Here is the link to the schematic - http://www.drtube.com/schematics/marshall/2099u.gif (http://www.drtube.com/schematics/marshall/2099u.gif)
The story starts with a simple power cord change since the original cord needed to be replaced. During the R&R of the power cord, a (base) lead broke off the TO-3 socket on the heatsink, and was soldered back - this was the PNP output darlington BJT MJ2501. When the amp was powered up with a variac I noticed that as the line voltage was increased excessive current would start to flow. The amp was working without issue before the mains cable was changed. GGrrrrrrrr.
If the PNP output transistor is removed from the circuit the mains current does not increase when the line voltage is increased. When a new MJ2501 is put into the circuit the over current issue comes back. I even removed the socket from the heatsink and when the MJ2501 is off the heatsink BUT the attachment screw (connects TO-3 case to collector tab on socket) over current is observed with increasing mains voltage. I even tried a new socket but still have the issue. I went back and looked at the driver transistors in the circuit - BC184 and MPSA06 - even though they tested out okay with a DMM I replaced them with no change to over current situation. So to recap ...new output transistors, socket and correct wiring but still over current when PNP (MJ2501) is in circuit - no issue when it is removed or no connection to the collector of the MJ2501.
Finally, I have looked at the mains cable that was replaced and can find no issue with the cable. I noticed on the schematic that the "neutral" side of the the mains is switched and fused (black colored wires) while the brown (hot) side is not in series with the fuse. Would swapping out the mains cable leads, so the hot side of the mains is also in series with the fuse, cause an issue with the 120/240 transformer??
I know I am missing something very subtle and simple but what is it?
Bias.
Looks like when you disconnected that base and powered the amp some surge current fried the bias transistor or diodes.
Then even good power and driver transistors will pass overcurrent.
Think good new tubes ... without bias.
Please search and post schematic.
Quote from: gbonoThe amp was working without issue before the mains cable was changed. GGrrrrrrrr.
Take heart, this sort of thing happens to the best of us - nice simple little fix suddenly turns into a saga. It comes under the heading of "The sheer cussedness of inanimate objects" a.k.a. Murphy's Law.
Quote from: gbonoWould swapping out the mains cable leads, so the hot side of the mains is also in series with the fuse, cause an issue with the 120/240 transformer??
I think that is
highly unlikely (unless you goofed the lead replacement and have somehow managed to connect 220V to the 110V windings, but I'll assume from your post that you are savvy enough not to have done something like that).
Quote from: gbonoover current when PNP (MJ2501) is in circuit - no issue when it is removed or no connection to the collector of the MJ2501.
Which leaves us with a few options;
- the PNP Darlington is boofed
- there is a physical short, say between one of the mounting screws and the heatsink. Pay particular attention to where the screw passes through the heatsink. I've been known to put a bit of sleeving or (loose) heatshrink around the shank of these screws to be certain they can't contact the heatsink.
- it is okay but being switched on by base current via a preceding fault.
Quote from: gbonoI went back and looked at the driver transistors in the circuit - BC184 and MPSA06
As you seem to realise, unwanted PNP base current could be caused by the MPSA06 "south-bound driver" being shorted, or by the BC184 active bias transistor being open. Since you have replaced both of these (and I will again assume that you are savvy enough to have got the replacements in the right way around in the right places, but it would be wise to double check) then that leaves us with something odd happening in the active bias circuit, or the south-bound driver surrounds.
The active bias can be taken out of consideration for the moment by simply tacking a link across the C-E of the BC184. This will underbias the output pair and if the cause is the active bias the amp should come up without excessive current ('tho it may sound rough due to lack of bias).
The south-bound drive is actually a constant current sink, and would draw excessive PNP base current if the two diodes from its base to the -ve rail were open for some reason, so you need to check that these diodes are okay and haven't become de-soldered as an indirect result of you replacing the MPSA06 transistor. Similarly an accidental solder bridge shorting the 100 ohm in its emitter would have a similar effect.
HTH
I'm looking at the schematic and wonder if the bias on the BC184 transistor could be turned off by accidental "adjustment" of the 1K ohm trimmer resistor in the base/emitter?? Need to check this out when I'm back home...
BTW is this trimmer used to set the symmetry of the output signal? I also assume BC212/182 act as limiters? THX
Not
symmetry but output stage
idle bias current to minimise crossover distortion. This circuit should be inherently self-centering (as most are).
Yes, that might give you the same results, which is why I suggest linking the active bias circuit out. If that alone cures your overcurrent, and you are
certain that the transistor is okay, then it may be that the bias adjustment has been "fiddled" at some stage (but always better not to alter unless you are
certain it has been fiddled).
The
minimum bias current is when the 1k adjustment pot is at it
maximum resistance; as it is reduced the bias current will rise. A typical value is about 30mA through the output pair. This may be measured by the voltage drop across one of the 0.33 ohm output emitter resistors;
E=IR
0.03*0.33 = 0.01V or 10mV
But let's see if linking it out makes any difference first.
Quote from: gbonoI also assume BC212/182 act as limiters?
That is correct. They sense the current in the output across the 0.33 ohm resistors via the 1500 ohm resistors, and if too high start to conduct and clamp the drive into the bases. The idea is that if the output should get shorted these hold the output current in check, hopefully saving the output transistors while the fuse blows (then some twit typically sticks in a 25 amp headlight fuse and all bets are off. :trouble "There, that'll
make it work" Er no, it will turn a problem into a catastrophe by blowing something else up, often the power supply).
Not-so-small-detail: IF this is the amp with an output autotransformer, any output offset betond, say, 100mV will cause excessive currents through the transformer.
It presents 4 ohms impedance to the amp ... at audio frequencies ... but some 0.5 ohms (or less) at DC.
So amp with offset self shorts.
If so, disconnect the OT until the amp is repaired.
Here's what I've done....
Replaced BC184 with 2N4401 (must flip 2N# 180 degrees to have collector and emitter in the correct orientation) - I even applied a short between E and C and still I have high current in the output :grr
I tested the BC184 on a curve tracer and it is in perfect condition - I didn't think it was damaged to begin with and hate "shot-gunning" components........more self directed :grr
I also maxed out the trimmer, in both directions, with no change in over current on the output.
I will remove the auto transformer on the output - why do they need one in this design??
This amp is playing silly buggers with me.... :grr
(http://i284.photobucket.com/albums/ll33/bonosurf/2099u_zps34f48a2f.gif)
A solid state amp lets the load determine the output power. AN amp that puts 100 watts into 4 ohms will put 50 watts into 8 ohms and 25 watts into 16 ohms. Yes, I know in real life it is not quite that neat, but close enough. By using the output transformer, they can put full power into any impedance you select.
And please disconnect the transformer during tests.
You have excess current flowing? And none of the outputs are shorted, and the drivers and bias transistors seem OK too? Well remember it is a whole circuit. You pulled the negative output xstr and it killed the curent. ANy chance you tried the same trick with the positive output? In other words is it only the negative one that kills the current or will either do it?
You shorted the bias xstr E-C and no help. I'd short it again and leave it until we find the problem. SHorting that puts the two driver bases together, which is the coldest bias there is. The amp ought to work with some crossover distortion thrown in. But yours still draws. High current draw either means the two sides are on at the same time shunting the V+ and V-, or one of them is shunting to ground. I will believe your power supply is OK if it doesn't draw current at all times.
But does it really put the bases together? With that short across the bias resistor in place, power off, measure resistance from the base of the 3001 to the base of the 2501. You ought to get zero ohms. Or without the short, we ought to see 11k-12k, the resistance around that transistor. Measure right at the output bases, not anywhere else. We want to see if all the connections and copper traces are intact. If you do not get that continuity, then you have an open in the bias. So do not assume shorting the transistor is the same as shorting the output bases, even though it is on paper.
But do make sure power is right. You can remove the 2501 and the amp will sit there not drwing current? GReat, Is proper V- on the wire to its collector? What voltage is on that bias transistor - any leg, it doesn;t matter which. Normally it should be close to zero, within a volt or two.
Quote from: Enzo on August 26, 2014, 06:33:04 PM
And please disconnect the transformer during tests.
With the autotransformer
disconnected and both output transistors connected AND the CE short removed from BC184 there is NO over current present :o
QuoteYou have excess current flowing? And none of the outputs are shorted, and the drivers and bias transistors seem OK too? Well remember it is a whole circuit. You pulled the negative output xstr and it killed the curent. ANy chance you tried the same trick with the positive output? In other words is it only the negative one that kills the current or will either do it?
With auto transformer connected and only the NPN output transistor (MJ3001) connected I have NO over current condition.
QuoteYou shorted the bias xstr E-C and no help. I'd short it again and leave it until we find the problem. SHorting that puts the two driver bases together, which is the coldest bias there is. The amp ought to work with some crossover distortion thrown in. But yours still draws. High current draw either means the two sides are on at the same time shunting the V+ and V-, or one of them is shunting to ground. I will believe your power supply is OK if it doesn't draw current at all times.
But does it really put the bases together? With that short across the bias resistor in place, power off, measure resistance from the base of the 3001 to the base of the 2501. You ought to get zero ohms. Or without the short, we ought to see 11k-12k, the resistance around that transistor. Measure right at the output bases, not anywhere else. We want to see if all the connections and copper traces are intact. If you do not get that continuity, then you have an open in the bias. So do not assume shorting the transistor is the same as shorting the output bases, even though it is on paper.
With the BC184 shorted EC the two basses of the MJ2501/3001 have .1 ohm connecting them and with no CE short about 3K.
QuoteBut do make sure power is right. You can remove the 2501 and the amp will sit there not drwing current? GReat, Is proper V- on the wire to its collector? What voltage is on that bias transistor - any leg, it doesn;t matter which. Normally it should be close to zero, within a volt or two.
The MJ2501 and 3001 have +/- 40V at their respective collectors. Will measure the voltage around the BC184 and MPSA06 but the issue has to be DC offset like JMF mentioned??
With everything connected but the output transformer not connected, it sits there working? OK, then how much DC is on the output?
That is why we want the transformer gone, it is a dead short on the output as far as DC is concerned. That is why I wondered what voltage was all over the bias transistor and thus the output bases.
:dbtu:
Conditions: autotransformer disconnected - all other components including output transistors as per the schematic.
BC184
Vc = .85V
Ve= -1.1
Vc= -.43
MPSA06
Vc= -1.1V
Ve= -39.8
Vb= -39.2
MPSA56
Vc= .86V
Ve= 41.1
Vc= 40.5
Offset at output (measured at output inductor - autotransformer disconnected) = -10mV
Yeah, well, um, that all looks pretty smooth.
So, (Q) by disconnecting the output autotranny, and with all the output transistors etc otherwise fully connected, we now have an amp that is DC stable and not drawing excessive current???
(or am I missing something?)
At this point, I'd be suspecting the output transformer and connecting a speaker directly to the amp output to find out.
Any suggestions on how to test this type of transformer?
Quote from: Roly on August 29, 2014, 09:44:29 AM
Yeah, well, um, that all looks pretty smooth.
So, (Q) by disconnecting the output autotranny, and with all the output transistors etc otherwise fully connected, we now have an amp that is DC stable and not drawing excessive current???
(or am I missing something?)
Note in your marked up schematic the BC184 should not have a jumper across CE.............
By verifying the rest of the amp works without it. That would demonstrate that the transformer and only the transformer causes the amp to malfunction.
To test the part on its own, well, it is a transformer. Put some AC across two points on a winding and verify that appropriate AC results on the other taps. It appears to be an autotransformer, so there is only one large winding.
You can also go to RG Keen's Geofex web site and look up the transformer tester there and make one. It is a simple gadget.
Measure the resistance, between terminations (lowest ohms), and to the frame (very high ohms).
Shorted turns test;
(http://www.ozvalveamps.org/repairs/xform_test.gif)
Visit: http://www.ozvalveamps.org/repairs/trannytest.htm (http://www.ozvalveamps.org/repairs/trannytest.htm)
-
The jumper was to show how to disable the bias network, voltages are without jumper or autotranny (but it has been overtaken by the autotranny discovery).
Here is a picture of the transformer in situ. The gray wire was connected to the 4 ohm tap. If you look at the schematic one side of the transformer is connected to ground. So how does this work? Every "tap" is going to and does have continuity to ground. How should the output transistors connect to this autotransformer or the output impedance switch?
(http://i284.photobucket.com/albums/ll33/bonosurf/DSC01623_zps998c76d9.jpg)
see attach
This is used as a step-up tranny to match 8 and 16 ohm loads to the 4 ohm nominal output impedance, which the amp drives all the time.
This could be a more conventional double-wound type, but for economy, and because primary-secondary isolation isn't required, they can combine the primary with the secondary in an "auto-transformer".
As you have noticed, this presents an impedance (AC resistance) of 4 ohms to the amp output, but it also presents the winding resistance, which will only be a few tenths of an ohm.
In the event of any DC offset in the amp the output current will rise dramatically. I don't consider this a robust design without an output coupling cap to keep any DC off the autotranny in-built DC short (or at least put the fuse on the amp side).
Given how critical this offset might be, I only see trims for bias and AC gain. ??? (which is currently where your problem seems to be)
I may have observed of the 741 op-amps in the front that you can do a hell of a lot better for a few dollars these days. These are instrumentation quality, they were never intended or designed for amplifying audio. Today a buck will buy you a drop-in op-amp that is almost Hi-Fi perfect. The line starts at TL071 and gets better from there.
OK, with that wire off, the transformer is disconnected. Now does the amp power up OK? And if so, connect a speaker between the gray wire and ground, sound come out?
Quote from: Enzo on September 01, 2014, 07:23:12 PM
OK, with that wire off, the transformer is disconnected. Now does the amp power up OK? And if so, connect a speaker between the gray wire and ground, sound come out?
Yes the amp works fine with the autoxfmr disconnected. So with 10mV offset the issue has to be a shorted turn to chassis ground - no?
Doesn't have to be to ground, even just winding to winding. That will make it stop working like a transformer and become a hunk of wire.
10mv offset is nothing by itself, but short the amp output, and the amp circuit will try to correct it into the load.
A "shorted turn" is an AC condition and has no effect on DC operation; it would not cause excessive standing current draw, only under AC signal drive.
A "short to frame (ground)" is both a DC and AC condition, but it is also very rare and unlikely in this situation. Neon-flash test the autotranny by all means, but I don't think that is where the basic problem lies. At the moment it looks like it's voltage offset related, but there is a deeper issue.
{maybe it's just a phase I'm going through, but} I think this is a design flaw, that the guys at Marshal didn't think through the implications of a much greater sensitivity to any output offset voltage. I can only guess what the DC resistance of the bottom winding is, but I expect it wil be only about an ohm or two, maybe less, so we have an output current to offset of 1mA/mV, 1 amp per volt if it's one ohm, more if it's less. Don't like.}
It's a bit drastic but I'd be considering modding the amp by inserting a suitable large electro between the amp output and the auto tranny, and moving the fuse to the amp output side too (that isn't thought through either).
Where is this (minimal) 10mV offset coming from, anyway. (I'm sure my own Twin-50's have about three times this much offset).
In defense of the design, we do not see a steady stream of all the 2099s coming in for service. There are other SS amps with autoformer OTs. And this is not much diferent from a ordinary SS amp driving a CV transformer.
Quote from: gbono on August 31, 2014, 09:42:03 PM
Here is a picture of the transformer in situ. The gray wire was connected to the 4 ohm tap. If you look at the schematic one side of the transformer is connected to ground. So how does this work? Every "tap" is going to and does have continuity to ground. How should the output transistors connect to this autotransformer or the output impedance switch?
1) please post transformer physical size.
Just the "roof" rectangle is enough.
Millimeters preferred (as in 32mm x 94 mm or wharever) but worst case inches are acceptable ;)
2) since it´s an autoformer, it´s just one long tapped winding.
One end goes to ground/0 and the others provide different voltages but all have DC continuity.
I said DC because they carry different AC voltages.
The amp direct output is 100W into 4 ohms, so it goes to the 4 ohms inpit/output (transformers are perfectly reversible) and proper voltages appear at the other taps.
As in: the amp produces clean 20V RMS.
If you do the math, 20^2/4=100W RMS
The 8 ohms tap rises that to 28.3V RMS and the 16 ohms one to 40V RMS .
Once you "see" it that way, it´s logic and simple.
EDIT:
a) I don´t think the transformer is bad.
Very low voltages around it, the transistors have built in reverse protection diodes, etc.
b) I suspect you still use the lamp bulb limoter.
P suggest:
turn it on with no load applied
measure you have 10mV or some other real low offset
connect the grey wire to 4 ohms tap and nothing else.
I guess nothing bad will happen, at most offset will lower a bit. or nothing
if everything, connect the speaker to the proper rap
if everything fine, play some music at low volume (say 1 W)
if everything fine, plug the amp straight in the wall and repeat tests, I guess everything will be fine.
blast away :)
The core is 100mm x 80mm and 30mm "thick". The windings are about 66mm in diameter.............
Okay this is a bit embarrassing ...... when I hooked up the light bulb limiter I also monitored the output DC offset. The DC offset is quite large until the amp biases up and then quickly drops to 10mV. Amp works fine :dbtu:
This is with or without the transformer?
Quote from: gbono on September 08, 2014, 11:23:04 PM
The core is 100mm x 80mm and 30mm "thick". The windings are about 66mm in diameter.............
Thanks :)
Yes with the transformer connected.
:tu: