Topics

Tek 576 - step generator power amp - how to check?


 

Hi all,

I don't know if I made a mistake during alignment or if I just didn't notice it before, but the step generator is no longer working properly. In NPN mode it gives a constant +30V, in PNP it looks better but the levels are not correct (going through zero although offset is set to 0).

I traced the signal through the step generator, it looks OK until the output on pin AD where it goes to the power amplifier. Its four transistors are located on a heatsink behind the step generator. I took out the step generator to gain access to the power amp and tried to test the transistors in-situ. That wasn't conclusive. Next I applied voltage at the entry, i.e. the bases of Q172 anad Q176, using the power supply of the 576 (cautiously with 20% line voltage). I couldn't make any sense of my measurements. I find this circuit very confusing, everything is floating around plus there is a bias voltage derived from the +100 and -75V supply.

I would like to make an educated guess about which, if any, of the transistors is defect before unsoldering everything.

Any suggestion how to check this amp by applying and measuring voltages and/or waveforms?

cheers
Martin


Ozan
 

Hello,
I looked at the schematic briefly (I don't have this equipment). When it is stuck at 30V what voltages do you see at both terminals of diode D188 and D189, and node "AD"? The schematic has two different states, voltage mode and current mode. When output is stuck at 30V is the equipment in voltage or current mode?

Node "AD" is the output of differential amplifier, with such a mismatch between what feeds into node "AF" and what is at output (30V constant) Q150A should rail, which should rail node "AD". Railing of Q150A is collector voltage away from ~ 12V.


Ozan


 

Hi Ozan,

the step amplifier card is mounted in front of what I call the power amplifier on the heatsink. Its practically impossible to make measurements on the power amp when the step amp card is installed. There are wires connected to the step amp card on practically all sides, making it impossible to take it off with the connections still active.

That said: when I made measurements on the step amp, I recall having had something like 0,8V on the right base of Q150. I did not suspect the feedback loop because the +30V appear on NPN only, and in voltage and current mode.

I'm not 100% sure about about the step amp between AF and AD, but then again, its working close to normal in PNP mode so it does not make sense to suspect this part of the circuit. Except perhaps the parts around Q246 and Q250, as they are interchanged when changing mode.

I checked the voltage and current setting resistors, all OK. I might reconnect the step amplifier and note voltages when having the +30V output.

What I'd like to do before is testing the power-amp alone.
I am thinking of:
- removing relay K102 (effectively disconnecting the 50V supply)
- supplying a regulated voltage (i.e. 20V) at the collector of Q180 / emitter of Q184
- tying the emitter of Q184 to ground
- applying a signal on the bases of Q176 / ´Q172, with reference to ground
- measuring current flowing to ground through the current limiting resistors

If the output turns out symmetric and without distortion for all levels on the entry the amp itselfs should not be the problem, but I'm not sure if I overlooked something.

cheers
Martin

On 4. Feb 2021, at 07:15, Ozan <ozan_g@erdogan.us> wrote:

Hello,
I looked at the schematic briefly (I don't have this equipment). When it is stuck at 30V what voltages do you see at both terminals of diode D188 and D189, and node "AD"? The schematic has two different states, voltage mode and current mode. When output is stuck at 30V is the equipment in voltage or current mode?

Node "AD" is the output of differential amplifier, with such a mismatch between what feeds into node "AF" and what is at output (30V constant) Q150A should rail, which should rail node "AD". Railing of Q150A is collector voltage away from ~ 12V.


Ozan





Ozan
 

Hi Martin,
As I said before I don't have this equipment so please take the comments below with a grain of salt.

In NPN mode (+ mode) only Q176 and Q184 are active, Q172 and Q180 should be off. When Q176 and Q184 are not driving "-" terminal of 50V supply is pulled to -75V by R192. With the feedback loop Q176 and Q184 pull "-" terminal of 50V supply close to ground, raising the output voltage to the desired point. Note that Q184 only helps above ~ 15mA, until then only Q176 does the job.

Is 30V you see well regulated, if you pull some current, for example 10mA with a 3.3k 0.5W resistor to ground from output, does the voltage move? If it is firm, then I would look around feedback path. +50V would mean Q176/Q184 is short, -25V (-75V + 50V) would mean they are open. However, getting 30V is more complicated.


That said: when I made measurements on the step amp, I recall having had
something like 0,8V on the right base of Q150. I did not suspect the feedback
Base of Q150B should be ~ 0V in voltage mode. AMP balance (R224) might be off.

loop because the +30V appear on NPN only, and in voltage and current mode.

I'm not 100% sure about about the step amp between AF and AD, but then again,
its working close to normal in PNP mode so it does not make sense to suspect
this part of the circuit. Except perhaps the parts around Q246 and Q250, as
they are interchanged when changing mode.
I agree Q248 area is one of the differences. One more change in NPN mode is in ramp generator with K101-A relay contacts. Is node "AI" accessible? You can look at the ramp signal in NPN and PNP modes. In PNP node the block marked "INVERTER" provides isolation to the "CURRENT TO VOLTAGE AMPLIFIER".


What I'd like to do before is testing the power-amp alone.
In the comments below I assume "power-amp" is only the section with Q172-Q184, not the whole amp.

I am thinking of:
- removing relay K102 (effectively disconnecting the 50V supply)
- supplying a regulated voltage (i.e. 20V) at the collector of Q180 / emitter
of Q184
I assume +20V at Q180 and -20 at Q184. Note that they are more like a switch than linear amps. You need to limit the current, or be mindful of the voltage you apply to AD node. Going positive on node AD you should see (V_AD-1.4)/R_current_limit pulled from +20V. In the negative direction (V_AD-0.7)/R_current_limit from -20V.

- tying the emitter of Q184 to ground
Q184 expects a negative voltage at the emitter.

- applying a signal on the bases of Q176 / ´Q172, with reference to ground
Do not exceed +3.4V / -2.7V at node AD, these are about the ranges Q169 supplies.

- measuring current flowing to ground through the current limiting resistors
Or you can measure the current through your supply, it will be simpler.


If the output turns out symmetric and without distortion for all levels on the
entry the amp itselfs should not be the problem, but I'm not sure if I
overlooked something.
This stage is quite nonlinear, only expectation is current will be about the levels I wrote above. Linearity is achieved with the feedback loop around the amp.

Ozan


Ozan
 

Hi Martin,

That said: when I made measurements on the step amp, I recall having had
something like 0,8V on the right base of Q150. I did not suspect the
feedback

Base of Q150B should be ~ 0V in voltage mode. AMP balance (R224) might be off.
Thinking about it a little more I think this should be the first thing to correct (R224 trim, assuming it is only a trim issue not a circuit issue) . The offset you measured is pretty large, it can definitely be the reason for PNP side offset and may contribute to the NPN issue. Did you measure +0.8V or -0.8V?

Output of Q229 based amp is available at the shield of node "AB", it should be pretty close to 0V as well. If it is below -1.2V or above +1.2V it will turn on the voltage clamp.

Ozan


 

Hi Ozan,

thanks a lot for all these insights - it helps clearing up the fog in my head.

I admire how you get all these numbers by looking at the schematic alone. I think I understood the current computations as a function of AD voltage - it depends on the number of PN sections on the currents way to ground, I guess. But I didn't manage to see how you can predict the voltage range of +3.4V / -2.7V at AD point. The preceding transistor is hanging between -75 and +12 volts...

That said, I did make the current measurements on what I call the "power amp" alone, and it corresponds quite well with your predictions:
- all the resistor in series are approx. 65 Ohms
+ 2V at AD -> 12,7 mA (9,2 mA computed)
+ 3V at AD -> 26,8 mA (24,6 mA computed)
-1,3V at AD -> 10,5 mA
- 2,3V at AD -> 25,3 mA
Its not perfectly symmetrical, but thats seems normal given the design so I decided not to touch anything on this power amp and put the step generator back. Thats not an easy task, with all the dangling wires around and one must take care not to mix up the connectors. Thats where I am right now, fingers crossed, I will make more measurements on the step generator amp itself as you suggested.

I intend to post some pictures of the 576 with the step generator card removed. Haven't found any on the web so far. Looks a bit messy compared to the rest of the unit.

so far, cheers
Martin


 

It seems I found the culprit... Q241, probably open. So it was the feedback section, thanks Ozan for your help.

I made a bunch of measurements before, leading to very different readings depending on NPN or PNP mode. Below is the collection of readings.

Thats the 4th transistor I have to change since I began repairing some days ago... and I have the impression that 3 of them were not faulty at the beginning. How can that be? Is my presence that corrosive?

cheers
Martin

----------------------------------------------------------------------------------------------------------

On the 0 Volt line:
+12V in NPN
+1V in PNP
(now its 0,000V)

The base of Q241 was reading -12V

The collector of Q235 was reading
+ 67V in NPN
+ 51V in PNP

The emitter of Q233 was reading
- 0,42V in NPN
- 0,58V in PNP, rising to -8,8V beginning at 5uA/Step until 200mA/Step


On the output of the Step Generator main amp, readings with a 3k load:
#NPN current mode:
+ 0V @ 0,05uA/Step
+ 10V @ 0,1mA/Step
+ 32V @ 0,1mA/Step (thats where my 30V came from)
+ 43V @ 0,1mA/Step
Voltages not regulated, decrease when increasing load, increase when increasing current limit.

#NPN voltage mode:
+ 30V @ all levels (thats where my 30V came from)
at 2V/Step the signal shows steps between 28 and 30V
Voltages not regulated, decrease when increasing load, increase when increasing current limit.

#PNP current mode:
approximately correct output, regulated

#PNP voltage mode:
- 10V @ 0,05V/Step, regulated
- 26V @ 2V/Step, regulated


Ozan
 

Hi Martin,
I think I understood the current computations as a function of AD voltage - it
depends on the number of PN sections on the currents way to ground, I guess.
The schematic shows 2A full scale with 1-ohm in the emitter (R182). In positive direction there are two Vbe's (Q172 & Q180). In the negative direction there is one Vbe (Q176) between "AD" and R182. At 2A we are dropping 2V across R182. Vbe is ~ 0.6-0.7V, I used 0.7V.
In positive direction 2*0.7+2V=3.4V, in negative direction -0.7-2=-2.7V

But I didn't manage to see how you can predict the voltage range of +3.4V /
-2.7V at AD point. The preceding transistor is hanging between -75 and +12
volts...
Two extremes of drive states are when Q164 is in cutoff and Q164 is in saturation.

When Q164 is in cut off:
-75V pulls bottom node of R167, this turns on diode D165 and clamps top of R166 to -0.7V. Current through R167 is
I_R167=(-0.7V-(-75V))/(1k+38.3k)=1.89mA
Voltage at the base of Q169 is
I_R167*38.3k+(-75V)=-2.6V
At emitter of Q169 = -2.6V-0.7V=-3.3V

When Q164 in saturation:
Collector of Q164 is at 12.5V (ignore Vce_sat).
I_R167=(12.5V-(-75V))/(2.55k+1k+38.3k)=2.09mA
Voltage at the base of Q169 is
I_R167*38.3k+(-75V)=5V
At emitter of Q169 = 5V-0.7V=4.3V

This range [-3.3V, 4.3V] is larger than the range based on the calculation from the output stage [-2.7V, 3.4V] but we made several assumptions in both (Vbe=0.7V, ignore base currents, ignore Vce_sat, didn't consider drop across R171, etc) so I recommended the narrower range to be safe.


That said, I did make the current measurements on what I call the "power amp"
alone, and it corresponds quite well with your predictions:
- all the resistor in series are approx. 65 Ohms
+ 2V at AD -> 12,7 mA (9,2 mA computed)
+ 3V at AD -> 26,8 mA (24,6 mA computed)
-1,3V at AD -> 10,5 mA
- 2,3V at AD -> 25,3 mA
Vbe is ~ 0.6V-0.7V depending on current through the transistor and Is of the transistor. I assumed 0.7V, if you use 0.6V for these low currents the prediction is closer.


Its not perfectly symmetrical, but thats seems normal given the design so I
I agree this stage looks fine. Most likely offset in Q229 amp is the cause. Step 22 on page 5-17 describes the trim. If the offset is large and negative it would turn on clamp Q248 and override node "AF" (step input), producing a constant voltage in NPN mode. In PNP mode the clamp Q250 would be off but you would get an offset.

Ozan


Ozan
 

Hi Martin,


It seems I found the culprit... Q241, probably open. So it was the feedback
Do you still see 30V in NPN mode after changing Q241 and adjusting the "AMP BAL" to 0V at the emitter of Q241? Any behavior change after replacing Q241?

Ozan


 

Hi Ozan,

Do you still see 30V in NPN mode after changing Q241 and adjusting the "AMP BAL" to 0V at the emitter of Q241? Any behavior change after replacing Q241?
everything (with one exception) is working as it should after replacing Q241. In particular, voltages match the circuit diagram and output is correct in both directions, NPN and PNP.

The one exception: when in voltage mode, the steps do not begin at zero even with the offset turned off. There is an offset, about 2-3 volts in NPN and 1-2 volts in PNP. Whereas in current mode they start exactly at zero volts.

I found out that when I probe voltages around Q235, 241 and 233, this offset goes away. Even when I approach my finger to the area around C229 the offset drops and eventually goes to zero when touching it. Touching it with probing lead disconnected from the DMM(!) is sufficient, too. Its not mechanically sensitive, pushing things with a plastic rod does not do anything.

When touching parts with the probe of the DMM, the effect comes and goes very quickly - in the area of milliseconds. Impossible to see a ramp on the scope.

I swapped Q239 against a BC639, that made the offset even worse in voltage mode. Strange. Could it be Q241 (replaced by BC640) has some special characteristics I must care for and that are too different in the replacement?

Any other ideas?

cheers
Martin


Ozan
 

Hi Martin,


I found out that when I probe voltages around Q235, 241 and 233, this offset
goes away. Even when I approach my finger to the area around C229 the offset
drops and eventually goes to zero when touching it. Touching it with probing
lead disconnected from the DMM(!) is sufficient, too. Its not mechanically
sensitive, pushing things with a plastic rod does not do anything.
This is usually a sign of oscillations, probably additional cap of the multimeter lead is enough to stop (or attenuate) it. If you look at base of Q150B with a scope probe does it fix the offset issue when probe touches, and if it doesn't (I am expecting it won't) do you see any oscillating waveform? Does the oscillation (if any) stop when you touch the Q235 area?

If all the voltages are correct (at least collector of Q233 at +13.3V, collector of Q235 at +36.2V, base of Q241 at -0.6V) only possibility I can think of is your replacement Q241 is much slower (lower ft, higher cap etc) than the original. C229 is the compensation cap, emitter of Q241 has a second (undesired) pole.

Does your replacement Q241 match the original part?


I swapped Q239 against a BC639, that made the offset even worse in voltage
I can't find Q239.

mode. Strange. Could it be Q241 (replaced by BC640) has some special
characteristics I must care for and that are too different in the replacement?
I didn't read this note while I was writing the above paragraph. Q241 is a 2N4250, its base and collector capacitances are much smaller than BC640. You want to replace Q241 with a transistor at least with similar (or higher is OK) ft, Cob, and Cib. If oscillation is the reason for offset (big if), larger Cob and Cib are the reason for the oscillation.

Ozan


Ozan
 

Just to clarify my previous comment:

want to replace Q241 with a transistor at least with similar (or higher is OK)
ft, Cob, and Cib.
Higher ft is OK but Cib and Cob should be similar.


 

Hi Ozan,

This is usually a sign of oscillations...
Spot on with everything!
BTW: I was meant Q235 when I wrote Q239 - sorry.


1. Oscillation
Yes, I could measure these at base of Q150B - I also see them on the output, see the pics that I added here:
https://groups.io/g/TekScopes2/album?id=260354
First pic shows the oscillations, second pic when touching C229. The red line is the step generator output during first step-up of its cycle. White line is base of Q150B...
Second pic shows normal reaction of the feedback line, I guess. And that may also be the trigger for the oscillations.

This explains why I had these thicker-as-usual lines on the analog scope connected to the output of the step generator.


2. wrong transistor
the fun-thing about socketed transistors is that you can swap them easily... I swapped the BC640 with the 2N4250 in socket Q164. No more oscillations... Q164 does not seem to be that picky.


Thanks again, Ozan, that was really great help!

I will try to understand your computations from the previous posts once I'm through recalibrating the step-amp.

cheers
Martin


Ozan
 

Hi Martin,
Great news and congratulations for fixing the 576.

the fun-thing about socketed transistors is that you can swap them easily... I
swapped the BC640 with the 2N4250 in socket Q164. No more oscillations...
Q164 does not seem to be that picky.
Q164 is also in a signal path but dominant pole is at the base of Q164 so more base capacitance likely made the output amplifier a little slower. Probably not a big issue. If you see any side effects you could try using Q250 as a donor, it is just a voltage clamp most of the time sitting idle. Next choice would be Q233 if Q250 is not an original.

Ozan


 

Hi Ozan,

I realigned the step generator of the 576 to specifications without problem. Everything seems fine now, its sitting on a desk awaiting the Metrix tube analyzer so they can play together :-)

For fun, I examined the defective transistors (see the pics here: https://groups.io/g/TekScopes2/album?id=260354).

Q241 (2N 4250 in the step generator):
seems quite open until 9V CE, then resistive with some kOhms, and is totally insensitive to base voltages...

Q725 and Q727 (NPN with Tek number 151-0190-00 in the 75V power supply):
these look OK, but quite different (double the current with same settings). As soon as I put them back the 75V volt supply shuts off. Their function is to protect against all sorts of shorts against other voltages.

cheers
Martin


Ozan
 

Hi Martin,

It is interesting to test dead transistors of 576 with itself, kind of circle of life ...

For fun, I examined the defective transistors (see the pics here:
https://groups.io/g/TekScopes2/album?id=260354).
I can't see these pictures, your link points to "TekScopes2" group (note the "2" at the end). There is such a group but Is it an active group?

Ozan


 

Hi Ozan,

It is interesting to test dead transistors of 576 with itself, kind of circle of life ...
:-)


I can't see these pictures, your link points to "TekScopes2" group (note the "2" at the end). There is such a group but Is it an active group?
The link is working, perhaps you're not registered to view the group? Anyway, it was a mistake of mine to put it on a different group, I didn't notice.
I'll make a copy in the "...2" group.

cheers
Martin


 

The link is working, perhaps you're not registered to view the group? Anyway, it was a mistake of mine to put it on a different group, I didn't notice.
I'll make a copy in the "...2" group.
I meant in the "..." without "2" group... its done.

Here it is: https://groups.io/g/TekScopes/album?id=260560

cheers
Martin


Ozan
 

I'll make a copy in the "...2" group.
I meant in the "..." without "2" group... its done.

Here it is: https://groups.io/g/TekScopes/album?id=260560
After seeing the pictures I want my own 576.

What is the difference between TekScopes and TekScopes2 groups?

Ozan


Dave Daniel
 

The TekScopes2 group was purportedly started by some TekScopes members who became disgruntled about something to do with the TekScopes forum. I don't know what precisely the issue was - it was before I joined either group.

DaveD

On 2/10/2021 4:10 PM, Ozan wrote:
I'll make a copy in the "...2" group.
I meant in the "..." without "2" group... its done.

Here it is: https://groups.io/g/TekScopes/album?id=260560
After seeing the pictures I want my own 576.

What is the difference between TekScopes and TekScopes2 groups?

Ozan



--
This email has been checked for viruses by Avast antivirus software.
https://www.avast.com/antivirus