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

March 28, 2024, 03:04:55 PM

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

#45
  Well guys, I'm happy to report, the B+ preamp supply voltage is clocking in at a cool 14.75V. The hum is no longer.
  I'd like to thank everyone who responded to my call. I couldn't have gotten this far without you. And a special thank you to Roly for your patience, and walking me through step by step.
  So after sorting out the supply voltage, I hooked up the speakers and fired her up for a victory lap. Plugged in my daughters bass and nothing. Had signal in, no sound out.
WHAT?!
Unfettered, and with a newfound electrical prowess, I began to poke and wiggle components on the board. When performing this technical troubleshooting, I found CN404 ( a plugged connection) was loose where it fastened to the board. Not sure how it ties in,but a closer look revealed another broken trace. I tightened her up and tried again.
   Excelsior! Self induced noise and no hum. That's one for the good guys.
  My victory was short lived,I regret. The sound was fuzzy,.particularly when playing the high strings. That's with the EQ on or off. I'm wondering if it might be the cab,but I haven't been able to try another yet. Thoughts?

Roly

Quote from: js1970And a special thank you to Roly for your patience, and walking me through step by step.

My pleasure as always.  None of us were born knowing any of this stuff.  Very well done on getting this far.   :dbtu:

Quote from: js1970I'm wondering if it might be the cab,but I haven't been able to try another yet. Thoughts?

Well you should try another cab if possible to eliminate that as a possibility.  If you get exactly the same result then we need to look back at the amp, otherwise at the cab and particularly the driver.  Alternatively if you have another amp you could try driving the cab with that and see if it sounds the same or clean.
If you say theory and practice don't agree you haven't applied enough theory.

js1970

   Hey guys. I finally got to try the 3500 on a known working cab.
Its still a little sick.
When I play, the sound is fuzzy/distorted. Otherwise, its silent, no buzz or hiss or hum. I ran a tone through the  FX return with the same result. What do you think?
Will I be able to fault find with DMM, or will I inevitably need a scope or some other test equipment?

phatt

If the Amp has plenty of power and all seems ok then maybe crossover distortion.
Easy to hear it with the volume low and with your ear near the speaker play one single note on your guitar and let it ring out. It sounds like a torn speaker a fizzy buzzing sound behind the signal. SVR302 may need adjustment, I'll leave better minds to add more.

Unlikely but there is always the outside chance of RL301 having failing contacts.
Phil.

Roly

Quote from: js1970Will I be able to fault find with DMM?

There's still a fair bit you can check with your trusty meter.   :dbtu:


With the volume controls at zero (no signal condition), measure the voltage between TP301 and TP302.  There is about 1 ohm between these two points and the idle current is flowing through them producing a voltage related to the idle current of one millivolt per milliamp.  I would expect an idle current somewhere around 25-50mA per transistor pair, so you should see a voltage of something like 25 to 50mV.

You would also be wise to check the individual voltages across R324 to R313 (all 0.47ohm 5watt) which should all be roughly equal (which indicates that all the output transistors are passing similar idle currents).  These might not be exactly identical, but should still be within about 20% of each other.

If phatt is right then you may find that the idle current is low or even zero.  The idle current is set by Q308 and its surrounding network including trimpot SVR302.  Normally we wouldn't touch SVR302 because it rarely needs resetting, but if you have low or no idle current you first need to investigate the Collector-Emitter voltage on Q308.

Do we still have a fault, i.e. Q308 shorted, or is it only that the new power transistors require a different bias setting?  Since it is uncommon to need to reset the bias control after transistor replacement I would proceed with caution and not touch SVR302 until we are sure that the bias setting transistor is healthy (excessive bias could be a snake that lands you right back at square one with a blown output stage - be careful here).

The voltage on the Collector of Q308 will be the sum of the voltage across the emitter resistors, plus the EB drop of the power transistors, plus the EB drop of the driver transistors, somewhere in the range;

Low
2 * ( 12 + 500 + 500 ) = 2024mV (~2 volts)

High
2 * ( 25 + 700 + 700 ) = 2850mV (~3 volts)

If phatt's guess is right then you might find only two volts across Q308, and perhaps a lot less if there is still a fault.

If you have sensible voltages you should try sticking your finger on thermistor TH302 to heat it a bit while observing the idle current, and as it warms the idle current should trend down a bit (indicating that the thermal negative feedback is working).


So firstly, run the amp at idle and measure the voltages between TP301 and TP302, and the voltage to ground from those two points, post them, and we'll see what they suggest.  (you should now be running without the Limiting Lamp).


(BTW trim pot SVR301 is for trimming the output idle output voltage to zero - it also should not be touched without good reason).

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

g1

 I've seen it said there should be 25mV between TP301 and 302.  I've also seen it said 2.5mV, however the early versions were mosfet output devices and I believe the 2.5mV was for the mosfet version.
  Can anyone confirm or opine whether that sounds likely?

js1970

  When you say between the TPs, are you referring to the halfrail voltage?


Roly

No, between the test points, meter +ve to TP301 and meter -ve to TP302. 
If you say theory and practice don't agree you haven't applied enough theory.

js1970

  Here's the values you asked for:
TP 301 to ground- 23.2 mV
TP 302 to ground- 24.0 mV
TP 301 to TP 302- 0.00 mV
You guys nailed that one, eh. Does this mean only one half of the power amp is working? Or, that both halves work, just not together? Why the fuzz sound?

Roly

Wherever you have push-pull output stages the signal is split between the upper and lower signal halves, and it is important that the transition between one side of the output conducting and the other side conducting is a smooth hand-over, otherwise crossover distortion results.  That is what the bias setting in solid-state amps is for; to slightly turn on both parts so that each side starts to pick up the ball before the other side drops it.

With zero bias, as you seem to have, one side stops conducting before the other side starts, so there is a small dead zone in the middle of the waveform.  As it happens the human ear is quite sensitive to this crossover distortion.



What worries me here is that you don't seem to have any bias at all, so you need to carry out the measurements I mentioned above to see if the bias network transistor Q308 has some voltage across it, or may have been shorted by the original fault (or there is some other problem in this area).
If you say theory and practice don't agree you haven't applied enough theory.

js1970

 Cool, thanks for the info. I will get after Q308 asap. Happy Thanksgiving to you all

js1970

Hey guys. I metered Q308 under the diode setting, and I'm pretty sure its faulty. I couldn't get an OL reading with any combination of the leads. I didn't get an exact replacement of C3200, instead, I got a C2240 transistor. Its supposed to be comparable.
  While I was waiting for my tranny's to arrive. I tried to meter the components Roly had suggested. At least, how I comprehended Roly's suggestions.
  So , my voltage on the collector of Q308, with com. lead on ground and pos. lead on the collector, I read 1.7V.
  When I tried to measure across the emitter resistors, I get 00.0 mV. That was with com. on one leg of the resistor and pos. on the other leg of the same resistor.  However, when I held the com. on earth ground, I was able to read each side of those resistors. I'm unsure if the readings I got are what Roly had in mind,but I tried his equation anyway:
    Sum of voltage across emitter resistors-- 1.6 mV
+ emitter/ base drop of power tran.-- .4mV
+ emitter/base drop of driver tran.-- .3mV
If I'm reading the equation right 2( 1.6+ 400+300)=1403.2mV (~1.4V)
Assuming all readings are correct, does this mean the bias is simply running  cool as was suggested earlier?

js1970

    Finally got replacement transistor for the faulty Q308. I'll get to work asap.
  Happy holidays to you all!

js1970

   Finally got some free time to tinker with the hartke. I swapped the transistor at Q308 from the scheme. I must say, I was a little apprehensive about getting  the BEC configuration correct, but after some deliberation I went for it. When the tranny was soldered in,I did the tests that were suggested from previous posts.
No pops or smoke upon power up. I metered  TP 301&302 both to ground and btween. Here's what I found:
          TP301 to G - 65mV
          TP302 to G - 63mV
          Btw TP301and TP302 - 2.3 mV
         Voltage at collector of Q308 -  the DMM flashes 2.3V breifly then holds about
         1.2V
    I haven't fired amp up with instrument or load yet. With the voltages I posted, does it seem as though things are in order? Are there other things I should check before reassembly?   
       

Roly

Quote from: js1970TP301 to G - 65mV
          TP302 to G - 63mV
          Btw TP301and TP302 - 2.3 mV
         Voltage at collector of Q308 -  the DMM flashes 2.3V breifly then holds about
         1.2V

The actual half-rail offset voltage will be half way between these;

(63+((65-63)/2) = +64mV DC offset - good.

As a check;

(65-63) = 2 (should be 2.3V) - good agreement.

Idle current (and we have to resort to Ohms Law).

Vbtw = 2.3mV
Rbtw = 2 * 0.47 = 0.94 ohms

I = E/R
2.3/0.94 = 2.45mA - is something, but it's not much.  About one-tenth or less of expected 20-50mA.

In this context the voltage across bias transistor Q308 also seems a bit low at 1.2V.

At a minimum I'd expect four times Vbe plus some more for the emitter resistor voltage drop.

In (round) numbers;

4 * 0.6 = 2.4 + (0.02 * 0.94) = 2.42V

So at the very least it has to be 2 volts to even start current flow, and more like somewhere around 2.4V for correct idle.

However we adjust this by working out the voltage at the test points that will give us the desired idle current, then with the meter clipped on and holding our breath, we twist the tail of the tiger, SVR302.  {if you badly overshoot on this critical adjustment you can easily melt the output stage.  Initially you move the pot a little bit and observe what happens.  (also: keep a note of the pot initial position)}

Let's aim low first, 20mA idle.  At this the voltage between the Test Points will be;

E=I*R
0.02*0.94=0.02V or 20mV

So as they have arranged it, the voltage between Test Points is almost exactly one millivolt per milliamp - just hook up the meter on DC mV and read it as mA idle.
If you say theory and practice don't agree you haven't applied enough theory.