Sony STR-6045 Diagnostic - help needed

Hi guys,

Lately, I have developed some interest in electronic repair and fixed a Celestion Ditton 662. But I wanted to learn more and my current project is now a small amplifier from the 70's, the Sony STR-6045.

When I purchased it, it had no power and a blown fuse on the Power Supply Board (a soldered one). I took the following steps to resolve the problem and I have some newb questions that I hope won't generate too many responses like "if you don't know electronic, hire a professional or go take some lessons online first". This is a hobby, I just want to learn.

1. I verified every components on the PSU board by unsoldering the necessary legs and measured resistors, diodes, capacitors and transistors. All were good except a 47uf capacitor that was 34% over specification. Any way, I changed the 3 electrolytics since there were not many and they were so old.

Question 1: The service manual seems to indicate that electrolytic capacitors can be 100% over specification. Is this normal and true?

2. I also cleaned the board heavily with alcohol and acetone as it seems to have oxidation or flux that were never cleaned up.

3. While everything was disconnected, I measured the voltage at the secondary terminals of the power transformer: 54.6 vac.
Question 2, is this normal for such and amp? The service manual does not specify what it is.


3. I put it back together and powered on the system (using a dim-bulb tester). Got a big shiny lamp. Half success. My diy current limiter is working but the amp is still not!

4. I then disconnected all wire except those going to the power transformer. And this time no light. And I then tried to measure the voltage as shown on the service manual.
Very stupid Question 3: How do I do this? I tried to clip the black prob to the chassis and to prob with the red one at the indicated points. I could not get any significant voltage.
Sony STR-6045 PS Board - Components Side.png

Sony STR-6045 PS Board - Conductor Side.png

5. Because there was no short when the power amplifier board was disconnected, I thought the problem was further down on the power amplifier board. But while trying to measure voltage it started to make smoke! R806 being super hot.
Question 4: Is it because I should have left the connections to the filter capacitors or is there a problem with the board I have repaired?

The full service manual is available online at many places (here for instance) for those interested although the scan is of poor quality. Here is an excerpt of the schematic diagram - power sections.


Thank you all in advance for your valuable advises.
Joined 2007
Paid Member
Question 1: The service manual seems to indicate that electrolytic capacitors can be 100% over specification. Is this normal and true?
For that era yes, it is correct. I used to work to +50% and -20%. Modern parts are closer to the marked value. Its not an issue.

Question 2, is this normal for such and amp? The service manual does not specify what it is.
Probably :)

The DC voltage rails are shown as -/+36 volts (which is 72 volts DC in total).

Divide 36 volts DC by 1.414 (which is root 2) and you get about 26 volts AC. That is the AC voltage needed to provide a 36 volt DC rail. So we have two 26 volt AC windings which is 52 volts AC in total. Just about what you measure.

Question 3: How do I do this? I tried to clip the black prob to the chassis and to prob with the red one at the indicated points. I could not get any significant voltage.
Bit hard to understand what you are trying to do there...

Question 4: Is it because I should have left the connections to the filter capacitors or is there a problem with the board I have repaired?
R806 is a series feed resistor to a 15 volt regulator. Sounds like you have a short or near short on the regulator output.

My advice would always be to try and diagnose by careful measurement (voltage and basic resistance checks) and not replace or remove any parts at first. There is so much that can go wrong just randomly replacing and taking parts out :) such as incorrect replacements, accidental shorts, fitting the wrong way around etc.

I would recommend you build the amp back up with everything connected, see where it is at and then begin with careful voltage checks.

Its good you are using a bulb tester (y)
  • Like
Reactions: 1 users
On point number 4, I wanted to isolate the short. The power section is made of two boards (the power supply and the power amplifier). I said to myself that if I disconnect the power amplifier and get no short, that would be because the problem is on the power amplifier board. Is that correct?

As for measuring voltage, this is the part that I'm struggling with. For instance, If i want to measure the voltage of 34.0 V on the collector of Q801:

  1. I set my fluke to DC testing ?
  2. I clip the black test lead to what?
  3. I touch the body of the transistor (or the right screw in the back) with my red test lead?
And I assume that all of this is only relevant if my bulb tester is not ON as there should be very limited voltage on the amplifier. Right?

Thank you so much for your help.
Joined 2007
Paid Member
On point number 4, I wanted to isolate the short.
I would isolate the line marked '23' that comes from the output of the regulator (the emitter of Q801). See if the short clears and check the emitter volts are correct at approx 15 volts.

You use DC volts for these tests.
Black lead goes to the zero volt line which ultimately is the junction of the two large reservoir caps (the 3300uF caps... can't just make the numbers out). That point should also be 'chassis' which is where I would measure from. If there are doubts then with the amp OFF check for continuity from the caps to chassis. 99.9% of the time chassis is ground and connected to the zero volt line.

The collector of Q801 is the metal body so that is where you measure to. The body may or may not be insulated from the heatsink. Looking at the picture it may not be for a single transistor on a heatsink but for multiple transistors sharing a heatsink they will likely be insulated with mica washers and insulating bushes on the screws. The case of the transistor is always (99.9999% for these style of transistor) the collector.
Thank you, Mooly, for all your answers. It really helped me.

I have been looking at this challenge further this week and today. I retested the 2 transistors Q801 (power transistor on heat sink) and Q802 (small separate transistor). They both tested good.

I followed your advice Mooly and put everything back together. When disconnecting line 23 (the orange wire to the tuner board), I still got a big shinny lamp on my bulb tester. So I disconnected every board one by one (leaving the ground wires this time - I think this is what cause the ps board to heat). Nothing would help.

I then disconnected the wire to the big filter capacitors. I then got no light on the bulb tester. I was even able to measure some voltage on the PS board (26 V at the collector of Q801 and close to 15V at base and emitter). Then strangely, I lost some voltage. Not sure it was a good idea to leave it on without the filter capacitors...

I am at the point where I suspect the two 3300 ufd filter capacitors. My multimeter is giving me some strange inconsistent reading (I made sure the capacitor were discharged with a screwdriver first) when measuring in circuit. However, this is such a nice wiring and I could not de-solder any of the wires at the terminals. This is why I would like to have some opinion before cutting the wires.

Notice I also checked Q707 and Q757 on the power amplifier board where the wires that go from the filter capacitors go to. They tested good.

Is it possible I have bad filter capacitors? Should I do more checks before cutting them out?

Here are some pictures of my suspects and their wiring:


Ok, I decided to bite the bullet and cut the filter capacitor out. One is definitely bad.

While i was creating my shopping list, I also decided to desolder the output power transistors Q901 and Q902 and to test them. Curiously the schematic and part list indicate two but there are 4 in reality. And guess what? 2 of the power transistors are also bad.

Now, I have a bigger issue. These transistors are Sony 2SD290 and are super difficult to find. My next question is what are the characteristics that can be changed (for instance can power or max collector to base voltage be higher?)

See substitute list here.

Joined 2007
Paid Member
Curiously the schematic and part list indicate two but there are 4 in reality. And guess what? 2 of the power transistors are also bad.
There are two per channel, so four in total :)

The 2SD290 are shown in my old Towers book as being a TO66 package which is smaller than a T03 so you need to check that. I wouldn't anticipate any real issues with alternatives but we need to know what size they are first. Another option is to fit flatpack types. And just looking at your picture and they do look to be T066 (ish) in size. That's an obsolete package these days.

Voltage rating, current rating and power rating must equal or exceed the original. It is possible to run into other issues due to modern parts being faster but I suspect this design will be fine. If flatpaks could be fitted then the TIP41C could be an option.
Thank you again for taking the time to read and answer my questions. I leaning toward fitting flatpacks as they are 1/20 the cost of TO66s.

Here are the specs of the 2SD290:

Material = Si
Struct = NPN
Pc > 20W
Vcb > 30V
Vce > 20V
Veb > 10V
Ic > 3A
Tj > 125C
Ft > 1MHz
Cc < 0pF
Hfe > 20
Caps = T066 + similar ones
R1 = 0 kOhm
R2 = 0 kOhm
R1/R2 = 0

Some more questions:

  1. TIP41C is rated 5V Vebo while this site ( says >10V. Strangely, if I filter on Mouser for >10V, I get a lot of results but none of them show the Vebo in the specification. Is the emitter base voltage important?
  2. On Mouser, the results I get have very high voltages VCEO and VCBO. Like 400 to 1000 V. This is far from the original specification. Any problem with this?
  3. To fit a flatpack, I just cut the collector lead and bend the other lead through the heat sink hole?
  4. Do I apply some thermal paste between the heatsink and the plastic case?
Joined 2007
Paid Member
Those 5 and 10 volt figures are reverse breakdown voltages of the B/E junction and are a region you never normally operate the transistor in. All the transistors in this amp only ever operate in the forward bias state and the B/E voltage is fixed by by the properties of silicon transistors and will always be around 0.6 to 0.8 volts.

The important figures are power dissipation, the collector current rating and the collector to emitter voltage rating.

Ft is the frequency where the current gain falls to 1 (one) and can be important in some designs. Although the figure is way over the audio frequency range huge differences in Ft can cause issues with stability in some designs... and that is hard to predict but I don't see any issues looking at this one.

The Sony is really a text book design and by and large and should work with pretty much anything.

You need to also make sure the driver transistors are OK and also all the low value resistors around the output stage.

Screenshot 2023-06-12 171222.png
I have found a local source where I can get 2N3054 transistors in TO-66 package. There are small differences but I think they could work and make my amp look a little bit less like a Frankenstein ;):

Type: NPN
Collector-Emitter Voltage, max: 55 V
Collector-Base Voltage, max: 90 V
Emitter-Base Voltage, max: 7 V
Collector Current − Continuous, max: 4 A
Collector Dissipation: 25 W
DC Current Gain (hfe): 25 to 150
Transition Frequency, min: 3 MHz
Operating and Storage Junction Temperature Range: -65 to +200 °C
Package: TO-66

The important figures are slightly under spec compared to the Sony. VCE = 55 V rather than 60 V, Collector current max 4 A rather than 5 A and FT min of 3 MHz instead of 4 MHz.

Should I change the 4 or leave the good channel with the original Sonys?
Joined 2007
Paid Member
All you can do is try them.

If you look at your amp the rails are shown as -/+36 volts. When idling with no signal the output is at zero volts and so each transistor sees 36 volts across it. As the signal rises and falls one transistor see less voltage and the other more. So in theory you could get close to 70 volts across one or other transistor under full signal drive although in practice you won't get close to that in normal listening.

What happens in those cases is something called 'secondary breakdown' where a transistor can fail even though none of the figures have been exceeded.

:xfingers: and hopefully all will be OK.

Remember to set the bias preset to its maximum resistance before powering up as that will give minimum bias current. Always use the bulb for initial testing.
Joined 2007
Paid Member
Bulb tester is absolutely perfect and tbh preferred in lots of situations. We 'had' to have all that sort of test gear to get Sony accreditation years ago but in practice it was rarely used. The bulb is quicker and easier and works just as well in cases like this.
Here is an update on the project. It took a while to get the parts and I had some vacations but it has now risen from the dead.

A big Thanks to you @Mooly.

It is now playing some radio fine despite I stupidly put my hot iron on the dial cord and broke it. :mad:

However I noticed that the voltages are incorrect on the PS board. And before I de-solder a bunch of components again, I could take some more advises (voltages in red are as per service manual, voltages in blue are the ones I measured):


Mooly said:

Divide 36 volts DC by 1.414 (which is root 2) and you get about 26 volts AC. That is the AC voltage needed to provide a 36 volt DC rail. So we have two 26 volt AC windings which is 52 volts AC in total. Just about what you measure.

  1. How can I get 34V at the collector of Q801 if the voltage is already dropped to 31.5V after the bridge rectifier and that it also goes through a 18 ohm resistor?
  2. The problem is definitely on the top trace where all voltages are at 0V. I'm not quite sure how it works. The voltage should flow from the base of Q801 and go through C803. Therefore there should be some voltage there. Bad capacitor or reverse capacitor? However shouldn't some voltage flow through R805 and C802 as well? They can't all be bad at the same time.
Thanks again for any insights that you can provide.
Joined 2007
Paid Member
Pleased to hear its working (y).

Drive cords, nightmare ;) I think all those involved in servicing have done that at one point or another. You need the manual which I assume gives a 'stringing guide' for a new one.

The voltages after the bridge are all 'raw' and unregulated. It also very common for actual voltages to be a little different to the manual. The actual voltage also depends on your AC mains voltage. If the AC mains has a tolerance of -/+10% then the secondary voltages from the transformer will also vary by that amount.

Provided the regulated rails are correct and the amp is not drawing excessive current then all is well.

The top trace is shown as being ground :) Look at the top right where it says 'WHT/BLACK Ground To Tuner Board'

Providing you have the +15 volts (and that is approx as well) on the emitter of Q801 then its fine.

The regulator works by having a fixed reference on Q802 emitter (the Zener diode). The 6.8k and 9.1k form a potential divider and set the voltage needed on the emitter of Q801 that is required to bring Q802 into conduction. That voltage is the Zener voltage plus approx 0.6 volts. When that happens Q801 conducts and holds the base of Q802 at a fixed voltage.

None of the caps pass any current at DC and should not have any influence on the DC conditions.
Joined 2007
Paid Member
Think about it in blocks with just these parts added one at a time.

Q801 has the unregulated supply on the collector. With Q801 alone that gives zero volts on the emitter. Q801 is OFF.

We add R807, the 2.2k. That turns Q801 on fully and we get the same voltage on the emitter as on the collector.

We now add Q802 and the 8.2 volt Zener. Q802 is OFF and nothing changes.

We add R804, the 6.8k. That turns on Q802 fully and it effectively ties the base of Q801 to the Zener. The output voltage at the emitter of Q801 is now approximately equal to the Zener voltage.

Now the clever bit. We add R805, the 9.1k.

Lets do the maths...
We need to apply at least the Zener voltage PLUS the base/emitter volt drop of Q802 (0.6 volts) in order to get Q802 to begin to conduct. Lets work backwards. We know the voltage needed on Q802 base which is 8.2 + 0.6 =8.8 volts.

The designer knows the rail voltage they want which is 15 volts.

We decide a value for R805, in this case 9.1k. We could choose 10k, we could choose 3k. We want it high enough not to cause excess power draw and low enough to ensure that it is dominant in setting the current in the resistor chain. So 9.1k was chosen.

We know we have 8.8 volts across 9100 ohm which is a current flow in the resistor of 0.96 milliamps. We can now calculate the other value of resistor (R804) needed to give the 15 volts we wanr. R=V/I and so we have (15-8.8)/0.00096 which is 6458 ohms. The designer used 6800 ohms. The current tapped off by the base of Q802 is minimal but the reason why the resistor chain has to be low enough to dominate the calculation. If we used 1meg then there wouldn't be enough base current available for example.

The theory is fixed but there are slight variables with exact Zener voltage and exact base/emitter volts. You can see how it works though and how its an exact calculation.

We could use a different Zener voltage and different resistor value for either of those resistors and simply calculate the value of the other resistor as needed.
  • Like
Reactions: 1 user
My teacher is awesome indeed! I can't thank him enough for all the details he is giving me.

@Mooly, not sure if you understood that I don't have 8.2 V at the emitter of Q802 but 23 V. But your explanations could only lead me to test again D801 (not a real zener but a transistor with only the base and the emitter connected to the circuit). And it shows OL on all legs. This little guy here:


One more trip to the store to buy a zener diode and I will confirm the new voltages.
  • Like
Reactions: 1 user
Joined 2007
Paid Member
Thanks for the kind words :)

That part (Zener) is definitely duff if it has 23 volts across it. It would be very unusual for a fault to have caused its failure... meaning its just a one off... but you should be aware of the possibility. The only viable high current path would be a failure of Q801 between C and B that allowed current to flow into Q801 and the Zener.