Welcome to Solid State Guitar Amp Forum | DIY Guitar Amplifiers. Please login or sign up.

March 29, 2024, 12:44:58 AM

Login with username, password and session length

Recent Posts

 

hartke3500 blown transistor

Started by js1970, October 12, 2014, 08:00:21 PM

Previous topic - Next topic

js1970

  I went to increase the idle as suggested. When I turned the dial, there was no change in the value. Its currently set about halfway, as from the factory, and I turned in both directions with no change as the result. Is there something I'm missing?  Should I run the amp to get good and warm before adjusting? I can't help but think there is still a component that is faulty. As I'm posting this, I'm thinking I will meter Q308 again. This time from base to emitter, rather than just collector. To be sure. Could the trimmer be bad,and showing high resistance? That is, if all other values are within reason.

Roly

Quote from: js1970When I turned the dial, there was no change in the value.

This is not right.  {I hope you are turning the correct trim pot, SVR302, and not the Offset Adjustment SVR301}

There might be a small difference between cold and warm but the action of the bias setting should be obvious at all temperatures.  (yes, bias should normally be set with the amp at warm idle, but seeing that you can't set it at all it's a rather moot point at the moment; but that's how you should finally tweek it when you get it working correctly.)


Quote from: js1970Is there something I'm missing?

Let us instead say that there is still some problem yet to be found.   8|

Quote from: js1970I'm thinking I will meter Q308 again. This time from base to emitter, rather than just collector.

Actually the voltage we are interested in the Q308 bias transistor Collector to Emitter voltage, the voltage across the transistor, not to ground.  You will see on the circuit that this sets the voltage between the driver Bases of Q309 and Q310 and should be variable from very roughly 1 volt up to a maximum of about 2.5/3.0 volts.


{This particular bias arrangement has a serious weakness; if the wiper of the trim pot should lose contact with its track (a not uncommon thing to happen) the bias transistor Q308 will lose its Base current injection and switch off.  This will cause the idle current setting to go to maximum current - not a healthy situation.  For this reason whenever you are messing with a s.s. bias adjustment you should have the amp on a limiting lamp to avoid destroying the output transistors.  There are some smarter arrangements where the wiper going open causes the current to go to minimum, a much safer arrangement.}


Try clipping your multimeter across Q308, Collector to Emitter (or the Base of Q308 to the Base of Q310 driver transistors, which is the same thing).  We are looking for around 2-and-a-bit volts, and this should change when trimmer SV302 is varied.  As it doesn't seem to be having any effect on the idle current it would seem that it isn't making the 2.4-odd volts required to start turning on the output stage.

The question now is, is it the bias circuit that isn't developing enough voltage, or is it something following there (e.g. a driver transistor) that is failing to transmit the change to the output transistors?

If the voltage across Q308 varies in sympathy with the pot then it's something downstream.  But if the voltage across Q308 doesn't change with the pot then it's something in the bias circuit, Q308 and surrounds.

Please check and report.
If you say theory and practice don't agree you haven't applied enough theory.

js1970

   I metered the voltage btw collecter and emitter of Q308 to be 1.9V. I metered btw Q309 and Q310 to be 1.8V.Then, I was going to meter btw bases of Q308 and Q310 to compare. When attempting this, I shorted across the legs of Q310 and popped the fuse. :-\ ( boy it didnt take much!) I didn't get the chance to try the trim pot while metering Q308. Hopefully, I didn't cause any damage when this happened. I'll try to acquire a fuse tomorrow, from the shack,and get back to the business at hand.

Roly

Oh dear.  A bit of a setback.

{In a previous life I serviced Variable Speed drives up to 100HP, rows of Silicon Controlled Rectifiers as big as beer cans directly connected to 415V three phase mains, and supposedly "protected" by expensive ultra-fast "solid-state fuses" specifically intended to blow before the SCR it was "protecting".  The trouble was that it was neck-and-neck which would melt first, and typically the SCR would short and then blow the fuse (so any old fuse would have done that job).  If you found a blown fuse it was more than likely that the "protected" SCR had shorted and had to be replace as well.

Generally I try to be hopeful, and if you had the limiting lamp in series you might get lucky, but you will need to take a step backward and treat the output stage as if it has just had a blow-up and bring it back up gently (lowest wattage limiting lamp) until you are satisfied all the transistors actually survived.

Try not to be too hard on yourself, this sort of accident happens to the best of us and is very annoying, but we just have to dust ourselves off and recover the situation.  This is why I said "clip the meter probes" - if you make your test connections using grip probes or clip leads before you power up for the test the chances of a slip are much less, but even then clips have been known to suddenly let go and float around.}

Anyway we've got something, 1.8-1.9V doesn't seem enough so this is suggesting that the problem is in the bias circuit around Q308.

Once you replace the fuse bring the amp up with the lowest wattage limiting lamp and check that the half-rail is very close to ground, and if it is you have got lucky, but if not you will have to reapply your cold resistance tests to the driver and output transistors until you find the one that has failed.  Fingers crossed.
If you say theory and practice don't agree you haven't applied enough theory.

js1970

    I tested the driver and output transistors out of circuit as was described on page one of this thread. I found one of the output transistors (Q313 to be exact) to be faulty. I was unable to get an OL display on the meter in any combo of the leads. I have an exact replacement on order. Should I get a matched pair, or will just replacing the offending transistor be sufficient?

Roly

Quote from: js1970Should I get a matched pair, or will just replacing the offending transistor be sufficient?


"Matched pair"?  Man that's as flash as a rat with a gold tooth.  No, a simple replacement should do just fine.

I hope you are quite certain about the health of the other transistors, in particular the associated driver transistor.
If you say theory and practice don't agree you haven't applied enough theory.

js1970

   I'm fairly certain the other transistors are fine. I tested them out of circuit using the diode mode on my DMM. The respective output transistors were consistent with each other and the drivers checked out too. The funny thing is, the driver transistor I shorted across (Q310) is opposite the output transistor I found faulty (Q313). At least, I thought it odd. Is this cause for concern?
   

Roly

Well that is a little odd, but it's hard to say, even knowing exactly what got shorted to what.

What we can say is that like any blow-up repair it's important to "hasten slowly", use your limiting lamp, carefully check supply and half rail voltages, be suspicious, don't assume, when in doubt stop and have a good think.  Be alert to the possibility of another fault still lurking undetected.
If you say theory and practice don't agree you haven't applied enough theory.

js1970

 I've got the new output transistor in, and everything powered up. When all said and done, all voltages are back to where they were previously. Except the halfrail voltage. That went up to 69.4mV.
  Satisfied, I checked the voltage across collector/emitter to find the same 1.8V. There was no change in value when adjusting the trimmer. While I had some downtime, I discovered a diode I believe faulty. It is the D306 diode off of the scheme,and it had a drop of .091V one way and .109V another. Pretty sure that's a no good. All other similar diodes had a drop of .581V one direction only. These were D303,304 and 305. It doesn't seem as though this would effect my cold bias problem does it? Seems more in line with my little blow out earlier.
 

Roly

Quote from: js1970D306 diode off of the scheme,and it had a drop of .091V one way and .109V another. Pretty sure that's a no good.

I'm pretty sure you are right, that ain't no diode no more.

It's one of the gating diodes into the protection circuit, so I would replace it and check the other gating diode D305, and protection transistors Q312 and Q311.

Yes this may have been because of your slip, still, fingers crossed eh?

Once SVR302 starts to vary the voltage between the bases of Q309 and Q310 you are really getting somewhere. (but keep that limiting lamp in circuit until you get the bias under control - don't want to go blowing any more transistors).
If you say theory and practice don't agree you haven't applied enough theory.

js1970

  Hey everyone.I finally got my hands on the transistors I needed. After I found I had a faulty diode(D306), I checked transistors Q311,312 as Roly had suggested. I  found Q312 to be bad.
  I swapped the offending components and did some metering. Here's what I found
     TP1 to g~ 16.1 mV
     TP2 to g~  -1.5mV
     TP1 to TP2~  19.2 mV
     Halfrail ~ 9.8 mV
   V btw Q309 & 310 ~ 2.2 V   this varies with trimmer. This voltage trended down to     2.15V after a bit. That is to say, with thermistor 302 btw my fingers, and feeling the transistors for warmth. All ran well with Q314 & 319 getting warmest, almost hot.
   This was  all while plugged into my limiter lamp.
   From what was discussed in previous posts, the 2.2 V btw the bases of Q309 & 310 has me feeling optimistic. The other measurements, however, seem a little inconsistent with each other to my untrained eye. What do you think?
   

Roly

Quote from: js1970TP1 to TP2~  19.2 mV
     Halfrail ~ 9.8 mV
   V btw Q309 & 310 ~ 2.2 V   this varies with trimmer. This voltage trended down to     2.15V after a bit. That is to say, with thermistor 302 btw my fingers, and feeling the transistors for warmth. All ran well with Q314 & 319 getting warmest, almost hot.

I think that is looking a lot more healthy.

Idle Current:

19.2mV across 2 * 0.47 ohms

I = E/R

0.0192/(2*0.47) = 0.0204 = 20mA

That looks very reasonable to me.

9.8mV half rail offset is good.

Response to trimmer and finger heat is good.

Uneven heating is not so good.  You should check that you have about 10mV across each of the 0.47 ohm emitter resistors (to confirm that they are all intact, none open circuit).

If all seems to be well electrically it could be due to these two devices having worked a bit loose/not being firmly clamped down, or that the heatsink compound under them has dried out and become ineffective.

On a low level test (with limiting lamp) this should now sound a whole lot cleaner than before because the output stage is now correctly biased.


You may now care to move up in steps with your limiting lamp, confirming all is still well (which I expect it will be) then to full mains, and finally a high power soak test into your dummy load.  You will need to keep a close eye on it to make sure there is no inclination to go into thermal runaway.
If you say theory and practice don't agree you haven't applied enough theory.

js1970

#72
   I checked the emitter resistors. This is what I found in mV:
     R325~ 14.4        R324~12.8
     R327~6.2           R326~8.6
     R329~ 5.3          R328~2.4
     R331~ 5.1          R330~10.2
   These were taken at idle,no load,no signal.
 
   In a previous post, I mentioned that a couple output transistors seemed warmer than others, almost hot. After some research, I found I should have fastened everything back to the heatsink. I have since done this.
   All is running well with bias voltage starting at 2.388V and after an hour at idle I'm at 2.350,with all hardware at even temperatures.
  Curious, I plugged the bass in and started thumping away. That is, signal in,no load and still watching my bias. With the volume up I noticed the bias Voltage jumping erraticly as I hit thee strings, sometimes going as low as .8V. Is that normal? Im hoping I stumbled across a normal audio phenomenon. What do you think?
   Still curious, I plugged into load. Sure enough, when the bais went below 2V or so, it resulted in crossover distortion. When I stopped playing the bias went back to a comfortable 2.343V. It went as far as opening the relay at a certain point. Where to look next?

Roly

Quote from: js1970After some research, I found I should have fastened everything back to the heatsink.

Arrrggghh...   :(


Quote from: js1970With the volume up I noticed the bias Voltage jumping erraticly as I hit thee strings, sometimes going as low as .8V. Is that normal? Im hoping I stumbled across a normal audio phenomenon. What do you think?

The bias is a static setting, so when you superimpose tens of volts of signal you could get anything at all.


Quote from: js1970Still curious, I plugged into load. Sure enough, when the bias went below 2V or so, it resulted in crossover distortion. When I stopped playing the bias went back to a comfortable 2.343V. It went as far as opening the relay at a certain point. Where to look next?

Yes that still doesn't sound 100% healthy, however it may be an artifact of trying to drive it too hard via your limiting lamp and the supplies are sagging.  The relay is detecting DC offset on the output half rail. (so at least we now know that is working)

Since it now seems to be idle biased and maintaining the half rail I'd be inclined to try a larger wattage limiting lamp and see if your symptoms get better.

Another reason I suggest using a sig gen or MP3 player as a test signal source is that it is generally fairly compressed and a lot less peaky that whacking guitar strings.

If you say theory and practice don't agree you haven't applied enough theory.

js1970

    I ran though it again to ,hopefully, get a different result. I ran the hartke through a 100W floodlight bulb, and into the 4x10 cab. For a signal, I have a Roland micro cube that has a tone gen. function. I plugged the signal into both the active and passive inputs on the head, both having the same result.
    Power on, fired up nice, circuit protection relay clicked closed.
    Again, bias started at 2.440V measured btw driver bases.
    At very low volume, situation normal; even signal, minor fluctuations of the bias voltage, limiter bulb holding steady.
    As volume increases; bias voltage drops sharply while my limiter bulb sympathetically, if not proportionately, gains briteness. As the tone decayed so did the bulbs britness to its steady low volume illumination.
    Needless to say, I'm reluctant to go full mains.
    Why does the bias drop like that with increased signal?
    You said the bias is a static setting, which i take to mean it is set  for the power amp to run most efficiently to amplify a signal, but with the signal effectively stealing voltage from this setting, how can the power amp be working sufficiently, let alone efficiently?
    Perhaps my apprehension is due to my inexperience, and if it is I do apologize.
Just trying to make sense of it all