If the bulb goes out the bridge is OK, there is no short.
I would concentrate on fixing one channel., lets say the one that does not light the bulb.
First check is to confirm the rail voltage is correct on the upper output transistor collector. The circuit neatly splits into two and all you need concentrate on is this:
Measure and record the voltages on the transistors and post the result here. You can copy and paste this format and add your readings. The voltages in brackets are typical of what would be expected.
Lets call them:
'Upper NPN Output' which is the top right.
C= (49)
B= (26)
E= (25.4)
'Upper driver' which is top left.
C= (49)
B= (26.6)
E= (26)
'lower NPN Output' which is lower right
C= (25)
B= (1.4)
E= (0.8)
'Lower Driver' which is the middle lower one.
C= (25)
B= (2)
E= (1.4)
'Current Regulator' which is lower left one.
C= (2.3)
B= (0.6)
E= Ground
I was referring to the bridge on the right channel, which is making the bulb bright and blowing fuses.
on it, boss. Thanks for all the advice, I'll report when I have news.
Ah, I was thinking it was a shared bridge. If a diode in the bridge has failed then that usually shows by simple diode checks on the bridge, even testing in circuit. Measure from each leg in turn to all the other three.
If its blowing the fuse with the 2.5A rail fuse removed then yes, bridge or reservoir cap is suspect.
Sorry - just to clarify a couple things here... I have tested the bridge already using diode mode - all measures fine, but @Extreme_Boky said that theirs measured fine passively but failed under load, so I wondered if that could be the case here.
Have replaced one reservoir cap, but not the other, maybe I should do that. When you say 'blowing the fuse with the 2.5a rail fuse removed' I should also clarify that it was the rail fuse that blew
No problem 🙂
OK, so if the rail fuse blew (the 2.5A on the diagram) then you need to check first of all that with the fuse removed you have at least the 49 volts DC rail voltage. That will be measurable on one side of the empty fuse holder or on that 33k shown across the cap. Generally bridges are pretty reliable and failure 'catastrophic and permanent' when it occurs.
If the voltage check is OK then switch off and measure on ohms from chassis to the other side of the fuse holder which is the rail as it goes to the amp. Does it read short? The 33k is to discharge the cap when the fuse is open but it will be quite slow to do so.
Also do a resistance check (with the amp OFF) from the emitter of the upper output transistor to chassis (ground) and make sure there is no short there. The emitter is the lead going to the 0.56 ohm and the speaker coupling cap.
Beyond that we will have to get creative with simple checks to determine what is happening. It should be fixable as it is quite a standard circuit really.
OK, so if the rail fuse blew (the 2.5A on the diagram) then you need to check first of all that with the fuse removed you have at least the 49 volts DC rail voltage. That will be measurable on one side of the empty fuse holder or on that 33k shown across the cap. Generally bridges are pretty reliable and failure 'catastrophic and permanent' when it occurs.
If the voltage check is OK then switch off and measure on ohms from chassis to the other side of the fuse holder which is the rail as it goes to the amp. Does it read short? The 33k is to discharge the cap when the fuse is open but it will be quite slow to do so.
Also do a resistance check (with the amp OFF) from the emitter of the upper output transistor to chassis (ground) and make sure there is no short there. The emitter is the lead going to the 0.56 ohm and the speaker coupling cap.
Beyond that we will have to get creative with simple checks to determine what is happening. It should be fixable as it is quite a standard circuit really.
I've had chance to throw the circuit into a simulation... very interesting seeing how it behaves... but it also gives us something to work to.
I've attached the sim, you will have to use your own models and delete my test load speaker but default 2N3055's for outputs and 2N5551 and 2N5401 essentially work just as well here.
I've put the key voltages on the diagram. The lower three transistors that form the current sink should limit any possible fault current to around 1A even even something like the upper output was short.
I've attached the sim, you will have to use your own models and delete my test load speaker but default 2N3055's for outputs and 2N5551 and 2N5401 essentially work just as well here.
I've put the key voltages on the diagram. The lower three transistors that form the current sink should limit any possible fault current to around 1A even even something like the upper output was short.
Hello hello!
I have news. Firstly - @Mooly thank you so much for your help, and the above-and-beyond work you did in creating the simulation, appreciate it so much. Thank you everyone else who chimed in too. I appear to have fixed the two power amps...
The principal problem with the left channel was that it had bad solder joints on the bridge and simply wasn't running. I resoldered the bridge and it came to life. I did notice that one of the filter caps was swollen in this channel too, again reading fine but noticably bulged, so I replaced it. Next, I took the fuse out and used my amp-meter to check current draw for each channel. Left channel was surging to 2.7a, before settling down to 1.5a, right channel was surging to 3.8a (which was enough to blow the time delay fuse) before settling to about 2.1a. I adjusted the potentiometers on both channels to a resting current draw of 1.4a - both start up great now with no more blown fuses 😀.
Sorry I didn't make more use of your sim Mooly! Thanks again.
I have news. Firstly - @Mooly thank you so much for your help, and the above-and-beyond work you did in creating the simulation, appreciate it so much. Thank you everyone else who chimed in too. I appear to have fixed the two power amps...
The principal problem with the left channel was that it had bad solder joints on the bridge and simply wasn't running. I resoldered the bridge and it came to life. I did notice that one of the filter caps was swollen in this channel too, again reading fine but noticably bulged, so I replaced it. Next, I took the fuse out and used my amp-meter to check current draw for each channel. Left channel was surging to 2.7a, before settling down to 1.5a, right channel was surging to 3.8a (which was enough to blow the time delay fuse) before settling to about 2.1a. I adjusted the potentiometers on both channels to a resting current draw of 1.4a - both start up great now with no more blown fuses 😀.
Sorry I didn't make more use of your sim Mooly! Thanks again.
That's brilliant news 👍 well done. Poor joints doesn't sound good for Sugden 🙁
The circuit seems to show 0.8 amp as a quiescent current draw, just mentioning that we you saying you have set to 1.4 amp. That circuit is all I have to go on although the earlier A21 I think also used 0.8 amp.
The circuit seems to show 0.8 amp as a quiescent current draw, just mentioning that we you saying you have set to 1.4 amp. That circuit is all I have to go on although the earlier A21 I think also used 0.8 amp.
Interesting, so I wonder if the circuit in post #6 really is of the A21SE.
It actually partly answers a question I had and that was that the sim has pretty limited 'real world' output current capability with a simulated tough but not unreasonable load (my B&W 703's actually).
You can see the current limitation here with load current and voltage across the speaker. This is at 0.8 A (The blue trace doesn't show well in an image)
This is at 1.4A bias current. The undistorted peak to peak current tops out at -/+4 amps here.
And 8 ohms resistive allows -/+21 volts swing (which pulls around -/+2.6 amp)
It actually partly answers a question I had and that was that the sim has pretty limited 'real world' output current capability with a simulated tough but not unreasonable load (my B&W 703's actually).
You can see the current limitation here with load current and voltage across the speaker. This is at 0.8 A (The blue trace doesn't show well in an image)
This is at 1.4A bias current. The undistorted peak to peak current tops out at -/+4 amps here.
And 8 ohms resistive allows -/+21 volts swing (which pulls around -/+2.6 amp)
I've done some comparisons with the pcbs, and it looks like this circuit is similar, but not quite the same. Thanks again for everything you do.
Buongiorno, vorrei ricollegarmi a questa discussione chiedendo a Mooly se nello schema della simulazione pubblicato i transistor inseriti sono delle alternative a quelli dell schema originale. Ho un problema nel reperire il 2SC3668, nello schema della simulazione sono inseriti MJE15032 sono dei validi sostituti. Confrontando alcuni datasheet ho trovato questo: 2SD2177, ha solo Emettitore e Base invertiti. Grazie
Tommaso
Tommaso
Please post in English.
Good morning, I would like to reconnect to this discussion by asking Mooly if in the published simulation scheme the transistors inserted are alternatives to those in the original scheme. I have a problem finding the 2SC3668, in the simulation scheme MJE15032 are inserted, they are valid substitutes. Comparing some datasheets I found this: 2SD2177, it only has inverted Emitter and Base. Thanks
Tommaso
I think you will find most suitably rated devices will work just fine in a circuit like this. Also make sure you get guaranteed parts from an authorised distributer rather than risking parts from other places (like eBay).
The 2SD2177 from what I can see might be marginal (I wouldn't use it) because it appears to have a 50 volt vce rating. That is only the same as the Sugden's main supply voltage... so its not really enough. The transistor sees the peak voltage that the output is delivering across C and E.
I would stick to common well known generic types for this amp.
The 2SD2177 from what I can see might be marginal (I wouldn't use it) because it appears to have a 50 volt vce rating. That is only the same as the Sugden's main supply voltage... so its not really enough. The transistor sees the peak voltage that the output is delivering across C and E.
I would stick to common well known generic types for this amp.
Grazie della tempestiva risposta, ho cercato il 2SD2177 in quanto anche il 2SC3668 ha una tensione nominale di 50V, infatti il suo Collettore si trova proprio collegato alla tensione dei 49V di alimentazione. Farò altra ricerca. Grazie
Tommaso
Tommaso
Please post in English. Last warning. Further non English posts will be deleted
Translation:
Thanks for the prompt reply, I looked for the 2SD2177 because the 2SC3668 also has a nominal voltage of 50V, in fact its Collector is connected to the 49V supply voltage. I will do more research.
Thanks Tommaso
This is what happens. This shows the supply in Red and the emitter volts of the upper NPN driver in Blue. At high output it sees nearly the full rail voltage across it. You need to have a safety margin of at least 30% in the Vce rating to be safe from the effects of secondary breakdown.
The lower driver (and output) also see the same happen.
Thanks, the reasoning is correct, the doubt remains because they mounted a transistor at the voltage limit. I find others. Thanks
Tommaso
Tommaso