Ampeg BT15

[gbono]

I received this amp in a box - it has the original "fliptop" speaker enclosure with the Altec 418 option - even has the detachable dolly. I replaced the filter caps, output cap and many of the bypass capacitors and replaced the output transistors. The amp worked okay but I noticed that it was not putting out the rated 65W more like 50W (tested with 1KHz sine wave - just before clipping - 4 ohms resistive load) and I had a few bad solder joints that need to be reflowed.

After reflowing the solder joints I have noticed a pretty strong hum that wasn't presnt before  :grr (need to check if its 60 or 120Hz) and measuring the voltage at the filter caps I'm getting 56VDC.

I'll scope the supply to see what kind of ripple is present but I'm a bit baffled why I don't see at least 78VDC at the caps - the secondary measures 58V RMS. I will pull one leg of the 6.8K resistor to see if something is loading down the supply - no bright light when using a light bulb limiter though. Any other ideas - suggestions?

Also what is the function of the diode (D1 connected to Q10 emitter and 300 ohm power resistors - crossover distortion compensation?

[Roly]

and replaced the output transistors

Why?  Were they blown?

After reflowing the solder joints I have noticed a pretty strong hum that wasn't presnt before  :grr

If it ain't busted, don't fix it.

This is a paradoxical problem with "routine" or "preventative" maintenance - there is always a finite chance you are going to introduce a new fault into a previously working assembly.  Three little O-rings (interestingly part of a safety monitoring system).

As well as "do no harm" we need to remember "if it's working, leave it alone".   8|

what is the function of the diode (D1 connected to Q10 emitter and 300 ohm power resistors

That is a bloody good question.   :dbtu:

{BTW Q10 is misdrawn, it should be a PNP with the Emitter up, pointing downward to the Base.  It looks like very primitive CAD so they may not have had a PNP symbol.}

What you've got is (roughly) arrangement A;


Points to note;

- the upper an lower cells are not voltage gain matched.  The upper cell is a Darlington or (super) emitter-follower with unity voltage gain, and the lower cell is a buffered grounded-emitter amplifier with voltage gain.

- The driver emitters are returned to the next stage 1 ohm "emitter" resistors to provide local DC and AC feedback.

(another local negative feedback resistor of about 47 ohms is sometimes inserted in the emitter of PNP driver Q3 to try to better balance the cell voltage gains.)

What Ampeg are doing is taking the Emitter of their PNP driver to a point that is 0.7V more positive than the output half rail (where it would normally be going).

Why?

The back-to-back electros tell us that they only wanted a DC offset, not AC.

The effect will be that the driver can still deliver Base injection to Q13 and Q14 even when they are saturated on negative peaks, otherwise if Q10's Emitter went directly to the OP Collectors the output swing available would be slightly less.

{The Collector voltage in a saturated transistor, Vce(sat), can go lower than the Base voltage, Vbe.  In a specific switching transistor you can have Vce=0.2V while Vbe=0.7V.  The venerable old 2N3055 is no switching transistor, but at these current its Vce(sat) is about 1.0V.  This would be an impossible situation for Q10 if its Emitter was connected back to the output Collectors, so they seem to have biased it up a bit to give it some drive headroom on signal peaks to the rails.}

Re: low voltage, what is your mains voltage and what does the rating plate expect?

[J M Fahey]

I have noticed a pretty strong hum that wasn't presnt before  :grr (need to check if its 60 or 120Hz) and measuring the voltage at the filter caps I'm getting 56VDC.

I'll scope the supply to see what kind of ripple is present but I'm a bit baffled why I don't see at least 78VDC at the caps - the secondary measures 58V RMS.

Symptoms match having lost the main filter cap, or its connections, either to +V or to ground (same end result).

What you measure matches average value of unfiltered rectified 50/60Hz secondary sinewave.

[teemuk]

"Factory" schematic:

http://wiki.talkbass.com/images/d/d5/%2769_pcb_BT-15_Schem.jpg

[Roly]

teemuk, JMF, et al, Any other thoughts on that diode?  I'm not sure I understand what they were thinking there. 

[J M Fahey]

It worried me from way back then, I started making Guitar amps when I bought Jack Darr's "Repair your electric guitar amplifier" book in 1969 and proceeded to build many of the schematics at the end section ...including this one.

I was lured by the caption (still remember it by heart) : "this amplifier produces 65W RMS into 8 ohms and 100W RMS into 4 ohms, be sure your speakers can handle that much power" which of course was true, regular 12"guitar speakers were 20W RMS, period, and a few 15"were 30W RMS.
The mighty Altecs which were an upgrade option, were 60W RMS and the JBL not far from that.

That era was the very beginning of Transistor amplification and, of course, Ampeg always did things "the right way" .

Quasi complementary stages are not symmetrical, and one of the problems is that the Darlington upper stage Q9/11/12 is two diode drops above the output/speaker rail while the quasi complementary lower stage Q10/13/14 is just one diode drop below.

Nowadays most nobody cares, Q10 emitter is straight connected to Q13/14 collectors and they call it a day; DIY-audio-Forum type guys which are more anal add a diode and a parallel resistor between Q10 emitter and those collectors so now you have an extra diode drop in the lower stage; Ampeg Engineers went a step beyond, added the diode (bypassing it with caps) and guaranteed that it would always be forward biased, no matter what , by passing a lot of current through it, fed from 300 ohm 7W resistors which run real hot.

I guess it also guarantees that Q10 emitter is always 0.7V above Q13/14 collectors, so it acts like some kind of floating supply to ensure they get better saturated ... since you may gain extra 0.7V peak on the lower swing no one else does it.

I also suspect the amp did not clip symmetrically, probably the negative swing clipped **a little** earlier than the top one (we see that very often, no big deal) and somebody was determined to solve it for good.

[gbono]

I had the light bulb limiter connected to the amp so retesting I get 78V DC at the last filter capacitor. The only issue now is that the amp has about 40mV of AC on the output and it's quite loud hum. I either didn't hear this before (though I had the amp running for several hours) or something is "different" after resoldering the cold joints that were causing intermittent pops/noise when the amp heated up.

There are two new filter capacitors (2700uF) on the output of the bridge - vaguely remember measuring 100mV ripple at the filter caps.

The PCBs are not assembled into the chassis and cover but I don't think that is the problem.

I will retest the ouput power and see what else I can do to eliminate the hum  :grr

Did notice that several of the pots (both treble and one channel 2 volume) would cause greater hum when I touched them.

[Loudthud]

There was a similar biasing scheme in the Dynaco Stereo 120 with the 300 Ohm resistors forcing the PNP driver to conduct. That amp didn't have a Vbe multiplyer, the power amp had 6 transistors total. The circuit never became very popular as technology was changing fast at the time and better PNP transistors became available.

Notice that Ampeg used the wrong symbol for input JFETs, they are actually P-channel Source followers.

Check the diodes in the rectifier bridge. One may have died with the new filter caps.

[gbono]

Yes I did replace the recifier diodes since I was initally getting an odd waveform out of the bridge - were this amps noisey? Not sure what else to try.

[Loudthud]

I had an amp very similar to this, I think it was a BT25 head. It didn't have the electroluminescent panel, but I remember the 300 Ohm resistors. Got rid of it in '83. It didn't have excessive hum. The Dynaco amps had a regulated power supply. They hum when the regulator pass transistor shorts out.

Peter Baxandall had a diode in some of his quasi-complimentary designs, but it was just in series with the PNP emitter and had a resistor across it. He claimed it improved thermal tracking. Some people started calling it the "Baxandall Diode".

[J M Fahey]

Did notice that several of the pots (both treble and one channel 2 volume) would cause greater hum when I touched them.

Well, they shouldn't, because metallic pot cases get grounded when bolted to front panel, so I guess you lost a ground somewhere.

@loudthud:  thanks for the Baxandall diode info

[gbono]

Baxandall compensation is shown on page 61 of Teemuk's book on SS MI design.......