|Rockhard||3/8/2018 9:44 AM|
|Marshall Dynamic Bass DBS200 72410 problem|
Hi folks, this is the first Marshall DBS200 I've ever worked on and its got an interesting problem.
There is neg voltage on the speaker out, so I first tested the poweramp components for shorts but none of the diodes, drivers or output devices are shorted. What I did find was that the following components were literally toasted and I replaced them:
TR20,TR21, TR22, R19, R46, R47, R48, R49, C30, C28 - I replaced TR11, 12 &15 too for good measure although they weren't shorted at all.
Problem is the R47, R49 & C30 immediately heat up when I turn the variac up to less than half of mains input voltage (40vac). Light bulb limiter glows faintly, and nothing burns up spectacularly at this mains voltage input, so I took all my voltages from that reference point.
Voltage drop over R47 is 34.2vdc and R49 21.12v and needless to say being only 1/4w, they get real hot.
Voltages are as follow:
Mains input 40vac
HT +/- 32v
MT +/- 17.5vdc
LT +/- 6.5vdc
Speaker out -29.5v
Voltage drop over the mosfet drain resistors are: R25 -10v & R27 -15.5v, no voltage drop on R16 & R17
Anybody have any idea where to start?
|Enzo||3/8/2018 11:41 AM|
|Any voltage readings are essentially meaningless at 40v on the mains. Also, if using a variac AND a bulb limiter, realize that at 40v mains even a DEAD short at the amp will result in only a dim bulb glow.|
Recognize that your output transistors are TR3,4,8,9.
I see no reason for C30 to heat up, unless somehow we greatly exceed its voltage rating.
For R47,49 to burn up, we have to have TR21,22 turning on hard. Or be shorted. Since you replaced them, did you use exact parts? Or did you make substitutions?
TR1,2,13, and the equivalent ones on the negative side, are your commutators, switching in the higher voltage rails when needed. The MOSFETs control those.
|Rockhard||3/8/2018 2:08 PM|
|Hi Enzo, thank you for the reply.|
The only reason I set the mains voltage as low as 40v was merely to be able to take some basic readings without frying the tiny quarter watt resistors again - even at this low voltage they heat up enough to burn skin.
I see now after looking at the diagram again that you're correct that the outputs are in fact the transistors, not the mosfets.
C30 was replaced with a 250v rated cap, and I think it might be heating up because it is adjacent to R49
Regarding Tr21,22 - I replaced them with the exact parts. The originals had exploded quite literally, I forgot to mention that.
What would your suggestion be -should I replace all the semiconductors regardless if they measure shorted and be done with it, or not? Also, why would R25 and 27 be heating up as well - is it the mosfets drawing too much current through them, or more likely that tr13,14 are the culprits?
|Enzo||3/8/2018 3:28 PM|
|I understand why you do it, I just mean that it is hard to use the numbers. Like adjusting your car steering alignment with flat tires.|
OK, real parts. I was concerned maybe subbing a US type for an asian type which has legs in different order. So we avoided that.
R25,27 burning? Any burning resistor needs a current path through it. And the only paths I see for those resistors is through their MOSFETs. So either the MOSFETs are faulty or something is turning them on hard so something else can draw current through them.
I am not really a believer in shotgun methods. Are we testing transistor with diode test and checking both base junctions as well as for E-C shorts?
|g1||3/8/2018 8:14 PM|
|Are these tests being done with no load connected?|
|Rockhard||3/14/2018 1:19 AM|
|Okay I'm on this again today - Ihavent removed the output transistors yet to do proper tests yet, only tested for short circuts while in place.|
R25,27 aren't burning up because they are 5w ceramics, but they do get warm compared to their complimentary pair - I'll do some measurements with full line voltage today, burnt resistors be damned.
@g1 - speaker load has been disconnected, shall I take readings with a dummy load in place?
|Jazz P Bass||3/14/2018 10:00 AM|
|"speaker load has been disconnected, shall I take readings with a dummy load in place? "|