Troubleshooting Wurlitzer 200a amp board for bias and crossover notch distortion

[corgan4321]

I recently recapped and replaced all the transistors on my wurlitzer 200a board and it's sounding much less noisy now. Problem is, I'm really hearing the crossover notch distortion which makes the headphone out and the internal speakers not very useful for recording purposes.

I've tried to bias TR9 as per the schematic: changing R58 to get 1-10mv across R37/R38, but when I remove R58 I still have crossover distortion AND the voltage across R37 is 20mv. When I have R58 at 1k, I have a combined resistance of 130 ohms (R58 and R34 in parallel), and my voltage across R37 is 8mv. My R34 measure about 150 ohms. So it seems like I can't safely bias the circuit properly...

Two strange things is that the PNP output driver TR-11 measures 24.5V on both the base and the emitter, and the base of TR10 measures .48 volts instead of the marked .65V

I'm going to go ahead and change out all the rectifier diodes with 1N4007s and see if I can bring the supply voltage down a bit, as well as change the resistors and coupling caps in this output section, but I wondered if anyone else has run into similar issues where biasing to remove crossover distortion has seemed impossible?
Is there something else I should look at?
Thank you!

 

[JohnRoberts]

I recently recapped and replaced all the transistors on my wurlitzer 200a board and it's sounding much less noisy now. Problem is, I'm really hearing the crossover notch distortion which makes the headphone out and the internal speakers not very useful for recording purposes.

I've tried to bias TR9 as per the schematic: changing R58 to get 1-10mv across R37/R38, but when I remove R58 I still have crossover distortion AND the voltage across R37 is 20mv. When I have R58 at 1k, I have a combined resistance of 130 ohms (R58 and R34 in parallel), and my voltage across R37 is 8mv. My R34 measure about 150 ohms. So it seems like I can't safely bias the circuit properly...

Two strange things is that the PNP output driver TR-11 measures 24.5V on both the base and the emitter, and the base of TR10 measures .48 volts instead of the marked .65V

I'm going to go ahead and change out all the rectifier diodes with 1N4007s and see if I can bring the supply voltage down a bit, as well as change the resistors and coupling caps in this output section, but I wondered if anyone else has run into similar issues where biasing to remove crossover distortion has seemed impossible?
Is there something else I should look at?
Thank you!

Without a schematic I can't infer anything useful from your notes other than about TR-11. If base-emitter junction has 0V across it either the transistor is broken (shorted b-e) or the drive to that base is faulty.

JR

 

[moamps]

Is there something else I should look at?

It looks to me that your amp works as it should. The crossover distortion are not so obvious if NFB is active so I think the problem is somewhere else. Did you measured DC voltage at output?

 

[BYacey]

I think you need to look closer at TR10 and TR11, it sounds like TR11 isn't conducting. What's the DC voltage on the collector of Q11, and what is the dc voltage on the speaker output at idle?

 

[JohnRoberts]

+1 check the base voltage of the npn to be +.65V above the output 0V if the NPN it turned on and working properly it will pull down the base of the pnp a similar -.65V below the + rail.  The problem could be in the bias network too.

JR

 

[corgan4321]

Thank you all for the replies!

Here's what I've just measured:

24.5 on both base and emitter of tr11

-0.01v on speaker output and collector of tr11

0.39V base of tr10

Thank you

 

[s2udio]

As you have recapped it........was it working without the problem you describe before the operation ?
Check the test points,plenty of them,they are there for a reason.
Random voltage testing across transistor DC voltage drops will lead you up the garden path.......seriously bad practice, in audio fault finding IMHO
That regulator chip , some of those voltages are critical, these should be checked.
Slightly depleted leaky old devices, will keep you searching forever. IE caps and semiconductors
A Signal gen and scope may  help.
Recheck your work, not being conderscending,  you may have created a problem, while solving  another, as we all have done at some time or another !

 

[corgan4321]

Yes it exhibited crossover distortion before the recap and the transistor swaps.

I figured replacing the transistors would take care of it... I've replaced the driver transistors and the output transistors, as well as the coupling caps in the bias section. I also swapped the bias resistance over to a 180R resistor which seems to have helped a little bit
What's weird now is that I'm not getting any voltage across r37 or r38 with the speakers unplugged.
All the voltages from the regulator are pretty close expect for the 22V output reads as 24.5V but I figured +10% is close enough to spec. The other voltages are within 10% too

 

[JohnRoberts]

Thank you all for the replies!

Here's what I've just measured:

24.5 on both base and emitter of tr11

This is not good valid voltages for operating circuit. base of TR11 should be 22V as noted in the schematic.

-0.01v on speaker output and collector of tr11

good

0.39V base of tr10

this voltage is low, according to schematic should be +.65V  . If tr10 is not turned on if can't turn on tr11.
======

For a simple circuit explanation, tr9 or the bias control transistor is a Vbe multiplier to generate roughly two diode drops across it (like a variable zener diode with the voltage tweaked by the resistor ratio), to just barely turn on both the + and - power stages at idle (that is the class A part of being class AB).

Ignore where the schematic provides two different voltages for the same node (0.57 and 0.65 at collector tr9/base tr10 same node).

There should be roughly 1.3V across TR9 equally spaced above and below 0V (at idle).

For class A current you can measure the V drop across R37 and R 38. The schematic says 10mA so  V=IxR or .01 x .47 = 5mV. You should measure roughly 5mA across both R37 and R38, both should measure the same with no speaker connected.

Thank you

Good luck

don't short anything out making hot measurements.

JR

 

[corgan4321]

Thanks John Roberts,

I think the 0.39V on the base of TR10 is too low too, what could be causing this? Ive replaced TR10 as well as the output transistors and still have  the same issue.  The voltage reads as 24.5V because thats the supply voltage. TR13 measures correctly with 0.4V- 0.5V between base and emitter. 

The schematic calls for 1-10millivolts across R36 or R38 and I measure 0mV right now

Thank you thank you!

 

[JohnRoberts]

Thanks John Roberts,

I think the 0.39V on the base of TR10 is too low too, what could be causing this?

There is global negative feedback so the fault could be almost anywhere,

The schematic is good about listing typical voltages, check around the circuit until you find things not right.

Ive replaced TR10 as well as the output transistors and still have  the same issue.  The voltage reads as 24.5V because thats the supply voltage. TR13 measures correctly with 0.4V- 0.5V between base and emitter. 

OK this is evidence that tr13 is turning on and working (pulling down) at idle, while tr11 is not.
I already asked these questions but does TR9 the bias transistor have roughly two diode drops across it?

If the top base of tr10 is only  +0.39V, then the bottom or base of tr12 should be more than -0.65V called for in the schematic, which could turn on the pull down stage too hard.

Either tr8 does not have 2 diode drops across it, or tr12 is flaky, or both (but fact you see base drive at tr13 suggest tr12 may be working. ).

Repairing gear is not just swapping out parts with no plan... that can get expensive with bad results, or break something else if not careful.

if TR9 is low voltage, figure out why.. a Vbe multipler is one of the simplest transistor circuits around. The voltage from base to emitter is 1 diode drop (0.5-0.6V), the voltage emitter to collector will be increased by the ratio of r34/r35 so Vce= (1+r34/r35) x Vbe.  For those values ignoring the optional r58 that reduces class A bias, the ratio is (1+(160/220)) times Vbe... or 1.12V for 0.65 Vbe. With r58 in place the voltage will be even less (1.05V for 0.65Vbe). All these voltages scale up and down with device temperature so could vary from cold to hot (will be smaller when hot).

The schematic calls for 1-10millivolts across R36 or R38 and I measure 0mV right now

Thank you thank you!

Again more evidence that tr9 may be delivering low class AB bias voltage.

It may be as simple as getting tr9 up to correct voltage, but be careful, too much voltage across tr9 could burn up the amp from mutual conduction (both pull up and pull down at the same time).

Good luck

JR

 

[corgan4321]

I am in fact getting two voltage drops across TR9, it measures 1.15V across it. Looking at my values all on one page, nothing looks off to me except the voltages on TR12. Should I go ahead and increase the resistance at TR9 to get more voltage and see what happens? Burning things up is all I'm worried about as you mentioned... I heard I could I possibly use a potentiometer to find the correct bias for that resistor. Could it be as simple as putting R34 in series with another low value resistor in parallel with a 1k pot? Is this a safe procedure?

I've measured all the voltages for the transistors in the power section:

TRx
E
B
C

TR 7
0.60
0.01
-23.8

TR8
0.60
0.00
-24.5

TR9
0.62
0.0
-0.533

TR10
-0.11
0.53
24.3

TR11
24.3
24.3
0.0

TR12
-0.01
-0.62
-24.0

TR13
-24.4
-24.0
0.0

TR14
-24.4
-24
-0.62

 

[JohnRoberts]

two weeks later this is like start over...

I am in fact getting two voltage drops across TR9, it measures 1.15V across it.

sounds close

Looking at my values all on one page, nothing looks off to me except the voltages on TR12. Should I go ahead and increase the resistance at TR9 to get more voltage and see what happens? Burning things up is all I'm worried about as you mentioned... I heard I could I possibly use a potentiometer to find the correct bias for that resistor. Could it be as simple as putting R34 in series with another low value resistor in parallel with a 1k pot? Is this a safe procedure?

I've measured all the voltages for the transistors in the power section:

TRx
E
B
C

TR 7
0.60
0.01
-23.8

ok

TR8
0.60
0.00
-24.5

ok

TR9
0.62
0.0
-0.533

ok

TR10
-0.11
0.53
24.3

not sure about negative voltage at emitter

TR11
24.3
24.3
0.0

Collector of TR11 is connected to emitter of TR10 so should be the same voltage.

TR12
-0.01
-0.62
-24.0

ok

TR13
-24.4
-24.0
0.0

again slight disagreement between two measurements of same node. but I'll ignore 0.01V

TR14
-24.4
-24
-0.62

OK...
=======
Before changing the design, lets figure out why TR11 is cut off...

with 0.64 V base emitter of TR10 there should be some drive current pulled from base of TR11

In proper operation there should be about 2 mA in R36 and some modest current in base of TR11 to turn it on.

0V b-e of TR11 suggests that either TR11 is shorted base-emitter (or solder bridged). OR TR10 is faulty.

Based on this data set, I would confirm with a VOM (with amp turned off) that you measure roughly 270 ohms across base-emitter of TR11 (R36).

Problem could be as simple as a solder bridge (base-emitter TR11), or a faulty TR11. If TR11 checks out good, then TR10 is suspect.

I would not mess with TR9 bias*** until we confirm TR11 is healthy and happy.

JR

*** if you do try to adjust class A bias with a trim, R35 might be the better one to mess with. An open trim pot in that leg would result in a low class A bias condition not a max bias condition that could release smoke,,, target a modest range of adjustment, you can blow stuff up with this trim.

 

[Gene Pink]

To OP:

If we are sure that TR-10 and 11 are good, check the voltage at the node between R-32 and R-33, should be about +11.5V, or a bit more since your PS voltage is a bit high. R-32 and R-33 should have the same voltage across them, +/- resistor tolerances.

If not, check those resistor values, and C-12 for leakage. Doesn't happen often, but sometimes even new caps are leaky. More likely, is it in the wrong way?

A leaky or reversed cap could cause these problems, reducing the current drive to TR-10 base to cutoff, yet all voltages would be close, except for TR-11 base, as TR-10 collector isn't pulling any current through R-36 at idle.

Also, small point, the voltages you have listed for TR-9 seem to have C and E transposed.

JR, I wouldn't worry much about slightly different voltages on the same node. While they are a cause for a second look, that node may well have drifted around a bit between the various measurements that the OP took. Plus, note the 0.08 voltage between TR-9 collector and TR-10 base per the voltages printed on the original schematic, and they are on the same node. We don't know how long and thin the traces are, or the wiring from the board to the heatsink where the outputs are mounted.

Gene

 

[JohnRoberts]

To OP:

If we are sure that TR-10 and 11 are good, check the voltage at the node between R-32 and R-33, should be about +11.5V, or a bit more since your PS voltage is a bit high. R-32 and R-33 should have the same voltage across them, +/- resistor tolerances.

yes,,

If not, check those resistor values, and C-12 for leakage. Doesn't happen often, but sometimes even new caps are leaky. More likely, is it in the wrong way?

A leaky or reversed cap could cause these problems, reducing the current drive to TR-10 base to cutoff, yet all voltages would be close, except for TR-11 base, as TR-10 collector isn't pulling any current through R-36 at idle.

Also, small point, the voltages you have listed for TR-9 seem to have C and E transposed.

yes

JR, I wouldn't worry much about slightly different voltages on the same node. While they are a cause for a second look, that node may well have drifted around a bit between the various measurements that the OP took. Plus, note the 0.08 voltage between TR-9 collector and TR-10 base per the voltages printed on the original schematic, and they are on the same node. We don't know how long and thin the traces are, or the wiring from the board to the heatsink where the outputs are mounted.

Gene

Yes i saw the obvious schematic voltage discrepancy. Sometimes the act of measuring a circuit changes voltages slightly, or things drift around a little.

The problem looks to me located in TR10/TR11, while a compromised bootstrap pull up might explain the lack of drive but  TR10 Vbe appears to be turned on. 

JR

PS In physics there is an "observer effect", where looking at something can change it.  :-\

 

[corgan4321]

I thank you guys so much for staying with this so long and in helping me, especially John, thank you!!!

I've taken TR10 and TR11 out of circuit and tested them, they both test properly with my diode  meter. Also, I checked resistor values in the output section already and all are within spec.

I get 12.44V at R32/R33 which is in the area of 11.5V considering my PSU is a couple volts high.

Sorry about the TR9 misprint. I actually had the negative sign in the wrong place, but the value was correct. I measured again:

TR9
-0.62
0.0
0.54

Just to clarify, you mention about R36 having voltage across it, which is does. With the speaker unplugged it has about 0.111V across it, with the speaker plugged in it has .04V across it. However, the schematic references R37/R38 needing to have 0.01~0.10 mV across them with the speaker disconnected. With the speaker disconnected, I measure 0.00 volts across both R37 and R38.
I mentioned earlier I measure -.011V on  the amp output. Wouldn't it be necessary to measure 0.0 volts here to get the proper bias across R37/R38? Even if TR9 is biasing properly, wherever that -.011 voltage is coming from pulls that positive voltage bias from TR9 down to 0V??

In other words, the reason I measure 0 volts across R37 is because the positive bias created by the bias circuit is being pulled down to 0.0V by the negative voltage present at the output?

 

[JohnRoberts]

I thank you guys so much for staying with this so long and in helping me, especially John, thank you!!!

I've taken TR10 and TR11 out of circuit and tested them, they both test properly with my diode  meter. Also, I checked resistor values in the output section already and all are within spec.

I get 12.44V at R32/R33 which is in the area of 11.5V considering my PSU is a couple volts high.

OK

Sorry about the TR9 misprint. I actually had the negative sign in the wrong place, but the value was correct. I measured again:

TR9
-0.62
0.0
0.54

Just to clarify, you mention about R36 having voltage across it, which is does. With the speaker unplugged it has about 0.111V across it, with the speaker plugged in it has .04V across it.

but you listed TR11 as having same voltage at base and emitter which should  mean 0V Vbe, or the two ends of R36 are not connected to the base and emitter of R11.  Can't be both 0V and .11V  at same time?

However, the schematic references R37/R38 needing to have 0.01~0.10 mV across them with the speaker disconnected. With the speaker disconnected, I measure 0.00 volts across both R37 and R38.
I mentioned earlier I measure -.011V on  the amp output. Wouldn't it be necessary to measure 0.0 volts here to get the proper bias across R37/R38? Even if TR9 is biasing properly, wherever that -.011 voltage is coming from pulls that positive voltage bias from TR9 down to 0V??

Tr11 is cut off so there is no pull up current.  If working properly the negative feedback will tell the amp to make output 0V.

0.11V across 270 ohms is only 400 uA so not much current  if TR10 is operating properly and TR11 base-emitter is "not" shorted. TR10 base emitter >0.6 V so appears turned on more than hundreds of microamps. But still could be faulty TR10 or faulty TR11.

If you remove TR11 do you see more than 0.11V across R36.. do you measure 270 ohms across R36 with and without TR11 in place (power off).

if yes voltage across R36 increases with TR11 removed it may be faulty, if voltage still low TR10 is suspect.

While solder bridges, broken traces, or mechanical faults could also be involved.   

In other words, the reason I measure 0 volts across R37 is because the positive bias created by the bias circuit is being pulled down to 0.0V by the negative voltage present at the output?

no problem is still most likely in tr11/tr10 no active pull up means no current in bias string.

In normal operation you should see a similar Vbe across TR11 and TR13 base emitter,  just opposite polarity from opposite rails.  The output devices and driver stage is identical just flipped polarity. The pull down side looks like it is working correctly, the pull up side not so much.

TR12 has same Vbe as TR10
TR13 has 0.4V Vbe, TR11  has 0.11 Vbe ????


JR

 

[corgan4321]

Okay so the confusion about Vbe comes from the poor resolution of my multimeter. With speaker plugged in, Vbe on tr11 is .01, I just can't see that resolution when I measure in the >20V range. But measuring across R36 ensures that is the case.

Removing TR11 drops the voltage down to 0.06V across R36, showing that TR11 is in fact doing something .

Let me ask this question: is MPSA06 a reasonable substitute for TR10? According to online sources it should be but maybe that's just not true?

I understand now about what you're saying, if both sides of the amp, pull up and pull down, are functioning, the output should read 0V.

I've replaced TR10 since this issue has been unresolved, so I'm skeptical that it could be bad, especially because I took it out of circuit and measured it to be okay.
But I do have an extra MPSA06 so I could go ahead and replace it again.

 

[JohnRoberts]

Okay so the confusion about Vbe comes from the poor resolution of my multimeter. With speaker plugged in, Vbe on tr11 is .01, I just can't see that resolution when I measure in the >20V range. But measuring across R36 ensures that is the case.

OK

Removing TR11 drops the voltage down to 0.06V across R36, showing that TR11 is in fact doing something .

Not sure it means that at all... With TR11 removed, ASSuming there is only a 270 ohm resistor there (I think I've requested that you measure that before) 0.06/270=222 uA... Not much current if TR10 is turned on pretty hard as the Vbe suggests. .

Let me ask this question: is MPSA06 a reasonable substitute for TR10? According to online sources it should be but maybe that's just not true?

I am not familiar with what a 203719 is  but if the previous device is a plastic to-92 the MPSA06 is a decent NPN.  I used to use lots of A06 and A56 for jelly bean NPN/PNP applications. 

I understand now about what you're saying, if both sides of the amp, pull up and pull down, are functioning, the output should read 0V.

negative feedback back to TR7/TR8 long tail pair, will actively drive the output up or down to 0V, with no signal in.  TR9 sets up some always on class A bias current, but exact output voltage is always controlled by global NF loop to 0V at idle.

I've replaced TR10 since this issue has been unresolved, so I'm skeptical that it could be bad, especially because I took it out of circuit and measured it to be okay.
But I do have an extra MPSA06 so I could go ahead and replace it again.

The Vbe voltage you report for TR10 is not consistent with only 200 uA of collector current, but again that ASSumes no solder bridges or stray connections on 270 ohm resistor.

I'd replace the tr10, just to see what happens.

JR

PS: To make more accurate base-emitter voltage measurements close to a PS rail you could measure them directly, then add to or subtract from the rail voltage. It's the relative voltage we care more about anyhow.  Likewise to measure the voltage drop across R36 just measure that directly.

 

[corgan4321]

The 270 ohm resistor measures 260 ohms when the amp is off and TR11 is in place. When I turn the amp off, it measures much lower and takes a while for the resistance to read 260. I believe when you'd asked me to remove TR11, It read closer to 266 ohms.

I replaced TR10 and still have the same measurements at TR11
TR10vbe is .560 with the emitter voltage being -0.007 and the base measuring .567

The voltage drop across R36 is 0.030 which is the same as when I measure directly across TR11 for Vbe