72V with no load is very low...there is something wrong with the circuit. Is this issue on both PSU channels?

Slowly raising means: I've turned on the unit with control pod and measured the 24V rails first after that i've measured 70V rail which showed actualy 86V, the voltage was not constant and was raising by 0.1V every 20-30 seconds. I can check with osciloscope and capture a record from whole voltage raising but only next weekend, don't have time for this now.

On my refurbished units It can reach near 90V or over with no load after some time (10-20 minutes) with no load.

For the translation don't worry. I understand what you mean.

Here you can check.

Yes, on both. About the same.
0.1 V every 20-30 seconds - this cannot be done on purpose. I just attribute this to warming up the power unit.
I could not find feedback circuits in this power supply to stabilize the output voltage, and the behavior of your serviceable kit shows that there is no stabilization there. 0.1 In every 20-30 seconds - I will be surprised if this does not turn out to be a warming effect.
I don’t think this is what needs to be checked. Here I have no doubt.
With respect, I may ask you to write down the waveform at several points in the circuit a bit later - to compare with mine. But this can happen only a little later, because this Saturday my oscilloscope broke, and I'm not sure when it will be operational!
Do not worry about it! ))) I also do not often get involved in this because of other things.
In addition, in the coming days I still have a broken oscilloscope.

In my refurbished unit the voltages with no load are raising a bit faster. Couple of minutes and the output voltages are 90V.

It is possible that here is no feedback for stabilizations...or it is achieved some strange way. I am not enough skilled for this.

OK when you will need help with the waveforms just ask me and i will do the measurements.

I looked at the oscilloscope for the pulse parameters on the primary and secondary windings of the T1 (T2) transformer, and I did not see any change in their duty cycle there when the load changed (tenths of a percent), as it should be with stabilized switching power supplies. Only a few percent decrease in amplitude — these percentages can be explained by an increase in the voltage drop across the transistor channel and the resistance of the transformer windings with increasing current passing through them, but these few percent cannot explain such a large voltage drop in my power source. And knowing how large a voltage drop occurs in your working circuit under load and not seeing the circuitry for feedback in the circuit, I am convinced that voltage stabilization is not provided here.
I see in this diagram only the elements that serve to synchronize the operation of all the arms and channels of these two power sources, but I still can not understand the principle of operation of this synchronization - there was very little time to look at the signals at different points, and their binding to each other .

Thanks!! ))

[quote=MichalD,Dec 15 2019, 10:37 PM]
*except tenths of a percent

Hello!
MichalD! Can you take the oscillograms for me at the following points of any of the 70V channels of the power source under load: 1. between the terminals 3 and 4, 1 and 2, and 9 and 10 of the T1 transformer (carefully for the oscilloscope - the voltage can exceed 300V in amplitude! ) Only three plots at this point.
2. voltage between terminal 4 (= terminal 1) of the transformer
T1 and anode D22,
3. voltage between terminal 4 (= terminal 1) of the transformer
T1 and connection point R65, R62 and C56 (just one plot at this point).
4. on the series-connected D7 and D6 (this item is the most insignificant for me).

According to the oscillograms from point 1, I can see the degree of squareness of the pulses on the primary winding of the transformer in a working device. The non-squareness of the pulses at the edges of the fronts, which is observed in my device, directly affects the power loss in the transistor Q2, and accordingly goes to its increased heating. Here I can also evaluate the transformation coefficient of the transformer and in what place should I look for the voltage loss at the output of my power source - but this is only if your oscilloscope allows you to measure the amplitude values ​​of the pulses, and you can measure them.
The waveform in paragraphs 2 and 3 is the control signal and the pre-signal on the Q2 shutter. I have big doubts about the correspondence of the control signal in my device - it will be easy to find the faulty element.

Very important!!! If your oscilloscope does not have a differential input, then the common (ground) conclusion of this input will almost certainly have a connection (direct or through capacitors of the noise suppression filter) to one of the terminals of the network supplying the oscilloscope, and when taking measurements in a device that also does not have galvanic isolation from networks - as it is in our case, due to shunting of the oscilloscope circuit (or even direct short circuit) of the circuits of the device under test by the parasitic circuit there will be at least incorrect measurements, but most likely just a failure of one of the devices due to a short circuit!! Therefore, to avoid this, it is highly advisable to connect the oscilloscope to the network through a 230V / 230V isolation transformer (or two identical on-board transformers 230V / xxx + xxx / 230V, if there is no one 230V / 230V) for galvanic isolation.

Извини, что я написал так длинно и подробно!

*Sorry to write so long and in detail!

Hello, sorry for keeping you waiting for so long, i was sick, still i am but feel better. I've made the oscilograms today for you. Don't worry i have differential probe for this type of measurements.

Here check:

1. T1 1(4) - 3



T1 1(4) - 2



T1 9 - 10



2. T1 1(4) - Anode of D22



3. T1 1(4) - R65, 62, C56



4. D7 and D6, this trace is little weird, should be at both side the same but mirrored, maybee one of these is failing.




and for the "but this is only if your oscilloscope allows you to measure the amplitude values ​​of the pulses" i am not sure what exactly you mean...so i guessed:


1-2 edge ringing



1-3 edge ringing



9-10 edge ringing



This post has been edited by MichalD: Jan 5 2020, 02:50 AM

Thank you so much, MichalD!
The oscilloscope already works for me, and I just analyzed the operation of my power supply - and the only thing I was missing was the exemplary oscillograms ... And now I have them!!!
Sorry that I am not writing yet - just I'm in a hurry to start comparing your and your data. I do it, double-checking everything if possible - therefore, most likely, I will not be heard for several days - I apologize in advance for this too!

Thank you very much again! It is impossible to express how much your help makes my work easier!

MichalD, what is the standard outlet voltage in your country, and what is the voltage between points A and B of the circuit in your device?

This is the waveform 1. T1 1 (4) - 3 in my device (yellow) - and here you can see that the voltage span on the primary coil T1 is 40 volts less than in your working one.

*in the sockets

Power grid voltage in my country is 230V AC. Between A and B is rectified mains voltage of 230V AC waveform. After turning on control pod is there around 325V peak value of 230V AC filterd by 4 big caps. Look here:

A - B control pod in stanby mode (RED LED ON) measured at D1 + - outputs. 4 big filtering caps disconnected by relay RL1



A - B control pod in IDLE mode or ON mode (MASTER LED ON, volume bars LED showing) measured at D1 + - outputs. 4 big filtering caps connected by relay RL1, flat line of around 325V



Yes i see, you have for some reason 40V less and therefore your output voltage is lower too.

This post has been edited by MichalD: Jan 6 2020, 05:25 AM

Thanks! I saw and understood. It is important!
Unfortunately, now I do not have the opportunity to answer in detail, but in the evening I will be free and will write!

Good evening, MichalD!
I arrived and I only had a little time. I measured my power grid voltage with a good tester, and got a value of 215VAC. Its peak value, respectively, should be 304V. The voltage that my tester showed between points A and B of the circuit is 274VDC. Theoretically, it should be equal to 304VDC with serviceable capacitors of the line filter, and no load. Thus, somewhere in the circuit up to the capacitors, approximately 304-274 = 30VDC is lost, and in the next few days when there is time, I will find the reason for this.
I hope this will solve the problem with the low voltage at the outputs of my 70V power supplies, but there is still the problem of severely overheating them even without load. Such overheating in switching power supplies can only be due to losses in powerful keys when they are switched.
MichalD! Among those oscillograms that you made at my request, there are number 2. T1 1 (4) - Anode of D22. In my device, at a nearby point - on the shutter G Q3, it looks like this - a yellow curve:



On this waveform of mine, I am confused by some impulse surges at the moment of switching MOSFETs. Could you once again retake this waveform (No. 2) - if possible synchronized? And also I would be very curious to look at the voltage waveform corresponding to the above mine - between the source (S) and the gate (G) Q3.
And it’s very good if all the oscillograms were taken with the oscilloscope input open - at a constant voltage - so that you can watch the voltage levels together with the constant component.

Hello! Hmmmm you power grid voltage 215 VAC...strange. If there is somewhere the voltage lost in your circuit, try to check VR1, TH1, R1 and all caps at the input.

Yes i can retake the waweforms again:

Do you want from me to oscilogram for:

T1 1(4) - Anode of D22 and Q3 G - Q3 S...not Q2?

T1 1(4) is the same as Q2 S.

I am not so skilled in osciloscope measurements and i don't want to damage or blow up my scope so i want to ask you if i can connect to scope input channel 1 dif. probe + to Q2 G and GND to T1 1(4, Q2 S) and to input channel 2 passive probe (supplied with the scope) + anode of D23 and GND to T1 1(4, Q2 S) in the real time and measure both signals at once? The scope will be connected to power grid thru isolation transformer. Tomorrow afternoon i will go outside of my house for a week so i can take the oscilograms only during next weekend, or if you will catch to respond me in tomorrow morning i can take the oscilograms the same day till noon.

By the way i tooked som thermal images of the PSU, and the mosfets are not the hottest parts. Hottest are some resistors near D6, 7 and the standing ones behing the 4 big caps. Mosfets are around 70 °C. Pictures taken after one hour in idle mode and the internals of the subwoofer were in open air.

Not. I have a power grid voltage of 220VAC +/- 5% in my mill, but I use a 220 / 220VAC isolation transformer here for debugging - for galvanic isolation from the network (for electrical safety when accidentally touching the high-voltage parts of the device), the output of which is, as I noticed yesterday , the voltage turned out to be slightly lower: 215VAC. It is not important for such work.
I also use the optional isolation transformer 220 / 220VAC for galvanic isolation from the network of my oscilloscope. This is not necessary if there is already an isolation transformer in the power circuit of the device being debugged.
Thanks! Of course, I will check all these elements, and still other possible reasons ...
Not. This is not necessary now. Maybe later, if you need to more subtly understand the operation of the device.
The following oscillograms are much more interesting to me now:
1. Between the source (S) Q2 and the anode (G) Q2.
2. Between the source (S) Q2 and the gate (G) Q2.
AND
On Q2 it is more interesting than on Q3 because in this case there will be no influence of the trigger circuit D6, D7 and C54 on the waveform.
Understand how perfect the differential inputs (or external differential probes) of your oscilloscope are, you need to read the documentation for them. Now there is no need to measure waveforms at different points of the device at the same time - just use one channel!
Just in case for the future: please measure with a good tester the resistance between the terminals of your dif. probe. In all four possible options. If the tester shows a short circuit in at least one of the measurement options (or even a resistance of hundreds of kilo-ohms), then these dif. probe cannot simultaneously record oscillograms in different — not having a common point — device locations !! In addition, when measuring in devices with high voltages, you must make sure in the manual that your dif. probe withstand this voltage difference !!
If you look carefully, the oscillograms that I post here to explain my thoughts are all necessarily taken with respect to the same common point for both channels, because I do not have real differential inputs at my oscilloscope (and this sometimes makes it very difficult understanding of the operation of the entire device. I have to somehow get out )
This is very interesting for me! Especially the description of the state of the measurement! )))
Unfortunately, I am not as "armed" as you are, and I do not have a thermal imager. Therefore, I can not lay out my thermograms. I have to focus only on my sensations of temperatures in the device (though they are very close to yours).I will be very glad if you manage to take for me those two oscillograms from the beginning of this message of mine. But please do not worry about the rest !!!

*Between the source (S) Q2 and the anode *D22.

OK i will do the measurements today, it looks that i will stay at home one more day. I will post them today later.

I was pleased to read this! )))

MichalD, I just made a reverse translation of my previous message - in order to understand how correctly Google translates technical texts, and saw the problem areas (at least in the reverse translation in some places I myself did not understand what I said! ) And besides, because of the rush today, I could not check everything.
Some places that I would like you to understand correctly:

1. To understand how perfect the differential inputs (or external differential probes) of your oscilloscope, I need to read the documentation from them. If you are interested, I could try to understand this if you send me such a description, or the brand of your product.

2. Just in case in the future: please measure the resistance between the terminals of both of your differential probes after connecting them to the oscilloscope - all four possible cross options (between their minuses, between their pluses and two more measurements between the minus of one and the plus of the other and vice versa). If the tester shows a short circuit in at least one of the measurement options (or even tens of kilo-ohms), then this is either not a differential probe at all, or a bad differential probe! Such probes cannot (!) Watch oscillograms if all four connection points of these probes are in different places of the device (there is no common ground for both probes)!! In addition, when measuring in high voltage instruments, the probe manual must make sure that it is designed for such a voltage with a margin!!