to the speaker terminals and slowly increased the volume while measuring AC voltage on the speaker outputs with a multimeter.

 

Good news: the Beomaster did not blow up this time Smile

More good news: I measured an AC voltage that was increasing while adjusting the volume on both speaker outputs. 

The bad news: the voltage measured was in the range of 0-20mV. Sad

 

I plugged in my headphones and indeed the signal was so small that I could not hear anything. So somewhere in the signal path there is a problem... more troubleshooting to do. 

Page 7 of 12 (169 items) « First ... < Previous 5 6 7 8 9 Next > ... Last » | RSS
Beoworld Security Certificate

SSL