Swr sm-900 dc offset

[J M Fahey]

What are the causes of DC offset? This being a 2 channel amp, one side settles back to 0.0 v and the other side goes to about 65v DC offset.

I think the DC on the output is taken care of.

But you do not tell us what you did or how you solved it

As I run up the variac though this amp is drawing quite a bit of current.
Thus far I have stopped short of turning it all the way up but at about 70% it's drawing
about an amp through the variac.

Disregard the part above where the current is high above. It is under control.

But you do not tell us what you did or how you solved it

Not sure this repair thread will be useful to others if you omit this essential information.

[nevetslab]

It sounds like you're at the point where you can power up the amp without ramping it up via Variac....just set the Variac for 110-120VAC, so you still have the Variac's Ammeter to show you current draw from the amp. I'd power it up without loads connected, then, once the amp is stable and the DC offset has become a stable value, THEN connect your load, and see if that offset and/or AC mains current changes.

On biasing, since SWR doesn't include their method, I'd set bias for a reading of 400mV across the driver transistor's emitter resistors....R19 on the top half for Q7 (100 ohm) and R21 for Q8 on the bottom half. I think the voltage drop on both R21 and R22 will be the same. What you're doing is setting the Vbe level of the output stage just below turn-on...setting each half for 400mV. With this level, you'd have 2V across the bias Xstr (0.6V + 0.4V + 0.4V + 0.6V).

Have you confirmed you can adjust the bias pot so you can both turn Q4 on hard (near 0V between C-E of Q4) to greater than 2V across Q4. What you find for the bias range on your 'working' module, is what you're looking for on this module you're trying to restore.

Now, as to the offset. it's possible the LF feedback cap is involved...C4. There are ways to trim the input stage to force the DC output level to 0V, such as selecting a high value resistor (to be determined) from Q14 Collector/R7/R8 junction to the base of Q2. This can be tricky. I normally have a resistor decade box with short leads in place, starting with 1.0000M and trim downwards in value and watch the DC output level closely, rather than start soldering in resistor values and guess. I'm still suspicious of C4. I also don't quite understand the role of C7 and C8 in this circuit. Perhaps one of the other members following this thread can shed some light on their role. If those caps are leaky, they may influence the DC level of the upper driver Q7 base.

[bigdrums925]

To solve the original problem of 65v on the output I believe I found a bad output trans at that point and changed it.
Next changed the B-C resistor on Q4, then Q4 itself. Then changed the B-C resistor again to match the working side.


I had rebuilt this amp about a year ago after someone else had worked on it and was being thrown out. I pieced it back together.
I had to buy a bunch of output trans and some that I got overseas were counterfeit. So pretty much all the output trans were suspect.
For whatever reason the tests on each of these about a year ago was not conclusive so I just bagged the ones I wasn't sure about.
The ones that were counterfeit I marked clearly to help prevent future problems like this.
So if I remember correctly changing one of the outputs fixed the 65v present on the output.
But I had this unusually high current draw shown on the variac once I turned it up past half way.
Normal before 50% so I was pretty sure it wasn't a shorted device at that point.
But rose quickly after the 50% point. A short would show itself before 1/2 way. Had to be something after some transistor was turning on.
With the clear guidance of nevetslab, we zeroed in on Q4 and the associated resistors
I changed Q4 B-C resistor from 2.2K to the 2.7k called out on the schematic and also the Q4 trans. Still too much draw.
The schematic in this case was misleading me. The resistor values shown were not the values present on either side of the amp.
At that point I abandoned using the schematic and just put the same value resistor as the working side. (1.5k instead of the 2.7k called out on the schematic)
That's when the current meter started acting normally, that is a slight jump in current draw each time I bumped the variac up, then the current would settle back down quickly.
As long as it was not settling back, I knew something was wrong and not to push the current to high.
At that point I could see the amp meter rise to about an amp at about 75% variac then suddenly drop as Enzo suggested.
In other words I knew there was a threshold I had to get over, but when the meter rose too quickly I needed to back off.
I now see between .5 and .6 amp draw at idle. With a load connected the rise is more dramatic, but settles after that.

And thanks to everyone, especially nevetslab for sticking with this to completion

[g1]

This is a little late, but it may help you or others in the future. Attached is an older power amp version that shows the 1K5 rather than the 2K7 in the Q4 circuit.